Research On Anti-occlusion Algorithm Of Tracking Moving Object

Target tracking as an important research area in the applications of computer science, it plays a very important role in our daily lives. Such as in the military area of warning in the air and object precision strike. In its civil area, target tracking has been used to control the traffic and handle the criminal cases and also plays important role in these area. Signal processing is the basic of graphic tracking, it combined with the result of the research, which contain the artificial intelligence and computer control. With the development of these technologies, the effects of target tracking have been proved most. The essence of the target tracking is calculating the target's information of state and position, and tracking the target by itself. The difficult is how to recognize the target properly and effectively in the complex background. Especially, the occlusion problem, which brings the adverse effects, then gradually became the limitation of the practical tracking algorithm.In this paper analysis the algorithm which based on traditional correlation, then consider the satiable of tracking from correlation and template update. We prove the s ability through the search method. To deal with the difficult in moving Object tracking in the clutter background, this paper handle it from speed of object, and at the same time we combine Kalman filer to argue the object moving. After various arithmetic of tracking is researched, a better adaptability multi-template tracking arithmetic based on template self-correlation is presented. And a new distilling template algorithm of replacing temple guide line, object Occlusion judgment guide line and etc is provided. It is a useful arithmetic in background that is proved effective and feasible by a series of experiments.