一种基于颜色分布的混合视频跟踪方法

A Hybrid Algorithm of Object Tracking Based on Color Distribution

传统的Mean shift算法虽然具有简单快速的特点,但在目标被遮挡的情况下无法进行有效的跟踪。与此同时,以Monte Carlo随机模拟理论为基础的粒子滤波虽然可以很好地解决这一问题,但是由于结果的好坏与粒子的数目成正比,计算耗时无法满足系统的实时性要求。该文基于颜色直方图分布,引入自适应选择粒子样本数的采样策略,有效地融合传统Mean shift算法的简单快速和粒子滤波跟踪算法的抗遮挡的优点,在保证跟踪效果的同时减少了跟踪的总体时间花费,有效提高了设计的跟踪系统的实时性。实验证明,该方法在实际的目标跟踪中是有效和稳健的。

Though traditional meanshift method has the virtue of simplicity and availability, it does not work well when the target gets an occlusion. In the meanwhile, particle filter can solve this problem easily. Unfortunately, its performance relies heavily on the numbers of the used particles. It makes the tracking technique by particle filtering have difficulty in satisfying the requirement of real time computing, To settle the problem, this article brings about a hybrid algorithm by combining the mean shift and the particle filter tracking technique on the basis of the color histogram distribution. By adopting the strategy that the number of particles is adaptively determined, it amalgamates the virtues of the two techniques. It reduces the computational cost and ensures the performance simultaneously. The experimental results show that the proposed method is effective and robust.