具有遮挡鲁棒性的目标中心距离加权的跟踪算法

本文提出了一种实时跟踪算法,以目标中心距离加权的目标图像直方图作为模板,采用mean-shift迭代方法进行目标定位;当目标被部分遮挡时,用"分块匹配"的方法提高算法鲁棒性;对于超过一定像素的较大目标,本算法进行"降采样",大大减小运算量,从而实现了对大尺度目标的实时跟踪.实时视频流的实际跟踪系统验证了该方法的有效性.

A robust occluded object tracking based on object center distance-weighting

A real-time image tracking algorithm is proposed,which gives small weights to pixels farther from the object center and uses the quantized image color scales as a template. It identifies the target's location by mean-shift iteration method and uses block match method to improve the algorithm in occluded object tracking. A decimation method is proposed to track large-sized targets. Real-time experimental results verify the effectiveness of the proposed algorithm.