def main(): args = parse_args() cam = Camera(args) cam.open() # import pdb # pdb.set_trace() if not cam.is_opened: sys.exit('[INFO] Failed to open camera!') cls_dict = get_cls_dict('coco') yolo_dim = int(args.model.split('-')[-1]) # 416 or 608 trt_yolov3 = TrtYOLOv3(args.model, (yolo_dim, yolo_dim)) print('[INFO] Camera: starting') cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'TensorRT YOLOv3 Detector') vis = BBoxVisualization(cls_dict) loop_and_detect(cam, args.runtime, trt_yolov3, conf_th=0.3, vis=vis, window_name=WINDOW_NAME) print('[INFO] Program: stopped') cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() if args.category_num <= 0: raise SystemExit('ERROR: bad category_num (%d)!' % args.category_num) if not os.path.isfile('yolo/%s.trt' % args.model): raise SystemExit('ERROR: file (yolo/%s.trt) not found!' % args.model) cam = Camera(args) cam.open() if not cam.is_opened: raise SystemExit('ERROR: failed to open camera!') cls_dict = get_cls_dict(args.category_num) yolo_dim = args.model.split('-')[-1] if 'x' in yolo_dim: dim_split = yolo_dim.split('x') if len(dim_split) != 2: raise SystemExit('ERROR: bad yolo_dim (%s)!' % yolo_dim) w, h = int(dim_split[0]), int(dim_split[1]) else: h = w = int(yolo_dim) if h % 32 != 0 or w % 32 != 0: raise SystemExit('ERROR: bad yolo_dim (%s)!' % yolo_dim) trt_yolo = TrtYOLO(args.model, (h, w), args.category_num) cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT YOLO Demo') vis = BBoxVisualization(cls_dict) loop_and_detect(cam, trt_yolo, conf_th=0.3, vis=vis) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') cls_dict = get_cls_dict('coco') yolo_dim = int(args.model.split('-')[-1]) # 416 or 608 trt_yolov3 = TrtYOLOv3(args.model, (yolo_dim, yolo_dim)) cam.start() # open_window(WINDOW_NAME, args.image_width, args.image_height, # 'Camera TensorRT YOLOv3 Demo') vis = BBoxVisualization(cls_dict) # for video # loop_and_detect(cam, trt_yolov3, conf_th=0.3, vis=vis) # for single file detect_demo(cam, trt_yolov3, conf_th=0.3, vis=vis) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') cls_dict = get_cls_dict(args.model) trt_ssd = TrtSSD(args.model, INPUT_HW) cam.start() if args.use_console: loop_and_detect_console(cam, trt_ssd, conf_th=0.3, loop=args.loop, cls_dict=cls_dict) else: open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT SSD Demo for Jetson Nano') vis = BBoxVisualization(cls_dict) loop_and_detect(cam, trt_ssd, conf_th=0.3, vis=vis) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() if args.category_num <= 0: raise SystemExit('Bad category_num: %d!' % args.category_num) cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') cls_dict = get_cls_dict(args.category_num) yolo_dim = int(args.model.split('-')[-1]) if yolo_dim not in (288, 416, 608): raise SystemExit('Bad yolo_dim: %d!\nPlease make sure the model file name contains the correct dimension...' % yolo_dim) trt_yolov3 = TrtYOLOv3(args.model, (yolo_dim, yolo_dim), args.category_num) cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT YOLOv3 Demo') vis = BBoxVisualization(cls_dict) loop_and_detect(cam, trt_yolov3, conf_th=0.3, vis=vis) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() #YOLO INIT #cls_dict = get_cls_dict('coco') cls_dict = get_cls_dict('deepfamily') print("classes count : ", len(cls_dict)) yolo_dim = int(args.model.split('-')[-1]) # 416 or 608 print("yolo_dim : ", yolo_dim) trt_yolov3 = TrtYOLOv3(args.model, (yolo_dim, yolo_dim)) #CAMERA cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') cam.start() #CAM-WINDOW open_window(WINDOW_NAME, args.image_width, args.image_height, 'DEEPFAMILY PROJECT - TensorRT YOLOv3') vis = BBoxVisualization(cls_dict) #DETECT-LOOP loop_and_detect(cam, trt_yolov3, conf_th=0.95, vis=vis) #loop_and_detect(cam, trt_yolov3, conf_th=0.95) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') trt_ssd = TrtSSD(args.model) cam.start() # initialize bot logger.info('initialize robot') robot = Robot() logger.info('starting to loop and detect') loop_and_detect(cam=cam, trt_ssd=trt_ssd, conf_th=0.3, robot=robot, model=args.model) logger.info('cleaning up') robot.stop() cam.stop() cam.release()
def main(): logging.basicConfig(level=logging.INFO) logger = logging.getLogger(__name__) # Ask tensorflow logger not to propagate logs to parent (which causes # duplicated logging) logging.getLogger('tensorflow').propagate = False global args args = parse_args() logger.info('called with args: %s' % args) # build the class (index/name) dictionary from labelmap file logger.info('reading label map') cls_dict = read_label_map(args.labelmap_file) pb_path = './data/{}_trt.pb'.format(args.model) log_path = './logs/{}_trt'.format(args.model) if args.do_build: logger.info('building TRT graph and saving to pb: %s' % pb_path) build_trt_pb(args.model, pb_path) logger.info('opening camera device/file') cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') logger.info('loading TRT graph from pb: %s' % pb_path) trt_graph = load_trt_pb(pb_path) logger.info('starting up TensorFlow session') tf_config = tf.ConfigProto() tf_config.gpu_options.allow_growth = True #tf_sess = tf.Session(config=tf_config, graph=trt_graph) -- Vincent #Solve : "unable to satfisfy explicit device /dev/CPU:0 -- Vincent tf_sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True, log_device_placement=True), graph=trt_graph) if args.do_tensorboard: logger.info('writing graph summary to TensorBoard') write_graph_tensorboard(tf_sess, log_path) logger.info('warming up the TRT graph with a dummy image') od_type = 'faster_rcnn' if 'faster_rcnn' in args.model else 'ssd' dummy_img = np.zeros((720, 1280, 3), dtype=np.uint8) _, _, _ = detect(dummy_img, tf_sess, conf_th=.3, od_type=od_type) cam.start() # ask the camera to start grabbing images # grab image and do object detection (until stopped by user) logger.info('starting to loop and detect') vis = BBoxVisualization(cls_dict) open_display_window(cam.img_height, cam.img_width) result = loop_and_detect(cam, tf_sess, args.conf_th, vis, od_type=od_type) logger.info('cleaning up') cam.stop() # terminate the sub-thread in camera tf_sess.close() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') mtcnn = TrtMtcnn() cam.start() open_window(WINDOW_NAME, width=640, height=480, title='MTCNN Window') detect_faces(cam, mtcnn) cam.stop() cam.release() cv2.destroyAllWindows() del mtcnn
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: raise SystemExit('ERROR: failed to open camera!') cls_dict = get_cls_dict(args.model.split('_')[-1]) trt_ssd = TrtSSD(args.model, INPUT_HW) cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT SSD Demo for Jetson Nano') vis = BBoxVisualization(cls_dict) loop_and_detect(cam, trt_ssd, conf_th=0.3, vis=vis) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() labels = np.loadtxt('googlenet/synset_words.txt', str, delimiter='\t') cam = Camera(args) cam.open() if not cam.is_opened: raise SystemExit('ERROR: failed to open camera!') # initialize the tensorrt googlenet engine net = PyTrtGooglenet(DEPLOY_ENGINE, ENGINE_SHAPE0, ENGINE_SHAPE1) cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT GoogLeNet Demo for Jetson Nano') loop_and_classify(cam, net, labels, args.crop_center) cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') mtcnn = TrtMtcnn() cam.start() open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT MTCNN Demo for Jetson TX2') loop_and_detect(cam, mtcnn, args.minsize) cam.stop() cam.release() cv2.destroyAllWindows() del (mtcnn)
def main(): args = parse_args() labels = np.loadtxt('googlenet/synset_words.txt', str, delimiter='\t') cam = Camera(args) cam.open() if not cam.is_opened: sys.exit('Failed to open camera!') cam.start() # let camera start grabbing frames open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT GoogLeNet Demo for Jetson Nano') condition = threading.Condition() trt_thread = TrtGooglenetThread(condition, cam, labels, args.crop_center) trt_thread.start() # start the child thread loop_and_display(condition) trt_thread.stop() # stop the child thread cam.stop() cam.release() cv2.destroyAllWindows()
def main(): args = parse_args() cam = Camera(args) cam.open() if not cam.is_opened: raise SystemExit('ERROR: failed to open camera!') cls_dict = get_cls_dict(args.model.split('_')[-1]) cuda.init() # init pycuda driver cam.start() # let camera start grabbing frames open_window(WINDOW_NAME, args.image_width, args.image_height, 'Camera TensorRT SSD Demo for Jetson Nano') vis = BBoxVisualization(cls_dict) condition = threading.Condition() trt_thread = TrtThread(condition, cam, args.model, conf_th=0.3) trt_thread.start() # start the child thread loop_and_display(condition, vis) trt_thread.stop() # stop the child thread cam.stop() cam.release() cv2.destroyAllWindows()
def loop_and_detect(cam, tf_sess, conf_th, vis, od_type): """Loop, grab images from camera, and do object detection. # Arguments cam: the camera object (video source). tf_sess: TensorFlow/TensorRT session to run SSD object detection. conf_th: confidence/score threshold for object detection. vis: for visualization. """ show_fps = True full_scrn = False fps = 0.0 #tic = time.time() tic = 0 toc = 0 global rects, ct, temp global person_x, person_y global leave_zone global leave_zone_2 global leave_zone_counter global enter_zone zone_x_bed = 0 zone_y_bed = 0 zone_x_clean = 0 zone_y_clean = 0 #Boundary boxes for RTSP (low resolution) zone_x_min_bed, zone_y_min_bed, zone_x_max_bed, zone_y_max_bed = 366, 369, 521, 667 zone_x_min_clean, zone_y_min_clean, zone_x_max_clean, zone_y_max_clean = 194, 300, 330, 420 zone_x_min_door, zone_y_min_door, zone_x_max_door, zone_y_max_door = 620, 151, 674, 414 zone_x_min_alchol, zone_y_min_alchol, zone_x_max_alchol, zone_y_max_alchol = 430, 329, 470, 356 distance_thres_bed = 165 distance_thres_clean = 105 distance_thres_alchol = 100 counter_msg = 0 fail_msg = 0 pass_msg = 0 global hand_wash_status, args, client hd = 0 wash_delay = 0 invalid_id = [] invalid_id.append(999) enter, leave = False, False restart_flag = False #restart issue backup_label = None #restart issue none_buff = 0 #restart issue previous_id = 999 personal_status = [] for i in range(0, 1000): personal_status.append(0) #CSV Log File if (5 > 2): while True: #if cv2.getWindowProperty(WINDOW_NAME, 0) < 0: # Check to see if the user has closed the display window. # If yes, terminate the while loop. # break if (restart_flag == True): cam = Camera(args) cam.open() cam.start() print("Camera is opened!") dummy_img = np.zeros((720, 1280, 3), dtype=np.uint8) _, _, _ = detect(dummy_img, tf_sess, conf_th=.3, od_type=od_type) print("Loading dummy image!") restart_flag = False rects = [] img = cam.read() if img is not None: img = cv2.flip(img, 0) #check mqtt status #client.on_message = on_message #client.loop_forever() box, conf, cls = detect(img, tf_sess, conf_th, od_type=od_type) img, hd = vis.draw_bboxes(img, box, conf, cls) cv2.rectangle(img, (zone_x_min_bed, zone_y_min_bed), (zone_x_max_bed, zone_y_max_bed), (255, 102, 255), 2) zone_x_bed = int((zone_x_min_bed + zone_x_max_bed) / 2.0) zone_y_bed = int((zone_y_min_bed + zone_y_max_bed) / 2.0) cv2.circle(img, (zone_x_bed, zone_y_bed), 4, (255, 102, 255), -1) cv2.putText(img, "Patient", (zone_x_bed - 40, zone_y_bed - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 102, 255), 1) cv2.rectangle(img, (zone_x_min_clean, zone_y_min_clean), (zone_x_max_clean, zone_y_max_clean), (255, 255, 51), 2) zone_x_clean = int((zone_x_min_clean + zone_x_max_clean) / 2.0) zone_y_clean = int((zone_y_min_clean + zone_y_max_clean) / 2.0) cv2.circle(img, (zone_x_clean, zone_y_clean), 4, (255, 255, 51), -1) cv2.putText(img, "CLEANING ZONE", (zone_x_clean - 110, zone_y_clean - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 51), 1) cv2.rectangle(img, (zone_x_min_door, zone_y_min_door), (zone_x_max_door, zone_y_max_door), (127, 0, 255), 2) zone_x_door = int((zone_x_min_door + zone_x_max_door) / 2.0) zone_y_door = int((zone_y_min_door + zone_y_max_door) / 2.0) cv2.circle(img, (zone_x_door, zone_y_door), 4, (127, 0, 255), -1) cv2.putText(img, "ENTRANCE", (zone_x_door - 40, zone_y_door - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (127, 0, 255), 1) cv2.rectangle(img, (zone_x_min_alchol, zone_y_min_alchol), (zone_x_max_alchol, zone_y_max_alchol), (255, 255, 51), 2) zone_x_alchol = int( (zone_x_min_alchol + zone_x_max_alchol) / 2.0) zone_y_alchol = int( (zone_y_min_alchol + zone_y_max_alchol) / 2.0) cv2.circle(img, (zone_x_alchol, zone_y_alchol), 4, (255, 255, 51), -1) cv2.putText(img, "CLEANING ZONE", (zone_x_alchol - 35, zone_y_alchol - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 51), 1) #ROI cv2.rectangle(img, (160, 130), (697, 721), (0, 255, 255), 2) distance_bed = 0 distance_clean = 0 #Detection Zone objects, valid_checker = ct.update(rects) flag = False leave_zone_counter = 0 for ((objectID, centroid), (objectID, valid)) in zip(objects.items(), valid_checker.items()): # draw both the ID of the object and the centroid of the # object on the output frame #text_id = "ID {}".format(objectID) text_id = "id = 0" backup_label = str(objectID) text = "staff" cv2.putText(img, text, (centroid[0] - 10, centroid[1] - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2) cv2.putText(img, text_id, (centroid[0] - 10, centroid[1] - 50), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 0), 2) cv2.circle(img, (centroid[0], centroid[1]), 4, (0, 255, 0), -1) distance_bed = int( math.sqrt((centroid[0] - zone_x_bed)**2 + (centroid[1] - zone_y_bed)**2)) distance_clean = int( math.sqrt((centroid[0] - zone_x_clean)**2 + (centroid[1] - zone_y_clean)**2)) distance_clean_alchol = int( math.sqrt((centroid[0] - zone_x_alchol)**2 + (centroid[1] - zone_y_alchol)**2)) enter = ct.display_enter_status(objectID) leave = ct.display_leave_status(objectID) flag = ct.display_hygiene(objectID) if (distance_clean_alchol <= distance_thres_alchol): cv2.line(img, (centroid[0], centroid[1]), (zone_x_alchol, zone_y_alchol), (0, 0, 255), 1) # if(hand_wash_status == 1): # personal_status[objectID] = 1 # ct.update_wash(True,objectID) if (distance_bed <= distance_thres_bed): #personal_status[0] = 0 hand_wash_status = 0 cv2.line(img, (centroid[0], centroid[1]), (zone_x_bed, zone_y_bed), (0, 255, 0), 1) #Update Hygiene Status as the staff is originally cleaned ct.update_hygiene(False, objectID) """ #Never enter if(enter == False): ct.update_enter(True,objectID) #If the staff did not wash hand and go to patient directly hand_wash_flag = ct.display_wash(objectID) if(hand_wash_flag == False): ct.update_valid(False,objectID) m = 0 match = True #Check whether this ID is marked as fail or not (on the screen) while(m<len(invalid_id)): if(objectID==invalid_id[m]): #ObjectID is found in the invalid bank match = True else: match = False m+=1 #If it is not in the bank,then mark it and fail counter + 1 if(match == False): fail_msg +=1 invalid_id.append(objectID) #Enter again with uncleaned => invalid else: if(flag == False): #Re-enter the patient zone if((enter == True)and(leave == True)): ct.update_valid(False,objectID) m = 0 match = True #Check whether this ID is marked as fail or not (on the screen) while(m<len(invalid_id)): if(objectID==invalid_id[m]): #ObjectID is found in the invalid bank match = True else: match = False m+=1 #If it is not in the bank,then mark it and fail counter + 1 if(match == False): fail_msg +=1 invalid_id.append(objectID) else: if(enter == True): ct.update_leave(True,objectID) """ if ((distance_clean <= distance_thres_clean) or (distance_clean_alchol <= distance_thres_alchol)): if (distance_clean <= distance_thres_clean): cv2.line(img, (centroid[0], centroid[1]), (zone_x_clean, zone_y_clean), (51, 255, 255), 1) #if(hand_wash_status == 1): # personal_status[0] = 1 #hand_wash_status = 1 #Update Hygiene Status ct.update_hygiene(True, objectID) #Reset IN/OUT Mechanism ct.update_enter(False, objectID) ct.update_leave(False, objectID) ct.update_wash(True, objectID) #Return hygiene status flag = ct.display_hygiene(objectID) #if(previous_id!=objectID): #hand_wash_status = 0 #personal_status = 0 previous_id = objectID with open('./path_analyzer/path_log.csv', 'a', newline='') as csv_log_file: log_writer = csv.writer(csv_log_file) log_writer.writerow([ objectID, centroid[0], centroid[1], int(distance_bed), int(distance_clean), int(distance_clean_alchol), int(0), int(0) ]) #log_writer.writerow([objectID,centroid[0],centroid[1],int(distance_bed),int(distance_clean),int(distance_clean_alchol),int(hand_wash_status),int(personal_status[objectID])]) if(((centroid[0]>=zone_x_min_door)and(centroid[0]<=zone_x_max_door)) and \ ((centroid[1]>=zone_y_min_door)and(centroid[1]<=zone_y_max_door))): person_x = centroid[0] person_y = centroid[1] leave_zone_counter += 1 if ((leave_zone_counter == 0) and (enter_zone == True) and (leave_zone == False)): if (leave_zone_2 == False): print("LEAVE 1!!!\r\n") leave_zone = True else: print("LAEVE 2!!!\r\n") enter_zone = False leave_zone_2 = False #print("leave_zone_counter:",leave_zone_counter) #print("leave_zone_2:",leave_zone_2) #print("enter_zone:",enter_zone) #print("leave_zone:",leave_zone) #If any id passed through entrance if (((person_x > 0) and (person_y > 0)) and (leave_zone_counter > 0)): font = cv2.FONT_HERSHEY_PLAIN line = cv2.LINE_AA if (enter_zone == False): cv2.putText(img, "START", (40, 100), font, 3.0, (255, 0, 0), 4, line) print("START\r\n") enter_zone = True start_time = time.time() pp = ' { "sys_status" :"' + str(10) + '"}' client.publish("MDSSCC/STATUS", pp) else: if (leave_zone == True): cv2.putText(img, "COUNT", (40, 100), font, 3.0, (255, 0, 0), 4, line) print("COUNT\r\n") pp = ' { "sys_status" :"' + str(30) + '"}' client.publish("MDSSCC/STATUS", pp) leave_zone = False leave_zone_2 = True leave_zone_counter = 0 person_x = 0 person_y = 0 print() if (restart_flag == False): if show_fps: img = draw_help_and_fps(img, fps) #img = img[200:721, 160:697] cv2.imshow(WINDOW_NAME, img) cv2.setMouseCallback( WINDOW_NAME, pixel_info) #Receive mouse click on HSV_Picker #toc = time.time() #curr_fps = 1.0 / (toc - tic) # calculate an exponentially decaying average of fps number #fps = curr_fps if fps == 0.0 else (fps*0.9 + curr_fps*0.1) #tic = toc else: print("None Image --> None Buff = {}".format(none_buff)) none_buff += 1 if (none_buff == 1000): print("[SYSTEM] VSTARCAMERA Restart") cam.stop() # terminate the sub-thread in camera #tf_sess.close() #tf.reset_default_graph() #tf.contrib.keras.backend.clear_session() cam.release() restart_flag = True none_buff = 0 #img = None cv2.destroyAllWindows() hand_wash_status = 0 if (restart_flag == False): key = cv2.waitKey(1) if key == 27: # ESC key: quit program break elif key == ord('H') or key == ord('h'): # Toggle help/fps show_fps = not show_fps elif key == ord('F') or key == ord('f'): # Toggle fullscreen full_scrn = not full_scrn set_full_screen(full_scrn) #client.loop_start() #client.loop_forever() client.reconnect()
def loop_and_detect(cam, tf_sess, conf_th, vis, od_type): """Loop, grab images from camera, and do object detection. # Arguments cam: the camera object (video source). tf_sess: TensorFlow/TensorRT session to run SSD object detection. conf_th: confidence/score threshold for object detection. vis: for visualization. """ show_fps = True full_scrn = False fps = 0.0 tic = time.time() feature_params = dict(maxCorners=1000, qualityLevel=0.1, minDistance=4, blockSize=7) lk_params = dict(winSize=(15, 15), maxLevel=3, criteria=(cv2.TERM_CRITERIA_EPS | cv2.TERM_CRITERIA_COUNT, 10, 0.02)) tracks = [] track_len = 8 frame_idx = 0 detect_interval = 10 global rects,ct,temp,args,optical_enable #restart issue frame_buff = 0 none_buff = 0 restart_flag = False #restart issue backup_label = None #restart issue while True: #if cv2.getWindowProperty(WINDOW_NAME, 0) < 0: # Check to see if the user has closed the display window. # If yes, terminate the while loop. # break if(restart_flag == True): cam = Camera(args) cam.open() cam.start() print("Camera is opened!") #pb_path = './data/{}_trt.pb'.format(args.model) #log_path = './logs/{}_trt'.format(args.model) #trt_graph = load_trt_pb(pb_path) #tf_config = tf.ConfigProto() #tf_config.gpu_options.allow_growth = True #tf_sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True,log_device_placement=True),graph=trt_graph) #od_type = 'faster_rcnn' if 'faster_rcnn' in args.model else 'ssd' dummy_img = np.zeros((720, 1280, 3), dtype=np.uint8) _, _, _ = detect(dummy_img, tf_sess, conf_th=.3, od_type=od_type) print("Loading dummy image!") restart_flag = False rects = [] img = cam.read() if img is not None: optical_flow_image = img box, conf, cls = detect(img, tf_sess, conf_th, od_type=od_type) img = vis.draw_bboxes(img, box, conf, cls) #Optical Flow if (optical_enable==True): frame_gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) if len(tracks) > 0: img0 , img1 = prev_gray, frame_gray p0 = np.float32([tr[-1] for tr in tracks]).reshape(-1,1,2) p1, st, err = cv2.calcOpticalFlowPyrLK(img0, img1, p0, None, **lk_params) p0r, _, _ = cv2.calcOpticalFlowPyrLK(img1, img0, p1, None, **lk_params) d = abs(p0-p0r).reshape(-1,2).max(-1) good = d < 1 new_tracks = [] for i, (tr, (x, y), flag) in enumerate(zip(tracks, p1.reshape(-1, 2), good)): if not flag: continue tr.append((x,y)) if len(tr)> track_len: del tr[0] new_tracks.append(tr) cv2.circle(img, (x, y), 2, (0, 255, 0), -1) tracks = new_tracks cv2.polylines(img, [np.int32(tr) for tr in tracks], False, (0, 255, 0), 1) for box_coord in box: y_min, x_min, y_max, x_max = box_coord[0], box_coord[1], box_coord[2], box_coord[3] for tr in tracks: tail = len(tr) start_point = tr[0] end_point = tr[tail-1] if((start_point[0]<=x_max)and(start_point[0]>=x_min))and((end_point[0]<=x_max)and(end_point[0]>=x_min)): if((start_point[1]<=y_max)and(start_point[1]>=y_min))and((end_point[1]<=y_max)and(end_point[1]>=y_min)): if(end_point[0]-start_point[0])>40: cv2.putText(img, "ENTER", (x_min+10,y_min+50),cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,153),2,cv2.LINE_AA) elif(start_point[0]-end_point[0])>40: cv2.putText(img, "LEAVE", (x_min+10,y_min+50),cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,153),2,cv2.LINE_AA) #elif(abs(start_point[0]-end_point[0])<10)and(abs(start_point[1]-end_point[1])<10): # cv2.putText(img, "IDLE", (x_min+10,y_min-20),cv2.FONT_HERSHEY_TRIPLEX,1,(102,255,178),2,cv2.LINE_AA) if frame_idx % detect_interval==0: mask = np.zeros_like(frame_gray) mask[:] = 255 if frame_idx !=0: for x,y in [np.int32(tr[-1]) for tr in tracks]: cv2.circle(mask, (x, y), 5, 0, -1) p = cv2.goodFeaturesToTrack(frame_gray, mask=mask, **feature_params) if p is not None: for x, y in np.float32(p).reshape(-1,2): tracks.append([(x, y)]) frame_idx+=1 prev_gray = frame_gray #Optical Flow done cv2.rectangle(img, (0,980),(1920,1075),(0,0,0),-1) objects = ct.update(rects) for (objectID, centroid) in objects.items(): # draw both the ID of the object and the centroid of the # object on the output frame text = "ID {}".format(objectID) cv2.putText(img, text, (centroid[0] - 10, centroid[1] - 10),cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255,0,255), 2) cv2.putText(img, str(objectID), (330,1035),cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,255),2,cv2.LINE_AA) cv2.circle(img, (centroid[0], centroid[1]), 4, (255,0,255), -1) backup_label = str(objectID) if(optical_enable == True): cv2.putText(img, "ON", (1380,1035), cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,255),2,cv2.LINE_AA) else: cv2.putText(img, "OFF", (1380,1035), cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,255),2,cv2.LINE_AA) sys_clock = str(datetime.datetime.now()) cv2.putText(img, backup_label, (330,1035),cv2.FONT_HERSHEY_TRIPLEX,1,(255,255,255),2,cv2.LINE_AA) cv2.putText(img, sys_clock, (20,950), cv2.FONT_HERSHEY_TRIPLEX,1,(0,0,255),2,cv2.LINE_AA) cv2.putText(img, "Traffic Counter: ", (20,1035), cv2.FONT_HERSHEY_TRIPLEX,1,(0,0,255),2,cv2.LINE_AA) cv2.putText(img, "Detector Type: Human", (400,1035), cv2.FONT_HERSHEY_TRIPLEX,1,(0,0,255),2,cv2.LINE_AA) cv2.putText(img, "Real Time Optical Trace :", (900,1035), cv2.FONT_HERSHEY_TRIPLEX,1,(0,0,255),2,cv2.LINE_AA) #print(sys_clock) if(restart_flag == False): if show_fps: img = draw_help_and_fps(img, fps) cv2.moveWindow(WINDOW_NAME,0,0) #restart issue cv2.imshow(WINDOW_NAME, img) toc = time.time() curr_fps = 1.0 / (toc - tic) # calculate an exponentially decaying average of fps number fps = curr_fps if fps == 0.0 else (fps*0.9 + curr_fps*0.1) tic = toc else: print("None Image --> None Buff = {}".format(none_buff)) none_buff+=1 if(none_buff == 1000): print("[SYSTEM] VSTARCAMERA Restart") cam.stop() # terminate the sub-thread in camera #tf_sess.close() #tf.reset_default_graph() #tf.contrib.keras.backend.clear_session() cam.release() restart_flag = True none_buff = 0 #img = None cv2.destroyAllWindows() if(restart_flag== False): key = cv2.waitKey(1) if key == 27: # ESC key: quit program break elif key == ord('H') or key == ord('h'): # Toggle help/fps show_fps = not show_fps elif key == ord('F') or key == ord('f'): # Toggle fullscreen full_scrn = not full_scrn set_full_screen(full_scrn) elif key == ord ('P') or key == ord('p'): if(optical_enable == True): optical_enable = False else: optical_enable = True
def loop_and_detect(cam, tf_sess, conf_th, vis, od_type): """Loop, grab images from camera, and do object detection. # Arguments cam: the camera object (video source). tf_sess: TensorFlow/TensorRT session to run SSD object detection. conf_th: confidence/score threshold for object detection. vis: for visualization. """ show_fps = True full_scrn = False fps = 0.0 tic = time.time() tracks = [] global rects, ct, temp, args frame_buff = 0 none_buff = 0 restart_flag = False backup_label = None while True: #if cv2.getWindowProperty(WINDOW_NAME, 0) < 0: # Check to see if the user has closed the display window. # If yes, terminate the while loop. # break if (restart_flag == True): cam = Camera(args) cam.open() cam.start() #pb_path = './data/{}_trt.pb'.format(args.model) #log_path = './logs/{}_trt'.format(args.model) #trt_graph = load_trt_pb(pb_path) #tf_config = tf.ConfigProto() #tf_config.gpu_options.allow_growth = True #tf_sess = tf.Session(config=tf.ConfigProto(allow_soft_placement=True,log_device_placement=True),graph=trt_graph) #od_type = 'faster_rcnn' if 'faster_rcnn' in args.model else 'ssd' dummy_img = np.zeros((720, 1280, 3), dtype=np.uint8) _, _, _ = detect(dummy_img, tf_sess, conf_th=.3, od_type=od_type) restart_flag = False rects = [] img = cam.read() optical_flow_image = img if img is not None: box, conf, cls = detect(img, tf_sess, conf_th, od_type=od_type) img = vis.draw_bboxes(img, box, conf, cls) objects = ct.update(rects) cv2.rectangle(img, (0, 980), (1920, 1075), (0, 0, 0), -1) for (objectID, centroid) in objects.items(): # draw both the ID of the object and the centroid of the # object on the output frame text = "ID {}".format(objectID) cv2.putText(img, text, (centroid[0] - 10, centroid[1] - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 0, 255), 2) cv2.circle(img, (centroid[0], centroid[1]), 4, (255, 0, 255), -1) backup_label = str(objectID) cv2.putText(img, backup_label, (330, 1035), cv2.FONT_HERSHEY_TRIPLEX, 1, (255, 255, 255), 2, cv2.LINE_AA) sys_clock = str( datetime.datetime.now()) + " Frame_buff=" + str(frame_buff) print(sys_clock) cv2.putText(img, sys_clock, (20, 950), cv2.FONT_HERSHEY_TRIPLEX, 1, (0, 0, 255), 2, cv2.LINE_AA) cv2.putText(img, "Traffic Counter: ", (20, 1035), cv2.FONT_HERSHEY_TRIPLEX, 1, (0, 0, 255), 2, cv2.LINE_AA) cv2.putText(img, "Detector Type: Human", (400, 1035), cv2.FONT_HERSHEY_TRIPLEX, 1, (0, 0, 255), 2, cv2.LINE_AA) cv2.putText(img, "Real Time Optical Trace :", (900, 1035), cv2.FONT_HERSHEY_TRIPLEX, 1, (0, 0, 255), 2, cv2.LINE_AA) cv2.putText(img, "OFF", (1380, 1035), cv2.FONT_HERSHEY_TRIPLEX, 1, (255, 255, 255), 2, cv2.LINE_AA) """ if(frame_buff == 2000): print("[SYSTEM] VSTARCAMERA Restart") cam.stop() # terminate the sub-thread in camera #tf_sess.close() #tf.reset_default_graph() #tf.contrib.keras.backend.clear_session() cam.release() restart_flag = True frame_buff = 0 img = None cv2.destroyAllWindows() frame_buff += 1 """ if (restart_flag == False): if show_fps: img = draw_help_and_fps(img, fps) #set_full_screen(full_scrn) cv2.moveWindow(WINDOW_NAME, 0, 0) cv2.imshow(WINDOW_NAME, img) toc = time.time() curr_fps = 1.0 / (toc - tic) # calculate an exponentially decaying average of fps number fps = curr_fps if fps == 0.0 else (fps * 0.9 + curr_fps * 0.1) tic = toc else: print("None Image --> None Buff = {}".format(none_buff)) none_buff += 1 if (none_buff == 500): print("[SYSTEM] VSTARCAMERA Restart") cam.stop() # terminate the sub-thread in camera #tf_sess.close() #tf.reset_default_graph() #tf.contrib.keras.backend.clear_session() cam.release() restart_flag = True none_buff = 0 img = None cv2.destroyAllWindows() if (restart_flag == False): key = cv2.waitKey(1) if key == 27: # ESC key: quit program break elif key == ord('H') or key == ord('h'): # Toggle help/fps show_fps = not show_fps elif key == ord('F') or key == ord('f'): # Toggle fullscreen full_scrn = not full_scrn set_full_screen(full_scrn)