How to increase the size of history_observations.

[AyushSarangi]

Search before asking

• I have searched the Yolo Tracking issues and found no similar bug report.

Question

How to increase the size of history_observations. i have tried updating STrack.history_observations = deque([],maxlen = 100) changes are not reflecting in output. what to do?

from boxmot import BoTSORT
from boxmot.trackers.botsort.bot_sort import STrack
from collections import deque

Initialize the tracker

STrack.history_observations = deque([],maxlen=100)

tracker = BoTSORT(
model_weights = Path('osnet_x0_25_msmt17.pt'), # which ReID model to use
device = 'cuda:0',
fp16 = False,
)

Initialize YOLO model

yolo_model = YOLOv10('/content/weights/yolov10b.pt')

Open the input video

input_video_path = '/content/5330828-hd_1920_1080_30fps.mp4'
vid = cv2.VideoCapture(input_video_path)

if not vid.isOpened():
print("Error: Could not open input video.")
exit()

Define the output video path

output_video_path = '/content/BoTSORT_output_tracking_video.mp4'

Create the video writer

out = create_video_writer(vid, output_video_path)
frame_count = 500

trail_path = {}

while frame_count:
frame_count -= 1
ret, im = vid.read()
if not ret:
break

try:
    # Run the YOLO model on the frame
    results = yolo_model(im)

    # Convert the detections to the required format: N X (x, y, x, y, conf, cls)
    dets = []
    for result in results:
        for detection in result.boxes.data.cpu().numpy():
            x1, y1, x2, y2, conf, cls = detection
            if int(cls)==0:
              dets.append([x1, y1, x2, y2, conf, int(cls)])
    dets = np.array(dets)

    # Update tracker with detections
    tracker.update(dets, im)
    tracker.plot_results(im, show_trajectories = True)

    # trail_hist
    for a in tracker.active_tracks:
      if a.cls == 0:
        if len(a.history_observations)>2:
          trail_path[a.id] = a.history_observations

    # Write the frame to the output video
    out.write(im)

except Exception as e:
    print(f"An error occurred: {e}")
    break

vid.release()
out.release()
cv2.destroyAllWindows()

print(f"Tracking video saved to {output_video_path}")

[YoungjaeDev]

Isn't that MAX_AGE?

[mikel-brostrom]

You can now use MAX_AGE. I implemented dynamic deques based on this value 🚀 . Available in the new release!