Add object detection, tracking, mobile notifications, and search to any security camera.
from detection.yolov9 import YOLOv9, fetch
from utils.helpers import draw_bounding_boxes, jit_infer
import cv2
from tinygrad import Tensor
from tinygrad.dtype import dtypes
import numpy as np
from pathlib import Path
if __name__ == "__main__":
from ocsort_tracker import ocsort
ocs_tracker = ocsort.OCSort(max_age=60)
jit_cache = {}
size = "t"
draw = True
class_labels = fetch('https://raw.githubusercontent.com/pjreddie/darknet/master/data/coco.names').read_text().split("\n")
Path('./test_outputs').mkdir(parents=True, exist_ok=True)
cap = cv2.VideoCapture("test/videos/MOT16-03.mp4")
w, h = int(cap.get(3)), int(cap.get(4))
if draw: out = cv2.VideoWriter(f"test_outputs/out_{size}.mp4", cv2.VideoWriter_fourcc(*"mp4v"), 30, (w, h))
model = YOLOv9(size, res=960)
expected = 156
i = 0
ppl = set()
while True:
ret, im0 = cap.read()
if not ret: break
im = im0
im = Tensor(im).cast(dtypes.float32)
pred = jit_infer(model, x=im, jit_cache=jit_cache).numpy()
online_targets = ocs_tracker.update(pred, 0.25)
preds = []
for x in online_targets:
if x.tracklet_len < 1 or x.speed < 2.5: continue
if x.class_id == 0 and x.track_id not in ppl: ppl.add(x.track_id)
preds.append(np.array([x.tlwh[0], x.tlwh[1], x.tlwh[0] + x.tlwh[2], x.tlwh[1] + x.tlwh[3], x.score, x.class_id, x.track_id]))
print("ppl =",len(ppl))
if draw:
_, buffer = cv2.imencode(".jpg", im0)
out.write(draw_bounding_boxes(buffer, preds, class_labels))
i+=1
print("frame",i)
cap.release()
if draw: out.release()
assert len(ppl) == expected