Robust Human Tracking Algorithm Applied for Occlusion Handling
Jin Zhang, Hongguang Sun, Weizhou Guan, Jialiang Wang, Yannan Xie, Bingnan Shang · 2010
The robust tracking algorithm is presented that can automatically track the human including partial or total occlusion in the paper. Firstly, as a result of the use of Camshift, the effectiveness of particle is improved, and the tracking window can change scale with the target adaptively; Secondly, the multiple cues integration technique is brought into the framework of Camshift guided particle filter method to improve tracking robustness; Finally, an occlusion strategy is proposed to deal with the full occlusion for a long time. The occlusion strategy based on the direct color-motion integration mechanism is adopted, which can realize steady target tracking, and reinitialize tracking automatically when the lost object reappears. The experiments show that the algorithm is effective to moving target tracking with occlusion. Its effectiveness is proved through quantitative analysis.