A Novel Algorithm for Background Updating and Target Tracking
Hui-Zhong Zhuang, Han Li · 2008
A novel algorithm of three-layer probability model based on synchronous detection lines is proposed to solve the problems of background updating and target tracking. Firstly, the state of each pixel point on the same synchronous detection line is judged in the probability model of pixel points. Secondly, the state of each synchronous detection line is estimated by maximum likelihood estimation so that the problem of background updating can be solved. Thirdly, the problem of target tracking is solved by the state distribution of the synchronous detection line. Finally, the coordination and shadow problems for multiple detection region groups are also discussed in this paper. The results of a practical system are shown that this algorithm has better effect to solve the problems of background updating and target tracking.