C++ API Reference

namespace dai

Typedefs

using Clock = std::chrono::steady_clock

Enums

enum CameraBoardSocket

Which Camera socket to use.

AUTO denotes that the decision will be made by device

Values:

enumerator AUTO

enumerator RGB

enumerator LEFT

enumerator RIGHT

enum CameraImageOrientation

Camera sensor image orientation / pixel readout. This exposes direct sensor settings. 90 or 270 degrees rotation is not available.

AUTO denotes that the decision will be made by device (e.g. on OAK-1/megaAI: ROTATE_180_DEG).

Values:

enumerator AUTO

enumerator NORMAL

enumerator HORIZONTAL_MIRROR

enumerator VERTICAL_FLIP

enumerator ROTATE_180_DEG

enum CameraModel

Which CameraModel to initialize the calibration with.

Values:

enumerator Perspective

enumerator Fisheye

enumerator Equirectangular

enumerator RadialDivision

enum ProcessorType

On which processor the node will be placed

Enum specifying processor

Values:

enumerator LEON_CSS

enumerator LEON_MSS

enum UsbSpeed

Get USB Speed

Values:

enumerator UNKNOWN

enumerator LOW

enumerator FULL

enumerator HIGH

enumerator SUPER

enumerator SUPER_PLUS

enum DatatypeEnum

Values:

enumerator Buffer

enumerator ImgFrame

enumerator NNData

enumerator ImageManipConfig

enumerator CameraControl

enumerator ImgDetections

enumerator SpatialImgDetections

enumerator SystemInformation

enumerator SpatialLocationCalculatorConfig

enumerator SpatialLocationCalculatorData

enumerator EdgeDetectorConfig

enumerator AprilTagConfig

enumerator AprilTags

enumerator Tracklets

enumerator IMUData

enumerator StereoDepthConfig

enumerator FeatureTrackerConfig

enumerator TrackedFeatures

enum SpatialLocationCalculatorAlgorithm

Values:

enumerator AVERAGE

enumerator MIN

enumerator MAX

enum MedianFilter

Median filter config for disparity post-processing

Values:

enumerator MEDIAN_OFF

enumerator KERNEL_3x3

enumerator KERNEL_5x5

enumerator KERNEL_7x7

enum LogLevel

Values:

enumerator TRACE

enumerator DEBUG

enumerator INFO

enumerator WARN

enumerator ERR

enumerator CRITICAL

enumerator OFF

enum IMUSensor

Available IMU sensors. More details about each sensor can be found in the datasheet:

https://www.ceva-dsp.com/wp-content/uploads/2019/10/BNO080_085-Datasheet.pdf

Values:

enumerator ACCELEROMETER_RAW

Section 2.1.1

Acceleration of the device without any postprocessing, straight from the sensor. Units are [m/s^2]

enumerator ACCELEROMETER

Section 2.1.1

Acceleration of the device including gravity. Units are [m/s^2]

enumerator LINEAR_ACCELERATION

Section 2.1.1

Acceleration of the device with gravity removed. Units are [m/s^2]

enumerator GRAVITY

Section 2.1.1

Gravity. Units are [m/s^2]

enumerator GYROSCOPE_RAW

Section 2.1.2

The angular velocity of the device without any postprocessing, straight from the sensor. Units are [rad/s]

enumerator GYROSCOPE_CALIBRATED

Section 2.1.2

The angular velocity of the device. Units are [rad/s]

enumerator GYROSCOPE_UNCALIBRATED

Section 2.1.2

Angular velocity without bias compensation. Units are [rad/s]

enumerator MAGNETOMETER_RAW

Section 2.1.3

Magnetic field measurement without any postprocessing, straight from the sensor. Units are [uTesla]

enumerator MAGNETOMETER_CALIBRATED

Section 2.1.3

The fully calibrated magnetic field measurement. Units are [uTesla]

enumerator MAGNETOMETER_UNCALIBRATED

Section 2.1.3

The magnetic field measurement without hard-iron offset applied. Units are [uTesla]

enumerator ROTATION_VECTOR

Section 2.2

The rotation vector provides an orientation output that is expressed as a quaternion referenced to magnetic north and gravity. It is produced by fusing the outputs of the accelerometer, gyroscope and magnetometer. The rotation vector is the most accurate orientation estimate available. The magnetometer provides correction in yaw to reduce drift and the gyroscope enables the most responsive performance.

enumerator GAME_ROTATION_VECTOR

Section 2.2

The game rotation vector is an orientation output that is expressed as a quaternion with no specific reference for heading, while roll and pitch are referenced against gravity. It is produced by fusing the outputs of the accelerometer and the gyroscope (i.e. no magnetometer). The game rotation vector does not use the magnetometer to correct the gyroscopes drift in yaw. This is a deliberate omission (as specified by Google) to allow gaming applications to use a smoother representation of the orientation without the jumps that an instantaneous correction provided by a magnetic field update could provide. Long term the output will likely drift in yaw due to the characteristics of gyroscopes, but this is seen as preferable for this output versus a corrected output.

enumerator GEOMAGNETIC_ROTATION_VECTOR

Section 2.2

The geomagnetic rotation vector is an orientation output that is expressed as a quaternion referenced to magnetic north and gravity. It is produced by fusing the outputs of the accelerometer and magnetometer. The gyroscope is specifically excluded in order to produce a rotation vector output using less power than is required to produce the rotation vector of section 2.2.4. The consequences of removing the gyroscope are: Less responsive output since the highly dynamic outputs of the gyroscope are not used More errors in the presence of varying magnetic fields.

enumerator ARVR_STABILIZED_ROTATION_VECTOR

Section 2.2

Estimates of the magnetic field and the roll/pitch of the device can create a potential correction in the rotation vector produced. For applications (typically augmented or virtual reality applications) where a sudden jump can be disturbing, the output is adjusted to prevent these jumps in a manner that takes account of the velocity of the sensor system.

enumerator ARVR_STABILIZED_GAME_ROTATION_VECTOR

Section 2.2

While the magnetometer is removed from the calculation of the game rotation vector, the accelerometer itself can create a potential correction in the rotation vector produced (i.e. the estimate of gravity changes). For applications (typically augmented or virtual reality applications) where a sudden jump can be disturbing, the output is adjusted to prevent these jumps in a manner that takes account of the velocity of the sensor system. This process is called AR/VR stabilization.

enum TrackerType

Values:

enumerator SHORT_TERM_KCF

Kernelized Correlation Filter tracking.

enumerator SHORT_TERM_IMAGELESS

Short term tracking without using image data.

enumerator ZERO_TERM_IMAGELESS

Ability to track the objects without accessing image data.

enumerator ZERO_TERM_COLOR_HISTOGRAM

Tracking using image data too.

enum TrackerIdAssignmentPolicy

Values:

enumerator UNIQUE_ID

Always take a new, unique ID.

enumerator SMALLEST_ID

Take the smallest available ID.

enum SerializationType

Values:

enumerator LIBNOP

enumerator JSON

enumerator JSON_MSGPACK

Functions

bool initialize()

bool initialize(std::string additionalInfo, bool installSignalHandler = true, void *javavm = nullptr)

bool initialize(const char *additionalInfo, bool installSignalHandler = true, void *javavm = nullptr)

bool initialize(void *javavm)

DEPTHAI_SERIALIZE_EXT(ChipTemperature, css, mss, upa, dss, average)

DEPTHAI_SERIALIZE_EXT(CpuUsage, average, msTime)

DEPTHAI_SERIALIZE_EXT(DetectionParserOptions, nnFamily, confidenceThreshold, classes, coordinates, anchors, anchorMasks, iouThreshold)

DEPTHAI_SERIALIZE_OPTIONAL_EXT(EepromData, version, boardCustom, boardName, boardRev, boardConf, hardwareConf, productName, batchName, batchTime, boardOptions, cameraData, stereoRectificationData, imuExtrinsics, miscellaneousData)

DEPTHAI_SERIALIZE_EXT(MemoryInfo, remaining, used, total)

DEPTHAI_SERIALIZE_EXT(Point2f, x, y)

DEPTHAI_SERIALIZE_EXT(Point3f, x, y, z)

DEPTHAI_SERIALIZE_EXT(Rect, x, y, width, height)

DEPTHAI_SERIALIZE_EXT(RotatedRect, center, size, angle)

DEPTHAI_SERIALIZE_EXT(Size2f, width, height)

DEPTHAI_SERIALIZE_EXT(StereoRectification, rectifiedRotationLeft, rectifiedRotationRight, leftCameraSocket, rightCameraSocket)

DEPTHAI_SERIALIZE_EXT(TensorInfo, order, dataType, numDimensions, dims, strides, name, offset)

DEPTHAI_SERIALIZE_EXT(Timestamp, sec, nsec)

bool isDatatypeSubclassOf(DatatypeEnum parent, DatatypeEnum children)

DEPTHAI_SERIALIZE_EXT(AprilTag, id, hamming, decisionMargin, topLeft, topRight, bottomRight, bottomLeft)

DEPTHAI_SERIALIZE_EXT(EdgeDetectorConfigData, sobelFilterHorizontalKernel, sobelFilterVerticalKernel)

DEPTHAI_SERIALIZE_EXT(IMUReport, sequence, accuracy, timestamp, tsDevice)

DEPTHAI_SERIALIZE_EXT(IMUReportAccelerometer, x, y, z, sequence, accuracy, timestamp, tsDevice)

DEPTHAI_SERIALIZE_EXT(IMUReportGyroscope, x, y, z, sequence, accuracy, timestamp, tsDevice)

DEPTHAI_SERIALIZE_EXT(IMUReportMagneticField, x, y, z, sequence, accuracy, timestamp, tsDevice)

DEPTHAI_SERIALIZE_EXT(IMUReportRotationVectorWAcc, i, j, k, real, rotationVectorAccuracy, sequence, accuracy, timestamp, tsDevice)

DEPTHAI_SERIALIZE_EXT(IMUPacket, acceleroMeter, gyroscope, magneticField, rotationVector)

DEPTHAI_SERIALIZE_EXT(ImgDetection, label, confidence, xmin, ymin, xmax, ymax)

DEPTHAI_SERIALIZE_EXT(SpatialImgDetection, label, confidence, xmin, ymin, xmax, ymax, spatialCoordinates)

DEPTHAI_SERIALIZE_EXT(SpatialLocationCalculatorConfigThresholds, lowerThreshold, upperThreshold)

DEPTHAI_SERIALIZE_EXT(SpatialLocationCalculatorConfigData, roi, depthThresholds, calculationAlgorithm)

DEPTHAI_SERIALIZE_EXT(SpatialLocations, config, depthAverage, depthMin, depthMax, depthAveragePixelCount, spatialCoordinates)

DEPTHAI_SERIALIZE_EXT(TrackedFeature, position, id, age, harrisScore, trackingError)

DEPTHAI_SERIALIZE_EXT(BoardConfig::USB, vid, pid, flashBootedVid, flashBootedPid, maxSpeed)

DEPTHAI_SERIALIZE_EXT(BoardConfig::Network, mtu, xlinkTcpNoDelay)

DEPTHAI_SERIALIZE_EXT(BoardConfig::GPIO, mode, direction, level, pull, drive, schmitt, slewFast)

DEPTHAI_SERIALIZE_EXT(BoardConfig::UART, tmp)

DEPTHAI_SERIALIZE_EXT(BoardConfig, usb, network, sysctl, watchdogTimeoutMs, watchdogInitialDelayMs, gpio, uart, pcieInternalClock, usb3PhyInternalClock)

DEPTHAI_SERIALIZE_EXT(LogMessage, nodeIdName, level, time, colorRangeStart, colorRangeEnd, payload)

DEPTHAI_SERIALIZE_EXT(NodeConnectionSchema, node1Id, node1OutputGroup, node1Output, node2Id, node2InputGroup, node2Input)

DEPTHAI_SERIALIZE_EXT(NodeIoInfo, group, name, type, blocking, queueSize, waitForMessage)

DEPTHAI_SERIALIZE_EXT(NodeObjInfo, id, name, properties, ioInfo)

DEPTHAI_SERIALIZE_EXT(PipelineSchema, connections, globalProperties, nodes)

DEPTHAI_SERIALIZE_EXT(AprilTagProperties, initialConfig, inputConfigSync)

DEPTHAI_SERIALIZE_EXT(ColorCameraProperties, initialControl, boardSocket, imageOrientation, colorOrder, interleaved, fp16, previewHeight, previewWidth, videoWidth, videoHeight, stillWidth, stillHeight, resolution, fps, sensorCropX, sensorCropY, previewKeepAspectRatio, ispScale)

DEPTHAI_SERIALIZE_EXT(DetectionNetworkProperties, blobSize, blobUri, numFrames, numThreads, numNCEPerThread, parser)

DEPTHAI_SERIALIZE_EXT(DetectionParserProperties, numFramesPool, networkInputs, parser)

DEPTHAI_SERIALIZE_EXT(EdgeDetectorProperties, initialConfig, outputFrameSize, numFramesPool)

DEPTHAI_SERIALIZE_EXT(FeatureTrackerProperties, initialConfig, numShaves, numMemorySlices)

DEPTHAI_SERIALIZE_EXT(GlobalProperties, leonCssFrequencyHz, leonMssFrequencyHz, pipelineName, pipelineVersion, cameraTuningBlobSize, cameraTuningBlobUri, calibData, xlinkChunkSize)

DEPTHAI_SERIALIZE_EXT(IMUSensorConfig, sensitivityEnabled, sensitivityRelative, changeSensitivity, reportRate, sensorId)

DEPTHAI_SERIALIZE_EXT(IMUProperties, imuSensors, batchReportThreshold, maxBatchReports, enableFirmwareUpdate)

DEPTHAI_SERIALIZE_EXT(ImageManipProperties, initialConfig, outputFrameSize, numFramesPool)

DEPTHAI_SERIALIZE_EXT(MonoCameraProperties, initialControl, boardSocket, imageOrientation, resolution, fps)

DEPTHAI_SERIALIZE_EXT(NeuralNetworkProperties, blobSize, blobUri, numFrames, numThreads, numNCEPerThread)

DEPTHAI_SERIALIZE_EXT(ObjectTrackerProperties, trackerThreshold, maxObjectsToTrack, detectionLabelsToTrack, trackerType, trackerIdAssignmentPolicy)

DEPTHAI_SERIALIZE_EXT(SPIInProperties, streamName, busId, maxDataSize, numFrames)

DEPTHAI_SERIALIZE_EXT(SPIOutProperties, streamName, busId)

DEPTHAI_SERIALIZE_EXT(ScriptProperties, scriptUri, scriptName, processor)

DEPTHAI_SERIALIZE_EXT(SpatialDetectionNetworkProperties, blobSize, blobUri, numFrames, numThreads, numNCEPerThread, parser, detectedBBScaleFactor, depthThresholds, calculationAlgorithm)

DEPTHAI_SERIALIZE_EXT(SpatialLocationCalculatorProperties, roiConfig)

DEPTHAI_SERIALIZE_EXT(StereoDepthProperties, initialConfig, depthAlignCamera, enableRectification, rectifyEdgeFillColor, width, height, outWidth, outHeight, outKeepAspectRatio, mesh, enableRuntimeStereoModeSwitch, numFramesPool, numPostProcessingShaves, numPostProcessingMemorySlices, focalLengthFromCalibration, useHomographyRectification)

DEPTHAI_SERIALIZE_EXT(SystemLoggerProperties, rateHz)

DEPTHAI_SERIALIZE_EXT(VideoEncoderProperties, bitrate, keyframeFrequency, maxBitrate, numBFrames, numFramesPool, outputFrameSize, profile, quality, lossless, rateCtrlMode, frameRate)

DEPTHAI_SERIALIZE_EXT(XLinkInProperties, streamName, maxDataSize, numFrames)

DEPTHAI_SERIALIZE_EXT(XLinkOutProperties, maxFpsLimit, streamName, metadataOnly)

Variables

constexpr uint32_t BOARD_CONFIG_MAGIC1 = 0x78010000U

constexpr uint32_t BOARD_CONFIG_MAGIC2 = 0x21ea17e6U

constexpr const char *LOG_DEFAULT_PATTERN = "[%E.%e] [%n] [%^%l%$] %v"

static constexpr auto DEFAULT_SERIALIZATION_TYPE = SerializationType::LIBNOP

class ADatatype

#include <ADatatype.hpp>

Abstract message.

Subclassed by dai::Buffer

struct AprilTag

#include <RawAprilTags.hpp>

AprilTag structure.

Public Members

int id = 0

The decoded ID of the tag

int hamming = 0

How many error bits were corrected? Note: accepting large numbers of corrected errors leads to greatly increased false positive rates. As of this implementation, the detector cannot detect tags with a hamming distance greater than 2.

float decisionMargin = 0.f

A measure of the quality of the binary decoding process; the average difference between the intensity of a data bit versus the decision threshold. Higher numbers roughly indicate better decodes. This is a reasonable measure of detection accuracy only for very small tags not effective for larger tags (where we could have sampled anywhere within a bit cell and still gotten a good detection.

Point2f topLeft

The detected top left coordinates.

Point2f topRight

The detected top right coordinates.

Point2f bottomRight

The detected bottom right coordinates.

Point2f bottomLeft

The detected bottom left coordinates.

class AprilTagConfig : public dai::Buffer

#include <AprilTagConfig.hpp>

AprilTagConfig message.

Public Functions

AprilTagConfig()

Construct AprilTagConfig message.

AprilTagConfig &setFamily(Family family)

Parameters

• family: AprilTag family

AprilTagConfig &set(dai::RawAprilTagConfig config)

Set explicit configuration.

Parameters

• config: Explicit configuration

dai::RawAprilTagConfig get() const

Retrieve configuration data for AprilTag.

Return

config for stereo depth algorithm

struct AprilTagProperties : public dai::PropertiesSerializable<Properties, AprilTagProperties>

#include <AprilTagProperties.hpp>

Specify properties for AprilTag

Public Members

bool inputConfigSync = false

Whether to wait for config at ‘inputConfig’ IO.

class AprilTags : public dai::Buffer

#include <AprilTags.hpp>

AprilTags message.

Public Functions

AprilTags()

Construct AprilTags message.

struct Asset

#include <AssetManager.hpp>

Asset is identified with string key and can store arbitrary binary data.

class AssetManager

#include <AssetManager.hpp>

AssetManager can store assets and serialize.

Public Functions

void addExisting(std::vector<std::shared_ptr<Asset>> assets)

Adds all assets in an array to the AssetManager

Parameters

• assets: Vector of assets to add

std::shared_ptr<dai::Asset> set(Asset asset)

Adds or overwrites an asset object to AssetManager.

Return

Shared pointer to asset

Parameters

• asset: Asset to add

std::shared_ptr<dai::Asset> set(const std::string &key, Asset asset)

Adds or overwrites an asset object to AssetManager with a specified key. Key value will be assigned to an Asset as well

Return

Shared pointer to asset

Parameters

• key: Key under which the asset should be stored

• asset: Asset to store

std::shared_ptr<dai::Asset> set(const std::string &key, const dai::Path &path, int alignment = 64)

Loads file into asset manager under specified key.

Parameters

• key: Key under which the asset should be stored

• path: Path to file which to load as asset

• alignment: [Optional] alignment of asset data in asset storage. Default is 64B

std::shared_ptr<dai::Asset> set(const std::string &key, const std::vector<std::uint8_t> &data, int alignment = 64)

Loads file into asset manager under specified key.

Return

Shared pointer to asset

Parameters

• key: Key under which the asset should be stored

• data: Asset data

• alignment: [Optional] alignment of asset data in asset storage. Default is 64B

std::shared_ptr<const Asset> get(const std::string &key) const

Return

Asset assigned to the specified key or a nullptr otherwise

std::shared_ptr<Asset> get(const std::string &key)

Return

Asset assigned to the specified key or a nullptr otherwise

std::vector<std::shared_ptr<const Asset>> getAll() const

Return

All asset stored in the AssetManager

std::vector<std::shared_ptr<Asset>> getAll()

Return

All asset stored in the AssetManager

std::size_t size() const

Return

Number of asset stored in the AssetManager

void remove(const std::string &key)

Removes asset with key

Parameters

• key: Key of asset to remove

void serialize(AssetsMutable &assets, std::vector<std::uint8_t> &assetStorage, std::string prefix = "") const

Serializes.

class Assets

Subclassed by dai::AssetsMutable

class AssetsMutable : public dai::Assets

struct AssetView

struct BoardConfig

Public Members

std::vector<std::string> sysctl

Optional list of FreeBSD sysctl parameters to be set (system, network, etc.). For example: “net.inet.tcp.delayed_ack=0” (this one is also set by default)

std::unordered_map<std::int8_t, UART> uart

UART instance map.

struct GPIO

Public Types

enum Drive

Drive strength in mA (2, 4, 8 and 12mA)

Values:

enumerator MA_2

enumerator MA_4

enumerator MA_8

enumerator MA_12

struct Network

#include <BoardConfig.hpp>

Network configuration.

Public Members

uint16_t mtu = 0

Network MTU, 0 is auto (usually 1500 for Ethernet) or forwarded from bootloader (not yet implemented there). Note: not advised to increase past 1500 for now

bool xlinkTcpNoDelay = true

Sets the TCP_NODELAY option for XLink TCP sockets (disable Nagle’s algorithm), reducing latency at the expense of a small hit for max throughput. Default is true

struct UART

#include <BoardConfig.hpp>

UART instance config.

struct USB

class Buffer : public dai::ADatatype

#include <Buffer.hpp>

Base message - buffer of binary data.

Subclassed by dai::AprilTagConfig, dai::AprilTags, dai::CameraControl, dai::EdgeDetectorConfig, dai::FeatureTrackerConfig, dai::ImageManipConfig, dai::ImgDetections, dai::ImgFrame, dai::IMUData, dai::NNData, dai::SpatialImgDetections, dai::SpatialLocationCalculatorConfig, dai::SpatialLocationCalculatorData, dai::StereoDepthConfig, dai::SystemInformation, dai::TrackedFeatures, dai::Tracklets

Public Functions

Buffer()

Creates Buffer message.

std::vector<std::uint8_t> &getData() const

Get non-owning reference to internal buffer.

Return

Reference to internal buffer

void setData(const std::vector<std::uint8_t> &data)

Parameters

• data: Copies data to internal buffer

void setData(std::vector<std::uint8_t> &&data)

Parameters

• data: Moves data to internal buffer

class CalibrationHandler

#include <CalibrationHandler.hpp>

CalibrationHandler is an interface to read/load/write structured calibration and device data. The following fields are protected and aren’t allowed to be overriden by default:

• boardName

• boardRev

• boardConf

• hardwareConf

• batchName

• batchTime

• boardOptions

Public Functions

CalibrationHandler(dai::Path eepromDataPath)

Construct a new Calibration Handler object using the eeprom json file created from calibration procedure.

Parameters

• eepromDataPath: takes the full path to the json file containing the calibration and device info.

CalibrationHandler(dai::Path calibrationDataPath, dai::Path boardConfigPath)

Construct a new Calibration Handler object using the board config json file and .calib binary files created using gen1 calibration.

Parameters

• calibrationDataPath: Full Path to the .calib binary file from the gen1 calibration. (Supports only Version 5)

• boardConfigPath: Full Path to the board config json file containing device information.

CalibrationHandler(EepromData eepromData)

Construct a new Calibration Handler object from EepromData object.

Parameters

• eepromData: EepromData data structure containing the calibration data.

dai::EepromData getEepromData() const

Get the Eeprom Data object

Return

EepromData object which contains the raw calibration data

std::vector<std::vector<float>> getCameraIntrinsics(CameraBoardSocket cameraId, int resizeWidth = -1, int resizeHeight = -1, Point2f topLeftPixelId = Point2f(), Point2f bottomRightPixelId = Point2f())

Get the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Represents the 3x3 intrinsics matrix of the respective camera at the requested size and crop dimensions.

Parameters

• cameraId: Uses the cameraId to identify which camera intrinsics to return

• resizewidth: resized width of the image for which intrinsics is requested. resizewidth = -1 represents width is same as default intrinsics

• resizeHeight: resized height of the image for which intrinsics is requested. resizeHeight = -1 represents height is same as default intrinsics

• topLeftPixelId: (x, y) point represents the top left corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

• bottomRightPixelId: (x, y) point represents the bottom right corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

std::vector<std::vector<float>> getCameraIntrinsics(CameraBoardSocket cameraId, Size2f destShape, Point2f topLeftPixelId = Point2f(), Point2f bottomRightPixelId = Point2f())

Get the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Represents the 3x3 intrinsics matrix of the respective camera at the requested size and crop dimensions.

Parameters

• cameraId: Uses the cameraId to identify which camera intrinsics to return

• destShape: resized width and height of the image for which intrinsics is requested.

• topLeftPixelId: (x, y) point represents the top left corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

• bottomRightPixelId: (x, y) point represents the bottom right corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

std::vector<std::vector<float>> getCameraIntrinsics(CameraBoardSocket cameraId, std::tuple<int, int> destShape, Point2f topLeftPixelId = Point2f(), Point2f bottomRightPixelId = Point2f())

Get the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Represents the 3x3 intrinsics matrix of the respective camera at the requested size and crop dimensions.

Parameters

• cameraId: Uses the cameraId to identify which camera intrinsics to return

• destShape: resized width and height of the image for which intrinsics is requested.

• topLeftPixelId: (x, y) point represents the top left corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

• bottomRightPixelId: (x, y) point represents the bottom right corner coordinates of the cropped image which is used to modify the intrinsics for the respective cropped image

std::tuple<std::vector<std::vector<float>>, int, int> getDefaultIntrinsics(CameraBoardSocket cameraId)

Get the Default Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Represents the 3x3 intrinsics matrix of the respective camera along with width and height at which it was calibrated.

Parameters

• cameraId: Uses the cameraId to identify which camera intrinsics to return

std::vector<float> getDistortionCoefficients(CameraBoardSocket cameraId)

Get the Distortion Coefficients object

Return

the distortion coefficients of the requested camera in this order: [k1,k2,p1,p2,k3,k4,k5,k6,s1,s2,s3,s4,τx,τy]

Parameters

• cameraId: Uses the cameraId to identify which distortion Coefficients to return.

float getFov(CameraBoardSocket cameraId, bool useSpec = true)

Get the Fov of the camera

Return

field of view of the camera with given cameraId.

Parameters

• cameraId: of the camera of which we are fetching fov.

• useSpec: Disabling this bool will calculate the fov based on intrinsics (focal length, image width), instead of getting it from the camera specs

uint8_t getLensPosition(CameraBoardSocket cameraId)

Get the lens position of the given camera

Return

lens position of the camera with given cameraId at which it was calibrated.

Parameters

• cameraId: of the camera with lens position is requested.

CameraModel getDistortionModel(CameraBoardSocket cameraId)

Get the distortion model of the given camera

Return

lens position of the camera with given cameraId at which it was calibrated.

Parameters

• cameraId: of the camera with lens position is requested.

std::vector<std::vector<float>> getCameraExtrinsics(CameraBoardSocket srcCamera, CameraBoardSocket dstCamera, bool useSpecTranslation = false)

Get the Camera Extrinsics object between two cameras from the calibration data if there is a linked connection between any two cameras then the relative rotation and translation (in centimeters) is returned by this function.

Matrix representation of transformation matrix

\[\begin{split} \text{Transformation Matrix} = \left [ \begin{matrix} r_{00} & r_{01} & r_{02} & T_x \\ r_{10} & r_{11} & r_{12} & T_y \\ r_{20} & r_{21} & r_{22} & T_z \\ 0 & 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

a transformationMatrix which is 4x4 in homogeneous coordinate system

Parameters

• srcCamera: Camera Id of the camera which will be considered as origin.

• dstCamera: Camera Id of the destination camera to which we are fetching the rotation and translation from the SrcCamera

• useSpecTranslation: Enabling this bool uses the translation information from the board design data

std::vector<float> getCameraTranslationVector(CameraBoardSocket srcCamera, CameraBoardSocket dstCamera, bool useSpecTranslation = true)

Get the Camera translation vector between two cameras from the calibration data.

Return

a translation vector like [x, y, z] in centimeters

Parameters

• srcCamera: Camera Id of the camera which will be considered as origin.

• dstCamera: Camera Id of the destination camera to which we are fetching the translation vector from the SrcCamera

• useSpecTranslation: Disabling this bool uses the translation information from the calibration data (not the board design data)

float getBaselineDistance(CameraBoardSocket cam1 = CameraBoardSocket::RIGHT, CameraBoardSocket cam2 = CameraBoardSocket::LEFT, bool useSpecTranslation = true)

Get the baseline distance between two specified cameras. By default it will get the baseline between CameraBoardSocket.RIGHT and CameraBoardSocket.LEFT.

Return

baseline distance in centimeters

Parameters

• cam1: First camera

• cam2: Second camera

• useSpecTranslation: Enabling this bool uses the translation information from the board design data (not the calibration data)

std::vector<std::vector<float>> getCameraToImuExtrinsics(CameraBoardSocket cameraId, bool useSpecTranslation = false)

Get the Camera To Imu Extrinsics object From the data loaded if there is a linked connection between IMU and the given camera then there relative rotation and translation from the camera to IMU is returned.

Matrix representation of transformation matrix

\[\begin{split} \text{Transformation Matrix} = \left [ \begin{matrix} r_{00} & r_{01} & r_{02} & T_x \\ r_{10} & r_{11} & r_{12} & T_y \\ r_{20} & r_{21} & r_{22} & T_z \\ 0 & 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Returns a transformationMatrix which is 4x4 in homogeneous coordinate system

Parameters

• cameraId: Camera Id of the camera which will be considered as origin. from which Transformation matrix to the IMU will be found

• useSpecTranslation: Enabling this bool uses the translation information from the board design data

std::vector<std::vector<float>> getImuToCameraExtrinsics(CameraBoardSocket cameraId, bool useSpecTranslation = false)

Get the Imu To Camera Extrinsics object from the data loaded if there is a linked connection between IMU and the given camera then there relative rotation and translation from the IMU to Camera is returned.

Matrix representation of transformation matrix

\[\begin{split} \text{Transformation Matrix} = \left [ \begin{matrix} r_{00} & r_{01} & r_{02} & T_x \\ r_{10} & r_{11} & r_{12} & T_y \\ r_{20} & r_{21} & r_{22} & T_z \\ 0 & 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Return

Returns a transformationMatrix which is 4x4 in homogeneous coordinate system

Parameters

• cameraId: Camera Id of the camera which will be considered as destination. To which Transformation matrix from the IMU will be found.

• useSpecTranslation: Enabling this bool uses the translation information from the board design data

std::vector<std::vector<float>> getStereoRightRectificationRotation()

Get the Stereo Right Rectification Rotation object

Return

returns a 3x3 rectification rotation matrix

std::vector<std::vector<float>> getStereoLeftRectificationRotation()

Get the Stereo Left Rectification Rotation object

Return

returns a 3x3 rectification rotation matrix

dai::CameraBoardSocket getStereoLeftCameraId()

Get the camera id of the camera which is used as left camera of the stereo setup

Return

cameraID of the camera used as left camera

dai::CameraBoardSocket getStereoRightCameraId()

Get the camera id of the camera which is used as right camera of the stereo setup

Return

cameraID of the camera used as right camera

bool eepromToJsonFile(dai::Path destPath) const

Write raw calibration/board data to json file.

Return

True on success, false otherwise

Parameters

• destPath: Full path to the json file in which raw calibration data will be stored

nlohmann::json eepromToJson() const

Get JSON representation of calibration data

Return

JSON structure

void setBoardInfo(std::string boardName, std::string boardRev)

Set the Board Info object

Parameters

• version: Sets the version of the Calibration data(Current version is 6)

• boardName: Sets your board name.

• boardRev: set your board revision id.

void setBoardInfo(std::string productName, std::string boardName, std::string boardRev, std::string boardConf, std::string hardwareConf, std::string batchName, uint64_t batchTime, uint32_t boardOptions, std::string boardCustom = "")

Set the Board Info object. Creates version 7 EEPROM data

Parameters

• productName: Sets product name (alias).

• boardName: Sets board name.

• boardRev: Sets board revision id.

• boardConf: Sets board configuration id.

• hardwareConf: Sets hardware configuration id.

• batchName: Sets batch name.

• batchTime: Sets batch time (unix timestamp).

• boardCustom: Sets a custom board (Default empty string).

void setCameraIntrinsics(CameraBoardSocket cameraId, std::vector<std::vector<float>> intrinsics, Size2f frameSize)

Set the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Parameters

• cameraId: CameraId of the camera for which Camera intrinsics are being loaded

• intrinsics: 3x3 intrinsics matrix

• frameSize: Represents the width and height of the image at which intrinsics are calculated.

void setCameraIntrinsics(CameraBoardSocket cameraId, std::vector<std::vector<float>> intrinsics, int width, int height)

Set the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Parameters

• cameraId: CameraId of the camera for which Camera intrinsics are being loaded

• intrinsics: 3x3 intrinsics matrix

• width: Represents the width of the image at which intrinsics are calculated.

• height: Represents the height of the image at which intrinsics are calculated.

void setCameraIntrinsics(CameraBoardSocket cameraId, std::vector<std::vector<float>> intrinsics, std::tuple<int, int> frameSize)

Set the Camera Intrinsics object

Matrix representation of intrinsic matrix

\[\begin{split} \text{Intrinsic Matrix} = \left [ \begin{matrix} f_x & 0 & c_x \\ 0 & f_y & c_y \\ 0 & 0 & 1 \end{matrix} \right ] \end{split}\]

Parameters

• cameraId: CameraId of the camera for which Camera intrinsics are being loaded

• intrinsics: 3x3 intrinsics matrix

• frameSize: Represents the width and height of the image at which intrinsics are calculated.

void setDistortionCoefficients(CameraBoardSocket cameraId, std::vector<float> distortionCoefficients)

Sets the distortion Coefficients obtained from camera calibration

Parameters

• cameraId: Camera Id of the camera for which distortion coefficients are computed

• distortionCoefficients: Distortion Coefficients of the respective Camera.

void setFov(CameraBoardSocket cameraId, float hfov)

Set the Fov of the Camera

Parameters

• cameraId: Camera Id of the camera

• hfov: Horizontal fov of the camera from Camera Datasheet

void setLensPosition(CameraBoardSocket cameraId, uint8_t lensPosition)

Sets the distortion Coefficients obtained from camera calibration

Parameters

• cameraId: Camera Id of the camera

• lensPosition: lens posiotion value of the camera at the time of calibration

void setCameraType(CameraBoardSocket cameraId, CameraModel cameraModel)

Set the Camera Type object

Parameters

• cameraId: CameraId of the camera for which cameraModel Type is being updated.

• cameraModel: Type of the model the camera represents

void setCameraExtrinsics(CameraBoardSocket srcCameraId, CameraBoardSocket destCameraId, std::vector<std::vector<float>> rotationMatrix, std::vector<float> translation, std::vector<float> specTranslation = {0, 0, 0})

Set the Camera Extrinsics object

Parameters

• srcCameraId: Camera Id of the camera which will be considered as relative origin.

• destCameraId: Camera Id of the camera which will be considered as destination from srcCameraId.

• rotationMatrix: Rotation between srcCameraId and destCameraId origins.

• translation: Translation between srcCameraId and destCameraId origins.

• specTranslation: Translation between srcCameraId and destCameraId origins from the design.

void setImuExtrinsics(CameraBoardSocket destCameraId, std::vector<std::vector<float>> rotationMatrix, std::vector<float> translation, std::vector<float> specTranslation = {0, 0, 0})

Set the Imu to Camera Extrinsics object

Parameters

• destCameraId: Camera Id of the camera which will be considered as destination from IMU.

• rotationMatrix: Rotation between srcCameraId and destCameraId origins.

• translation: Translation between IMU and destCameraId origins.

• specTranslation: Translation between IMU and destCameraId origins from the design.

void setStereoLeft(CameraBoardSocket cameraId, std::vector<std::vector<float>> rectifiedRotation)

Set the Stereo Left Rectification object

Homography of the Left Rectification = Intrinsics_right * rectifiedRotation * inv(Intrinsics_left)

Parameters

• cameraId: CameraId of the camera which will be used as left Camera of stereo Setup

• rectifiedRotation: Rectification rotation of the left camera required for feature matching

void setStereoRight(CameraBoardSocket cameraId, std::vector<std::vector<float>> rectifiedRotation)

Set the Stereo Right Rectification object

Homography of the Right Rectification = Intrinsics_right * rectifiedRotation * inv(Intrinsics_right)

Parameters

• cameraId: CameraId of the camera which will be used as left Camera of stereo Setup

• rectifiedRotation: Rectification rotation of the left camera required for feature matching

bool validateCameraArray()

Using left camera as the head it iterates over the camera extrinsics connection to check if all the camera extrinsics are connected and no loop exists.

Return

true on proper connection with no loops.

Public Static Functions

CalibrationHandler fromJson(nlohmann::json eepromDataJson)

Construct a new Calibration Handler object from JSON EepromData.

Parameters

• eepromDataJson: EepromData as JSON

class CallbackHandler

class CameraControl : public dai::Buffer

#include <CameraControl.hpp>

CameraControl message. Specifies various camera control commands like:

• Still capture

• Auto focus

• Anti banding

• Auto white balance

• Scene

• Effect

• …

Public Functions

CameraControl()

Construct CameraControl message.

CameraControl &setCaptureStill(bool capture)

Set a command to capture a still image

CameraControl &setStartStreaming()

Set a command to start streaming

CameraControl &setStopStreaming()

Set a command to stop streaming

CameraControl &setAutoFocusMode(AutoFocusMode mode)

Set a command to specify autofocus mode

CameraControl &setAutoFocusTrigger()

Set a command to trigger autofocus

CameraControl &setAutoFocusRegion(uint16_t startX, uint16_t startY, uint16_t width, uint16_t height)

Set a command to specify focus region in pixels

Parameters

• startX: X coordinate of top left corner of region

• startY: Y coordinate of top left corner of region

• width: Region width

• height: Region height

CameraControl &setManualFocus(uint8_t lensPosition)

Set a command to specify manual focus position

Parameters

• lensPosition: specify lens position 0..255

CameraControl &setAutoExposureEnable()

Set a command to enable auto exposure

CameraControl &setAutoExposureLock(bool lock)

Set a command to specify lock auto exposure

Parameters

• lock: Auto exposure lock mode enabled or disabled

CameraControl &setAutoExposureRegion(uint16_t startX, uint16_t startY, uint16_t width, uint16_t height)

Set a command to specify auto exposure region in pixels

Parameters

• startX: X coordinate of top left corner of region

• startY: Y coordinate of top left corner of region

• width: Region width

• height: Region height

CameraControl &setAutoExposureCompensation(int compensation)

Set a command to specify auto exposure compensation

Parameters

• compensation: Compensation value between -9..9

CameraControl &setAntiBandingMode(AntiBandingMode mode)

Set a command to specify auto banding mode

Parameters

• mode: Auto banding mode to use

CameraControl &setManualExposure(uint32_t exposureTimeUs, uint32_t sensitivityIso)

Set a command to manually specify exposure

Parameters

• exposureTimeUs: Exposure time in microseconds

• sensitivityIso: Sensitivity as ISO value, usual range 100..1600

CameraControl &setAutoWhiteBalanceMode(AutoWhiteBalanceMode mode)

Set a command to specify auto white balance mode

Parameters

• mode: Auto white balance mode to use

CameraControl &setAutoWhiteBalanceLock(bool lock)

Set a command to specify auto white balance lock

Parameters

• lock: Auto white balance lock mode enabled or disabled

CameraControl &setManualWhiteBalance(int colorTemperatureK)

Set a command to manually specify white-balance color correction

Parameters

• colorTemperatureK: Light source color temperature in kelvins, range 1000..12000

CameraControl &setBrightness(int value)

Set a command to adjust image brightness

Parameters

• value: Brightness, range -10..10

CameraControl &setContrast(int value)

Set a command to adjust image contrast

Parameters

• value: Contrast, range -10..10

CameraControl &setSaturation(int value)

Set a command to adjust image saturation

Parameters

• value: Saturation, range -10..10

CameraControl &setSharpness(int value)

Set a command to adjust image sharpness

Parameters

• value: Sharpness, range 0..4

CameraControl &setLumaDenoise(int value)

Set a command to adjust luma denoise amount

Parameters

• value: Luma denoise amount, range 0..4

CameraControl &setChromaDenoise(int value)

Set a command to adjust chroma denoise amount

Parameters

• value: Chroma denoise amount, range 0..4

CameraControl &setSceneMode(SceneMode mode)

Set a command to specify scene mode

Parameters

• mode: Scene mode

CameraControl &setEffectMode(EffectMode mode)

Set a command to specify effect mode

Parameters

• mode: Effect mode

bool getCaptureStill() const

Check whether command to capture a still is set

Return

True if capture still command is set

struct CameraInfo

#include <CameraInfo.hpp>

CameraInfo structure.

struct ChipTemperature

#include <ChipTemperature.hpp>

Chip temperature information.

Multiple temperature measurement points and their average

Public Members

float css

CPU Subsystem

float mss

Media Subsystem

float upa

Shave Array

float dss

DRAM Subsystem

float average

Average of measurements

struct ColorCameraProperties : public dai::PropertiesSerializable<Properties, ColorCameraProperties>

#include <ColorCameraProperties.hpp>

Specify properties for ColorCamera such as camera ID, …

Public Types

enum SensorResolution

Select the camera sensor resolution

Values:

enumerator THE_1080_P

enumerator THE_4_K

enumerator THE_12_MP

enumerator THE_13_MP

enum ColorOrder

For 24 bit color these can be either RGB or BGR

Values:

enumerator BGR

enumerator RGB

Public Members

CameraBoardSocket boardSocket = CameraBoardSocket::AUTO

Which socket will color camera use

CameraImageOrientation imageOrientation = CameraImageOrientation::AUTO

Camera sensor image orientation / pixel readout

ColorOrder colorOrder = ColorOrder::BGR

For 24 bit color these can be either RGB or BGR

bool interleaved = true

Are colors interleaved (R1G1B1, R2G2B2, …) or planar (R1R2…, G1G2…, B1B2)

bool fp16 = false

Are values FP16 type (0.0 - 255.0)

uint32_t previewHeight = 300

Preview frame output height

uint32_t previewWidth = 300

Preview frame output width

int32_t videoWidth = AUTO

Preview frame output width

int32_t videoHeight = AUTO

Preview frame output height

int32_t stillWidth = AUTO

Preview frame output width

int32_t stillHeight = AUTO

Preview frame output height

SensorResolution resolution = SensorResolution::THE_1080_P

Select the camera sensor resolution

float fps = 30.0

Camera sensor FPS

float sensorCropX = AUTO

Initial sensor crop, -1 signifies center crop

bool previewKeepAspectRatio = true

Whether to keep aspect ratio of input (video size) or not

IspScale ispScale

Configure scaling for isp output.

struct IspScale

template<typename T>
class copyable_unique_ptr : public std::unique_ptr<T>

#include <copyable_unique_ptr.hpp>

A smart pointer with deep copy semantics.

This is similar to std::unique_ptr in that it does not permit shared ownership of the contained object. However, unlike std::unique_ptr, copyable_unique_ptr supports copy and assignment operations, by insisting that the contained object be “copyable”. To be copyable, the class must have either an accessible copy constructor, or it must have an accessible clone method with signature

std::unique_ptr<Foo> Clone() const;

where Foo is the type of the managed object. By “accessible” we mean either that the copy constructor or clone method is public, or friend copyable_unique_ptr<Foo>; appears in Foo’s class declaration.

Generally, the API is modeled as closely as possible on the C++ standard std::unique_ptr API and copyable_unique_ptr<T> is interoperable with unique_ptr<T> wherever that makes sense. However, there are some differences:

1. It always uses a default deleter.

2. There is no array version.

3. To allow for future copy-on-write optimizations, there is a distinction between writable and const access, the get() method is modified to return only a const pointer, with get_mutable() added to return a writable pointer. Furthermore, derefencing (operator*()) a mutable pointer will give a mutable reference (in so far as T is not declared const), and dereferencing a const pointer will give a const reference.

This class is entirely inline and has no computational or space overhead except when copying is required; it contains just a single pointer and does no reference counting.

Usage

In the simplest use case, the instantiation type will match the type of object it references, e.g.:

copyable_unique_ptr<Foo> ptr = make_unique<Foo>(...);

In this case, as long Foo is deemed compatible, the behavior will be as expected, i.e., when ptr copies, it will contain a reference to a new instance of Foo.

copyable_unique_ptr can also be used with polymorphic classes a copyable_unique_ptr, instantiated on a base class, references an instance of a derived class. When copying the object, we would want the copy to likewise contain an instance of the derived class. For example:

copyable_unique_ptr<Base> cu_ptr = make_unique<Derived>();
copyable_unique_ptr<Base> other_cu_ptr = cu_ptr;           // Triggers a copy.
is_dynamic_castable<Derived>(other_cu_ptr.get());          // Should be true.

This works for well-designed polymorphic classes.

• The Base class’s Clone() implementation does not invoke the Derived class’s implementation of a suitable virtual method.

Warning

Ill-formed polymorphic classes can lead to fatal type slicing of the referenced object, such that the new copy contains an instance of Base instead of Derived. Some mistakes that would lead to this degenerate behavior:

Template Parameters

• T: The type of the contained object, which must be copyable as defined above. May be an abstract or concrete type.

Constructors

copyable_unique_ptr() noexcept

Default constructor stores a nullptr. No heap allocation is performed. The empty() method will return true when called on a default-constructed copyable_unique_ptr.

copyable_unique_ptr(T *raw) noexcept

Given a raw pointer to a writable heap-allocated object, take over ownership of that object. No copying occurs.

copyable_unique_ptr(const T &value)

Constructs a unique instance of T as a copy of the provided model value.

copyable_unique_ptr(const copyable_unique_ptr &cu_ptr)

Copy constructor is deep; the new copyable_unique_ptr object contains a new copy of the object in the source, created via the source object’s copy constructor or Clone() method. If the source container is empty this one will be empty also.

template<typename U>
copyable_unique_ptr(const std::unique_ptr<U> &u_ptr)

Copy constructor from a standard unique_ptr of compatible type. The copy is deep; the new copyable_unique_ptr object contains a new copy of the object in the source, created via the source object’s copy constructor or Clone() method. If the source container is empty this one will be empty also.

copyable_unique_ptr(copyable_unique_ptr &&cu_ptr) noexcept

Move constructor is very fast and leaves the source empty. Ownership is transferred from the source to the new copyable_unique_ptr. If the source was empty this one will be empty also. No heap activity occurs.

copyable_unique_ptr(std::unique_ptr<T> &&u_ptr) noexcept

Move constructor from a standard unique_ptr. The move is very fast and leaves the source empty. Ownership is transferred from the source to the new copyable_unique_ptr. If the source was empty this one will be empty also. No heap activity occurs.

template<typename U>
copyable_unique_ptr(std::unique_ptr<U> &&u_ptr) noexcept

Move construction from a compatible standard unique_ptr. Type U* must be implicitly convertible to type T*. Ownership is transferred from the source to the new copyable_unique_ptr. If the source was empty this one will be empty also. No heap activity occurs.

Assignment

copyable_unique_ptr &operator=(T *raw) noexcept

This form of assignment replaces the currently-held object by the given source object and takes over ownership of the source object. The currently-held object (if any) is deleted.

copyable_unique_ptr &operator=(const T &ref)

This form of assignment replaces the currently-held object by a heap-allocated copy of the source object, created using its copy constructor or Clone() method. The currently-held object (if any) is deleted.

copyable_unique_ptr &operator=(const copyable_unique_ptr &cu_ptr)

Copy assignment from copyable_unique_ptr replaces the currently-held object by a copy of the object held in the source container, created using the source object’s copy constructor or Clone() method. The currently-held object (if any) is deleted. If the source container is empty this one will be empty also after the assignment. Nothing happens if the source and destination are the same container.

template<typename U>
copyable_unique_ptr &operator=(const copyable_unique_ptr<U> &cu_ptr)

Copy assignment from a compatible copyable_unique_ptr replaces the currently-held object by a copy of the object held in the source container, created using the source object’s copy constructor or Clone() method. The currently-held object (if any) is deleted. If the source container is empty this one will be empty also after the assignment. Nothing happens if the source and destination are the same container.

copyable_unique_ptr &operator=(const std::unique_ptr<T> &src)

Copy assignment from a standard unique_ptr replaces the currently-held object by a copy of the object held in the source container, created using the source object’s copy constructor or Clone() method. The currently-held object (if any) is deleted. If the source container is empty this one will be empty also after the assignment. Nothing happens if the source and destination are the same container.

template<typename U>
copyable_unique_ptr &operator=(const std::unique_ptr<U> &u_ptr)

Copy assignment from a compatible standard unique_ptr replaces the currently-held object by a copy of the object held in the source container, created using the source object’s copy constructor or Clone() method. The currently-held object (if any) is deleted. If the source container is empty this one will be empty also after the assignment. Nothing happens if the source and destination are the same container.

copyable_unique_ptr &operator=(copyable_unique_ptr &&cu_ptr) noexcept

Move assignment replaces the currently-held object by the source object, leaving the source empty. The currently-held object (if any) is deleted. The instance is not copied. Nothing happens if the source and destination are the same containers.

template<typename U>
copyable_unique_ptr &operator=(copyable_unique_ptr<U> &&cu_ptr) noexcept

Move assignment replaces the currently-held object by the compatible source object, leaving the source empty. The currently-held object (if any) is deleted. The instance is not copied. Nothing happens if the source and destination are the same containers.

copyable_unique_ptr &operator=(std::unique_ptr<T> &&u_ptr) noexcept

Move assignment replaces the currently-held object by the source object, leaving the source empty. The currently-held object (if any) is deleted. The instance is not copied. Nothing happens if the source and destination are the same containers.

template<typename U>
copyable_unique_ptr &operator=(std::unique_ptr<U> &&u_ptr) noexcept

Move assignment replaces the currently-held object by the compatible source object, leaving the source empty. The currently-held object (if any) is deleted. The instance is not copied. Nothing happens if the source and destination are the same containers.

Observers

bool empty() const noexcept

Return true if this container is empty, which is the state the container is in immediately after default construction and various other operations.

const T *get() const noexcept

Return a const pointer to the contained object if any, or nullptr. Note that this is different than get() for the standard smart pointers like std::unique_ptr which return a writable pointer. Use get_mutable() here for that purpose.

T *get_mutable() noexcept

Return a writable pointer to the contained object if any, or nullptr. Note that you need write access to this container in order to get write access to the object it contains.

Warning

If copyable_unique_ptr is instantiated on a const template parameter (e.g., copyable_unique_ptr<const Foo>), then get_mutable() returns a const pointer.

const T &operator*() const

Return a const reference to the contained object. Note that this is different from std::unique_ptr::operator*() which would return a non-const reference (if T is non-const), even if the container itself is const. For a const copyable_unique_ptr will always return a const reference to its contained value.

Warning

Currently copyable_unique_ptr is a std::unique_ptr. As such, a const copyable_unique_ptr<Foo> can be upcast to a const unique_ptr<Foo> and the parent’s behavior will provide a mutable reference. This is strongly discouraged and will break as the implementation of this class changes to shore up this gap in the const correctness protection.

Pre

this != nullptr reports true.

T &operator*()

Return a writable reference to the contained object (if T is itself not const). Note that you need write access to this container in order to get write access to the object it contains.

We strongly recommend, that, if dereferencing a copyable_unique_ptr without the intention of mutating the underlying value, prefer to dereference a const copyable_unique_ptr (or use *my_ptr.get()) and not a mutable copyable_unique_ptr. As “copy-on-write” behavior is introduced in the future, this recommended practice will prevent unwanted copies of the underlying value.

If copyable_unique_ptr is instantiated on a const template parameter (e.g., copyable_unique_ptr<const Foo>), then operator*() must return a const reference.

Pre

this != nullptr reports true.

Related

template<class charT, class traits, class T>
std::basic_ostream<charT, traits> &operator<<(std::basic_ostream<charT, traits> &os, const copyable_unique_ptr<T> &cu_ptr)

Output the system-dependent representation of the pointer contained in a copyable_unique_ptr object. This is equivalent to os << p.get();.

struct CpuUsage

#include <CpuUsage.hpp>

CpuUsage structure

Average usage in percent and time span of the average (since last query)

Public Members

float average

Average CPU usage, expressed with a normalized value (0-1)

int32_t msTime

Time span in which the average was calculated in milliseconds

class DataInputQueue

#include <DataQueue.hpp>

Access to send messages through XLink stream

Public Functions

bool isClosed() const

Check whether queue is closed

void close()

Closes the queue and the underlying thread

void setMaxDataSize(std::size_t maxSize)

Sets maximum message size. If message is larger than specified, then an exception is issued.

Parameters

• maxSize: Maximum message size to add to queue

std::size_t getMaxDataSize()

Gets maximum queue size.

Return

Maximum message size

void setBlocking(bool blocking)

Sets queue behavior when full (maxSize)

Parameters

• blocking: Specifies if block or overwrite the oldest message in the queue

bool getBlocking() const

Gets current queue behavior when full (maxSize)

Return

True if blocking, false otherwise

void setMaxSize(unsigned int maxSize)

Sets queue maximum size

Parameters

• maxSize: Specifies maximum number of messages in the queue

unsigned int getMaxSize() const

Gets queue maximum size

Return

Maximum queue size

std::string getName() const

Gets queues name

Return

Queue name

void send(const std::shared_ptr<RawBuffer> &rawMsg)

Adds a raw message to the queue, which will be picked up and sent to the device. Can either block if ‘blocking’ behavior is true or overwrite oldest

Parameters

• rawMsg: Message to add to the queue

void send(const std::shared_ptr<ADatatype> &msg)

Adds a message to the queue, which will be picked up and sent to the device. Can either block if ‘blocking’ behavior is true or overwrite oldest

Parameters

• msg: Message to add to the queue

void send(const ADatatype &msg)

Adds a message to the queue, which will be picked up and sent to the device. Can either block if ‘blocking’ behavior is true or overwrite oldest

Parameters

• msg: Message to add to the queue

bool send(const std::shared_ptr<RawBuffer> &rawMsg, std::chrono::milliseconds timeout)

Adds message to the queue, which will be picked up and sent to the device. Can either block until timeout if ‘blocking’ behavior is true or overwrite oldest

Parameters

• rawMsg: Message to add to the queue

• timeout: Maximum duration to block in milliseconds

bool send(const std::shared_ptr<ADatatype> &msg, std::chrono::milliseconds timeout)

Adds message to the queue, which will be picked up and sent to the device. Can either block until timeout if ‘blocking’ behavior is true or overwrite oldest

Parameters

• msg: Message to add to the queue

• timeout: Maximum duration to block in milliseconds

bool send(const ADatatype &msg, std::chrono::milliseconds timeout)

Adds message to the queue, which will be picked up and sent to the device. Can either block until timeout if ‘blocking’ behavior is true or overwrite oldest

Parameters

• msg: Message to add to the queue

• timeout: Maximum duration to block in milliseconds

class DataOutputQueue

#include <DataQueue.hpp>

Access to receive messages coming from XLink stream

Public Types

using CallbackId = int

Alias for callback id.

Public Functions

bool isClosed() const

Check whether queue is closed

void close()

Closes the queue and the underlying thread

void setBlocking(bool blocking)

Sets queue behavior when full (maxSize)

Parameters

• blocking: Specifies if block or overwrite the oldest message in the queue

bool getBlocking() const

Gets current queue behavior when full (maxSize)

Return

True if blocking, false otherwise

void setMaxSize(unsigned int maxSize)

Sets queue maximum size

Parameters

• maxSize: Specifies maximum number of messages in the queue

unsigned int getMaxSize() const

Gets queue maximum size

Return

Maximum queue size

std::string getName() const

Gets queues name

Return

Queue name

CallbackId addCallback(std::function<void(std::string, std::shared_ptr<ADatatype>)>)

Adds a callback on message received

Return

Callback id

Parameters

• callback: Callback function with queue name and message pointer

CallbackId addCallback(std::function<void(std::shared_ptr<ADatatype>)>)

Adds a callback on message received

Return

Callback id

Parameters

• callback: Callback function with message pointer

CallbackId addCallback(std::function<void()> callback)

Adds a callback on message received

Return

Callback id

Parameters

• callback: Callback function without any parameters

bool removeCallback(CallbackId callbackId)

Removes a callback

Return

True if callback was removed, false otherwise

Parameters

• callbackId: Id of callback to be removed

template<class T>
bool has()

Check whether front of the queue has message of type T

Return

True if queue isn’t empty and the first element is of type T, false otherwise

bool has()

Check whether front of the queue has a message (isn’t empty)

Return

True if queue isn’t empty, false otherwise

template<class T>
std::shared_ptr<T> tryGet()

Try to retrieve message T from queue. If message isn’t of type T it returns nullptr

Return

Message of type T or nullptr if no message available

std::shared_ptr<ADatatype> tryGet()

Try to retrieve message from queue. If no message available, return immediately with nullptr

Return

Message or nullptr if no message available

template<class T>
std::shared_ptr<T> get()

Block until a message is available.

Return

Message of type T or nullptr if no message available

std::shared_ptr<ADatatype> get()

Block until a message is available.

Return

Message or nullptr if no message available

template<class T>
std::shared_ptr<T> front()

Gets first message in the queue.

Return

Message of type T or nullptr if no message available

std::shared_ptr<ADatatype> front()

Gets first message in the queue.

Return

Message or nullptr if no message available

template<class T, typename Rep, typename Period>
std::shared_ptr<T> get(std::chrono::duration<Rep, Period> timeout, bool &hasTimedout)

Block until a message is available with a timeout.

Return

Message of type T otherwise nullptr if message isn’t type T or timeout occurred

Parameters

• timeout: Duration for which the function should block

• [out] hasTimedout: Outputs true if timeout occurred, false otherwise

template<typename Rep, typename Period>
std::shared_ptr<ADatatype> get(std::chrono::duration<Rep, Period> timeout, bool &hasTimedout)

Block until a message is available with a timeout.

Return

Message of type T otherwise nullptr if message isn’t type T or timeout occurred

Parameters

• timeout: Duration for which the function should block

• [out] hasTimedout: Outputs true if timeout occurred, false otherwise

template<class T>
std::vector<std::shared_ptr<T>> tryGetAll()

Try to retrieve all messages in the queue.

Return

Vector of messages which can either be of type T or nullptr

std::vector<std::shared_ptr<ADatatype>> tryGetAll()

Try to retrieve all messages in the queue.

Return

Vector of messages

template<class T>
std::vector<std::shared_ptr<T>> getAll()

Block until at least one message in the queue. Then return all messages from the queue.

Return

Vector of messages which can either be of type T or nullptr

std::vector<std::shared_ptr<ADatatype>> getAll()

Block until at least one message in the queue. Then return all messages from the queue.

Return

Vector of messages

template<class T, typename Rep, typename Period>
std::vector<std::shared_ptr<T>> getAll(std::chrono::duration<Rep, Period> timeout, bool &hasTimedout)

Block for maximum timeout duration. Then return all messages from the queue.

Return

Vector of messages which can either be of type T or nullptr

Parameters

• timeout: Maximum duration to block

• [out] hasTimedout: Outputs true if timeout occurred, false otherwise

template<typename Rep, typename Period>
std::vector<std::shared_ptr<ADatatype>> getAll(std::chrono::duration<Rep, Period> timeout, bool &hasTimedout)

Block for maximum timeout duration. Then return all messages from the queue.

Return

Vector of messages

Parameters

• timeout: Maximum duration to block

• [out] hasTimedout: Outputs true if timeout occurred, false otherwise

struct DetectionNetworkProperties : public dai::PropertiesSerializable<NeuralNetworkProperties, DetectionNetworkProperties>

#include <DetectionNetworkProperties.hpp>

Specify properties for DetectionNetwork

Subclassed by dai::PropertiesSerializable< DetectionNetworkProperties, SpatialDetectionNetworkProperties >

struct DetectionParserOptions

#include <DetectionParserOptions.hpp>

DetectionParserOptions

Specifies how to parse output of detection networks

Public Members

DetectionNetworkType nnFamily

Generic Neural Network properties.

int classes

YOLO specific network properties.

struct DetectionParserProperties : public dai::PropertiesSerializable<Properties, DetectionParserProperties>

#include <DetectionParserProperties.hpp>

Specify properties for DetectionParser

Public Members

int numFramesPool = 8

Num frames in output pool.

std::unordered_map<std::string, TensorInfo> networkInputs

Network inputs.

DetectionParserOptions parser

Options for parser.

class Device : public dai::DeviceBase

#include <Device.hpp>

Represents the DepthAI device with the methods to interact with it. Implements the host-side queues to connect with XLinkIn and XLinkOut nodes

Public Functions

Device(const Pipeline &pipeline)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
Device(const Pipeline &pipeline, T usb2Mode)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• usb2Mode: (bool) Boot device using USB2 mode firmware

Device(const Pipeline &pipeline, UsbSpeed maxUsbSpeed)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• maxUsbSpeed: Maximum allowed USB speed

Device(const Pipeline &pipeline, const dai::Path &pathToCmd)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• pathToCmd: Path to custom device firmware

Device(const Pipeline &pipeline, const DeviceInfo &devInfo)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
Device(const Pipeline &pipeline, const DeviceInfo &devInfo, T usb2Mode)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• usb2Mode: (bool) Boot device using USB2 mode firmware

Device(const Pipeline &pipeline, const DeviceInfo &devInfo, UsbSpeed maxUsbSpeed)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• maxUsbSpeed: Maximum allowed USB speed

Device(const Pipeline &pipeline, const DeviceInfo &devInfo, const dai::Path &pathToCmd)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• pathToCmd: Path to custom device firmware

Device()

Connects to any available device with a DEFAULT_SEARCH_TIME timeout. Uses OpenVINO version Pipeline::DEFAULT_OPENVINO_VERSION

~Device() override

dtor to close the device

std::shared_ptr<DataOutputQueue> getOutputQueue(const std::string &name)

Gets an output queue corresponding to stream name. If it doesn’t exist it throws

Return

Smart pointer to DataOutputQueue

Parameters

• name: Queue/stream name, created by XLinkOut node

std::shared_ptr<DataOutputQueue> getOutputQueue(const std::string &name, unsigned int maxSize, bool blocking = true)

Gets a queue corresponding to stream name, if it exists, otherwise it throws. Also sets queue options

Return

Smart pointer to DataOutputQueue

Parameters

• name: Queue/stream name, set in XLinkOut node

• maxSize: Maximum number of messages in queue

• blocking: Queue behavior once full. True specifies blocking and false overwriting of oldest messages. Default: true

std::vector<std::string> getOutputQueueNames() const

Get all available output queue names

Return

Vector of output queue names

std::shared_ptr<DataInputQueue> getInputQueue(const std::string &name)

Gets an input queue corresponding to stream name. If it doesn’t exist it throws

Return

Smart pointer to DataInputQueue

Parameters

• name: Queue/stream name, set in XLinkIn node

std::shared_ptr<DataInputQueue> getInputQueue(const std::string &name, unsigned int maxSize, bool blocking = true)

Gets an input queue corresponding to stream name. If it doesn’t exist it throws. Also sets queue options

Return

Smart pointer to DataInputQueue

Parameters

• name: Queue/stream name, set in XLinkOut node

• maxSize: Maximum number of messages in queue

• blocking: Queue behavior once full. True: blocking, false: overwriting of oldest messages. Default: true

std::vector<std::string> getInputQueueNames() const

Get all available input queue names

Return

Vector of input queue names

std::vector<std::string> getQueueEvents(const std::vector<std::string> &queueNames, std::size_t maxNumEvents = std::numeric_limits<std::size_t>::max(), std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until any of specified queues has received a message

Return

Names of queues which received messages first

Parameters

• queueNames: Names of queues for which to block

• maxNumEvents: Maximum number of events to remove from queue - Default is unlimited

• timeout: Timeout after which return regardless. If negative then wait is indefinite - Default is -1

std::vector<std::string> getQueueEvents(std::string queueName, std::size_t maxNumEvents = std::numeric_limits<std::size_t>::max(), std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until specified queue has received a message

Return

Names of queues which received messages first

Parameters

• queueName: Name of queues for which to wait for

• maxNumEvents: Maximum number of events to remove from queue. Default is unlimited

• timeout: Timeout after which return regardless. If negative then wait is indefinite. Default is -1

std::vector<std::string> getQueueEvents(std::size_t maxNumEvents = std::numeric_limits<std::size_t>::max(), std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until any queue has received a message

Return

Names of queues which received messages first

Parameters

• maxNumEvents: Maximum number of events to remove from queue. Default is unlimited

• timeout: Timeout after which return regardless. If negative then wait is indefinite. Default is -1

std::string getQueueEvent(const std::vector<std::string> &queueNames, std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until any of specified queues has received a message

Return

Queue name which received a message first

Parameters

• queueNames: Names of queues for which to wait for

• timeout: Timeout after which return regardless. If negative then wait is indefinite. Default is -1

std::string getQueueEvent(std::string queueName, std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until specified queue has received a message

Return

Queue name which received a message

Parameters

• queueNames: Name of queues for which to wait for

• timeout: Timeout after which return regardless. If negative then wait is indefinite. Default is -1

std::string getQueueEvent(std::chrono::microseconds timeout = std::chrono::microseconds(-1))

Gets or waits until any queue has received a message

Return

Queue name which received a message

Parameters

• timeout: Timeout after which return regardless. If negative then wait is indefinite. Default is -1

Public Static Attributes

constexpr std::size_t EVENT_QUEUE_MAXIMUM_SIZE = {2048}

Maximum number of elements in event queue.

class DeviceBase

#include <DeviceBase.hpp>

The core of depthai device for RAII, connects to device and maintains watchdog, timesync, …

Subclassed by dai::Device

Public Functions

DeviceBase(const Pipeline &pipeline)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
DeviceBase(const Pipeline &pipeline, T usb2Mode)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• usb2Mode: Boot device using USB2 mode firmware

DeviceBase(const Pipeline &pipeline, UsbSpeed maxUsbSpeed)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• maxUsbSpeed: Maximum allowed USB speed

DeviceBase(const Pipeline &pipeline, const dai::Path &pathToCmd)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• pipeline: Pipeline to be executed on the device

• pathToCmd: Path to custom device firmware

DeviceBase(const Pipeline &pipeline, const DeviceInfo &devInfo)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
DeviceBase(const Pipeline &pipeline, const DeviceInfo &devInfo, T usb2Mode)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• usb2Mode: Boot device using USB2 mode firmware

DeviceBase(const Pipeline &pipeline, const DeviceInfo &devInfo, UsbSpeed maxUsbSpeed)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• maxUsbSpeed: Maximum allowed USB speed

DeviceBase(const Pipeline &pipeline, const DeviceInfo &devInfo, const dai::Path &pathToCmd)

Connects to device specified by devInfo.

Parameters

• pipeline: Pipeline to be executed on the device

• devInfo: DeviceInfo which specifies which device to connect to

• pathToCmd: Path to custom device firmware

DeviceBase()

Connects to any available device with a DEFAULT_SEARCH_TIME timeout. Uses OpenVINO version Pipeline::DEFAULT_OPENVINO_VERSION

DeviceBase(OpenVINO::Version version)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• version: OpenVINO version which the device will be booted with.

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
DeviceBase(OpenVINO::Version version, T usb2Mode)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• version: OpenVINO version which the device will be booted with

• usb2Mode: Boot device using USB2 mode firmware

DeviceBase(OpenVINO::Version version, UsbSpeed maxUsbSpeed)

Connects to device specified by devInfo.

Parameters

• version: OpenVINO version which the device will be booted with

• maxUsbSpeed: Maximum allowed USB speed

DeviceBase(OpenVINO::Version version, const dai::Path &pathToCmd)

Connects to any available device with a DEFAULT_SEARCH_TIME timeout.

Parameters

• version: OpenVINO version which the device will be booted with

• pathToCmd: Path to custom device firmware

DeviceBase(OpenVINO::Version version, const DeviceInfo &devInfo)

Connects to device specified by devInfo.

Parameters

• version: OpenVINO version which the device will be booted with

• devInfo: DeviceInfo which specifies which device to connect to

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
DeviceBase(OpenVINO::Version version, const DeviceInfo &devInfo, T usb2Mode)

Connects to device specified by devInfo.

Parameters

• version: OpenVINO version which the device will be booted with

• devInfo: DeviceInfo which specifies which device to connect to

• usb2Mode: Boot device using USB2 mode firmware

DeviceBase(OpenVINO::Version version, const DeviceInfo &devInfo, UsbSpeed maxUsbSpeed)

Connects to device specified by devInfo.

Parameters

• version: OpenVINO version which the device will be booted with

• devInfo: DeviceInfo which specifies which device to connect to

• maxUsbSpeed: Maximum allowed USB speed

DeviceBase(OpenVINO::Version version, const DeviceInfo &devInfo, const dai::Path &pathToCmd)

Connects to device specified by devInfo.

Parameters

• version: OpenVINO version which the device will be booted with

• devInfo: DeviceInfo which specifies which device to connect to

• pathToCmd: Path to custom device firmware

DeviceBase(Config config)

Connects to any available device with custom config.

Parameters

• config: Device custom configuration to boot with

DeviceBase(Config config, const DeviceInfo &devInfo)

Connects to device ‘devInfo’ with custom config.

Parameters

• devInfo: DeviceInfo which specifies which device to connect to

• config: Device custom configuration to boot with

~DeviceBase()

Device destructor

Note

In the destructor of the derived class, remember to call close()

bool isPipelineRunning()

Checks if devices pipeline is already running

Return

True if running, false otherwise

bool startPipeline()

Starts the execution of the devices pipeline

Return

True if pipeline started, false otherwise

bool startPipeline(const Pipeline &pipeline)

Starts the execution of a given pipeline

Return

True if pipeline started, false otherwise

Parameters

• pipeline: OpenVINO version of the pipeline must match the one which the device was booted with.

void setLogLevel(LogLevel level)

Sets the devices logging severity level. This level affects which logs are transferred from device to host.

Parameters

• level: Logging severity

LogLevel getLogLevel()

Gets current logging severity level of the device.

Return

Logging severity level

void setXLinkChunkSize(int sizeBytes)

Sets the chunk size for splitting device-sent XLink packets. A larger value could increase performance, and 0 disables chunking. A negative value is ignored. Device defaults are configured per protocol, currently 64*1024 for both USB and Ethernet.

Parameters

• sizeBytes: XLink chunk size in bytes

int getXLinkChunkSize()

Gets current XLink chunk size.

Return

XLink chunk size in bytes

DeviceInfo getDeviceInfo() const

Get the Device Info object o the device which is currently running

Return

DeviceInfo of the current device in execution

std::string getMxId()

Get MxId of device

Return

MxId of connected device

void setLogOutputLevel(LogLevel level)

Sets logging level which decides printing level to standard output. If lower than setLogLevel, no messages will be printed

Parameters

• level: Standard output printing severity

LogLevel getLogOutputLevel()

Gets logging level which decides printing level to standard output.

Return

Standard output printing severity

bool setIrLaserDotProjectorBrightness(float mA, int mask = -1)

Sets the brightness of the IR Laser Dot Projector. Limits: up to 765mA at 30% duty cycle, up to 1200mA at 6% duty cycle. The duty cycle is controlled by left camera STROBE, aligned to start of exposure. The emitter is turned off by default

Return

True on success, false if not found or other failure

Parameters

• mA: Current in mA that will determine brightness, 0 or negative to turn off

• mask: Optional mask to modify only Left (0x1) or Right (0x2) sides on OAK-D-Pro-W-DEV

bool setIrFloodLightBrightness(float mA, int mask = -1)

Sets the brightness of the IR Flood Light. Limits: up to 1500mA at 30% duty cycle. The duty cycle is controlled by the left camera STROBE, aligned to start of exposure. If the dot projector is also enabled, its lower duty cycle limits take precedence. The emitter is turned off by default

Return

True on success, false if not found or other failure

Parameters

• mA: Current in mA that will determine brightness, 0 or negative to turn off

• mask: Optional mask to modify only Left (0x1) or Right (0x2) sides on OAK-D-Pro-W-DEV

std::vector<std::tuple<std::string, int, int>> getIrDrivers()

Retrieves detected IR laser/LED drivers.

Return

Vector of tuples containing: driver name, I2C bus, I2C address. For OAK-D-Pro it should be [{"LM3644", 2, 0x63}]

int addLogCallback(std::function<void(LogMessage)> callback)

Add a callback for device logging. The callback will be called from a separate thread with the LogMessage being passed.

Return

Id which can be used to later remove the callback

Parameters

• callback: Callback to call whenever a log message arrives

bool removeLogCallback(int callbackId)

Removes a callback

Return

True if callback was removed, false otherwise

Parameters

• callbackId: Id of callback to be removed

void setSystemInformationLoggingRate(float rateHz)

Sets rate of system information logging (“info” severity). Default 1Hz If parameter is less or equal to zero, then system information logging will be disabled

Parameters

• rateHz: Logging rate in Hz

float getSystemInformationLoggingRate()

Gets current rate of system information logging (“info” severity) in Hz.

Return

Logging rate in Hz

std::vector<CameraBoardSocket> getConnectedCameras()

Get cameras that are connected to the device

Return

Vector of connected cameras

std::unordered_map<CameraBoardSocket, std::string> getCameraSensorNames()

Get sensor names for cameras that are connected to the device

Return

Map/dictionary with camera sensor names, indexed by socket

MemoryInfo getDdrMemoryUsage()

Retrieves current DDR memory information from device

Return

Used, remaining and total ddr memory

MemoryInfo getCmxMemoryUsage()

Retrieves current CMX memory information from device

Return

Used, remaining and total cmx memory

MemoryInfo getLeonCssHeapUsage()

Retrieves current CSS Leon CPU heap information from device

Return

Used, remaining and total heap memory

MemoryInfo getLeonMssHeapUsage()

Retrieves current MSS Leon CPU heap information from device

Return

Used, remaining and total heap memory

ChipTemperature getChipTemperature()

Retrieves current chip temperature as measured by device

Return

Temperature of various onboard sensors

CpuUsage getLeonCssCpuUsage()

Retrieves average CSS Leon CPU usage

Return

Average CPU usage and sampling duration

CpuUsage getLeonMssCpuUsage()

Retrieves average MSS Leon CPU usage

Return

Average CPU usage and sampling duration

bool isEepromAvailable()

Check if EEPROM is available

Return

True if EEPROM is present on board, false otherwise

bool flashCalibration(CalibrationHandler calibrationDataHandler)

Stores the Calibration and Device information to the Device EEPROM

Return

true on successful flash, false on failure

Parameters

• calibrationObj: CalibrationHandler object which is loaded with calibration information.

void flashCalibration2(CalibrationHandler calibrationDataHandler)

Stores the Calibration and Device information to the Device EEPROM

Exceptions

• std::runtime_exception: if failed to flash the calibration

Parameters

• calibrationObj: CalibrationHandler object which is loaded with calibration information.

CalibrationHandler readCalibration()

Fetches the EEPROM data from the device and loads it into CalibrationHandler object If no calibration is flashed, it returns default

Return

The CalibrationHandler object containing the calibration currently flashed on device EEPROM

CalibrationHandler readCalibration2()

Fetches the EEPROM data from the device and loads it into CalibrationHandler object

Return

The CalibrationHandler object containing the calibration currently flashed on device EEPROM

Exceptions

• std::runtime_exception: if no calibration is flashed

CalibrationHandler readCalibrationOrDefault()

Fetches the EEPROM data from the device and loads it into CalibrationHandler object If no calibration is flashed, it returns default

Return

The CalibrationHandler object containing the calibration currently flashed on device EEPROM

void factoryResetCalibration()

Factory reset EEPROM data if factory backup is available.

Exceptions

• std::runtime_exception: If factory reset was unsuccessful

void flashFactoryCalibration(CalibrationHandler calibrationHandler)

Stores the Calibration and Device information to the Device EEPROM in Factory area To perform this action, correct env variable must be set

Return

True on successful flash, false on failure

Exceptions

• std::runtime_exception: if failed to flash the calibration

CalibrationHandler readFactoryCalibration()

Fetches the EEPROM data from Factory area and loads it into CalibrationHandler object

Return

The CalibrationHandler object containing the calibration currently flashed on device EEPROM in Factory Area

Exceptions

• std::runtime_exception: if no calibration is flashed

CalibrationHandler readFactoryCalibrationOrDefault()

Fetches the EEPROM data from Factory area and loads it into CalibrationHandler object If no calibration is flashed, it returns default

Return

The CalibrationHandler object containing the calibration currently flashed on device EEPROM in Factory Area

UsbSpeed getUsbSpeed()

Retrieves USB connection speed

Return

USB connection speed of connected device if applicable. Unknown otherwise.

void close()

Explicitly closes connection to device.

Note

This function does not need to be explicitly called as destructor closes the device automatically

bool isClosed() const

Is the device already closed (or disconnected)

std::shared_ptr<XLinkConnection> getConnection()

Returns underlying XLinkConnection

std::shared_ptr<const XLinkConnection> getConnection() const

Returns underlying XLinkConnection

Public Static Functions

std::chrono::milliseconds getDefaultSearchTime()

Get the Default Search Time for finding devices.

Return

Default search time in milliseconds

std::tuple<bool, DeviceInfo> getAnyAvailableDevice(std::chrono::milliseconds timeout)

Waits for any available device with a timeout

Return

Tuple of bool and DeviceInfo. Bool specifies if device was found. DeviceInfo specifies the found device

Parameters

• timeout: duration of time to wait for the any device

std::tuple<bool, DeviceInfo> getAnyAvailableDevice()

Gets any available device

Return

Tuple of bool and DeviceInfo. Bool specifies if device was found. DeviceInfo specifies the found device

std::tuple<bool, DeviceInfo> getAnyAvailableDevice(std::chrono::milliseconds timeout, std::function<void()> cb)

Waits for any available device with a timeout

Return

Tuple of bool and DeviceInfo. Bool specifies if device was found. DeviceInfo specifies the found device

Parameters

• timeout: duration of time to wait for the any device

• cb: callback function called between pooling intervals

std::tuple<bool, DeviceInfo> getFirstAvailableDevice(bool skipInvalidDevice = true)

Gets first available device. Device can be either in XLINK_UNBOOTED or XLINK_BOOTLOADER state

Return

Tuple of bool and DeviceInfo. Bool specifies if device was found. DeviceInfo specifies the found device

std::tuple<bool, DeviceInfo> getDeviceByMxId(std::string mxId)

Finds a device by MX ID. Example: 14442C10D13EABCE00

Return

Tuple of bool and DeviceInfo. Bool specifies if device was found. DeviceInfo specifies the found device

Parameters

• mxId: MyraidX ID which uniquely specifies a device

std::vector<DeviceInfo> getAllAvailableDevices()

Returns all connected devices

Return

Vector of connected devices

std::vector<std::uint8_t> getEmbeddedDeviceBinary(bool usb2Mode, OpenVINO::Version version = OpenVINO::DEFAULT_VERSION)

Gets device firmware binary for a specific OpenVINO version

Return

Firmware binary

Parameters

• usb2Mode: USB2 mode firmware

• version: Version of OpenVINO which firmware will support

std::vector<std::uint8_t> getEmbeddedDeviceBinary(Config config)

Gets device firmware binary for a specific configuration

Return

Firmware binary

Parameters

• config: FW with applied configuration

Public Static Attributes

constexpr std::chrono::seconds DEFAULT_SEARCH_TIME = {3}

Default search time for constructors which discover devices.

constexpr float DEFAULT_SYSTEM_INFORMATION_LOGGING_RATE_HZ = {1.0f}

Default rate at which system information is logged.

constexpr UsbSpeed DEFAULT_USB_SPEED = {UsbSpeed::SUPER}

Default UsbSpeed for device connection.

struct Config

#include <DeviceBase.hpp>

Device specific configuration

class DeviceBootloader

#include <DeviceBootloader.hpp>

Represents the DepthAI bootloader with the methods to interact with it.

Public Functions

DeviceBootloader(const DeviceInfo &devInfo)

Connects to or boots device in bootloader mode depending on devInfo state; flashing not allowed

Parameters

• devInfo: DeviceInfo of which to boot or connect to

template<typename T, std::enable_if_t<std::is_same<T, bool>::value, bool> = true>
DeviceBootloader(const DeviceInfo &devInfo, T allowFlashingBootloader)

Connects to or boots device in bootloader mode depending on devInfo state.

Parameters

• devInfo: DeviceInfo of which to boot or connect to

• allowFlashingBootloader: (bool) Set to true to allow flashing the devices bootloader

DeviceBootloader(const DeviceInfo &devInfo, Type type, bool allowFlashingBootloader = false)

Connects to device in bootloader of specified type. Throws if it wasn’t possible. This constructor will automatically boot into specified bootloader type if not already running

Parameters

• devInfo: DeviceInfo of which to boot or connect to

• type: Type of bootloader to boot/connect to.

• allowFlashingBootloader: Set to true to allow flashing the devices bootloader. Defaults to false

DeviceBootloader(const DeviceInfo &devInfo, const dai::Path &pathToBootloader, bool allowFlashingBootloader = false)

Connects to or boots device in bootloader mode depending on devInfo state with a custom bootloader firmware.

Parameters

• devInfo: DeviceInfo of which to boot or connect to

• pathToBootloader: Custom bootloader firmware to boot

• allowFlashingBootloader: Set to true to allow flashing the devices bootloader. Defaults to false

~DeviceBootloader()

Destroy the Device Bootloader object.

std::tuple<bool, std::string> flash(std::function<void(float)> progressCallback, const Pipeline &pipeline, bool compress = false, )

Flashes a given pipeline to the device.

Parameters

• progressCallback: Callback that sends back a value between 0..1 which signifies current flashing progress

• pipeline: Pipeline to flash to the board

std::tuple<bool, std::string> flash(const Pipeline &pipeline, bool compress = false)

Flashes a given pipeline to the device.

Parameters

• pipeline: Pipeline to flash to the board

std::tuple<bool, std::string> flashDepthaiApplicationPackage(std::function<void(float)> progressCallback, std::vector<uint8_t> package, )

Flashes a specific depthai application package that was generated using createDepthaiApplicationPackage or saveDepthaiApplicationPackage

Parameters

• progressCallback: Callback that sends back a value between 0..1 which signifies current flashing progress

• package: Depthai application package to flash to the board

std::tuple<bool, std::string> flashDepthaiApplicationPackage(std::vector<uint8_t> package)

Flashes a specific depthai application package that was generated using createDepthaiApplicationPackage or saveDepthaiApplicationPackage

Parameters

• package: Depthai application package to flash to the board

std::tuple<bool, std::string> flashBootloader(std::function<void(float)> progressCallback, const dai::Path &path = {}, )

Flashes bootloader to the current board

Parameters

• progressCallback: Callback that sends back a value between 0..1 which signifies current flashing progress

• path: Optional parameter to custom bootloader to flash

std::tuple<bool, std::string> flashBootloader(Memory memory, Type type, std::function<void(float)> progressCallback, const dai::Path &path = {}, )

Flash selected bootloader to the current board

Parameters

• memory: Memory to flash

• type: Bootloader type to flash

• progressCallback: Callback that sends back a value between 0..1 which signifies current flashing progress

• path: Optional parameter to custom bootloader to flash

nlohmann::json readConfigData(Memory memory = Memory::AUTO, Type type = Type::AUTO)

Flash arbitrary data at custom offset in specified memory

Return

Unstructured configuration data

Parameters

• memory: Memory to flash

• offset: Offset at which to flash the given data in bytes

• progressCallback: Callback that sends back a value between 0..1 which signifies current flashing progress

• data: Data to flash Reads configuration data from bootloader

Parameters

• memory: Optional - from which memory to read configuration data

• type: Optional - from which type of bootloader to read configuration data

std::tuple<bool, std::string> flashConfigData(nlohmann::json configData, Memory memory = Memory::AUTO, Type type = Type::AUTO)

Flashes configuration data to bootloader

Parameters

• configData: Unstructured configuration data

• memory: Optional - to which memory flash configuration

• type: Optional - for which type of bootloader to flash configuration

std::tuple<bool, std::string> flashConfigFile(const dai::Path &configPath, Memory memory = Memory::AUTO, Type type = Type::AUTO)

Flashes configuration data to bootloader

Parameters

• configPath: Unstructured configuration data

• memory: Optional - to which memory flash configuration

• type: Optional - for which type of bootloader to flash configuration

std::tuple<bool, std::string> flashConfigClear(Memory memory = Memory::AUTO, Type type = Type::AUTO)

Clears configuration data

Parameters

• memory: Optional - on which memory to clear configuration data

• type: Optional - for which type of bootloader to clear configuration data

Config readConfig(Memory memory = Memory::AUTO, Type type = Type::AUTO)

Reads configuration from bootloader

Return

Configuration structure

Parameters

• memory: Optional - from which memory to read configuration

• type: Optional - from which type of bootloader to read configuration

std::tuple<bool, std::string> flashConfig(const Config &config, Memory memory = Memory::AUTO, Type type = Type::AUTO)

Flashes configuration to bootloader

Parameters

• configData: Configuration structure

• memory: Optional - to which memory flash configuration

• type: Optional - for which type of bootloader to flash configuration

void bootMemory(const std::vector<uint8_t> &fw)

Boots a custom FW in memory

Parameters

• fw:

Exceptions

• A: runtime exception if there are any communication issues

void bootUsbRomBootloader()

Boots into integrated ROM bootloader in USB mode

Exceptions

• A: runtime exception if there are any communication issues

Version getVersion() const

Return

Version of current running bootloader

bool isEmbeddedVersion() const

Return

True when bootloader was booted using latest bootloader integrated in the library. False when bootloader is already running on the device and just connected to.

Type getType() const

Return

Type of currently connected bootloader

bool isAllowedFlashingBootloader() const

Return

True if allowed to flash bootloader

void close()

Explicitly closes connection to device.

Note

This function does not need to be explicitly called as destructor closes the device automatically

bool isClosed() const

Is the device already closed (or disconnected)

Public Static Functions

std::tuple<bool, DeviceInfo> getFirstAvailableDevice()

Searches for connected devices in either UNBOOTED or BOOTLOADER states and returns first available.

Return

Tuple of boolean and DeviceInfo. If found boolean is true and DeviceInfo describes the device. Otherwise false

std::vector<DeviceInfo> getAllAvailableDevices()

Searches for connected devices in either UNBOOTED or BOOTLOADER states.

Return

Vector of all found devices

std::vector<uint8_t> createDepthaiApplicationPackage(const Pipeline &pipeline, const dai::Path &pathToCmd = {}, bool compress = false)

Creates application package which can be flashed to depthai device.

Return

Depthai application package

Parameters

• pipeline: Pipeline from which to create the application package

• pathToCmd: Optional path to custom device firmware

• compress: Optional boolean which specifies if contents should be compressed

std::vector<uint8_t> createDepthaiApplicationPackage(const Pipeline &pipeline, bool compress)

Creates application package which can be flashed to depthai device.

Return

Depthai application package

Parameters

• pipeline: Pipeline from which to create the application package

• compress: Specifies if contents should be compressed

void saveDepthaiApplicationPackage(const dai::Path &path, const Pipeline &pipeline, const dai::Path &pathToCmd = {}, bool compress = false)

Saves application package to a file which can be flashed to depthai device.

Parameters

• path: Path where to save the application package

• pipeline: Pipeline from which to create the application package

• pathToCmd: Optional path to custom device firmware

• compress: Optional boolean which specifies if contents should be compressed

void saveDepthaiApplicationPackage(const dai::Path &path, const Pipeline &pipeline, bool compress)

Saves application package to a file which can be flashed to depthai device.

Parameters

• path: Path where to save the application package

• pipeline: Pipeline from which to create the application package

• compress: Specifies if contents should be compressed

Version getEmbeddedBootloaderVersion()

Return

Embedded bootloader version

std::vector<std::uint8_t> getEmbeddedBootloaderBinary(Type type = DEFAULT_TYPE)

Return

Embedded bootloader binary

Public Static Attributes

constexpr const Type DEFAULT_TYPE = {Type::USB}

Default Bootloader type.

struct Config : public dai::bootloader::Config

Public Functions

void setStaticIPv4(std::string ip, std::string mask, std::string gateway)

Setting a static IPv4 won’t start DHCP client.

void setDynamicIPv4(std::string ip, std::string mask, std::string gateway)

Setting a dynamic IPv4 will set that IP as well as start DHCP client.

bool isStaticIPV4()

Get if static IPv4 configuration is set.

std::string getIPv4()

Get IPv4.

std::string getIPv4Mask()

Get IPv4 mask.

std::string getIPv4Gateway()

Get IPv4 gateway.

void setDnsIPv4(std::string dns, std::string dnsAlt = "")

Set IPv4 DNS options.

std::string getDnsIPv4()

Get primary IPv4 DNS server.

std::string getDnsAltIPv4()

Get alternate IPv4 DNS server.

void setUsbTimeout(std::chrono::milliseconds ms)

Set USB timeout.

std::chrono::milliseconds getUsbTimeout()

Get USB timeout.

void setNetworkTimeout(std::chrono::milliseconds ms)

Set NETWOR timeout.

std::chrono::milliseconds getNetworkTimeout()

Get NETWORK timeout.

void setMacAddress(std::string mac)

Set MAC address if not flashed on controller.

std::string getMacAddress()

Get MAC address if not flashed on controller.

void setUsbMaxSpeed(UsbSpeed speed)

Set maxUsbSpeed.

UsbSpeed getUsbMaxSpeed()

Get maxUsbSpeed.

struct Version

#include <DeviceBootloader.hpp>

Bootloader version structure.

Public Functions

Version(const std::string &v)

Construct Version from string.

Version(unsigned major, unsigned minor, unsigned patch)

Construct Version major, minor and patch numbers.

Version(unsigned major, unsigned minor, unsigned patch, std::string buildInfo)

Construct Version major, minor and patch numbers with buildInfo.

std::string toString() const

Convert Version to string.

std::string toStringSemver() const

Convert Version to semver string.

std::string getBuildInfo() const

Get build info.

Version getSemver() const

Retrieves semver version (no build information)

struct DeviceInfo

#include <XLinkConnection.hpp>

Describes a connected device

Public Functions

DeviceInfo(std::string mxidOrName)

Creates a DeviceInfo by checking whether supplied parameter is a MXID or IP/USB name

Parameters

• mxidOrName: Either MXID, IP Address or USB port name

class EdgeDetectorConfig : public dai::Buffer

#include <EdgeDetectorConfig.hpp>

EdgeDetectorConfig message. Carries sobel edge filter config.

Public Functions

EdgeDetectorConfig()

Construct EdgeDetectorConfig message.

void setSobelFilterKernels(const std::vector<std::vector<int>> &horizontalKernel, const std::vector<std::vector<int>> &verticalKernel)

Set sobel filter horizontal and vertical 3x3 kernels

Parameters

• horizontalKernel: Used for horizontal gradient computation in 3x3 Sobel filter

• verticalKernel: Used for vertical gradient computation in 3x3 Sobel filter

EdgeDetectorConfigData getConfigData() const

Retrieve configuration data for EdgeDetector

Return

EdgeDetectorConfigData: sobel filter horizontal and vertical 3x3 kernels

struct EdgeDetectorConfigData

#include <RawEdgeDetectorConfig.hpp>

EdgeDetectorConfigData configuration data structure.

Public Members

std::vector<std::vector<int>> sobelFilterHorizontalKernel

Used for horizontal gradient computation in 3x3 Sobel filter Format - 3x3 matrix, 2nd column must be 0 Default - +1 0 -1; +2 0 -2; +1 0 -1

std::vector<std::vector<int>> sobelFilterVerticalKernel

Used for vertical gradient computation in 3x3 Sobel filter Format - 3x3 matrix, 2nd row must be 0 Default - +1 +2 +1; 0 0 0; -1 -2 -1

struct EdgeDetectorProperties : public dai::PropertiesSerializable<Properties, EdgeDetectorProperties>

#include <EdgeDetectorProperties.hpp>

Specify properties for EdgeDetector

Public Members

RawEdgeDetectorConfig initialConfig

Initial edge detector config.

int outputFrameSize = 1 * 1024 * 1024

Maximum output frame size in bytes (eg: 300x300 BGR image -> 300*300*3 bytes)

int numFramesPool = 4

Num frames in output pool.

struct EepromData

#include <EepromData.hpp>

EepromData structure

Contains the Calibration and Board data stored on device

struct Extrinsics

#include <Extrinsics.hpp>

Extrinsics structure.

Public Members

Point3f translation

(x, y, z) pose of destCameraSocket w.r.t currentCameraSocket obtained through calibration

Point3f specTranslation

(x, y, z) pose of destCameraSocket w.r.t currentCameraSocket measured through CAD design

class FeatureTrackerConfig : public dai::Buffer

#include <FeatureTrackerConfig.hpp>

FeatureTrackerConfig message. Carries config for feature tracking algorithm

Public Functions

FeatureTrackerConfig()

Construct FeatureTrackerConfig message.

FeatureTrackerConfig &setCornerDetector(dai::FeatureTrackerConfig::CornerDetector::Type cornerDetector)

Set corner detector algorithm type.

Parameters

• cornerDetector: Corner detector type, HARRIS or SHI_THOMASI

FeatureTrackerConfig &setCornerDetector(dai::FeatureTrackerConfig::CornerDetector config)

Set corner detector full configuration.

Parameters

• config: Corner detector configuration

FeatureTrackerConfig &setOpticalFlow()

Set optical flow as motion estimation algorithm type.

FeatureTrackerConfig &setOpticalFlow(dai::FeatureTrackerConfig::MotionEstimator::OpticalFlow config)

Set optical flow full configuration.

Parameters

• config: Optical flow configuration

FeatureTrackerConfig &setHwMotionEstimation()

Set hardware accelerated motion estimation using block matching. Faster than optical flow (software implementation) but might not be as accurate.

FeatureTrackerConfig &setNumTargetFeatures(std::int32_t numTargetFeatures)

Set number of target features to detect.

Parameters

• numTargetFeatures: Number of features

FeatureTrackerConfig &setMotionEstimator(bool enable)

Enable or disable motion estimator.

Parameters

• enable:

FeatureTrackerConfig &setMotionEstimator(dai::FeatureTrackerConfig::MotionEstimator config)

Set motion estimator full configuration.

Parameters

• config: Motion estimator configuration

FeatureTrackerConfig &setFeatureMaintainer(bool enable)

Enable or disable feature maintainer.

Parameters

• enable:

FeatureTrackerConfig &setFeatureMaintainer(dai::FeatureTrackerConfig::FeatureMaintainer config)

Set feature maintainer full configuration.

Parameters

• config: feature maintainer configuration

FeatureTrackerConfig &set(dai::RawFeatureTrackerConfig config)

Set explicit configuration.

Parameters

• config: Explicit configuration

dai::RawFeatureTrackerConfig get() const

Retrieve configuration data for FeatureTracker.

Return

config for feature tracking algorithm

struct FeatureTrackerProperties : public dai::PropertiesSerializable<Properties, FeatureTrackerProperties>

#include <FeatureTrackerProperties.hpp>

Specify properties for FeatureTracker

Public Members

RawFeatureTrackerConfig initialConfig

Initial feature tracker config

std::int32_t numShaves = 1

Number of shaves reserved for feature tracking. Optical flow can use 1 or 2 shaves, while for corner detection only 1 is enough. Hardware motion estimation doesn’t require shaves. Maximum 2, minimum 1.

std::int32_t numMemorySlices = 1

Number of memory slices reserved for feature tracking. Optical flow can use 1 or 2 memory slices, while for corner detection only 1 is enough. Maximum number of features depends on the number of allocated memory slices. Hardware motion estimation doesn’t require memory slices. Maximum 2, minimum 1.

struct GlobalProperties : public dai::PropertiesSerializable<Properties, GlobalProperties>

#include <GlobalProperties.hpp>

Specify properties which apply for whole pipeline

Public Members

double leonCssFrequencyHz = 700 * 1000 * 1000

Set frequency of Leon OS - Increasing can improve performance, at the cost of higher power draw

double leonMssFrequencyHz = 700 * 1000 * 1000

Set frequency of Leon RT - Increasing can improve performance, at the cost of higher power draw

tl::optional<dai::EepromData> calibData

Calibration data sent through pipeline

tl::optional<std::uint32_t> cameraTuningBlobSize

Camera tuning blob size in bytes

std::string cameraTuningBlobUri

Uri which points to camera tuning blob

int32_t xlinkChunkSize = -1

Chunk size for splitting device-sent XLink packets, in bytes. A larger value could increase performance, with 0 disabling chunking. A negative value won’t modify the device defaults - configured per protocol, currently 64*1024 for both USB and Ethernet.

class ImageManipConfig : public dai::Buffer

#include <ImageManipConfig.hpp>

ImageManipConfig message. Specifies image manipulation options like:

• Crop

• Resize

• Warp

• …

Public Functions

ImageManipConfig()

Construct ImageManipConfig message.

ImageManipConfig &setCropRect(float xmin, float ymin, float xmax, float ymax)

Specifies crop with rectangle with normalized values (0..1)

Parameters

• xmin: Top left X coordinate of rectangle

• ymin: Top left Y coordinate of rectangle

• xmax: Bottom right X coordinate of rectangle

• ymax: Bottom right Y coordinate of rectangle

ImageManipConfig &setCropRect(std::tuple<float, float, float, float> coordinates)

Specifies crop with rectangle with normalized values (0..1)

Parameters

• coordinates: Coordinate of rectangle

ImageManipConfig &setCropRotatedRect(RotatedRect rr, bool normalizedCoords = true)

Specifies crop with rotated rectangle. Optionally as non normalized coordinates

Parameters

• rr: Rotated rectangle which specifies crop

• normalizedCoords: If true coordinates are in normalized range (0..1) otherwise absolute

ImageManipConfig &setCenterCrop(float ratio, float whRatio = 1.0f)

Specifies a centered crop.

Parameters

• ratio: Ratio between input image and crop region (0..1)

• whRatio: Crop region aspect ratio - 1 equals to square, 1.7 equals to 16:9, …

ImageManipConfig &setWarpTransformFourPoints(std::vector<Point2f> pt, bool normalizedCoords)

Specifies warp by supplying 4 points in either absolute or normalized coordinates

Parameters

• pt: 4 points specifying warp

• normalizedCoords: If true pt is interpreted as normalized, absolute otherwise

ImageManipConfig &setWarpTransformMatrix3x3(std::vector<float> mat)

Specifies warp with a 3x3 matrix

Parameters

• mat: 3x3 matrix

ImageManipConfig &setWarpBorderReplicatePixels()

Specifies that warp replicates border pixels

ImageManipConfig &setWarpBorderFillColor(int red, int green, int blue)

Specifies fill color for border pixels. Example:

• setWarpBorderFillColor(255,255,255) -> white

• setWarpBorderFillColor(0,0,255) -> blue

Parameters

• red: Red component

• green: Green component

• blue: Blue component

ImageManipConfig &setRotationDegrees(float deg)

Specifies clockwise rotation in degrees

Parameters

• deg: Rotation in degrees

ImageManipConfig &setRotationRadians(float rad)

Specifies clockwise rotation in radians

Parameters

• rad: Rotation in radians

ImageManipConfig &setResize(int w, int h)

Specifies output image size. After crop stage the image will be stretched to fit.

Parameters

• w: Width in pixels

• h: Height in pixels

ImageManipConfig &setResize(std::tuple<int, int> size)

Specifies output image size. After crop stage the image will be stretched to fit.

Parameters

• size: Size in pixels

ImageManipConfig &setResizeThumbnail(int w, int h, int bgRed = 0, int bgGreen = 0, int bgBlue = 0)

Specifies output image size. After crop stage the image will be resized by preserving aspect ration. Optionally background can be specified.

Parameters

• w: Width in pixels

• h: Height in pixels

• bgRed: Red component

• bgGreen: Green component

• bgBlue: Blue component

ImageManipConfig &setResizeThumbnail(std::tuple<int, int> size, int bgRed = 0, int bgGreen = 0, int bgBlue = 0)

Specifies output image size. After crop stage the image will be resized by preserving aspect ration. Optionally background can be specified.

Parameters

• size: Size in pixels

• bgRed: Red component

• bgGreen: Green component

• bgBlue: Blue component

ImageManipConfig &setFrameType(ImgFrame::Type name)

Specify output frame type.

Parameters

• name: Frame type

ImageManipConfig &setHorizontalFlip(bool flip)

Specify horizontal flip

Parameters

• flip: True to enable flip, false otherwise

ImageManipConfig &setReusePreviousImage(bool reuse)

Instruct ImageManip to not remove current image from its queue and use the same for next message.

Parameters

• reuse: True to enable reuse, false otherwise

ImageManipConfig &setSkipCurrentImage(bool skip)

Instructs ImageManip to skip current image and wait for next in queue.

Parameters

• skip: True to skip current image, false otherwise

ImageManipConfig &setKeepAspectRatio(bool keep)

Specifies to whether to keep aspect ratio or not

float getCropXMin() const

Return

Top left X coordinate of crop region

float getCropYMin() const

Return

Top left Y coordinate of crop region

float getCropXMax() const

Return

Bottom right X coordinate of crop region

float getCropYMax() const

Return

Bottom right Y coordinate of crop region

int getResizeWidth() const

Return

Output image width

int getResizeHeight() const

Return

Output image height

CropConfig getCropConfig() const

Return

Crop configuration

ResizeConfig getResizeConfig() const

Return

Resize configuration

FormatConfig getFormatConfig() const

Return

Format configuration

bool isResizeThumbnail() const

Return

True if resize thumbnail mode is set, false otherwise

struct ImageManipProperties : public dai::PropertiesSerializable<Properties, ImageManipProperties>

#include <ImageManipProperties.hpp>

Specify properties for ImageManip

Public Members

RawImageManipConfig initialConfig

Initial configuration for ImageManip node.

int outputFrameSize = 1 * 1024 * 1024

Maximum output frame size in bytes (eg: 300x300 BGR image -> 300*300*3 bytes)

int numFramesPool = 4

Num frames in output pool.

struct ImgDetection

#include <RawImgDetections.hpp>

ImgDetection structure.

Subclassed by dai::SpatialImgDetection

class ImgDetections : public dai::Buffer

#include <ImgDetections.hpp>

ImgDetections message. Carries normalized detection results

Public Functions

ImgDetections()

Construct ImgDetections message.

ImgDetections &setTimestamp(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Sets image timestamp related to dai::Clock::now()

ImgDetections &setTimestampDevice(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Sets image timestamp related to dai::Clock::now()

ImgDetections &setSequenceNum(int64_t sequenceNum)

Retrieves image sequence number

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestamp() const

Retrieves image timestamp related to dai::Clock::now()

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestampDevice() const

Retrieves image timestamp directly captured from device’s monotonic clock, not synchronized to host time. Used mostly for debugging

int64_t getSequenceNum() const

Retrieves image sequence number

Public Members

std::vector<ImgDetection> &detections

Detections.

class ImgFrame : public dai::Buffer

#include <ImgFrame.hpp>

ImgFrame message. Carries image data and metadata.

Public Functions

ImgFrame()

Construct ImgFrame message. Timestamp is set to now

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestamp() const

Retrieves image timestamp related to dai::Clock::now()

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestampDevice() const

Retrieves image timestamp directly captured from device’s monotonic clock, not synchronized to host time. Used mostly for debugging

unsigned int getInstanceNum() const

Retrieves instance number

unsigned int getCategory() const

Retrieves image category

int64_t getSequenceNum() const

Retrieves image sequence number

unsigned int getWidth() const

Retrieves image width in pixels

unsigned int getHeight() const

Retrieves image height in pixels

Type getType() const

Retrieves image type

ImgFrame &setTimestamp(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Retrieves image timestamp related to dai::Clock::now()

ImgFrame &setTimestampDevice(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Sets image timestamp related to dai::Clock::now()

ImgFrame &setInstanceNum(unsigned int instance)

Instance number relates to the origin of the frame (which camera)

Parameters

• instance: Instance number

ImgFrame &setCategory(unsigned int category)

Parameters

• category: Image category

ImgFrame &setSequenceNum(int64_t seq)

Specifies sequence number

Parameters

• seq: Sequence number

ImgFrame &setWidth(unsigned int width)

Specifies frame width

Parameters

• width: frame width

ImgFrame &setHeight(unsigned int height)

Specifies frame height

Parameters

• height: frame height

ImgFrame &setSize(unsigned int width, unsigned int height)

Specifies frame size

Parameters

• height: frame height

• width: frame width

ImgFrame &setSize(std::tuple<unsigned int, unsigned int> size)

Specifies frame size

Parameters

• size: frame size

ImgFrame &setType(Type type)

Specifies frame type, RGB, BGR, …

Parameters

• type: Type of image

ImgFrame &setFrame(cv::Mat frame)

Copies cv::Mat data to

ImgFrame buffer

Note

This API only available if OpenCV support is enabled

Parameters

• frame: Input cv::Mat frame from which to copy the data

cv::Mat getFrame(bool copy = false)

Retrieves data as cv::Mat with specified width, height and type

Note

This API only available if OpenCV support is enabled

Return

cv::Mat with corresponding to ImgFrame parameters

Parameters

• copy: If false only a reference to data is made, otherwise a copy

cv::Mat getCvFrame()

Retrieves cv::Mat suitable for use in common opencv functions.

ImgFrame is converted to color BGR interleaved or grayscale depending on type.

Note

This API only available if OpenCV support is enabled

A copy is always made

Return

cv::Mat for use in opencv functions

class IMUData : public dai::Buffer

#include <IMUData.hpp>

IMUData message. Carries normalized detection results

Public Functions

IMUData()

Construct IMUData message.

Public Members

std::vector<IMUPacket> &packets

Detections.

struct IMUPacket

#include <RawIMUData.hpp>

IMU output

Contains combined output for all possible modes. Only the enabled outputs are populated.

struct IMUProperties : public dai::PropertiesSerializable<Properties, IMUProperties>

struct IMUReport

Subclassed by dai::IMUReportAccelerometer, dai::IMUReportGyroscope, dai::IMUReportMagneticField, dai::IMUReportRotationVectorWAcc

Public Members

int32_t sequence = 0

The sequence number increments once for each report sent. Gaps in the sequence numbers indicate missing or dropped reports. Max value 255 after which resets to 0.

Accuracy accuracy = Accuracy::UNRELIABLE

Accuracy of sensor

Timestamp timestamp = {}

Generation timestamp, synced to host time

Timestamp tsDevice = {}

Generation timestamp, direct device monotonic clock

struct IMUReportAccelerometer : public dai::IMUReport

#include <RawIMUData.hpp>

Accelerometer.

Units are [m/s^2]

struct IMUReportGyroscope : public dai::IMUReport

#include <RawIMUData.hpp>

Gyroscope.

Units are [rad/s]

struct IMUReportMagneticField : public dai::IMUReport

#include <RawIMUData.hpp>

Magnetic field.

Units are [uTesla]

struct IMUReportRotationVectorWAcc : public dai::IMUReport

#include <RawIMUData.hpp>

Rotation Vector with Accuracy.

Contains quaternion components: i,j,k,real

Public Members

float i = 0

Quaternion component i.

float j = 0

Quaternion component j.

float k = 0

Quaternion component k.

float real = 0

Quaternion component, real.

float rotationVectorAccuracy = 0

Accuracy estimate [radians], 0 means no estimate.

struct IMUSensorConfig

Public Members

bool sensitivityRelative = false

Change reports relative (vs absolute)

uint16_t changeSensitivity = 0

Report-on-change threshold.

template<typename T>
class LockingQueue

struct LogMessage

struct MemoryInfo

#include <MemoryInfo.hpp>

MemoryInfo structure

Free, remaining and total memory stats

struct MonoCameraProperties : public dai::PropertiesSerializable<Properties, MonoCameraProperties>

#include <MonoCameraProperties.hpp>

Specify properties for MonoCamera such as camera ID, …

Public Types

enum SensorResolution

Select the camera sensor resolution: 1280×720, 1280×800, 640×400, 640×480

Values:

enumerator THE_720_P

enumerator THE_800_P

enumerator THE_400_P

enumerator THE_480_P

Public Members

CameraBoardSocket boardSocket = CameraBoardSocket::AUTO

Which socket will mono camera use

CameraImageOrientation imageOrientation = CameraImageOrientation::AUTO

Camera sensor image orientation / pixel readout

SensorResolution resolution = SensorResolution::THE_720_P

Select the camera sensor resolution

float fps = 30.0

Camera sensor FPS

struct NeuralNetworkProperties : public dai::PropertiesSerializable<Properties, NeuralNetworkProperties>

#include <NeuralNetworkProperties.hpp>

Specify properties for NeuralNetwork such as blob path, …

Subclassed by dai::PropertiesSerializable< NeuralNetworkProperties, DetectionNetworkProperties >

Public Members

tl::optional<std::uint32_t> blobSize

Blob binary size in bytes

std::string blobUri

Uri which points to blob

std::uint32_t numFrames = 8

Number of available output tensors in pool

std::uint32_t numThreads = 0

Number of threads to create for running inference. 0 = auto

std::uint32_t numNCEPerThread = 0

Number of NCE (Neural Compute Engine) per inference thread. 0 = auto

class NNData : public dai::Buffer

#include <NNData.hpp>

NNData message. Carries tensors and their metadata

Public Functions

NNData()

Construct NNData message.

NNData &setLayer(const std::string &name, std::vector<std::uint8_t> data)

Set a layer with datatype U8.

Parameters

• name: Name of the layer

• data: Data to store

NNData &setLayer(const std::string &name, const std::vector<int> &data)

Set a layer with datatype U8. Integers are cast to bytes.

Parameters

• name: Name of the layer

• data: Data to store

NNData &setLayer(const std::string &name, std::vector<float> data)

Set a layer with datatype FP16. Float values are converted to FP16.

Parameters

• name: Name of the layer

• data: Data to store

NNData &setLayer(const std::string &name, std::vector<double> data)

Set a layer with datatype FP16. Double values are converted to FP16.

Parameters

• name: Name of the layer

• data: Data to store

std::vector<std::string> getAllLayerNames() const

Return

Names of all layers added

std::vector<TensorInfo> getAllLayers() const

Return

All layers and their information

bool getLayer(const std::string &name, TensorInfo &tensor) const

Retrieve layers tensor information

Return

True if layer exists, false otherwise

Parameters

• name: Name of the layer

• [out] tensor: Outputs tensor information of that layer

bool hasLayer(const std::string &name) const

Checks if given layer exists

Return

True if layer exists, false otherwise

Parameters

• name: Name of the layer

bool getLayerDatatype(const std::string &name, TensorInfo::DataType &datatype) const

Retrieve datatype of a layers tensor

Return

True if layer exists, false otherwise

Parameters

• name: Name of the layer

• [out] datatype: Datatype of layers tensor

std::vector<std::uint8_t> getLayerUInt8(const std::string &name) const

Convenience function to retrieve U8 data from layer

Return

U8 binary data

Parameters

• name: Name of the layer

std::vector<float> getLayerFp16(const std::string &name) const

Convenience function to retrieve float values from layers FP16 tensor

Return

Float data

Parameters

• name: Name of the layer

std::vector<std::int32_t> getLayerInt32(const std::string &name) const

Convenience function to retrieve INT32 values from layers tensor

Return

INT32 data

Parameters

• name: Name of the layer

std::vector<std::uint8_t> getFirstLayerUInt8() const

Convenience function to retrieve U8 data from first layer

Return

U8 binary data

std::vector<float> getFirstLayerFp16() const

Convenience function to retrieve float values from first layers FP16 tensor

Return

Float data

std::vector<std::int32_t> getFirstLayerInt32() const

Convenience function to retrieve INT32 values from first layers tensor

Return

INT32 data

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestamp() const

Retrieves image timestamp related to dai::Clock::now()

std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> getTimestampDevice() const

Retrieves image timestamp directly captured from device’s monotonic clock, not synchronized to host time. Used mostly for debugging

int64_t getSequenceNum() const

Retrieves image sequence number

NNData &setTimestamp(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Sets image timestamp related to dai::Clock::now()

NNData &setTimestampDevice(std::chrono::time_point<std::chrono::steady_clock, std::chrono::steady_clock::duration> timestamp)

Sets image timestamp related to dai::Clock::now()

NNData &setSequenceNum(int64_t sequenceNum)

Retrieves image sequence number

class Node

#include <Node.hpp>

Abstract Node.

Subclassed by dai::NodeCRTP< Node, AprilTag, AprilTagProperties >, dai::NodeCRTP< Node, ColorCamera, ColorCameraProperties >, dai::NodeCRTP< Node, DetectionParser, DetectionParserProperties >, dai::NodeCRTP< Node, EdgeDetector, EdgeDetectorProperties >, dai::NodeCRTP< Node, FeatureTracker, FeatureTrackerProperties >, dai::NodeCRTP< Node, ImageManip, ImageManipProperties >, dai::NodeCRTP< Node, IMU, IMUProperties >, dai::NodeCRTP< Node, MonoCamera, MonoCameraProperties >, dai::NodeCRTP< Node, NeuralNetwork, NeuralNetworkProperties >, dai::NodeCRTP< Node, ObjectTracker, ObjectTrackerProperties >, dai::NodeCRTP< Node, Script, ScriptProperties >, dai::NodeCRTP< Node, SpatialLocationCalculator, SpatialLocationCalculatorProperties >, dai::NodeCRTP< Node, SPIIn, SPIInProperties >, dai::NodeCRTP< Node, SPIOut, SPIOutProperties >, dai::NodeCRTP< Node, StereoDepth, StereoDepthProperties >, dai::NodeCRTP< Node, SystemLogger, SystemLoggerProperties >, dai::NodeCRTP< Node, VideoEncoder, VideoEncoderProperties >, dai::NodeCRTP< Node, XLinkIn, XLinkInProperties >, dai::NodeCRTP< Node, XLinkOut, XLinkOutProperties >

Public Types

using Id = std::int64_t

Node identificator. Unique for every node on a single Pipeline.

Public Functions

std::unique_ptr<Node> clone() const = 0

Deep copy the node.

const char *getName() const = 0

Retrieves nodes name.

std::vector<Output> getOutputs()

Retrieves all nodes outputs.

std::vector<Input> getInputs()

Retrieves all nodes inputs.

std::vector<Output*> getOutputRefs()

Retrieves reference to node outputs.

std::vector<const Output*> getOutputRefs() const

Retrieves reference to node outputs.

std::vector<Input*> getInputRefs()

Retrieves reference to node inputs.

std::vector<const Input*> getInputRefs() const

Retrieves reference to node inputs.

const AssetManager &getAssetManager() const

Get node AssetManager as a const reference.

AssetManager &getAssetManager()

Get node AssetManager as a reference.

Public Members

const Id id

Id of node.

struct Connection

#include <Node.hpp>

Connection between an Input and Output.

struct NodeConnectionSchema

#include <NodeConnectionSchema.hpp>

Specifies a connection between nodes IOs

template<typename Base, typename Derived, typename Props>
class NodeCRTP : public Base

Public Members

Properties &properties

Underlying properties.

struct NodeIoInfo

#include <NodeIoInfo.hpp>

NodeIo informations such as name, type, …

struct NodeObjInfo

#include <NodeObjInfo.hpp>

NodeObj information structure.

struct IoInfoKey

struct ObjectTrackerProperties : public dai::PropertiesSerializable<Properties, ObjectTrackerProperties>

#include <ObjectTrackerProperties.hpp>

Specify properties for ObjectTracker

Public Members

float trackerThreshold = 0.0

Confidence threshold for tracklets. Above this threshold detections will be tracked. Default 0, all detections are tracked.

std::int32_t maxObjectsToTrack = 60

Maximum number of objects to track. Maximum 60 for SHORT_TERM_KCF, maximum 1000 for other tracking methods. Default 60.

std::vector<std::uint32_t> detectionLabelsToTrack

Which detections labels to track. Default all labels are tracked.

TrackerType trackerType = TrackerType::ZERO_TERM_IMAGELESS

Tracking method.

TrackerIdAssignmentPolicy trackerIdAssignmentPolicy = TrackerIdAssignmentPolicy::UNIQUE_ID

New ID assignment policy.

class OpenVINO

#include <OpenVINO.hpp>

Support for basic OpenVINO related actions like version identification of neural network blobs,…

Public Types

enum Version

OpenVINO Version supported version information.

Values:

enumerator VERSION_2020_3

enumerator VERSION_2020_4

enumerator VERSION_2021_1

enumerator VERSION_2021_2

enumerator VERSION_2021_3

enumerator VERSION_2021_4

enumerator VERSION_2022_1

Public Static Functions

std::vector<Version> getVersions()

Return

Supported versions

std::string getVersionName(Version version)

Returns string representation of a given version

Return

Name of a given version

Parameters

• version: OpenVINO version

Version parseVersionName(const std::string &versionString)

Creates Version from string representation. Throws if not possible.

Return

Version object if successful

Parameters

• versionString: Version as string

std::vector<Version> getBlobSupportedVersions(std::uint32_t majorVersion, std::uint32_t minorVersion)

Returns a list of potentially supported versions for a specified blob major and minor versions.

Return

Vector of potentially supported versions

Parameters

• majorVersion: Major version from OpenVINO blob

• minorVersion: Minor version from OpenVINO blob

Version getBlobLatestSupportedVersion(std::uint32_t majorVersion, std::uint32_t minorVersion)

Returns latest potentially supported version by a given blob version.

Return

Latest potentially supported version

Parameters

• majorVersion: Major version from OpenVINO blob

• minorVersion: Minor version from OpenVINO blob

bool areVersionsBlobCompatible(Version v1, Version v2)

Checks whether two blob versions are compatible

Public Static Attributes

constexpr const Version DEFAULT_VERSION = VERSION_2022_1

Main OpenVINO version.

struct Blob

#include <OpenVINO.hpp>

OpenVINO Blob.

Public Functions

Blob(std::vector<uint8_t> data)

Construct a new Blob from data in memory.

Parameters

• data: In memory blob

Blob(const dai::Path &path)

Construct a new Blob by loading from a filesystem path.

Parameters

• path: Filesystem path to the blob

Public Members

Version version

OpenVINO version.

std::unordered_map<std::string, TensorInfo> networkInputs

Map of input names to additional information.

std::unordered_map<std::string, TensorInfo> networkOutputs

Map of output names to additional information.

uint32_t stageCount = 0

Number of network stages.

uint32_t numShaves = 0

Number of shaves the blob was compiled for.

uint32_t numSlices = 0

Number of CMX slices the blob was compiled for.

std::vector<uint8_t> data

Blob data.

class Path

#include <Path.hpp>

Represents paths on a filesystem; accepts utf-8, Windows utf-16 wchar_t, or std::filesystem::path.

It is suitable for direct use with OS APIs. Features are limited to character-set conversion of paths. It is not intended as a full replacement for std::filesystem::path

Public Types

using value_type = char

character used by native-encoding of filesystem

Public Functions

Path(string_type &&source) noexcept

Construct Path object from source.

Parameters

• source: native-encoding character sequence; no conversion

Path(const string_type &source)

Construct Path object from source.

Parameters

• source: native-encoding character sequence; no conversion

Path(const value_type *source)

Construct Path object from source.

Parameters

• source: pointer to null-terminated native-encoding character sequence; no conversion

std::string string() const

Get path in native-encoding string; no conversion.

Return

std::string

std::string u8string() const

Get path in utf-8.

Return

std::string in utf-8

operator string_type() const noexcept

Implicitly convert to native-encoding string, suitable for use with OS APIs.

Return

std::string of utf-8 on most OSs, std::wstring of utf-16 on Windows

const string_type &native() const noexcept

Returns native-encoding string by const reference, suitable for use with OS APIs.

Return

const std::string& of utf-8 on most OSs, const std::wstring& of utf-16 on Windows

bool empty() const noexcept

Observes if path is empty (contains no string/folders/filename)

Return

bool true if the path is empty, false otherwise

template<typename T>
class Pimpl

class Pipeline

#include <Pipeline.hpp>

Represents the pipeline, set of nodes and connections between them.

Public Functions

Pipeline()

Constructs a new pipeline

Pipeline clone() const

Clone the pipeline (Creates a copy)

GlobalProperties getGlobalProperties() const

Return

Global properties of current pipeline

PipelineSchema getPipelineSchema(SerializationType type = DEFAULT_SERIALIZATION_TYPE) const

Return

Pipeline schema

nlohmann::json serializeToJson() const

Returns whole pipeline represented as JSON.

template<class N>
std::shared_ptr<N> create()

Adds a node to pipeline.

Node is specified by template argument N

void remove(std::shared_ptr<Node> node)

Removes a node from pipeline.

std::vector<std::shared_ptr<const Node>> getAllNodes() const

Get a vector of all nodes.

std::vector<std::shared_ptr<Node>> getAllNodes()

Get a vector of all nodes.

std::shared_ptr<const Node> getNode(Node::Id id) const

Get node with id if it exists, nullptr otherwise.

std::shared_ptr<Node> getNode(Node::Id id)

Get node with id if it exists, nullptr otherwise.

std::vector<Node::Connection> getConnections() const

Get all connections.

const NodeConnectionMap &getConnectionMap() const

Get a reference to internal connection representation.

const NodeMap &getNodeMap() const

Get a reference to internal node map.

void link(const Node::Output &out, const Node::Input &in)

Link output to an input. Both nodes must be on the same pipeline

Throws an error if they aren’t or cannot be connected

Parameters

• out: Nodes output to connect from

• in: Nodes input to connect to

void unlink(const Node::Output &out, const Node::Input &in)

Unlink output from an input.

Throws an error if link doesn’t exists

Parameters

• out: Nodes output to unlink from

• in: Nodes input to unlink to

const AssetManager &getAssetManager() const

Get pipelines AssetManager as reference.

AssetManager &getAssetManager()

Get pipelines AssetManager as reference.

void setOpenVINOVersion(OpenVINO::Version version)

Set a specific OpenVINO version to use with this pipeline.

void setCalibrationData(CalibrationHandler calibrationDataHandler)

Sets the calibration in pipeline which overrides the calibration data in eeprom

Parameters

• calibrationDataHandler: CalibrationHandler object which is loaded with calibration information.

CalibrationHandler getCalibrationData() const

gets the calibration data which is set through pipeline

Return

the calibrationHandler with calib data in the pipeline

OpenVINO::Version getOpenVINOVersion() const

Get possible OpenVINO version to run this pipeline.

tl::optional<OpenVINO::Version> getRequiredOpenVINOVersion() const

Get required OpenVINO version to run this pipeline. Can be none.

void setCameraTuningBlobPath(const dai::Path &path)

Set a camera IQ (Image Quality) tuning blob, used for all cameras.

void setXLinkChunkSize(int sizeBytes)

Set chunk size for splitting device-sent XLink packets, in bytes. A larger value could increase performance, with 0 disabling chunking. A negative value won’t modify the device defaults - configured per protocol, currently 64*1024 for both USB and Ethernet.

bool isOpenVINOVersionCompatible(OpenVINO::Version version) const

Checks whether a given OpenVINO version is compatible with the pipeline.

Device::Config getDeviceConfig() const

Get device configuration needed for this pipeline.

class PipelineImpl

struct PipelineSchema

#include <PipelineSchema.hpp>

Specifies whole pipeline, nodes, properties and connections between nodes IOs

struct Point2f

#include <Point2f.hpp>

Point2f structure

x and y coordinates that define a 2D point.

struct Point3f

#include <Point3f.hpp>

Point3f structure

x,y,z coordinates that define a 3D point.

struct Properties

#include <Properties.hpp>

Base Properties structure.

Subclassed by dai::PropertiesSerializable< Properties, AprilTagProperties >, dai::PropertiesSerializable< Properties, ColorCameraProperties >, dai::PropertiesSerializable< Properties, DetectionParserProperties >, dai::PropertiesSerializable< Properties, EdgeDetectorProperties >, dai::PropertiesSerializable< Properties, FeatureTrackerProperties >, dai::PropertiesSerializable< Properties, GlobalProperties >, dai::PropertiesSerializable< Properties, ImageManipProperties >, dai::PropertiesSerializable< Properties, IMUProperties >, dai::PropertiesSerializable< Properties, MonoCameraProperties >, dai::PropertiesSerializable< Properties, NeuralNetworkProperties >, dai::PropertiesSerializable< Properties, ObjectTrackerProperties >, dai::PropertiesSerializable< Properties, ScriptProperties >, dai::PropertiesSerializable< Properties, SpatialLocationCalculatorProperties >, dai::PropertiesSerializable< Properties, SPIInProperties >, dai::PropertiesSerializable< Properties, SPIOutProperties >, dai::PropertiesSerializable< Properties, StereoDepthProperties >, dai::PropertiesSerializable< Properties, SystemLoggerProperties >, dai::PropertiesSerializable< Properties, VideoEncoderProperties >, dai::PropertiesSerializable< Properties, XLinkInProperties >, dai::PropertiesSerializable< Properties, XLinkOutProperties >

template<typename Base, typename Derived>
struct PropertiesSerializable : public Base

#include <Properties.hpp>

Serializable properties.

struct RawAprilTagConfig : public dai::RawBuffer

#include <RawAprilTagConfig.hpp>

RawAprilTags configuration structure.

Public Types

enum Family

Supported AprilTag families.

Values:

enumerator TAG_36H11

enumerator TAG_36H10

enumerator TAG_25H9

enumerator TAG_16H5

enumerator TAG_CIR21H7

enumerator TAG_STAND41H12

Public Members

Family family = Family::TAG_36H11

AprilTag family.

std::int32_t quadDecimate = 4

Detection of quads can be done on a lower-resolution image, improving speed at a cost of pose accuracy and a slight decrease in detection rate. Decoding the binary payload is still done at full resolution.

float quadSigma = 0.0f

What Gaussian blur should be applied to the segmented image. Parameter is the standard deviation in pixels. Very noisy images benefit from non-zero values (e.g. 0.8).

bool refineEdges = true

When non-zero, the edges of the each quad are adjusted to “snap

to” strong gradients nearby. This is useful when decimation is employed, as it can increase the quality of the initial quad estimate substantially. Generally recommended to be on. Very computationally inexpensive. Option is ignored if quadDecimate = 1.

float decodeSharpening = 0.25f

How much sharpening should be done to decoded images? This can help decode small tags but may or may not help in odd lighting conditions or low light conditions. The default value is 0.25.

std::int32_t maxHammingDistance = 1

Max number of error bits that should be corrected. Accepting large numbers of corrected errors leads to greatly increased false positive rates. As of this implementation, the detector cannot detect tags with a hamming distance greater than 2.

QuadThresholds quadThresholds

AprilTag quad threshold parameters.

struct QuadThresholds

#include <RawAprilTagConfig.hpp>

AprilTag quad threshold parameters.

Public Members

std::int32_t minClusterPixels = 5

Reject quads containing too few pixels.

std::int32_t maxNmaxima = 10

How many corner candidates to consider when segmenting a group of pixels into a quad.

float criticalDegree = 10.f

Reject quads where pairs of edges have angles that are close to straight or close to 180 degrees. Zero means that no quads are rejected. (In degrees).

float maxLineFitMse = 10.f

When fitting lines to the contours, what is the maximum mean squared error allowed? This is useful in rejecting contours that are far from being quad shaped; rejecting these quads “early” saves expensive decoding processing.

std::int32_t minWhiteBlackDiff = 5

When we build our model of black & white pixels, we add an extra check that the white model must be (overall) brighter than the black model. How much brighter? (in pixel values: [0,255]).

bool deglitch = false

Should the thresholded image be deglitched? Only useful for very noisy images

struct RawAprilTags : public dai::RawBuffer

#include <RawAprilTags.hpp>

RawAprilTags structure.

struct RawBuffer

#include <RawBuffer.hpp>

RawBuffer structure.

Subclassed by dai::RawAprilTagConfig, dai::RawAprilTags, dai::RawCameraControl, dai::RawEdgeDetectorConfig, dai::RawFeatureTrackerConfig, dai::RawImageManipConfig, dai::RawImgDetections, dai::RawImgFrame, dai::RawIMUData, dai::RawNNData, dai::RawSpatialImgDetections, dai::RawSpatialLocationCalculatorConfig, dai::RawSpatialLocations, dai::RawStereoDepthConfig, dai::RawSystemInformation, dai::RawTrackedFeatures, dai::RawTracklets

struct RawCameraControl : public dai::RawBuffer

#include <RawCameraControl.hpp>

RawCameraControl structure.

Public Members

uint8_t lensPosition = 0

Lens/VCM position, range: 0..255. Used with autoFocusMode = OFF. With current IMX378 modules:

• max 255: macro focus, at 8cm distance

• infinite focus at about 120..130 (may vary from module to module)

• lower values lead to out-of-focus (lens too close to the sensor array)

struct ManualExposureParams

struct RegionParams

struct RawEdgeDetectorConfig : public dai::RawBuffer

#include <RawEdgeDetectorConfig.hpp>

RawEdgeDetectorConfig configuration structure.

struct RawFeatureTrackerConfig : public dai::RawBuffer

#include <RawFeatureTrackerConfig.hpp>

RawFeatureTrackerConfig configuration structure.

Public Members

CornerDetector cornerDetector

Corner detector configuration. Used for feature detection.

MotionEstimator motionEstimator

Motion estimator configuration. Used for feature reidentification between current and previous features.

FeatureMaintainer featureMaintainer

FeatureMaintainer configuration. Used for feature maintaining.

struct CornerDetector

#include <RawFeatureTrackerConfig.hpp>

Corner detector configuration structure.

Public Members

Type type = Type::HARRIS

Corner detector algorithm type.

std::int32_t cellGridDimension = 4

Ensures distributed feature detection across the image. Image is divided into horizontal and vertical cells, each cell has a target feature count = numTargetFeatures / cellGridDimension. Each cell has its own feature threshold. A value of 4 means that the image is divided into 4x4 cells of equal width/height. Maximum 4, minimum 1.

std::int32_t numTargetFeatures = 320

Target number of features to detect. Maximum number of features is determined at runtime based on algorithm type.

std::int32_t numMaxFeatures = AUTO

Hard limit for the maximum number of features that can be detected. 0 means auto, will be set to the maximum value based on memory constraints.

bool enableSobel = true

Enable 3x3 Sobel operator to smoothen the image whose gradient is to be computed. If disabled, a simple 1D row/column differentiator is used for gradient.

bool enableSorting = true

Enable sorting detected features based on their score or not.

Thresholds thresholds

Threshold settings. These are advanced settings, suitable for debugging/special cases.

struct Thresholds

#include <RawFeatureTrackerConfig.hpp>

Threshold settings structure for corner detector.

Public Members

float initialValue = AUTO

Minimum strength of a feature which will be detected. 0 means automatic threshold update. Recommended so the tracker can adapt to different scenes/textures. Each cell has its own threshold. Empirical value.

float min = AUTO

Minimum limit for threshold. Applicable when automatic threshold update is enabled. 0 means auto, 6000000 for HARRIS, 1200 for SHI_THOMASI. Empirical value.

float max = AUTO

Maximum limit for threshold. Applicable when automatic threshold update is enabled. 0 means auto. Empirical value.

float decreaseFactor = 0.9f

When detected number of features exceeds the maximum in a cell threshold is lowered by multiplying its value with this factor.

float increaseFactor = 1.1f

When detected number of features doesn’t exceed the maximum in a cell, threshold is increased by multiplying its value with this factor.

struct FeatureMaintainer

#include <RawFeatureTrackerConfig.hpp>

FeatureMaintainer configuration structure.

Public Members

bool enable = true

Enable feature maintaining or not.

float minimumDistanceBetweenFeatures = 50

Used to filter out detected feature points that are too close. Requires sorting enabled in detector. Unit of measurement is squared euclidean distance in pixels.

float lostFeatureErrorThreshold = 50000

Optical flow measures the tracking error for every feature. If the point can’t be tracked or it’s out of the image it will set this error to a maximum value. This threshold defines the level where the tracking accuracy is considered too bad to keep the point.

float trackedFeatureThreshold = 200000

Once a feature was detected and we started tracking it, we need to update its Harris score on each image. This is needed because a feature point can disappear, or it can become too weak to be tracked. This threshold defines the point where such a feature must be dropped. As the goal of the algorithm is to provide longer tracks, we try to add strong points and track them until they are absolutely untrackable. This is why, this value is usually smaller than the detection threshold.

struct MotionEstimator

#include <RawFeatureTrackerConfig.hpp>

Used for feature reidentification between current and previous features.

Public Members

bool enable = true

Enable motion estimation or not.

Type type = Type::LUCAS_KANADE_OPTICAL_FLOW

Motion estimator algorithm type.

OpticalFlow opticalFlow

Optical flow configuration. Takes effect only if MotionEstimator algorithm type set to LUCAS_KANADE_OPTICAL_FLOW.

struct OpticalFlow

#include <RawFeatureTrackerConfig.hpp>

Optical flow configuration structure.

Public Members

std::int32_t pyramidLevels = AUTO

Number of pyramid levels, only for optical flow. AUTO means it’s decided based on input resolution: 3 if image width <= 640, else 4. Valid values are either 3/4 for VGA, 4 for 720p and above.

std::int32_t searchWindowWidth = 5

Image patch width used to track features. Must be an odd number, maximum 9. N means the algorithm will be able to track motion at most (N-1)/2 pixels in a direction per pyramid level. Increasing this number increases runtime

std::int32_t searchWindowHeight = 5

Image patch height used to track features. Must be an odd number, maximum 9. N means the algorithm will be able to track motion at most (N-1)/2 pixels in a direction per pyramid level. Increasing this number increases runtime

float epsilon = 0.01f

Feature tracking termination criteria. Optical flow will refine the feature position on each pyramid level until the displacement between two refinements is smaller than this value. Decreasing this number increases runtime.

std::int32_t maxIterations = 9

Feature tracking termination criteria. Optical flow will refine the feature position maximum this many times on each pyramid level. If the Epsilon criteria described in the previous chapter is not met after this number of iterations, the algorithm will continue with the current calculated value. Increasing this number increases runtime.

struct RawImageManipConfig : public dai::RawBuffer

#include <RawImageManipConfig.hpp>

RawImageManipConfig structure.

struct CropConfig

struct CropRect

struct FormatConfig

struct ResizeConfig

Public Members

bool keepAspectRatio = true

Whether to keep aspect ratio of input or not

struct RawImgDetections : public dai::RawBuffer

#include <RawImgDetections.hpp>

RawImgDetections structure.

struct RawImgFrame : public dai::RawBuffer

#include <RawImgFrame.hpp>

RawImgFrame structure.

struct Specs

struct RawIMUData : public dai::RawBuffer

struct RawNNData : public dai::RawBuffer

#include <RawNNData.hpp>

RawNNData structure.

struct RawSpatialImgDetections : public dai::RawBuffer

#include <RawSpatialImgDetections.hpp>

RawSpatialImgDetections structure.

struct RawSpatialLocationCalculatorConfig : public dai::RawBuffer

#include <RawSpatialLocationCalculatorConfig.hpp>

RawSpatialLocation configuration structure.

struct RawSpatialLocations : public dai::RawBuffer

#include <RawSpatialLocations.hpp>

RawSpatialLocations structure.

struct RawStereoDepthConfig : public dai::RawBuffer

#include <RawStereoDepthConfig.hpp>

RawStereoDepthConfig configuration structure.

Public Members

AlgorithmControl algorithmControl

Controls the flow of stereo algorithm - left-right check, subpixel etc.

PostProcessing postProcessing

Controls the postprocessing of disparity and/or depth map.

CensusTransform censusTransform

Census transform settings.

CostMatching costMatching

Cost matching settings.

CostAggregation costAggregation

Cost aggregation settings.

struct AlgorithmControl

Public Types

enum DepthAlign

Align the disparity/depth to the perspective of a rectified output, or center it

Values:

enumerator RECTIFIED_RIGHT

enumerator RECTIFIED_LEFT

enumerator CENTER

enum DepthUnit

Measurement unit for depth data

Values:

enumerator METER

enumerator CENTIMETER

enumerator MILLIMETER

enumerator INCH

enumerator FOOT

enumerator CUSTOM

Public Members

DepthAlign depthAlign = DepthAlign::RECTIFIED_RIGHT

Set the disparity/depth alignment to the perspective of a rectified output, or center it

DepthUnit depthUnit = DepthUnit::MILLIMETER

Measurement unit for depth data. Depth data is integer value, multiple of depth unit.

float customDepthUnitMultiplier = 1000.f

Custom depth unit multiplier, if custom depth unit is enabled, relative to 1 meter. A multiplier of 1000 effectively means depth unit in millimeter.

bool enableLeftRightCheck = true

Computes and combines disparities in both L-R and R-L directions, and combine them. For better occlusion handling

bool enableExtended = false

Disparity range increased from 95 to 190, combined from full resolution and downscaled images. Suitable for short range objects

bool enableSubpixel = false

Computes disparity with sub-pixel interpolation (5 fractional bits), suitable for long range

std::int32_t leftRightCheckThreshold = 10

Left-right check threshold for left-right, right-left disparity map combine, 0..128 Used only when left-right check mode is enabled. Defines the maximum difference between the confidence of pixels from left-right and right-left confidence maps

std::int32_t subpixelFractionalBits = 3

Number of fractional bits for subpixel mode

Valid values: 3,4,5

Defines the number of fractional disparities: 2^x

Median filter postprocessing is supported only for 3 fractional bits

struct CensusTransform

#include <RawStereoDepthConfig.hpp>

The basic cost function used by the Stereo Accelerator for matching the left and right images is the Census Transform. It works on a block of pixels and computes a bit vector which represents the structure of the image in that block. There are two types of Census Transform based on how the middle pixel is used: Classic Approach and Modified Census. The comparisons that are made between pixels can be or not thresholded. In some cases a mask can be applied to filter out only specific bits from the entire bit stream. All these approaches are: Classic Approach: Uses middle pixel to compare against all its neighbors over a defined window. Each comparison results in a new bit, that is 0 if central pixel is smaller, or 1 if is it bigger than its neighbor. Modified Census Transform: same as classic Census Transform, but instead of comparing central pixel with its neighbors, the window mean will be compared with each pixel over the window. Thresholding Census Transform: same as classic Census Transform, but it is not enough that a neighbor pixel to be bigger than the central pixel, it must be significant bigger (based on a threshold). Census Transform with Mask: same as classic Census Transform, but in this case not all of the pixel from the support window are part of the binary descriptor. We use a ma sk “M” to define which pixels are part of the binary descriptor (1), and which pixels should be skipped (0).

Public Types

enum KernelSize

Census transform kernel size possible values.

Values:

enumerator AUTO

enumerator KERNEL_5x5

enumerator KERNEL_7x7

enumerator KERNEL_7x9

Public Members

KernelSize kernelSize = KernelSize::AUTO

Census transform kernel size.

uint64_t kernelMask = 0

Census transform mask, default - auto, mask is set based on resolution and kernel size. Disabled for 400p input resolution. Enabled for 720p. 0XA82415 for 5x5 census transform kernel. 0XAA02A8154055 for 7x7 census transform kernel. 0X2AA00AA805540155 for 7x9 census transform kernel. Empirical values.

bool enableMeanMode = true

If enabled, each pixel in the window is compared with the mean window value instead of the central pixel.

uint32_t threshold = 0

Census transform comparison threshold value.

struct CostAggregation

#include <RawStereoDepthConfig.hpp>

Cost Aggregation is based on Semi Global Block Matching (SGBM). This algorithm uses a semi global technique to aggregate the cost map. Ultimately the idea is to build inertia into the stereo algorithm. If a pixel has very little texture information, then odds are the correct disparity for this pixel is close to that of the previous pixel considered. This means that we get improved results in areas with low texture.

Public Members

uint8_t divisionFactor = 1

Cost calculation linear equation parameters.

uint16_t horizontalPenaltyCostP1 = defaultPenaltyP1

Horizontal P1 penalty cost parameter.

uint16_t horizontalPenaltyCostP2 = defaultPenaltyP2

Horizontal P2 penalty cost parameter.

uint16_t verticalPenaltyCostP1 = defaultPenaltyP1

Vertical P1 penalty cost parameter.

uint16_t verticalPenaltyCostP2 = defaultPenaltyP2

Vertical P2 penalty cost parameter.

struct CostMatching

#include <RawStereoDepthConfig.hpp>

The matching cost is way of measuring the similarity of image locations in stereo correspondence algorithm. Based on the configuration parameters and based on the descriptor type, a linear equation is applied to computing the cost for each candidate disparity at each pixel.

Public Types

enum DisparityWidth

Disparity search range: 64 or 96 pixels are supported by the HW.

Values:

enumerator DISPARITY_64

enumerator DISPARITY_96

Public Members

DisparityWidth disparityWidth = DisparityWidth::DISPARITY_96

Disparity search range, default 96 pixels.

bool enableCompanding = false

Disparity companding using sparse matching. Matching pixel by pixel for N disparities. Matching every 2nd pixel for M disparitites. Matching every 4th pixel for T disparities. In case of 96 disparities: N=48, M=32, T=16. This way the search range is extended to 176 disparities, by sparse matching. Note: when enabling this flag only depth map will be affected, disparity map is not.

uint8_t invalidDisparityValue = 0

Used only for debug purposes, SW postprocessing handled only invalid value of 0 properly.

uint8_t confidenceThreshold = 245

Disparities with confidence value under this threshold are accepted. Higher confidence threshold means disparities with less confidence are accepted too.

LinearEquationParameters linearEquationParameters

Cost calculation linear equation parameters.

struct LinearEquationParameters

#include <RawStereoDepthConfig.hpp>

The linear equation applied for computing the cost is: COMB_COST = α*AD + β*(CTC<<3). CLAMP(COMB_COST >> 5, threshold). Where AD is the Absolute Difference between 2 pixels values. CTC is the Census Transform Cost between 2 pixels, based on Hamming distance (xor). The α and β parameters are subject to fine tuning by the user.

struct PostProcessing

#include <RawStereoDepthConfig.hpp>

Post-processing filters, all the filters are applied in disparity domain.

Public Members

MedianFilter median = MedianFilter::KERNEL_5x5

Set kernel size for disparity/depth median filtering, or disable

std::int16_t bilateralSigmaValue = 0

Sigma value for bilateral filter. 0 means disabled. A larger value of the parameter means that farther colors within the pixel neighborhood will be mixed together.

SpatialFilter spatialFilter

Edge-preserving filtering: This type of filter will smooth the depth noise while attempting to preserve edges.

TemporalFilter temporalFilter

Temporal filtering with optional persistence.

ThresholdFilter thresholdFilter

Threshold filtering. Filters out distances outside of a given interval.

SpeckleFilter speckleFilter

Speckle filtering. Removes speckle noise.

DecimationFilter decimationFilter

Decimation filter. Reduces disparity/depth map x/y complexity, reducing runtime complexity for other filters.

struct DecimationFilter

#include <RawStereoDepthConfig.hpp>

Decimation filter. Reduces the depth scene complexity. The filter runs on kernel sizes [2x2] to [8x8] pixels.

Public Types

enum DecimationMode

Decimation algorithm type.

Values:

enumerator PIXEL_SKIPPING

enumerator NON_ZERO_MEDIAN

enumerator NON_ZERO_MEAN

Public Members

std::uint32_t decimationFactor = 1

Decimation factor. Valid values are 1,2,3,4. Disparity/depth map x/y resolution will be decimated with this value.

DecimationMode decimationMode = DecimationMode::