Design And Implementation Of Moving Object Tracking Algorithms Based On Mean Shift In The Visual Surveillance

In the modern time of pursuing for the high automatization, there is very strong need for a smart visual surveillance system, which can monitor the moving object in some surveillance scape automatically and consecutively. In most cases, this intelligent system has some functions, such as moving object detecting and tracking. This paper study the moving object tracking mainly. Object tracking achieves moving contrail of monitored object, and provides credible data information for moving analysis. On the basis of referring to algorithms of relating technologies, the main work in this paper is as follows.In this paper, we firstly analyzes the current situation of object tracking algorithms in the visual surveillance, and expounds the existent issues of the algorithm. Secondly, the theory of Mean Shift and its application in the object tracking algorithms are described in detail. Third-ly, after analyzing the advantages and disadvantages of the tracking algorithms of the Mean Shift, the paper raises a question that when the colors of the background and the object are alike, or the illumination changes, the tracking algorithm of Mean Shift is not roust. For this question, we propose a tracking algorithm of Mean Shift based on adaptive fusion of color cue and local binary patter(LBP) texture cue. This algorithm describes the shape of the object by local binary patter(LBP) texture cue. It supplys a gap of color cue. Finally, for this question that the conventional tracking algorithm of Mean Shift can not effectively tracking the object moving quickly, there proposes a tracking algorithm of Mean Shift based on Gray Model GM(1,1). According to the previous location information of the real objects, this peper uses GM(1,1) to forecast the next location, which is used to the object origination. And then, it uses the tracking algorithm of Mean Shift to find the real location of the objects in the neighborhood of the origination. Then the question about the tracking of objects moving quickly will be solved well.