Files

150 lines
3.3 KiB
Python

#!/usr/bin/env python3
import json
import os
import tempfile
import time
from pathlib import Path
from ultralytics import YOLO
IMAGE_PATH = Path("/dev/shm/robot_frame.jpg")
OUTPUT_PATH = Path("/dev/shm/r1_detections.json")
MODEL = "yolo26n.pt"
CONFIDENCE = 0.35
IMAGE_SIZE = 640
def atomic_write_json(path: Path, data):
with tempfile.NamedTemporaryFile(
mode="w",
dir=path.parent,
delete=False,
suffix=".tmp",
) as f:
json.dump(
data,
f,
ensure_ascii=False,
indent=2,
)
tmp = f.name
os.replace(tmp, path)
def main():
print(f"[YOLO] Loading {MODEL}")
model = YOLO(MODEL)
print("[YOLO] Ready")
print(f"[YOLO] Camera: {IMAGE_PATH}")
print(f"[YOLO] Output: {OUTPUT_PATH}")
last_mtime = None
while True:
try:
if not IMAGE_PATH.exists():
time.sleep(0.05)
continue
mtime = IMAGE_PATH.stat().st_mtime_ns
if mtime == last_mtime:
time.sleep(0.01)
continue
last_mtime = mtime
results = model.predict(
source=str(IMAGE_PATH),
conf=CONFIDENCE,
imgsz=IMAGE_SIZE,
verbose=False,
)
if not results:
continue
result = results[0]
height, width = result.orig_shape
detections = []
if result.boxes is not None:
for box in result.boxes:
class_id = int(box.cls[0])
confidence = float(box.conf[0])
x1, y1, x2, y2 = [
float(v)
for v in box.xyxy[0]
]
cx = (x1 + x2) / 2.0
cy = (y1 + y2) / 2.0
detections.append({
"class_id": class_id,
"class": result.names[class_id],
"confidence": confidence,
"box": {
"x1": x1,
"y1": y1,
"x2": x2,
"y2": y2,
},
"center": {
"x": cx,
"y": cy,
},
"normalized_center": {
"x": cx / width,
"y": cy / height,
},
})
data = {
"timestamp": time.time(),
"width": width,
"height": height,
"detections": detections,
}
atomic_write_json(
OUTPUT_PATH,
data,
)
objects = ", ".join(
f"{d['class']} {d['confidence']:.2f}"
for d in detections
)
if objects:
print(f"[YOLO] {objects}")
else:
print("[YOLO] no objects")
except KeyboardInterrupt:
print("\n[YOLO] stopped")
break
except Exception as e:
print(f"[YOLO] error: {e}")
time.sleep(0.25)
if __name__ == "__main__":
main()