基于粒子滤波和多特征融合的目标跟踪算法

Object Tracking Algorithm Based on Particle Filter and Multi-feature Fusion

为了克服单纯基于颜色特征的跟踪方法在复杂环境下易导致跟踪失败的缺点,提出了将颜色和结构信息相结合的跟踪方法.利用基于HSV颜色空间的加权颜色直方图表示目标的颜色模型,利用目标的灰度图像建立结构模型,并将两者融合于粒子滤波的框架中,结合的纽带就是粒子权值的计算,同时自适应的调整颜色和结构信息的融合系数.实验表明,该算法的稳定性较高,同时提高了跟踪的精度.

In order to solve the problem that object tracking only based on color information always falls under the conditions of cluttered backgrounds, a tracking algorithm combining the color and structural information was proposed. Weighted color histogram based on HSV was used to describe the color model of the target, a structural model was developed by using target gray level image. The two features were fused in the frame of particle filter, the link is the calculation of the particle weight. Meanwhile, the weighs of fusing for color and structural were adjusted adaptively. The experimental results show that the proposed algorithm has more stability and higher accuracy.