Research on Infrared Small Target Detection Algorithm Based on YOLOv5 and Enhanced KCF Algorithm with Kalman Filter

This paper mainly compares and studies three target detection algorithms-YOLOv5, KCF algorithm and the enhanced KCF algorithm with Kalman Filter. The features, advantages and disadvantages of these algorithms are summarized. Also, this paper proposes a combined target detection algorithm that uses YOLOv5 to identify the target in the first frame and enhanced KCF algorithm with Kalman Filter to track the target in the following frames. The proposed algorithm is tested and validated on an open-source infrared small target dataset. The results show that the proposed algorithm can achieve a high recognition rate without consuming too many computational resources, and it is especially robust compared to pure KCF algorithm when dealing with jerky videos.