基于边缘直方图的粒子滤波跟踪算法及其实现

以颜色直方图为特征的运动目标跟踪算法容易受到光线变化及视场内其它同色目标的干扰.采用运动目标的边缘方向直方图作为特征,利用序列重要性采样原理和粒子滤波算法实现了对人体运动目标的跟踪.实验显示了该算法在光线变化及存在同色目标干扰时能够有效跟踪目标.在算法实现过程中,采用积分图计算边缘方向直方图,减少了计算时间,提高了计算速度,达到了实时跟踪的效果.

Particle filter tracking algorithm based on edge histogram and it's implementation

The traditional tracking method based on color histogram of a moving target tends is easy to be affected,when the illumination changes and/or similar-color objects exist in the field.By choosing the edge orientation histogram of a moving target as the feature,sequential importance sampling theory and particle filter algorithm are used to solve the tracking problem of a human body target.Experiments show that the tracker works effectively even when illumination changes and/or similar target exists.During the implementation,integral gradient histogram maps are built to reduce the processing time,the calculation speed is much improved and real-time tracking is achieved.