Improved Single Target Tracking Learning Detection Algorithm

In order to improve the robustness and speed of single target tracking, this paper proposes an improved tracking learning method. The purpose is to improve the tracking module in the traditional tracking learning detection (TLD) algorithm. By introduced oriented fast and rotated brief (ORB) feature points and keep the original uniform distribution point to improve the robustness and speed up execution of tracking. The experiment shows that the improved TLD algorithm has strong robustness in different environments, and the feat can quickly and accurately track the single object. The proposed algorithm can overcome the tracking failures caused by objects with partial occlusion, fast motion and leave the tracking field of vision, and has better robustness. It is experimentally verified that it has the veracity and the execution speed, compared with the traditional TLD algorithm.