An Improved Moving Object Tracking Algorithm Based on Camshift for Dynamic Backg

Tracking moving objects from video frames is one of the key techniques in computer vision.An improved Camshift algorithm for dynamic background is presented in this paper. The algorithm uses Harriscorner matching to estimate background movement parameters, adopts two-frame difference to detectmoving objects, combines merged hue histogram with edge direction histogram to describe moving objectsand employs Camshift to track moving objects. The experiments show that the impr