Moving target tracking based on improved Meanshift and Kalman filter algorithm
Long Yu · 2018
The target tracking technology on video is an indispensable technology in the field of computer vision research. MeanShift algorithm is one of the many excellent moving target tracking algorithms. This paper mainly research focuses on MeanShift theory and traditional MeanShift target tracking algorithm. MeanShift does not need parameters and has the advantage of small computation. However, under some conditions the MeanShift does not work well. For the problem of the failure to tracking target when the target move fast or the target occluded happend, improve the MeanShift algorithm and Kalman filter algorithm is introduced to deal with the problem. The improved algorithm is tested, and experimental result show that this algorithm can effectively track the target in occlusion enviorment.