Sample Codes for NVIDIA L4T ML

opencv_csi-camera.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

##
## @file        opencv_csi-camera.py
## @brief       Display the image captured by the CSI camera
## @author      Keitetsu
## @date        2022/01/15
## @copyright   Copyright (c) 2022 Keitetsu
## @par         License
##              This software is released under the MIT License.
##

import cv2

GST_STR = "nvarguscamerasrc sensor-id=0 ! \
    video/x-raw(memory:NVMM), width=(int)1920, height=(int)1080, format=(string)NV12, framerate=(fraction)30/1 ! \
    nvvidconv ! \
    video/x-raw, width=(int)1920, height=(int)1080, format=(string)BGRx ! \
    videoconvert ! \
    appsink max-buffers=1 drop=true"

if __name__ == "__main__":
    capture = cv2.VideoCapture(GST_STR, cv2.CAP_GSTREAMER)

    if capture.isOpened() is False:
        raise IOError("Failed to open camera device.")

    while True:
        retVal, frame = capture.read()

        if retVal == False:
            continue

        cv2.imshow("camera", frame)
        
        key = cv2.waitKey(1)
        if key == 27:   # ESC Key
            break

    capture.release()
    cv2.destroyAllWindows()

opencv_face-detection_csi-camera.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

##
## @file        opencv_face-detection_csi-camera.py
## @brief       Face Detection
## @author      Keitetsu
## @date        2022/01/16
## @copyright   Copyright (c) 2022 Keitetsu
## @par         License
##              This software is released under the MIT License.
##

import cv2

GST_STR = "nvarguscamerasrc sensor-id=0 ! \
    video/x-raw(memory:NVMM), width=(int)1920, height=(int)1080, format=(string)NV12, framerate=(fraction)30/1 ! \
    nvvidconv ! \
    video/x-raw, width=(int)1920, height=(int)1080, format=(string)BGRx ! \
    videoconvert ! \
    appsink max-buffers=1 drop=true"

cascade_file_path = "/usr/share/opencv4/haarcascades/haarcascade_frontalface_default.xml"

if __name__ == "__main__":
    capture = cv2.VideoCapture(GST_STR, cv2.CAP_GSTREAMER)
    if capture.isOpened() is False:
        raise IOError("Failed to open camera device.")

    cascade = cv2.CascadeClassifier(cascade_file_path)

    face_rect_line_color = (255, 255, 255)

    while True:
        retVal, frame = capture.read()
        if retVal == False:
            continue

        # グレースケールに変換
        frame_gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

        # 顔検出
        faces = cascade.detectMultiScale(frame_gray, scaleFactor=1.1, minNeighbors=3, minSize=(192, 192))

        # 顔が1つ以上検出されている場合に矩形を描画
        if len(faces) > 0:
            for (x, y, width, height) in faces:
                cv2.rectangle(frame, (x, y), (x + width, y + height), face_rect_line_color, thickness=2)
        
        cv2.imshow("camera", frame)
        key = cv2.waitKey(1)
        if key == 27:   # ESC Key
            break

    capture.release()
    cv2.destroyAllWindows()

run.sh

#!/bin/bash -eu

xhost +

sudo docker run \
    -it \
    --rm \
    --net=host \
    --runtime nvidia \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix/:/tmp/.X11-unix \
    -v ${PWD}:/data \
    nvcr.io/nvidia/l4t-ml:r32.6.1-py3

run_with_csi-camera.sh

#!/bin/bash -eu

xhost +

sudo docker run \
    -it \
    --rm \
    --net=host \
    --runtime nvidia \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix/:/tmp/.X11-unix \
    -v /tmp/argus_socket:/tmp/argus_socket \
    -v ${PWD}:/data \
    nvcr.io/nvidia/l4t-ml:r32.6.1-py3