Script change pipeline flow

This example shows how you can change the flow of data inside your pipeline in runtime using the Script node. In this example, we send a message from the host to choose whether we want Script node to forwards color frame to the MobileNetDetectionNetwork.

Demo

Pipeline Graph

Setup

Please run the install script to download all required dependencies. Please note that this script must be ran from git context, so you have to download the depthai-python repository first and then run the script

Command Line

1git clone https://github.com/luxonis/depthai-python.git
2cd depthai-python/examples
3python3 install_requirements.py

For additional information, please follow the installation guide.

Source code

Python

C++

Python

GitHub

1#!/usr/bin/env python3
2import depthai as dai
3import cv2
4from pathlib import Path
5import numpy as np
6
7parentDir = Path(__file__).parent
8nnPath = str((parentDir / Path('../models/mobilenet-ssd_openvino_2021.4_5shave.blob')).resolve().absolute())
9
10pipeline = dai.Pipeline()
11
12cam = pipeline.createColorCamera()
13cam.setBoardSocket(dai.CameraBoardSocket.CAM_A)
14cam.setInterleaved(False)
15cam.setIspScale(2,3)
16cam.setVideoSize(720,720)
17cam.setPreviewSize(300,300)
18
19xoutRgb = pipeline.create(dai.node.XLinkOut)
20xoutRgb.setStreamName('rgb')
21cam.video.link(xoutRgb.input)
22
23script = pipeline.createScript()
24
25xin = pipeline.create(dai.node.XLinkIn)
26xin.setStreamName('in')
27xin.out.link(script.inputs['toggle'])
28
29cam.preview.link(script.inputs['rgb'])
30script.setScript("""
31    toggle = False
32    while True:
33        msg = node.io['toggle'].tryGet()
34        if msg is not None:
35            toggle = msg.getData()[0]
36            node.warn('Toggle! Perform NN inferencing: ' + str(toggle))
37
38        frame = node.io['rgb'].get()
39
40        if toggle:
41            node.io['nn'].send(frame)
42""")
43
44nn = pipeline.create(dai.node.MobileNetDetectionNetwork)
45nn.setBlobPath(nnPath)
46script.outputs['nn'].link(nn.input)
47
48xoutNn = pipeline.create(dai.node.XLinkOut)
49xoutNn.setStreamName('nn')
50nn.out.link(xoutNn.input)
51
52# Connect to device with pipeline
53with dai.Device(pipeline) as device:
54    inQ = device.getInputQueue("in")
55    qRgb = device.getOutputQueue("rgb")
56    qNn = device.getOutputQueue("nn")
57
58    runNn = False
59
60    def frameNorm(frame, bbox):
61        normVals = np.full(len(bbox), frame.shape[0])
62        normVals[::2] = frame.shape[1]
63        return (np.clip(np.array(bbox), 0, 1) * normVals).astype(int)
64
65    color = (255, 127, 0)
66    def drawDetections(frame, detections):
67        for detection in detections:
68            bbox = frameNorm(frame, (detection.xmin, detection.ymin, detection.xmax, detection.ymax))
69            cv2.putText(frame, f"{int(detection.confidence * 100)}%", (bbox[0] + 10, bbox[1] + 20), cv2.FONT_HERSHEY_TRIPLEX, 0.5, color)
70            cv2.rectangle(frame, (bbox[0], bbox[1]), (bbox[2], bbox[3]), color, 2)
71
72
73    while True:
74        frame = qRgb.get().getCvFrame()
75
76        if qNn.has():
77            detections = qNn.get().detections
78            drawDetections(frame, detections)
79
80        cv2.putText(frame, f"NN inferencing: {runNn}", (20,20), cv2.FONT_HERSHEY_TRIPLEX, 0.7, color)
81        cv2.imshow('Color frame', frame)
82
83        key = cv2.waitKey(1)
84        if key == ord('q'):
85            break
86        elif key == ord('t'):
87            runNn = not runNn
88            print(f"{'Enabling' if runNn else 'Disabling'} NN inferencing")
89            buf = dai.Buffer()
90            buf.setData(runNn)
91            inQ.send(buf)

Need assistance?

Head over to Discussion Forum for technical support or any other questions you might have.