Meanshift跟踪算法针对传统的MeanShift跟踪算法在目标发生遮挡时容易导致目标丢失的情况,提出了一种改进的MeanShift跟踪算法.将多尺度空间理论、Kalman滤波器与遮挡算法相结合,当目标发生遮挡后,利用Kalman估计目标信息量,能对目标尺寸有后续跟踪能力.实验结果表明,当目标发生遮挡后,改进的跟踪算法对目标无论增大或减小都能连续地、自动地选择大小合适的跟踪窗口.-Meanshift tracking algorithmMeanshift tracking algorithm for traditional Meanshift in target tracking algorithm when the shade is easy to cause the condition of the missing target, an improved Meanshift tracking algorithm is proposed. The multi-scale space theory, combined with block algorithm, Kalman filter the shade when the target, the target information is estimated based on the Kalman, follow-up tracking ability to target size. The experiment results show that when the target after shade, the improved algorithm of target tracking can both increase/decrease continuously and automatically　appropriate to the size of the tracking window.