# Mono & MobilenetSSD with spatial data

This example shows how to run MobileNetv2SSD on the rectified right input frame, and how to display both the preview, detections,
depth map and spatial information (X,Y,Z). It's similar to example [RGB &
MobilenetSSD](https://docs.luxonis.com/software/depthai/examples/rgb_mobilenet.md) except it has spatial data. X,Y,Z coordinates
are relative to the center of depth map.

setConfidenceThreshold - confidence threshold above which objects are detected

### Similar samples:

 * [Spatial location calculator](https://docs.luxonis.com/software/depthai/examples/spatial_location_calculator.md)
 * [Spatial object tracker on RGB](https://docs.luxonis.com/software/depthai/examples/spatial_object_tracker.md)
 * [RGB & MobilenetSSD with spatial data](https://docs.luxonis.com/software/depthai/examples/spatial_mobilenet.md)
 * [RGB & TinyYolo with spatial data](https://docs.luxonis.com/software/depthai/examples/spatial_tiny_yolo.md)

## Demo

## Setup

This example requires the DepthAI v3 API, see [installation instructions](https://docs.luxonis.com/software-v3/depthai.md).

## Source code

#### Python

```python
#!/usr/bin/env python3

from pathlib import Path
import sys
import cv2
import depthai as dai
import numpy as np
import time

'''
Mobilenet SSD device side decoding demo
  The "mobilenet-ssd" model is a Single-Shot multibox Detection (SSD) network intended
  to perform object detection. This model is implemented using the Caffe* framework.
  For details about this model, check out the repository <https://github.com/chuanqi305/MobileNet-SSD>.
'''

# Get argument first
nnPath = str((Path(__file__).parent / Path('../models/mobilenet-ssd_openvino_2021.4_6shave.blob')).resolve().absolute())
if len(sys.argv) > 1:
    nnPath = sys.argv[1]

if not Path(nnPath).exists():
    import sys
    raise FileNotFoundError(f'Required file/s not found, please run "{sys.executable} install_requirements.py"')

# MobilenetSSD label texts
labelMap = ["background", "aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow",
            "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]

syncNN = True

# Create pipeline
pipeline = dai.Pipeline()

# Define sources and outputs
monoLeft = pipeline.create(dai.node.MonoCamera)
monoRight = pipeline.create(dai.node.MonoCamera)
stereo = pipeline.create(dai.node.StereoDepth)
spatialDetectionNetwork = pipeline.create(dai.node.MobileNetSpatialDetectionNetwork)
imageManip = pipeline.create(dai.node.ImageManip)

xoutManip = pipeline.create(dai.node.XLinkOut)
nnOut = pipeline.create(dai.node.XLinkOut)
xoutDepth = pipeline.create(dai.node.XLinkOut)

xoutManip.setStreamName("right")
nnOut.setStreamName("detections")
xoutDepth.setStreamName("depth")

# Properties
imageManip.initialConfig.setResize(300, 300)
# The NN model expects BGR input. By default ImageManip output type would be same as input (gray in this case)
imageManip.initialConfig.setFrameType(dai.ImgFrame.Type.BGR888p)

monoLeft.setResolution(dai.MonoCameraProperties.SensorResolution.THE_400_P)
monoLeft.setCamera("left")
monoRight.setResolution(dai.MonoCameraProperties.SensorResolution.THE_400_P)
monoRight.setCamera("right")

# StereoDepth
stereo.setDefaultProfilePreset(dai.node.StereoDepth.PresetMode.HIGH_DENSITY)
stereo.setSubpixel(True)

# Define a neural network that will make predictions based on the source frames
spatialDetectionNetwork.setConfidenceThreshold(0.5)
spatialDetectionNetwork.setBlobPath(nnPath)
spatialDetectionNetwork.input.setBlocking(False)
spatialDetectionNetwork.setBoundingBoxScaleFactor(0.5)
spatialDetectionNetwork.setDepthLowerThreshold(100)
spatialDetectionNetwork.setDepthUpperThreshold(5000)

# Linking
monoLeft.out.link(stereo.left)
monoRight.out.link(stereo.right)

imageManip.out.link(spatialDetectionNetwork.input)
if syncNN:
    spatialDetectionNetwork.passthrough.link(xoutManip.input)
else:
    imageManip.out.link(xoutManip.input)

spatialDetectionNetwork.out.link(nnOut.input)

stereo.rectifiedRight.link(imageManip.inputImage)
stereo.depth.link(spatialDetectionNetwork.inputDepth)
spatialDetectionNetwork.passthroughDepth.link(xoutDepth.input)

# Connect to device and start pipeline
with dai.Device(pipeline) as device:

    # Output queues will be used to get the rgb frames and nn data from the outputs defined above
    previewQueue = device.getOutputQueue(name="right", maxSize=4, blocking=False)
    detectionNNQueue = device.getOutputQueue(name="detections", maxSize=4, blocking=False)
    depthQueue = device.getOutputQueue(name="depth", maxSize=4, blocking=False)

    rectifiedRight = None
    detections = []

    startTime = time.monotonic()
    counter = 0
    fps = 0
    color = (255, 255, 255)

    while True:
        inRectified = previewQueue.get()
        inDet = detectionNNQueue.get()
        inDepth = depthQueue.get()

        counter += 1
        currentTime = time.monotonic()
        if (currentTime - startTime) > 1:
            fps = counter / (currentTime - startTime)
            counter = 0
            startTime = currentTime

        rectifiedRight = inRectified.getCvFrame()

        depthFrame = inDepth.getFrame() # depthFrame values are in millimeters

        depth_downscaled = depthFrame[::4]
        if np.all(depth_downscaled == 0):
            min_depth = 0  # Set a default minimum depth value when all elements are zero
        else:
            min_depth = np.percentile(depth_downscaled[depth_downscaled != 0], 1)
        max_depth = np.percentile(depth_downscaled, 99)
        depthFrameColor = np.interp(depthFrame, (min_depth, max_depth), (0, 255)).astype(np.uint8)
        depthFrameColor = cv2.applyColorMap(depthFrameColor, cv2.COLORMAP_HOT)

        detections = inDet.detections

        # If the rectifiedRight is available, draw bounding boxes on it and show the rectifiedRight
        height = rectifiedRight.shape[0]
        width = rectifiedRight.shape[1]
        for detection in detections:
            roiData = detection.boundingBoxMapping
            roi = roiData.roi
            roi = roi.denormalize(depthFrameColor.shape[1], depthFrameColor.shape[0])
            topLeft = roi.topLeft()
            bottomRight = roi.bottomRight()
            xmin = int(topLeft.x)
            ymin = int(topLeft.y)
            xmax = int(bottomRight.x)
            ymax = int(bottomRight.y)
            cv2.rectangle(depthFrameColor, (xmin, ymin), (xmax, ymax), color, cv2.FONT_HERSHEY_SCRIPT_SIMPLEX)

            # Denormalize bounding box
            x1 = int(detection.xmin * width)
            x2 = int(detection.xmax * width)
            y1 = int(detection.ymin * height)
            y2 = int(detection.ymax * height)

            try:
                label = labelMap[detection.label]
            except:
                label = detection.label

            cv2.putText(rectifiedRight, str(label), (x1 + 10, y1 + 20), cv2.FONT_HERSHEY_TRIPLEX, 0.5, 255)
            cv2.putText(rectifiedRight, "{:.2f}".format(detection.confidence*100), (x1 + 10, y1 + 35), cv2.FONT_HERSHEY_TRIPLEX, 0.5, 255)
            cv2.putText(rectifiedRight, f"X: {int(detection.spatialCoordinates.x)} mm", (x1 + 10, y1 + 50), cv2.FONT_HERSHEY_TRIPLEX, 0.5, 255)
            cv2.putText(rectifiedRight, f"Y: {int(detection.spatialCoordinates.y)} mm", (x1 + 10, y1 + 65), cv2.FONT_HERSHEY_TRIPLEX, 0.5, 255)
            cv2.putText(rectifiedRight, f"Z: {int(detection.spatialCoordinates.z)} mm", (x1 + 10, y1 + 80), cv2.FONT_HERSHEY_TRIPLEX, 0.5, 255)

            cv2.rectangle(rectifiedRight, (x1, y1), (x2, y2), color, cv2.FONT_HERSHEY_SIMPLEX)

        cv2.putText(rectifiedRight, "NN fps: {:.2f}".format(fps), (2, rectifiedRight.shape[0] - 4), cv2.FONT_HERSHEY_TRIPLEX, 0.4, color)
        cv2.imshow("depth", depthFrameColor)
        cv2.imshow("rectified right", rectifiedRight)

        if cv2.waitKey(1) == ord('q'):
            break
```

#### C++

```cpp
#include <chrono>
#include <iostream>

// Includes common necessary includes for development using depthai library
#include "depthai/depthai.hpp"

static const std::vector<std::string> labelMap = {"background", "aeroplane", "bicycle",     "bird",  "boat",        "bottle", "bus",
                                                  "car",        "cat",       "chair",       "cow",   "diningtable", "dog",    "horse",
                                                  "motorbike",  "person",    "pottedplant", "sheep", "sofa",        "train",  "tvmonitor"};

static std::atomic<bool> syncNN{true};

int main(int argc, char** argv) {
    using namespace std;
    using namespace std::chrono;
    std::string nnPath(BLOB_PATH);

    // If path to blob specified, use that
    if(argc > 1) {
        nnPath = std::string(argv[1]);
    }

    // Print which blob we are using
    printf("Using blob at path: %s\n", nnPath.c_str());

    // Create pipeline
    dai::Pipeline pipeline;

    // Define sources and outputs
    auto monoLeft = pipeline.create<dai::node::MonoCamera>();
    auto monoRight = pipeline.create<dai::node::MonoCamera>();
    auto stereo = pipeline.create<dai::node::StereoDepth>();
    auto spatialDetectionNetwork = pipeline.create<dai::node::MobileNetSpatialDetectionNetwork>();
    auto imageManip = pipeline.create<dai::node::ImageManip>();

    auto xoutManip = pipeline.create<dai::node::XLinkOut>();
    auto nnOut = pipeline.create<dai::node::XLinkOut>();
    auto xoutDepth = pipeline.create<dai::node::XLinkOut>();

    xoutManip->setStreamName("right");
    nnOut->setStreamName("detections");
    xoutDepth->setStreamName("depth");

    // Properties
    imageManip->initialConfig.setResize(300, 300);
    // The NN model expects BGR input. By default ImageManip output type would be same as input (gray in this case)
    imageManip->initialConfig.setFrameType(dai::ImgFrame::Type::BGR888p);

    monoLeft->setResolution(dai::MonoCameraProperties::SensorResolution::THE_400_P);
    monoLeft->setCamera("left");
    monoRight->setResolution(dai::MonoCameraProperties::SensorResolution::THE_400_P);
    monoRight->setCamera("right");

    // StereoDepth
    stereo->setDefaultProfilePreset(dai::node::StereoDepth::PresetMode::HIGH_DENSITY);

    // Define a neural network that will make predictions based on the source frames
    spatialDetectionNetwork->setConfidenceThreshold(0.5f);
    spatialDetectionNetwork->setBlobPath(nnPath);
    spatialDetectionNetwork->input.setBlocking(false);
    spatialDetectionNetwork->setBoundingBoxScaleFactor(0.5);
    spatialDetectionNetwork->setDepthLowerThreshold(100);
    spatialDetectionNetwork->setDepthUpperThreshold(5000);

    // Linking
    monoLeft->out.link(stereo->left);
    monoRight->out.link(stereo->right);

    imageManip->out.link(spatialDetectionNetwork->input);
    if(syncNN) {
        spatialDetectionNetwork->passthrough.link(xoutManip->input);
    } else {
        imageManip->out.link(xoutManip->input);
    }

    spatialDetectionNetwork->out.link(nnOut->input);

    stereo->rectifiedRight.link(imageManip->inputImage);
    stereo->depth.link(spatialDetectionNetwork->inputDepth);
    spatialDetectionNetwork->passthroughDepth.link(xoutDepth->input);

    // Connect to device and start pipeline
    dai::Device device(pipeline);

    // Output queues will be used to get the rgb frames and nn data from the outputs defined above
    auto previewQueue = device.getOutputQueue("right", 4, false);
    auto detectionNNQueue = device.getOutputQueue("detections", 4, false);
    auto depthQueue = device.getOutputQueue("depth", 4, false);

    auto startTime = steady_clock::now();
    int counter = 0;
    float fps = 0;
    auto color = cv::Scalar(255, 255, 255);

    while(true) {
        auto inRectified = previewQueue->get<dai::ImgFrame>();
        auto inDet = detectionNNQueue->get<dai::SpatialImgDetections>();
        auto inDepth = depthQueue->get<dai::ImgFrame>();

        counter++;
        auto currentTime = steady_clock::now();
        auto elapsed = duration_cast<duration<float>>(currentTime - startTime);
        if(elapsed > seconds(1)) {
            fps = counter / elapsed.count();
            counter = 0;
            startTime = currentTime;
        }

        cv::Mat rectifiedRight = inRectified->getCvFrame();

        cv::Mat depthFrame = inDepth->getFrame();  // depthFrame values are in millimeters
        cv::Mat depthFrameColor;

        cv::normalize(depthFrame, depthFrameColor, 255, 0, cv::NORM_INF, CV_8UC1);
        cv::equalizeHist(depthFrameColor, depthFrameColor);
        cv::applyColorMap(depthFrameColor, depthFrameColor, cv::COLORMAP_HOT);

        auto detections = inDet->detections;

        for(auto& detection : detections) {
            auto roiData = detection.boundingBoxMapping;
            auto roi = roiData.roi;
            roi = roi.denormalize(depthFrameColor.cols, depthFrameColor.rows);
            auto topLeft = roi.topLeft();
            auto bottomRight = roi.bottomRight();
            auto xmin = (int)topLeft.x;
            auto ymin = (int)topLeft.y;
            auto xmax = (int)bottomRight.x;
            auto ymax = (int)bottomRight.y;
            cv::rectangle(depthFrameColor, cv::Rect(cv::Point(xmin, ymin), cv::Point(xmax, ymax)), color, cv::FONT_HERSHEY_SIMPLEX);

            int x1 = detection.xmin * rectifiedRight.cols;
            int y1 = detection.ymin * rectifiedRight.rows;
            int x2 = detection.xmax * rectifiedRight.cols;
            int y2 = detection.ymax * rectifiedRight.rows;

            uint32_t labelIndex = detection.label;
            std::string labelStr = to_string(labelIndex);
            if(labelIndex < labelMap.size()) {
                labelStr = labelMap[labelIndex];
            }
            cv::putText(rectifiedRight, labelStr, cv::Point(x1 + 10, y1 + 20), cv::FONT_HERSHEY_TRIPLEX, 0.5, 255);
            std::stringstream confStr;
            confStr << std::fixed << std::setprecision(2) << detection.confidence * 100;
            cv::putText(rectifiedRight, confStr.str(), cv::Point(x1 + 10, y1 + 35), cv::FONT_HERSHEY_TRIPLEX, 0.5, 255);

            std::stringstream depthX;
            depthX << "X: " << (int)detection.spatialCoordinates.x << " mm";
            cv::putText(rectifiedRight, depthX.str(), cv::Point(x1 + 10, y1 + 50), cv::FONT_HERSHEY_TRIPLEX, 0.5, 255);
            std::stringstream depthY;
            depthY << "Y: " << (int)detection.spatialCoordinates.y << " mm";
            cv::putText(rectifiedRight, depthY.str(), cv::Point(x1 + 10, y1 + 65), cv::FONT_HERSHEY_TRIPLEX, 0.5, 255);
            std::stringstream depthZ;
            depthZ << "Z: " << (int)detection.spatialCoordinates.z << " mm";
            cv::putText(rectifiedRight, depthZ.str(), cv::Point(x1 + 10, y1 + 80), cv::FONT_HERSHEY_TRIPLEX, 0.5, 255);

            cv::rectangle(rectifiedRight, cv::Rect(cv::Point(x1, y1), cv::Point(x2, y2)), color, cv::FONT_HERSHEY_SIMPLEX);
        }

        std::stringstream fpsStr;
        fpsStr << std::fixed << std::setprecision(2) << fps;
        cv::putText(rectifiedRight, fpsStr.str(), cv::Point(2, rectifiedRight.rows - 4), cv::FONT_HERSHEY_TRIPLEX, 0.4, color);

        cv::imshow("depth", depthFrameColor);
        cv::imshow("rectified right", rectifiedRight);

        int key = cv::waitKey(1);
        if(key == 'q' || key == 'Q') {
            return 0;
        }
    }
    return 0;
}
```

## Pipeline

### Need assistance?

Head over to [Discussion Forum](https://discuss.luxonis.com/) for technical support or any other questions you might have.