Improved Meanshift Tracking Algorithm Based on Optical flow

Abstract

The mean shift tracker has difficulty in tracking fast moving targets and suffers from local optimal problem. To overcome the limitation of the mean-shift tracking algorithm, a new approach is proposed by integrating the mean-shift algorithm and optical flow methods. Even with n the rough position, the mean-shift algorithm achieves precise tracking of the target. Several tracking experiments show that the proposed algorithm can effectively track fast moving target and overcome the tracking error cumulating problems.