Object detection is done using a YOLOv8 model trained on images of the various predators local to [donutorelse]. The model is running on a Raspberry Pi and getting images from a standard webcam.