结合帧差法与Mean Shift的抗遮挡跟踪算法

Anti-occlusion tracking algorithm combining frame difference method and Mean Shift

针对目标严重遮挡后,运动状态发生改变时,传统的基于运动预测的算法无法有效跟踪的问题,提出一种基于帧差法的改进算法。引入巴氏系数(Bhattacharyya)作为目标是否发生遮挡的判据;当发生遮挡时,帧差法检测目标,再次检测到目标时将此位置作为Mean Shift迭代的起始位置;最后正常跟踪时采用卡尔曼滤波预测目标位置,减少迭代次数。实验结果表明,当目标在严重遮挡后,运动状态改变时,基于运动预测的算法将无法跟踪目标,改进算法能够重新跟踪目标。

Since the traditional motion prediction based algorithm cannot conduct effective tracking when the moving state of the target changes after severe occlusion,an improved algorithm based on the frame difference method is proposed. Bhattacharyya is introduced as the judgment criterion of whether the target is occluded or not. The frame difference method is used to detect the target when occlusion occurs. The position where the target is again detected is taken as the starting position of the Mean Shift iteration. The Kalman filtering is used to predict the target position during normal tracking,so as to reduce iteration times. The experimental results show that the improved algorithm can track the target again while the motion prediction based algorithm cannot track the target when the moving state of the target changes after severe occlusion.