Research on Embedded Animal Recognition System Based on YOLO

A lot of manpower and material resources are consumed in the classification and identification of wild animals. The detection cycle is long and the real-time performance is poor. In this paper, the YOLOv1 algorithm is applied to animal detection, and the training model is embedded into K210 chip to design a portable detection product. Experimental results show that this design can effectively realize animal classification, fast target detection and high detection accuracy. The design can meet the demand of real-time detection of wild animals. It can also track the activity range and habits of the population, which provides a good scientific tool for ecological civilization construction and animal diversity research.