blakeblackshear.frigate/frigate/video.py

787 lines
26 KiB
Python
Raw Normal View History

2019-02-26 03:27:02 +01:00
import datetime
2020-11-04 13:31:25 +01:00
import itertools
2020-11-04 04:26:39 +01:00
import logging
2019-03-30 02:49:27 +01:00
import multiprocessing as mp
2020-11-04 13:31:25 +01:00
import queue
import random
2020-11-29 23:19:59 +01:00
import signal
2021-10-31 17:12:44 +01:00
import subprocess as sp
2020-11-04 13:31:25 +01:00
import threading
import time
from collections import defaultdict
2020-11-04 13:31:25 +01:00
import numpy as np
import cv2
2021-10-31 17:12:44 +01:00
from setproctitle import setproctitle
2020-11-04 13:31:25 +01:00
from frigate.config import CameraConfig, DetectConfig, PixelFormatEnum
from frigate.object_detection import RemoteObjectDetector
2020-12-04 13:59:03 +01:00
from frigate.log import LogPipe
2020-02-16 04:07:54 +01:00
from frigate.motion import MotionDetector
2020-11-04 13:31:25 +01:00
from frigate.objects import ObjectTracker
2021-02-17 14:23:32 +01:00
from frigate.util import (
EventsPerSecond,
FrameManager,
SharedMemoryFrameManager,
2021-10-31 17:12:44 +01:00
area,
2021-02-17 14:23:32 +01:00
calculate_region,
clipped,
2021-10-31 17:12:44 +01:00
intersection,
intersection_over_union,
2021-02-17 14:23:32 +01:00
listen,
yuv_crop_and_resize,
2021-02-17 14:23:32 +01:00
yuv_region_2_rgb,
yuv_region_2_bgr,
2021-02-17 14:23:32 +01:00
)
2019-02-26 03:27:02 +01:00
2020-11-04 04:26:39 +01:00
logger = logging.getLogger(__name__)
2021-02-17 14:23:32 +01:00
2021-01-15 14:52:28 +01:00
def filtered(obj, objects_to_track, object_filters):
2020-02-16 04:07:54 +01:00
object_name = obj[0]
object_score = obj[1]
object_box = obj[2]
object_area = obj[3]
object_ratio = obj[4]
2020-02-16 04:07:54 +01:00
if not object_name in objects_to_track:
return True
2021-02-17 14:23:32 +01:00
2020-02-16 04:07:54 +01:00
if object_name in object_filters:
obj_settings = object_filters[object_name]
# if the min area is larger than the
# detected object, don't add it to detected objects
if obj_settings.min_area > object_area:
2020-02-16 04:07:54 +01:00
return True
2021-02-17 14:23:32 +01:00
2020-02-16 04:07:54 +01:00
# if the detected object is larger than the
# max area, don't add it to detected objects
if obj_settings.max_area < object_area:
2020-02-16 04:07:54 +01:00
return True
2020-09-07 19:17:42 +02:00
# if the score is lower than the min_score, skip
if obj_settings.min_score > object_score:
return True
# if the object is not proportionally wide enough
if obj_settings.min_ratio > object_ratio:
return True
# if the object is proportionally too wide
if obj_settings.max_ratio < object_ratio:
2020-02-16 04:07:54 +01:00
return True
2021-02-17 14:23:32 +01:00
2021-01-15 14:52:28 +01:00
if not obj_settings.mask is None:
# compute the coordinates of the object and make sure
# the location isn't outside the bounds of the image (can happen from rounding)
object_xmin = object_box[0]
object_xmax = object_box[2]
object_ymax = object_box[3]
y_location = min(int(object_ymax), len(obj_settings.mask) - 1)
2021-02-17 14:23:32 +01:00
x_location = min(
int((object_xmax + object_xmin) / 2.0),
2021-02-17 14:23:32 +01:00
len(obj_settings.mask[0]) - 1,
)
2021-01-15 14:52:28 +01:00
# if the object is in a masked location, don't add it to detected objects
if obj_settings.mask[y_location][x_location] == 0:
return True
2021-02-17 14:23:32 +01:00
2020-09-07 19:17:42 +02:00
return False
2021-02-17 14:23:32 +01:00
def create_tensor_input(frame, model_config, region):
if model_config.input_pixel_format == PixelFormatEnum.rgb:
cropped_frame = yuv_region_2_rgb(frame, region)
elif model_config.input_pixel_format == PixelFormatEnum.bgr:
cropped_frame = yuv_region_2_bgr(frame, region)
else:
cropped_frame = yuv_crop_and_resize(frame, region)
2019-03-30 02:49:27 +01:00
# Resize if needed
if cropped_frame.shape != (model_config.height, model_config.width, 3):
2021-02-17 14:23:32 +01:00
cropped_frame = cv2.resize(
cropped_frame,
dsize=(model_config.height, model_config.width),
interpolation=cv2.INTER_LINEAR,
2021-02-17 14:23:32 +01:00
)
# Expand dimensions since the model expects images to have shape: [1, height, width, 3]
2020-02-16 04:07:54 +01:00
return np.expand_dims(cropped_frame, axis=0)
2021-02-17 14:23:32 +01:00
2020-12-04 13:59:03 +01:00
def stop_ffmpeg(ffmpeg_process, logger):
2020-11-29 23:19:59 +01:00
logger.info("Terminating the existing ffmpeg process...")
ffmpeg_process.terminate()
try:
logger.info("Waiting for ffmpeg to exit gracefully...")
ffmpeg_process.communicate(timeout=30)
except sp.TimeoutExpired:
logger.info("FFmpeg didnt exit. Force killing...")
ffmpeg_process.kill()
ffmpeg_process.communicate()
ffmpeg_process = None
2021-02-17 14:23:32 +01:00
def start_or_restart_ffmpeg(
ffmpeg_cmd, logger, logpipe: LogPipe, frame_size=None, ffmpeg_process=None
):
if ffmpeg_process is not None:
2020-12-04 13:59:03 +01:00
stop_ffmpeg(ffmpeg_process, logger)
2020-11-29 22:55:53 +01:00
if frame_size is None:
2021-02-17 14:23:32 +01:00
process = sp.Popen(
ffmpeg_cmd,
stdout=sp.DEVNULL,
stderr=logpipe,
stdin=sp.DEVNULL,
start_new_session=True,
)
2020-11-29 22:55:53 +01:00
else:
2021-02-17 14:23:32 +01:00
process = sp.Popen(
ffmpeg_cmd,
stdout=sp.PIPE,
stderr=logpipe,
stdin=sp.DEVNULL,
bufsize=frame_size * 10,
start_new_session=True,
)
return process
2021-02-17 14:23:32 +01:00
def capture_frames(
ffmpeg_process,
camera_name,
frame_shape,
frame_manager: FrameManager,
frame_queue,
fps: mp.Value,
skipped_fps: mp.Value,
current_frame: mp.Value,
):
2020-11-03 15:15:58 +01:00
frame_size = frame_shape[0] * frame_shape[1]
2020-10-25 16:05:21 +01:00
frame_rate = EventsPerSecond()
2020-10-26 13:59:22 +01:00
frame_rate.start()
2020-10-25 16:05:21 +01:00
skipped_eps = EventsPerSecond()
skipped_eps.start()
while True:
2020-10-25 16:05:21 +01:00
fps.value = frame_rate.eps()
skipped_fps = skipped_eps.eps()
2020-09-07 19:17:42 +02:00
current_frame.value = datetime.datetime.now().timestamp()
2020-10-24 18:36:04 +02:00
frame_name = f"{camera_name}{current_frame.value}"
frame_buffer = frame_manager.create(frame_name, frame_size)
try:
2020-12-12 16:12:15 +01:00
frame_buffer[:] = ffmpeg_process.stdout.read(frame_size)
except Exception as e:
2022-02-06 15:46:41 +01:00
logger.error(f"{camera_name}: Unable to read frames from ffmpeg process.")
2020-12-12 16:12:15 +01:00
if ffmpeg_process.poll() != None:
2022-02-06 15:46:41 +01:00
logger.error(
2021-02-17 14:23:32 +01:00
f"{camera_name}: ffmpeg process is not running. exiting capture thread..."
)
2020-12-12 16:12:15 +01:00
frame_manager.delete(frame_name)
break
continue
2020-10-25 16:05:21 +01:00
frame_rate.update()
# if the queue is full, skip this frame
if frame_queue.full():
2020-10-25 16:05:21 +01:00
skipped_eps.update()
2020-10-24 18:36:04 +02:00
frame_manager.delete(frame_name)
continue
2020-10-24 18:36:04 +02:00
# close the frame
frame_manager.close(frame_name)
# add to the queue
2020-09-07 19:17:42 +02:00
frame_queue.put(current_frame.value)
2021-02-17 14:23:32 +01:00
2020-10-25 16:05:21 +01:00
class CameraWatchdog(threading.Thread):
2021-02-17 14:23:32 +01:00
def __init__(
self, camera_name, config, frame_queue, camera_fps, ffmpeg_pid, stop_event
):
2020-10-25 16:05:21 +01:00
threading.Thread.__init__(self)
2020-12-04 13:59:03 +01:00
self.logger = logging.getLogger(f"watchdog.{camera_name}")
2020-11-04 13:28:07 +01:00
self.camera_name = camera_name
2020-10-25 16:05:21 +01:00
self.config = config
self.capture_thread = None
2020-11-29 22:55:53 +01:00
self.ffmpeg_detect_process = None
self.logpipe = LogPipe(f"ffmpeg.{self.camera_name}.detect")
2020-11-29 22:55:53 +01:00
self.ffmpeg_other_processes = []
2020-10-25 16:05:21 +01:00
self.camera_fps = camera_fps
self.ffmpeg_pid = ffmpeg_pid
2020-10-25 16:05:21 +01:00
self.frame_queue = frame_queue
2020-11-03 15:15:58 +01:00
self.frame_shape = self.config.frame_shape_yuv
self.frame_size = self.frame_shape[0] * self.frame_shape[1]
2020-11-29 23:19:59 +01:00
self.stop_event = stop_event
2020-10-25 16:05:21 +01:00
def run(self):
2020-11-29 22:55:53 +01:00
self.start_ffmpeg_detect()
for c in self.config.ffmpeg_cmds:
2021-02-17 14:23:32 +01:00
if "detect" in c["roles"]:
2020-11-29 22:55:53 +01:00
continue
2021-02-17 14:23:32 +01:00
logpipe = LogPipe(
f"ffmpeg.{self.camera_name}.{'_'.join(sorted(c['roles']))}"
2021-02-17 14:23:32 +01:00
)
self.ffmpeg_other_processes.append(
{
"cmd": c["cmd"],
"logpipe": logpipe,
"process": start_or_restart_ffmpeg(c["cmd"], self.logger, logpipe),
}
)
2020-10-25 16:05:21 +01:00
time.sleep(10)
while not self.stop_event.wait(10):
2020-10-25 16:05:21 +01:00
now = datetime.datetime.now().timestamp()
if not self.capture_thread.is_alive():
2021-08-14 21:04:00 +02:00
self.logger.error(
2022-02-06 15:46:41 +01:00
f"Ffmpeg process crashed unexpectedly for {self.camera_name}."
2021-08-14 21:04:00 +02:00
)
self.logger.error(
2021-08-16 14:38:53 +02:00
"The following ffmpeg logs include the last 100 lines prior to exit."
2021-08-14 21:04:00 +02:00
)
self.logpipe.dump()
2020-11-29 22:55:53 +01:00
self.start_ffmpeg_detect()
elif now - self.capture_thread.current_frame.value > 20:
2021-02-17 14:23:32 +01:00
self.logger.info(
f"No frames received from {self.camera_name} in 20 seconds. Exiting ffmpeg..."
)
2020-11-29 22:55:53 +01:00
self.ffmpeg_detect_process.terminate()
2020-10-25 16:05:21 +01:00
try:
2020-12-04 13:59:03 +01:00
self.logger.info("Waiting for ffmpeg to exit gracefully...")
2020-11-29 22:55:53 +01:00
self.ffmpeg_detect_process.communicate(timeout=30)
2020-10-25 16:05:21 +01:00
except sp.TimeoutExpired:
2020-12-04 13:59:03 +01:00
self.logger.info("FFmpeg didnt exit. Force killing...")
2020-11-29 22:55:53 +01:00
self.ffmpeg_detect_process.kill()
self.ffmpeg_detect_process.communicate()
2021-02-17 14:23:32 +01:00
2020-11-29 22:55:53 +01:00
for p in self.ffmpeg_other_processes:
2021-02-17 14:23:32 +01:00
poll = p["process"].poll()
if poll is None:
2020-11-29 22:55:53 +01:00
continue
2021-02-17 14:23:32 +01:00
p["logpipe"].dump()
p["process"] = start_or_restart_ffmpeg(
p["cmd"], self.logger, p["logpipe"], ffmpeg_process=p["process"]
)
stop_ffmpeg(self.ffmpeg_detect_process, self.logger)
for p in self.ffmpeg_other_processes:
stop_ffmpeg(p["process"], self.logger)
p["logpipe"].close()
self.logpipe.close()
2021-02-17 14:23:32 +01:00
2020-11-29 22:55:53 +01:00
def start_ffmpeg_detect(self):
2021-02-17 14:23:32 +01:00
ffmpeg_cmd = [
c["cmd"] for c in self.config.ffmpeg_cmds if "detect" in c["roles"]
][0]
self.ffmpeg_detect_process = start_or_restart_ffmpeg(
ffmpeg_cmd, self.logger, self.logpipe, self.frame_size
)
2020-11-29 22:55:53 +01:00
self.ffmpeg_pid.value = self.ffmpeg_detect_process.pid
2021-02-17 14:23:32 +01:00
self.capture_thread = CameraCapture(
self.camera_name,
self.ffmpeg_detect_process,
self.frame_shape,
self.frame_queue,
self.camera_fps,
)
2020-11-01 17:55:11 +01:00
self.capture_thread.start()
2020-10-25 16:05:21 +01:00
2021-02-17 14:23:32 +01:00
class CameraCapture(threading.Thread):
2020-11-04 13:28:07 +01:00
def __init__(self, camera_name, ffmpeg_process, frame_shape, frame_queue, fps):
threading.Thread.__init__(self)
2020-11-04 13:28:07 +01:00
self.name = f"capture:{camera_name}"
self.camera_name = camera_name
self.frame_shape = frame_shape
self.frame_queue = frame_queue
self.fps = fps
self.skipped_fps = EventsPerSecond()
self.frame_manager = SharedMemoryFrameManager()
self.ffmpeg_process = ffmpeg_process
2021-02-17 14:23:32 +01:00
self.current_frame = mp.Value("d", 0.0)
self.last_frame = 0
def run(self):
self.skipped_fps.start()
2021-02-17 14:23:32 +01:00
capture_frames(
self.ffmpeg_process,
self.camera_name,
self.frame_shape,
self.frame_manager,
self.frame_queue,
self.fps,
self.skipped_fps,
self.current_frame,
)
2020-11-03 15:15:58 +01:00
def capture_camera(name, config: CameraConfig, process_info):
2020-11-29 23:19:59 +01:00
stop_event = mp.Event()
2021-02-17 14:23:32 +01:00
2020-11-29 23:19:59 +01:00
def receiveSignal(signalNumber, frame):
stop_event.set()
2021-02-17 14:23:32 +01:00
2020-11-29 23:19:59 +01:00
signal.signal(signal.SIGTERM, receiveSignal)
signal.signal(signal.SIGINT, receiveSignal)
2021-02-17 14:23:32 +01:00
frame_queue = process_info["frame_queue"]
camera_watchdog = CameraWatchdog(
name,
config,
frame_queue,
process_info["camera_fps"],
process_info["ffmpeg_pid"],
stop_event,
)
2020-10-25 16:05:21 +01:00
camera_watchdog.start()
camera_watchdog.join()
2021-02-17 14:23:32 +01:00
def track_camera(
name,
config: CameraConfig,
model_config,
labelmap,
2021-02-17 14:23:32 +01:00
detection_queue,
result_connection,
detected_objects_queue,
process_info,
):
2020-11-29 23:19:59 +01:00
stop_event = mp.Event()
2021-02-17 14:23:32 +01:00
2020-11-29 23:19:59 +01:00
def receiveSignal(signalNumber, frame):
stop_event.set()
2021-02-17 14:23:32 +01:00
2020-11-29 23:19:59 +01:00
signal.signal(signal.SIGTERM, receiveSignal)
signal.signal(signal.SIGINT, receiveSignal)
2020-11-04 13:28:07 +01:00
threading.current_thread().name = f"process:{name}"
2021-01-03 20:41:02 +01:00
setproctitle(f"frigate.process:{name}")
listen()
2020-02-16 04:07:54 +01:00
2021-02-17 14:23:32 +01:00
frame_queue = process_info["frame_queue"]
detection_enabled = process_info["detection_enabled"]
motion_enabled = process_info["motion_enabled"]
improve_contrast_enabled = process_info["improve_contrast_enabled"]
motion_threshold = process_info["motion_threshold"]
motion_contour_area = process_info["motion_contour_area"]
2020-10-25 16:05:21 +01:00
2020-11-03 15:15:58 +01:00
frame_shape = config.frame_shape
objects_to_track = config.objects.track
object_filters = config.objects.filters
2020-02-16 04:07:54 +01:00
2022-04-16 15:42:44 +02:00
motion_detector = MotionDetector(
frame_shape,
config.motion,
improve_contrast_enabled,
motion_threshold,
motion_contour_area,
2022-04-16 15:42:44 +02:00
)
2021-02-17 14:23:32 +01:00
object_detector = RemoteObjectDetector(
name, labelmap, detection_queue, result_connection, model_config
2021-02-17 14:23:32 +01:00
)
2020-02-16 04:07:54 +01:00
object_tracker = ObjectTracker(config.detect)
frame_manager = SharedMemoryFrameManager()
2021-02-17 14:23:32 +01:00
process_frames(
name,
frame_queue,
frame_shape,
model_config,
config.detect,
2021-02-17 14:23:32 +01:00
frame_manager,
motion_detector,
object_detector,
object_tracker,
detected_objects_queue,
process_info,
objects_to_track,
object_filters,
detection_enabled,
motion_enabled,
2021-02-17 14:23:32 +01:00
stop_event,
)
2020-11-04 04:26:39 +01:00
logger.info(f"{name}: exiting subprocess")
2021-02-17 14:23:32 +01:00
2021-10-30 14:24:26 +02:00
def box_overlaps(b1, b2):
if b1[2] < b2[0] or b1[0] > b2[2] or b1[1] > b2[3] or b1[3] < b2[1]:
return False
return True
2021-11-04 17:57:26 +01:00
def reduce_boxes(boxes, iou_threshold=0.0):
2021-10-30 14:24:26 +02:00
clusters = []
for box in boxes:
matched = 0
for cluster in clusters:
2021-11-04 17:57:26 +01:00
if intersection_over_union(box, cluster) > iou_threshold:
2021-10-30 14:24:26 +02:00
matched = 1
cluster[0] = min(cluster[0], box[0])
cluster[1] = min(cluster[1], box[1])
cluster[2] = max(cluster[2], box[2])
cluster[3] = max(cluster[3], box[3])
if not matched:
clusters.append(list(box))
return [tuple(c) for c in clusters]
2021-02-17 14:23:32 +01:00
def intersects_any(box_a, boxes):
for box in boxes:
2021-10-30 14:24:26 +02:00
if box_overlaps(box_a, box):
2021-11-04 15:25:17 +01:00
return True
2021-10-30 14:24:26 +02:00
return False
2021-02-17 14:23:32 +01:00
def detect(
detect_config: DetectConfig,
object_detector,
frame,
model_config,
region,
objects_to_track,
object_filters,
2021-02-17 14:23:32 +01:00
):
tensor_input = create_tensor_input(frame, model_config, region)
detections = []
region_detections = object_detector.detect(tensor_input)
for d in region_detections:
box = d[2]
2021-02-17 14:23:32 +01:00
size = region[2] - region[0]
x_min = int(max(0, (box[1] * size) + region[0]))
y_min = int(max(0, (box[0] * size) + region[1]))
x_max = int(min(detect_config.width - 1, (box[3] * size) + region[0]))
y_max = int(min(detect_config.height - 1, (box[2] * size) + region[1]))
# ignore objects that were detected outside the frame
if (x_min >= detect_config.width - 1) or (y_min >= detect_config.height - 1):
continue
width = x_max - x_min
height = y_max - y_min
area = width * height
ratio = width / height
2021-02-17 14:23:32 +01:00
det = (
d[0],
d[1],
(x_min, y_min, x_max, y_max),
area,
ratio,
2021-02-17 14:23:32 +01:00
region,
)
# apply object filters
2021-01-15 14:52:28 +01:00
if filtered(det, objects_to_track, object_filters):
continue
detections.append(det)
return detections
2021-02-17 14:23:32 +01:00
def process_frames(
camera_name: str,
frame_queue: mp.Queue,
frame_shape,
model_config,
detect_config: DetectConfig,
2021-02-17 14:23:32 +01:00
frame_manager: FrameManager,
motion_detector: MotionDetector,
object_detector: RemoteObjectDetector,
object_tracker: ObjectTracker,
detected_objects_queue: mp.Queue,
process_info: dict,
objects_to_track: list[str],
2021-02-17 14:23:32 +01:00
object_filters,
detection_enabled: mp.Value,
motion_enabled: mp.Value,
2021-02-17 14:23:32 +01:00
stop_event,
exit_on_empty: bool = False,
):
fps = process_info["process_fps"]
detection_fps = process_info["detection_fps"]
current_frame_time = process_info["detection_frame"]
2020-10-25 16:05:21 +01:00
2020-02-16 04:07:54 +01:00
fps_tracker = EventsPerSecond()
fps_tracker.start()
2022-02-05 14:10:00 +01:00
startup_scan_counter = 0
while not stop_event.is_set():
2020-11-01 17:37:17 +01:00
if exit_on_empty and frame_queue.empty():
2020-11-04 04:26:39 +01:00
logger.info(f"Exiting track_objects...")
2020-11-01 17:37:17 +01:00
break
try:
frame_time = frame_queue.get(True, 10)
except queue.Empty:
continue
current_frame_time.value = frame_time
2021-02-17 14:23:32 +01:00
frame = frame_manager.get(
f"{camera_name}{frame_time}", (frame_shape[0] * 3 // 2, frame_shape[1])
)
if frame is None:
2020-11-04 04:26:39 +01:00
logger.info(f"{camera_name}: frame {frame_time} is not in memory store.")
continue
# look for motion if enabled
motion_boxes = motion_detector.detect(frame) if motion_enabled.value else []
2020-02-16 04:07:54 +01:00
regions = []
# if detection is disabled
if not detection_enabled.value:
object_tracker.match_and_update(frame_time, [])
else:
# get stationary object ids
# check every Nth frame for stationary objects
# disappeared objects are not stationary
# also check for overlapping motion boxes
stationary_object_ids = [
obj["id"]
for obj in object_tracker.tracked_objects.values()
# if there hasn't been motion for 10 frames
if obj["motionless_count"] >= 10
# and it isn't due for a periodic check
and (
detect_config.stationary.interval == 0
or obj["motionless_count"] % detect_config.stationary.interval != 0
)
# and it hasn't disappeared
and object_tracker.disappeared[obj["id"]] == 0
# and it doesn't overlap with any current motion boxes
and not intersects_any(obj["box"], motion_boxes)
]
# get tracked object boxes that aren't stationary
tracked_object_boxes = [
obj["box"]
for obj in object_tracker.tracked_objects.values()
if not obj["id"] in stationary_object_ids
]
# combine motion boxes with known locations of existing objects
combined_boxes = reduce_boxes(motion_boxes + tracked_object_boxes)
region_min_size = max(model_config.height, model_config.width)
# compute regions
regions = [
2022-02-05 14:10:00 +01:00
calculate_region(
frame_shape,
a[0],
a[1],
a[2],
a[3],
region_min_size,
multiplier=random.uniform(1.2, 1.5),
2022-02-05 14:10:00 +01:00
)
for a in combined_boxes
]
# consolidate regions with heavy overlap
regions = [
calculate_region(
frame_shape, a[0], a[1], a[2], a[3], region_min_size, multiplier=1.0
2021-02-17 14:23:32 +01:00
)
for a in reduce_boxes(regions, 0.4)
]
# if starting up, get the next startup scan region
if startup_scan_counter < 9:
ymin = int(frame_shape[0] / 3 * startup_scan_counter / 3)
ymax = int(frame_shape[0] / 3 + ymin)
xmin = int(frame_shape[1] / 3 * startup_scan_counter / 3)
xmax = int(frame_shape[1] / 3 + xmin)
regions.append(
calculate_region(
frame_shape,
xmin,
ymin,
xmax,
ymax,
region_min_size,
multiplier=1.2,
)
)
startup_scan_counter += 1
2021-02-17 14:23:32 +01:00
# resize regions and detect
# seed with stationary objects
detections = [
(
obj["label"],
obj["score"],
obj["box"],
obj["area"],
obj["ratio"],
obj["region"],
)
for obj in object_tracker.tracked_objects.values()
if obj["id"] in stationary_object_ids
]
for region in regions:
detections.extend(
detect(
detect_config,
object_detector,
frame,
model_config,
region,
objects_to_track,
object_filters,
)
)
#########
# merge objects, check for clipped objects and look again up to 4 times
#########
refining = len(regions) > 0
refine_count = 0
while refining and refine_count < 4:
refining = False
# group by name
detected_object_groups = defaultdict(lambda: [])
for detection in detections:
detected_object_groups[detection[0]].append(detection)
selected_objects = []
for group in detected_object_groups.values():
# apply non-maxima suppression to suppress weak, overlapping bounding boxes
# o[2] is the box of the object: xmin, ymin, xmax, ymax
# apply max/min to ensure values do not exceed the known frame size
boxes = [
(
o[2][0],
o[2][1],
o[2][2] - o[2][0],
o[2][3] - o[2][1],
)
for o in group
]
confidences = [o[1] for o in group]
idxs = cv2.dnn.NMSBoxes(boxes, confidences, 0.5, 0.4)
for index in idxs:
2022-02-21 13:32:17 +01:00
index = index if isinstance(index, np.int32) else index[0]
obj = group[index]
if clipped(obj, frame_shape):
box = obj[2]
# calculate a new region that will hopefully get the entire object
region = calculate_region(
frame_shape,
box[0],
box[1],
box[2],
box[3],
region_min_size,
2021-02-17 14:23:32 +01:00
)
regions.append(region)
selected_objects.extend(
detect(
detect_config,
object_detector,
frame,
model_config,
region,
objects_to_track,
object_filters,
)
)
refining = True
else:
selected_objects.append(obj)
# set the detections list to only include top, complete objects
# and new detections
detections = selected_objects
if refining:
refine_count += 1
## drop detections that overlap too much
consolidated_detections = []
# if detection was run on this frame, consolidate
if len(regions) > 0:
# group by name
detected_object_groups = defaultdict(lambda: [])
for detection in detections:
detected_object_groups[detection[0]].append(detection)
# loop over detections grouped by label
for group in detected_object_groups.values():
# if the group only has 1 item, skip
if len(group) == 1:
consolidated_detections.append(group[0])
continue
# sort smallest to largest by area
sorted_by_area = sorted(group, key=lambda g: g[3])
for current_detection_idx in range(0, len(sorted_by_area)):
current_detection = sorted_by_area[current_detection_idx][2]
overlap = 0
for to_check_idx in range(
min(current_detection_idx + 1, len(sorted_by_area)),
len(sorted_by_area),
):
to_check = sorted_by_area[to_check_idx][2]
# if 90% of smaller detection is inside of another detection, consolidate
if (
area(intersection(current_detection, to_check))
/ area(current_detection)
> 0.9
):
overlap = 1
break
if overlap == 0:
consolidated_detections.append(
sorted_by_area[current_detection_idx]
)
# now that we have refined our detections, we need to track objects
object_tracker.match_and_update(frame_time, consolidated_detections)
# else, just update the frame times for the stationary objects
else:
object_tracker.update_frame_times(frame_time)
2020-02-16 04:07:54 +01:00
2020-10-24 18:36:04 +02:00
# add to the queue if not full
2021-02-17 14:23:32 +01:00
if detected_objects_queue.full():
2021-01-16 03:52:59 +01:00
frame_manager.delete(f"{camera_name}{frame_time}")
continue
2020-10-24 18:36:04 +02:00
else:
2021-01-16 03:52:59 +01:00
fps_tracker.update()
fps.value = fps_tracker.eps()
2021-02-17 14:23:32 +01:00
detected_objects_queue.put(
(
camera_name,
frame_time,
object_tracker.tracked_objects,
motion_boxes,
regions,
)
)
2021-01-16 03:52:59 +01:00
detection_fps.value = object_detector.fps.eps()
frame_manager.close(f"{camera_name}{frame_time}")