diff --git a/detect_objects.py b/detect_objects.py index f93a22669..bcfa10a49 100644 --- a/detect_objects.py +++ b/detect_objects.py @@ -52,11 +52,12 @@ def main(): 'mask': region_mask, # Event for motion detection signaling 'motion_detected': mp.Event(), - # create shared array for storing 10 detected objects - # note: this must be a double even though the value you are storing - # is a float. otherwise it stops updating the value in shared - # memory. probably something to do with the size of the memory block - 'output_array': mp.Array(ctypes.c_double, 6*10) + # array for prepped frame with shape (1, 300, 300, 3) + 'prepped_frame_array': mp.Array(ctypes.c_uint8, 300*300*3), + # shared value for storing the prepped_frame_time + 'prepped_frame_time': mp.Value('d', 0.0), + # Lock to control access to the prepped frame + 'prepped_frame_lock': mp.Lock() }) # capture a single frame and check the frame shape so the correct array # size can be allocated in memory @@ -85,16 +86,6 @@ def main(): objects_parsed = mp.Condition() # Queue for detected objects object_queue = mp.Queue() - # array for prepped frame with shape (1, 300, 300, 3) - prepped_frame_array = mp.Array(ctypes.c_uint8, 300*300*3) - # shared value for storing the prepped_frame_time - prepped_frame_time = mp.Value('d', 0.0) - # Condition for notifying that a new prepped frame is ready - prepped_frame_ready = mp.Condition() - # Lock to control access to the prepped frame - prepped_frame_lock = mp.Lock() - # array for prepped frame box [x1, y1, x2, y2] - prepped_frame_box = mp.Array(ctypes.c_uint16, 4) # shape current frame so it can be treated as an image frame_arr = tonumpyarray(shared_arr).reshape(frame_shape) @@ -116,8 +107,8 @@ def main(): region['motion_detected'], frame_shape, region['size'], region['x_offset'], region['y_offset'], - prepped_frame_array, prepped_frame_time, prepped_frame_ready, - prepped_frame_lock, prepped_frame_box)) + region['prepped_frame_array'], region['prepped_frame_time'], + region['prepped_frame_lock'])) detection_prep_process.daemon = True detection_prep_processes.append(detection_prep_process) @@ -135,9 +126,12 @@ def main(): # create a process for object detection detection_process = mp.Process(target=detect_objects, args=( - prepped_frame_array, prepped_frame_time, - prepped_frame_lock, prepped_frame_ready, - prepped_frame_box, object_queue, DEBUG + [region['prepped_frame_array'] for region in regions], + [region['prepped_frame_time'] for region in regions], + [region['prepped_frame_lock'] for region in regions], + [[region['size'], region['x_offset'], region['y_offset']] for region in regions], + motion_changed, [region['motion_detected'] for region in regions], + object_queue, DEBUG )) detection_process.daemon = True diff --git a/frigate/object_detection.py b/frigate/object_detection.py index df77dc70a..da0375e96 100644 --- a/frigate/object_detection.py +++ b/frigate/object_detection.py @@ -19,48 +19,64 @@ def ReadLabelFile(file_path): ret[int(pair[0])] = pair[1].strip() return ret -def detect_objects(prepped_frame_array, prepped_frame_time, prepped_frame_lock, - prepped_frame_ready, prepped_frame_box, object_queue, debug): - prepped_frame_np = tonumpyarray(prepped_frame_array) +def detect_objects(prepped_frame_arrays, prepped_frame_times, prepped_frame_locks, + prepped_frame_boxes, motion_changed, motion_regions, object_queue, debug): + prepped_frame_nps = [tonumpyarray(prepped_frame_array) for prepped_frame_array in prepped_frame_arrays] # Load the edgetpu engine and labels engine = DetectionEngine(PATH_TO_CKPT) labels = ReadLabelFile(PATH_TO_LABELS) frame_time = 0.0 - region_box = [0,0,0,0] + region_box = [0,0,0] while True: - with prepped_frame_ready: - prepped_frame_ready.wait() + # while there is motion + while len([r for r in motion_regions if r.is_set()]) > 0: - # make a copy of the cropped frame - with prepped_frame_lock: - prepped_frame_copy = prepped_frame_np.copy() - frame_time = prepped_frame_time.value - region_box[:] = prepped_frame_box + # loop over all the motion regions and look for objects + for i, motion_region in enumerate(motion_regions): + # skip the region if no motion + if not motion_region.is_set(): + continue - # Actual detection. - objects = engine.DetectWithInputTensor(prepped_frame_copy, threshold=0.5, top_k=3) - # print(engine.get_inference_time()) - # put detected objects in the queue - if objects: - # assumes square - region_size = region_box[2]-region_box[0] - for obj in objects: - box = obj.bounding_box.flatten().tolist() - object_queue.put({ - 'frame_time': frame_time, - 'name': str(labels[obj.label_id]), - 'score': float(obj.score), - 'xmin': int((box[0] * region_size) + region_box[0]), - 'ymin': int((box[1] * region_size) + region_box[1]), - 'xmax': int((box[2] * region_size) + region_box[0]), - 'ymax': int((box[3] * region_size) + region_box[1]) - }) + # make a copy of the cropped frame + with prepped_frame_locks[i]: + prepped_frame_copy = prepped_frame_nps[i].copy() + frame_time = prepped_frame_times[i].value + region_box[:] = prepped_frame_boxes[i] + + # Actual detection. + objects = engine.DetectWithInputTensor(prepped_frame_copy, threshold=0.5, top_k=3) + # print(engine.get_inference_time()) + # put detected objects in the queue + if objects: + for obj in objects: + box = obj.bounding_box.flatten().tolist() + object_queue.put({ + 'frame_time': frame_time, + 'name': str(labels[obj.label_id]), + 'score': float(obj.score), + 'xmin': int((box[0] * region_box[0]) + region_box[1]), + 'ymin': int((box[1] * region_box[0]) + region_box[2]), + 'xmax': int((box[2] * region_box[0]) + region_box[1]), + 'ymax': int((box[3] * region_box[0]) + region_box[2]) + }) + else: + object_queue.put({ + 'frame_time': frame_time, + 'name': 'dummy', + 'score': 0.99, + 'xmin': int(0 + region_box[1]), + 'ymin': int(0 + region_box[2]), + 'xmax': int(10 + region_box[1]), + 'ymax': int(10 + region_box[2]) + }) + # wait for the global motion flag to change + with motion_changed: + motion_changed.wait() def prep_for_detection(shared_whole_frame_array, shared_frame_time, frame_lock, frame_ready, motion_detected, frame_shape, region_size, region_x_offset, region_y_offset, - prepped_frame_array, prepped_frame_time, prepped_frame_ready, prepped_frame_lock, - prepped_frame_box): + prepped_frame_array, prepped_frame_time, prepped_frame_lock): # shape shared input array into frame for processing shared_whole_frame = tonumpyarray(shared_whole_frame_array).reshape(frame_shape) @@ -94,9 +110,4 @@ def prep_for_detection(shared_whole_frame_array, shared_frame_time, frame_lock, # copy the prepped frame to the shared output array with prepped_frame_lock: shared_prepped_frame[:] = frame_expanded - prepped_frame_time = frame_time - prepped_frame_box[:] = [region_x_offset, region_y_offset, region_x_offset+region_size, region_y_offset+region_size] - - # signal that a prepped frame is ready - with prepped_frame_ready: - prepped_frame_ready.notify_all() \ No newline at end of file + prepped_frame_time.value = frame_time