一种基于Yolo V4-tiny和KCF的目标跟踪融合算法分析

Analysis of a Target Tracking Fusion Algorithm Based on Yolo V4-tiny and KCF

为解决由于目标尺度变换而让KCF跟踪算法在跟踪过程中产生目标丢失的问题,提出一种基于YoloV4-tiny和KCF的目标跟踪融合算法。在实际跟踪过程中,利用YoloV4-tiny目标检测算法形成追踪闭环控制,使算法获得对丢失目标进行重新跟踪的能力,将所提出的算法与传统的BOOSTING、TLD、CSRT以及KCF目标跟踪算法进行跟踪性能比对研究。分析表明,该算法具有较好的跟踪效果、实时性以及丢失目标后重新跟踪的能力,适用于需要高实时性,高跟踪精度的目标跟踪场景。

In order to solve the problem that the KCF tracking algorithm loses the target in the tracking process due to the target scale transformation,this paper proposes a target tracking fusion algorithm based on Yolo V4-tiny and KCF.In the actual tracking process,Yolo V4-tiny target detection algorithm is used to form tracking closed-loop control,so that the algorithm can obtain the ability to re track the lost target.The proposed algorithm is compared with the traditional BOOSTING,TLD,CSR T and KCF target tracking algorithms for tracking performance.The analysis shows that the algorithm has good tracking effect,real-time performance and the ability to re track after losing the target,and is suitable for target tracking scenes requiring high real-time performance and high tracking accuracy.