refactor resizing into generic priority queues

frigate/object_detection.py

@@ -27,6 +27,7 @@ class PreppedQueueProcessor(threading.Thread):
             # print(self.engine.get_inference_time())
 
             # parse and pass detected objects back to the camera
+            # TODO: just send this back with all the same info you received and objects as a new property
             parsed_objects = []
             for obj in objects:
                 parsed_objects.append({
@@ -92,3 +93,67 @@ class FramePrepper(threading.Thread):
                 })
             else:
                 print("queue full. moving on")
+
+class RegionRequester(threading.Thread):
+    def __init__(self, camera):
+        self.camera = camera
+
+    def run(self):
+        frame_time = 0.0
+        while True:
+            now = datetime.datetime.now().timestamp()
+
+            with self.camera.frame_ready:
+                # if there isnt a frame ready for processing or it is old, wait for a new frame
+                if self.camera.frame_time.value == frame_time or (now - self.camera.frame_time.value) > 0.5:
+                    self.camera.frame_ready.wait()
+            
+            # make a copy of the frame_time
+            frame_time = self.camera.frame_time.value
+            
+            for index, region in enumerate(self.camera.config['regions']):
+                # queue with priority 1
+                self.camera.resize_queue.put((1, {
+                    'camera_name': self.camera.name,
+                    'frame_time': frame_time,
+                    'region_id': index,
+                    'size': region['size'],
+                    'x_offset': region['x_offset'],
+                    'y_offset': region['y_offset']
+                }))
+
+class RegionPrepper(threading.Thread):
+    def __init__(self, frame_cache, resize_request_queue, prepped_frame_queue):
+
+        threading.Thread.__init__(self)
+        self.frame_cache = frame_cache
+        self.resize_request_queue = resize_request_queue
+        self.prepped_frame_queue = prepped_frame_queue
+
+    def run(self):
+        while True:
+
+            resize_request = self.resize_request_queue.get()
+
+            frame = self.frame_cache.get(resize_request['frame_time'], None)
+            
+            if frame is None:
+                print("RegionPrepper: frame_time not in frame_cache")
+                continue
+
+            # make a copy of the region
+            cropped_frame = frame[resize_request['y_offset']:resize_request['y_offset']+resize_request['size'], resize_request['x_offset']:resize_request['x_offset']+resize_request['size']].copy()
+            
+            # Resize to 300x300 if needed
+            if cropped_frame.shape != (300, 300, 3):
+                cropped_frame = cv2.resize(cropped_frame, dsize=(300, 300), interpolation=cv2.INTER_LINEAR)
+            # Expand dimensions since the model expects images to have shape: [1, 300, 300, 3]
+            frame_expanded = np.expand_dims(cropped_frame, axis=0)
+
+            # add the frame to the queue
+            if not self.prepped_frame_queue.full():
+                resize_request['frame'] = frame_expanded.flatten().copy()
+                # add to queue with priority 1
+                self.prepped_frame_queue.put((1, resize_request))
+            else:
+                print("queue full. moving on")