부산대학교 도서관

Understanding the behavior of honey bees is essential for maintaining a healthy bee colony. Hive monitoring systems are crucial for this purpose. In the last few years, computer vision and deep learning have become widely used in such systems. This paper uses a deep learning approach for detecting and tracking honey bees. Firstly, for detection, we employed the YOLO model using a data set of 1000 ground truth images. Secondly, for tracking, we used the StrongSORT approach. Results show that the detector performs well in both classes of honey bees (with or without pollen). The models based on this approach provide considerably good average tracking accuracy with low latency. Thus, this procedure is reliable and can be used in the future for real-time monitoring systems.