中文摘要：

针对目标存在遮挡、尺寸动态变化及背景影响情况下,传统Mean Shift目标跟踪算法精度低的问题,提出一种基于多图像块表示目标且融合SURF特征的改进Mean Shift跟踪算法。设计自适应分块方法,通过分块、判断遮挡、自适应融合3个步骤获得鲁棒的被跟踪目标动态模型;在跟踪过程中,考虑到背景对目标模板的影响,利用改进的权重计算方法更新待融合图像块的权重,保证跟踪的准确性;通过SURF特征匹配对初步跟踪结果做校正,提高算法的跟踪精度。仿真实验验证了该算法的有效性。

英文摘要：

Occlusion,dynamically size change and the effects of background cause the problem of low accuracy of the traditional Mean Shift tracking algorithm.In view of this situation,an improved Mean Shift tracking algorithm was proposed.The algorithm based on multi-image block indicating object and fusing SURF features.The method was designed to achieve adaptive block,a dynamic model of the tracked object was got by three steps：block,judge occlusion,adaptive fusion.In the tracking process,taking the impact of the background on the object template into account,the weights of image blocks to be fused were updated using the improved weight right calculation method to ensure the accuracy of tracking.SURF features were used to regulate the preliminary tracking results to improve tracking accuracy.Simulation results show the effectiveness of the proposed algorithm.