operators/object_detection.py

import numpy as np
import pyarrow as pa

from dora import DoraStatus
from ultralytics import YOLO

pa.array([])

CAMERA_WIDTH = 960
CAMERA_HEIGHT = 540

model = YOLO("yolov8n.pt")


class Operator:
    """
    Infering object from images
    """

    def on_event(
        self,
        dora_event,
        send_output,
    ) -> DoraStatus:
        if dora_event["type"] == "INPUT":
            frame = (
                dora_event["value"].to_numpy().reshape((CAMERA_HEIGHT, CAMERA_WIDTH, 3))
            )
            frame = frame[:, :, ::-1]  # OpenCV image (BGR to RGB)
            results = model(frame, verbose=False)  # includes NMS
            boxes = np.array(results[0].boxes.xyxy.cpu())
            conf = np.array(results[0].boxes.conf.cpu())
            label = np.array(results[0].boxes.cls.cpu())
            # concatenate them together
            arrays = np.concatenate((boxes, conf[:, None], label[:, None]), axis=1)

            send_output("bbox", pa.array(arrays.ravel()), dora_event["metadata"])

        return DoraStatus.CONTINUE