Application of the image detection technology on dairy goat management

The identification of goat breeds in dairy farms often relies on the experience of breeders. However, the traditional method of manual breed identification requires significant training and depends on the professional experience of breeder. This process is time-consuming and can lead to misjudgments, particularly when the breeder lacks experience. To address these challenges, this study applied YOLO technology to conduct research on the detection and classification of goat breeds. Several advantages were observed in this study: (1) The augmentation technique was integrated with the YOLO model, and the anchor boxes were adjusted by k-means to improve the recognition accuracy; (2) The knowledge and research data were collected from breeders in a dairy goat farm located in southern Taiwan. (3) The results demonstrated a substantial improvement in average precision (AP) values, with mAP reaching 90.14%. Based on the findings, the outcomes of this research are expected to serve as a valuable reference for future studies on classification and detection technologies.