Automation, sensing and robotics • Re: Tracking person by camera

Code:

#!/usr/bin/env python3"""Example module for Hailo Detection with tracking."""import argparseimport cv2from picamera2 import MappedArray, Picamera2, Previewfrom picamera2.devices import Hailofrom picamera2.encoders import H264Encoderfrom picamera2.outputs import CircularOutputfrom libcamera import controlsimport timeimport pantilthatimport datetimeimport osimport globimport shutil# objects to detectobjects = ["clock","person"]# centralise pan and tilt camerahposn = 0vposn = 0pantilthat.pan(hposn)pantilthat.tilt(vposn)# initialisev_width      = 1280  # video widthv_height     = 960   # video heightv_length     = 5     # seconds, minimum video lengthpre_frames   = 5     # seconds, defines length of pre-detection bufferfps          = 25    # video frame ratemp4_fps      = 25    # mp4 frame ratemp4_timer    = 10    # seconds, convert h264s to mp4s after this time if no detectionsmp4_anno     = 1     # show time on video, 1 = yes, 0 = noshow_detects = 1    # show detections on video, 1 = yes, 0 = noram_limit    = 150Users  = []Users.append(os.getlogin())user   = Users[0]h_user = "/home/" + os.getlogin( )# mp4_annotation parameterscolour    = (255, 255, 255)origin    = (184, int(v_height - 35))font      = cv2.FONT_HERSHEY_SIMPLEXscale     = 1thickness = 2def Camera_Version():    global cam1    if os.path.exists('/run/shm/libcams.txt'):        os.rename('/run/shm/libcams.txt', '/run/shm/oldlibcams.txt')    os.system("rpicam-vid --list-cameras >> /run/shm/libcams.txt")    time.sleep(0.5)    # read libcams.txt file    camstxt = []    with open("/run/shm/libcams.txt", "r") as file:        line = file.readline()        while line:            camstxt.append(line.strip())            line = file.readline()    cam1 = camstxt[2][4:10]Camera_Version()def extract_detections(hailo_output, w, h, class_names, threshold=0.5):    """Extract detections from the HailoRT-postprocess output."""    results = []    for class_id, detections in enumerate(hailo_output):        for detection in detections:            score = detection[4]            if score >= threshold:                y0, x0, y1, x1 = detection[:4]                bbox = (int(x0 * w), int(y0 * h), int(x1 * w), int(y1 * h))                results.append([class_names[class_id], bbox, score])    return results# apply timestamp to videosdef apply_timestamp(request):  global mp4_anno  if mp4_anno == 1:      timestamp = time.strftime("%Y/%m/%d %T")      with MappedArray(request, "main") as m:          lst = list(origin)          lst[0] += 365          lst[1] -= 20          end_point = tuple(lst)          cv2.rectangle(m.array, origin, end_point, (0,0,0), -1)           cv2.putText(m.array, timestamp, origin, font, scale, colour, thickness)def draw_objects(request):    current_detections = detections    if current_detections:        with MappedArray(request, "main") as m:            for class_name, bbox, score in current_detections:                x0, y0, x1, y1 = bbox                label = f"{class_name} %{int(score * 100)}"                cv2.rectangle(m.array, (x0, y0), (x1, y1), (0, 255, 0, 0), 2)                cv2.putText(m.array, label, (x0 + 5, y0 + 15),                            cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0, 0), 1, cv2.LINE_AA)if __name__ == "__main__":    # Parse command-line arguments.    parser = argparse.ArgumentParser(description="Detection Example")    parser.add_argument("-m", "--model", help="Path for the HEF model.",                        default="/usr/share/hailo-models/yolov8s_h8l.hef")    parser.add_argument("-l", "--labels", default="coco.txt",                        help="Path to a text file containing labels.")    parser.add_argument("-s", "--score_thresh", type=float, default=0.5,                        help="Score threshold, must be a float between 0 and 1.")    args = parser.parse_args()    # Get the Hailo model, the input size it wants, and the size of our preview stream.    with Hailo(args.model) as hailo:        model_h, model_w, _ = hailo.get_input_shape()        video_w, video_h = v_width, v_height        # Load class names from the labels file        with open(args.labels, 'r', encoding="utf-8") as f:            class_names = f.read().splitlines()        # The list of detected objects to draw.        detections = None        encoding     = False        # get free ram space        st = os.statvfs("/run/shm/")        freeram = (st.f_bavail * st.f_frsize)/1100000        # Configure and start Picamera2.        with Picamera2() as picam2:            main = {'size': (video_w, video_h), 'format': 'XRGB8888'}            lores = {'size': (model_w, model_h), 'format': 'RGB888'}            if cam1 == "imx708":                controls2 = {'FrameRate': fps,"AfMode": controls.AfModeEnum.Continuous,"AfTrigger": controls.AfTriggerEnum.Start}            else:                controls2 = {'FrameRate': fps}            config = picam2.create_preview_configuration(main, lores=lores, controls=controls2)            picam2.configure(config)            encoder = H264Encoder(2000000, repeat=True)            encoder.output = CircularOutput(buffersize = pre_frames * fps)            picam2.pre_callback = apply_timestamp            picam2.start_preview(Preview.QTGL, x=0, y=0, width=480, height=480)            picam2.start()            picam2.start_encoder(encoder)            encoding = False            if show_detects == 1:                picam2.pre_callback = draw_objects            start = time.monotonic()            startmp4 = time.monotonic()                        # Process each low resolution camera frame.            while True:                frame = picam2.capture_array('lores')                # Run inference on the preprocessed frame                results = hailo.run(frame)                # Extract detections from the inference results                detections = extract_detections(results, video_w, video_h, class_names, args.score_thresh)                # detection                if len(detections) != 0:                    for e in range(0,len(detections)):                        if detections[e][0] in objects:                            value = float(detections[e][2])                            obj = detections[e][0]                            if  value < 1 and value > args.score_thresh:                                start = time.monotonic()                                startrec = time.monotonic()                                startmp4 = time.monotonic()                                # start recording                                if not encoding and freeram > ram_limit:                                    now = datetime.datetime.now()                                    timestamp = now.strftime("%y%m%d_%H%M%S")                                    encoder.output.fileoutput = "/run/shm/" + str(timestamp) + '.h264'                                    encoder.output.start()                                    encoding = True                                    print("Start Record", timestamp)                                # determine object horzontal position                                hpos = int((detections[e][1][0] + detections[e][1][2])/2)                                hpos2 = hpos/400                                hpos2 = min(hpos2,1)                                if hpos > (v_width/2) + 50:                                    hposn -= hpos2                                elif hpos < (v_width/2) - 50:                                    hposn += hpos2                                if hposn > 90 or hposn < -90:                                    hposn = 0                                # determine object vertical position                                vpos = int((detections[e][1][1] + detections[e][1][3])/2)                                vpos2 = vpos/320                                vpos2 = min(vpos2,1)                                if vpos > (v_height/2) + 50:                                    vposn += vpos2                                elif vpos < (v_height/2) - 50:                                    vposn -= vpos2                                if vposn > 30 or vposn < -30:                                    vposn = 0                                # pan and tilt                                pantilthat.pan(hposn)                                pantilthat.tilt(vposn)                                                    else:                            # centralise if object lost for greater than 2 seconds                            if time.monotonic() - start > 2:                                print ("lost")                                pantilthat.pan(0)                                pantilthat.tilt(0)                                hposn = 0                                vposn = 0                                start = time.monotonic()                # stop recording                if encoding and (time.monotonic() - startrec > v_length or freeram <= ram_limit):                    now = datetime.datetime.now()                    timestamp2 = now.strftime("%y%m%d_%H%M%S")                    print("Stopped Record", timestamp2)                    encoder.output.stop()                    encoding = False                    startmp4 = time.monotonic()                # make mp4s                if time.monotonic() - startmp4 > mp4_timer and not encoding:                    startmp4 = time.monotonic()                    # convert h264 to mp4                    h264s = glob.glob('/run/shm/2*.h264')                    h264s.sort(reverse = False)                    for x in range(0,len(h264s)):                        print(h264s[x][:-5] + '.mp4')                        cmd = 'ffmpeg -framerate ' + str(mp4_fps) + ' -i ' + h264s[x] + " -c copy " + h264s[x][:-5] + '.mp4'                        os.system(cmd)                        os.remove(h264s[x])                        print("Saved",h264s[x][:-5] + '.mp4')                    Videos = glob.glob('/run/shm/*.mp4')                    Videos.sort()                    # move Video RAM mp4s to SD card                    for xx in range(0,len(Videos)):                        if not os.path.exists(h_user + "/" + '/Videos/' + Videos[xx]):                            shutil.move(Videos[xx], h_user + '/Videos/')