一种新的粒子滤波目标跟踪算法

A New Particle Filter Object Tracking Algorithm

为了进一步提高目标跟踪的性能,采用一种新的建议分布构造方法,即利用状态分割技术和平行扩展卡尔曼滤波技术构造建议分布.依据该方法构造的建议分布相对传统的方法提高了粒子滤波估计的准确性.同时,在新的跟踪算法框架中,将颜色模型和形状模型自适应地融合,并结合了一种新的模型更新方法.实验结果证明,该跟踪算法具有较强的适应性和有效性.

To improve the performance of object tracking, a particle filter algorithm was proposed which uses state partition technique and parallel extended kalman filter to construct proposal distribution. This proposal enhances the estimation accuracy compared to traditional filters. At the same time, color model and shape model are adaptively fused in the framework, a new model update scheme is also combined. The experimental results show the availability of the proposed algorithm.