An improved target tracking scheme via integrating mean-shift with TLD algorithm
Neng Fan, Na Huang, Fengmin Yu, Yaoxian Song, Shijie Zhao, Zhao Tan · 2017
Tracking-Learning-Detection (TLD) is an excellent long-term tracking method which has the advantages of high accuracy of tracking rate and self-detection mechanism. Noting that TLD algorithm is sensitive to illumination change and clutter results in drift even missing, and the corresponding tracker designed based on the pyramid Lucaks-Kanade optical flow method needs vast computation. To overcome these shortcomings, an improved target tracking scheme by integrating mean-shift and TLD algorithm is proposed. The designed scheme improves the ability of resistance to shade and increases the processing speed through setting the reasonable iterative starting point of mean-shift algorithm. Meanwhile, by combining self-detection with on-line learning mechanism, we can solve the problem of goals lost in tracking process. Finally, experimental results are provided to demonstrate that the proposed method can properly detect and accurately track a target in complex scenes.