基于梯度方向检测的自适应带宽mean shift目标跟踪算法

A mean shift tracking algorithm with adaptive bandwidth based on gradient-oriented detection

传统核窗宽固定的mean shift跟踪算法不能很好地对尺寸变化的目标进行有效的跟踪。在结合增量试探法和梯度方向检测的基础上,提出了一种适应带宽的mean shift目标跟踪算法。算法能够对逐渐放大和逐渐缩小的目标都能够进行有效的跟踪,解决了增量试探法难以很好地对放大目标进行自适应带宽跟踪的问题,提高了自适应带宽跟踪的准确性。两段不同场景下的运动目标跟踪实验,证实了该算法的有效性。

Classic mean shift based tracking algorithm uses fixed kernel-bandwidth, which limits the performance when the object scale exceeds the size of the tracking window. A mean shift tracking algorithm with adaptive bandwidth is proposed based successive bandwidth trials and gradient-oriented detection in this paper, The algorithm can track the target effectively when the target is changing in amplificatory or contractible size, and it solves the question that successive bandwidth trials can not track the target whose scale is magnifying.The experiments on two videos show that the proposed method improves the performance of mean shift tracking algorithm with adaptive bandwidth.