mirror of
https://github.com/blakeblackshear/frigate.git
synced 2026-02-07 03:35:26 +03:00
optimize motion and velocity estimation
This commit is contained in:
Josh Hawkins
parent
cd64399fe5
commit
3583163eff
@ -60,7 +60,7 @@ REQUEST_REGION_GRID = "request_region_grid"
|
||||
|
||||
# Autotracking
|
||||
|
||||
AUTOTRACKING_MAX_AREA_RATIO = 0.5
|
||||
AUTOTRACKING_MAX_AREA_RATIO = 0.6
|
||||
AUTOTRACKING_MOTION_MIN_DISTANCE = 20
|
||||
AUTOTRACKING_MOTION_MAX_POINTS = 500
|
||||
AUTOTRACKING_MAX_MOVE_METRICS = 500
|
||||
|
||||
@ -58,6 +58,8 @@ class PtzMotionEstimator:
|
||||
self.ptz_metrics = ptz_metrics
|
||||
self.ptz_start_time = self.ptz_metrics["ptz_start_time"]
|
||||
self.ptz_stop_time = self.ptz_metrics["ptz_stop_time"]
|
||||
self.last_update = 0
|
||||
self.update_interval = 1 / (self.camera_config.detect.fps / 3)
|
||||
|
||||
self.ptz_metrics["ptz_reset"].set()
|
||||
logger.debug(f"{config.name}: Motion estimator init")
|
||||
@ -85,13 +87,27 @@ class PtzMotionEstimator:
|
||||
)
|
||||
|
||||
self.coord_transformations = None
|
||||
self.last_update = 0
|
||||
|
||||
if ptz_moving_at_frame_time(
|
||||
ptz_moving = ptz_moving_at_frame_time(
|
||||
frame_time, self.ptz_start_time.value, self.ptz_stop_time.value
|
||||
)
|
||||
|
||||
if (
|
||||
self.camera_config.onvif.autotracking.zooming != ZoomingModeEnum.disabled
|
||||
and (
|
||||
(frame_time - self.last_update > self.update_interval and ptz_moving)
|
||||
or frame_time == self.ptz_start_time.value
|
||||
or frame_time == self.ptz_stop_time.value
|
||||
)
|
||||
) or (
|
||||
self.camera_config.onvif.autotracking.zooming == ZoomingModeEnum.disabled
|
||||
and ptz_moving
|
||||
):
|
||||
logger.debug(
|
||||
f"{camera}: Motion estimator running - frame time: {frame_time}"
|
||||
)
|
||||
self.last_update = frame_time
|
||||
|
||||
frame_id = f"{camera}{frame_time}"
|
||||
yuv_frame = self.frame_manager.get(
|
||||
@ -215,8 +231,8 @@ class PtzAutoTracker:
|
||||
maxlen=round(camera_config.detect.fps * 1.5)
|
||||
)
|
||||
self.tracked_object_metrics[camera] = {
|
||||
"max_target_box": 1
|
||||
- (AUTOTRACKING_MAX_AREA_RATIO ** self.zoom_factor[camera])
|
||||
"max_target_box": AUTOTRACKING_MAX_AREA_RATIO
|
||||
** (1 / self.zoom_factor[camera])
|
||||
}
|
||||
|
||||
self.calibrating[camera] = False
|
||||
@ -521,7 +537,11 @@ class PtzAutoTracker:
|
||||
camera_height = camera_config.frame_shape[0]
|
||||
|
||||
# Extract areas and calculate weighted average
|
||||
areas = [obj["area"] for obj in self.tracked_object_history[camera]]
|
||||
# grab the largest dimension of the bounding box and create a square from that
|
||||
areas = [
|
||||
max(obj["box"][2] - obj["box"][0], obj["box"][3] - obj["box"][1]) ** 2
|
||||
for obj in self.tracked_object_history[camera]
|
||||
]
|
||||
|
||||
filtered_areas = (
|
||||
remove_outliers(areas)
|
||||
@ -686,19 +706,20 @@ class PtzAutoTracker:
|
||||
camera_height = camera_config.frame_shape[0]
|
||||
camera_fps = camera_config.detect.fps
|
||||
|
||||
# estimate_velocity is a numpy array of bbox top,left and bottom,right velocities
|
||||
velocities = obj.obj_data["estimate_velocity"]
|
||||
logger.debug(f"{camera}: Velocities from norfair: {velocities}")
|
||||
|
||||
# if we are close enough to zero, return right away
|
||||
if np.all(np.round(velocities) == 0):
|
||||
return True, np.zeros((2, 2))
|
||||
return True, np.zeros((4,))
|
||||
|
||||
# Thresholds
|
||||
x_mags_thresh = camera_width / camera_fps / 2
|
||||
y_mags_thresh = camera_height / camera_fps / 2
|
||||
dir_thresh = 0.93
|
||||
delta_thresh = 12
|
||||
var_thresh = 5
|
||||
delta_thresh = 20
|
||||
var_thresh = 10
|
||||
|
||||
# Check magnitude
|
||||
x_mags = np.abs(velocities[:, 0])
|
||||
@ -722,7 +743,6 @@ class PtzAutoTracker:
|
||||
np.linalg.norm(velocities[0]) * np.linalg.norm(velocities[1])
|
||||
)
|
||||
dir_thresh = 0.6 if np.all(delta < delta_thresh / 2) else dir_thresh
|
||||
print(f"cosine sim: {cosine_sim}")
|
||||
invalid_dirs = cosine_sim < dir_thresh
|
||||
|
||||
# Combine
|
||||
@ -752,10 +772,10 @@ class PtzAutoTracker:
|
||||
)
|
||||
)
|
||||
# invalid velocity
|
||||
return False, np.zeros((2, 2))
|
||||
return False, np.zeros((4,))
|
||||
else:
|
||||
logger.debug(f"{camera}: Valid velocity ")
|
||||
return True, np.mean(velocities, axis=0)
|
||||
return True, velocities.flatten()
|
||||
|
||||
def _get_distance_threshold(self, camera, obj):
|
||||
# Returns true if Euclidean distance from object to center of frame is
|
||||
@ -836,7 +856,7 @@ class PtzAutoTracker:
|
||||
# ensure object is not moving quickly
|
||||
below_velocity_threshold = np.all(
|
||||
np.abs(average_velocity)
|
||||
< np.array([velocity_threshold_x, velocity_threshold_y])
|
||||
< np.tile([velocity_threshold_x, velocity_threshold_y], 2)
|
||||
) or np.all(average_velocity == 0)
|
||||
|
||||
below_area_threshold = (
|
||||
@ -938,7 +958,7 @@ class PtzAutoTracker:
|
||||
camera_height = camera_config.frame_shape[0]
|
||||
camera_fps = camera_config.detect.fps
|
||||
|
||||
average_velocity = np.zeros((2, 2))
|
||||
average_velocity = np.zeros((4,))
|
||||
predicted_box = obj.obj_data["box"]
|
||||
|
||||
centroid_x = obj.obj_data["centroid"][0]
|
||||
@ -966,7 +986,6 @@ class PtzAutoTracker:
|
||||
# this box could exceed the frame boundaries if velocity is high
|
||||
# but we'll handle that in _enqueue_move() as two separate moves
|
||||
current_box = np.array(obj.obj_data["box"])
|
||||
average_velocity = np.tile(average_velocity, 2)
|
||||
predicted_box = (
|
||||
current_box
|
||||
+ camera_fps * predicted_movement_time * average_velocity
|
||||
@ -1010,7 +1029,10 @@ class PtzAutoTracker:
|
||||
zoom = 0
|
||||
result = None
|
||||
current_zoom_level = self.ptz_metrics[camera]["ptz_zoom_level"].value
|
||||
target_box = obj.obj_data["area"] / (camera_width * camera_height)
|
||||
target_box = max(
|
||||
obj.obj_data["box"][2] - obj.obj_data["box"][0],
|
||||
obj.obj_data["box"][3] - obj.obj_data["box"][1],
|
||||
) ** 2 / (camera_width * camera_height)
|
||||
|
||||
# absolute zooming separately from pan/tilt
|
||||
if camera_config.onvif.autotracking.zooming == ZoomingModeEnum.absolute:
|
||||
@ -1061,24 +1083,15 @@ class PtzAutoTracker:
|
||||
< self.tracked_object_metrics[camera]["max_target_box"]
|
||||
else self.tracked_object_metrics[camera]["max_target_box"]
|
||||
)
|
||||
zoom = (
|
||||
2
|
||||
* (
|
||||
limit
|
||||
/ (
|
||||
self.tracked_object_metrics[camera]["target_box"]
|
||||
+ limit
|
||||
)
|
||||
)
|
||||
- 1
|
||||
)
|
||||
ratio = limit / self.tracked_object_metrics[camera]["target_box"]
|
||||
zoom = (ratio - 1) / (ratio + 1)
|
||||
logger.debug(f"{camera}: Zoom calculation: {zoom}")
|
||||
if not result:
|
||||
# zoom out with special condition if zooming out because of velocity, edges, etc.
|
||||
zoom = -(1 - zoom) if zoom > 0 else -(zoom + 1)
|
||||
zoom = -(1 - zoom) if zoom > 0 else -(zoom * 2 + 1)
|
||||
if result:
|
||||
# zoom in
|
||||
zoom = 1 - zoom if zoom > 0 else (zoom + 1)
|
||||
zoom = 1 - zoom if zoom > 0 else (zoom * 2 + 1)
|
||||
|
||||
logger.debug(f"{camera}: Zooming: {result} Zoom amount: {zoom}")
|
||||
|
||||
@ -1199,8 +1212,8 @@ class PtzAutoTracker:
|
||||
)
|
||||
self.tracked_object[camera] = None
|
||||
self.tracked_object_metrics[camera] = {
|
||||
"max_target_box": 1
|
||||
- (AUTOTRACKING_MAX_AREA_RATIO ** self.zoom_factor[camera])
|
||||
"max_target_box": AUTOTRACKING_MAX_AREA_RATIO
|
||||
** (1 / self.zoom_factor[camera])
|
||||
}
|
||||
|
||||
def camera_maintenance(self, camera):
|
||||
|
||||
Loading…
Reference in New Issue
Block a user