Research On Moving Object Tracking Based On CamShift Algorithm

With the computer software and hardware’s fast development, intelligent video surveillance technology based on computer vision is increasingly important. And moving target tracking technology is one of the most important intelligent video surveillance technologies, and is also widely used in many fields, such as Medical Diagnosis, Autopilot, Robot and Human-Computer Interaction. Cam Shift algorithm uses the target’s H component in HSV colorspace as target characteristic, so this algorithmhas low sensitivity to the target’s shape change. Besides this algorithm has low complexity and computation, good real-time performance, which make it be widely used in moving target tracking.This paper based on Cam Shift mainly gives two improvements. On one hand combining Cam Shiftalgorithmwith moving target detection algorithm, and to reduce the computation brought by detection algorithm, this paper uses small region image containing target onlyto calculate, this method greatly improves the operation speed of the whole algorithm.On the other hand, to solve the problem that tracking result will appears obvious shift when moving target is closing the occlusion having similar color with the target in Cam Shift algorithm, this paper proposes the solution that using single Gaussian algorithm to build the background model of the moving target It is proved by experiments that this solution can effectively solves the problem that in original Cam Shift algorithm when the target approaches the fixed obstruction having the similar color with it, tracking appears shift and losing, and increases the accuracy of original algorithm in some extent.