关键词: Meanshift,动态阈值,五帧差分,模板更新

Abstract: To improve the traditional frame difference of moving object detection method in which the hole edge tend to appear and to reduce the disadvantages of tracking object losing because traditional Meanshift algorithm used in video monitoring is easy to be disturbed by background,almost leading to the failture of tracking,a new method was proposed by employing dynamic threshold in the five frame difference method to detect the moving object and a improved Meanshift algorithm was designed to realize object tracking location by background-weighting and template updating.The new method improves the real-time performance and robustness of tracking.The results show that the method is feasible with advantages of detecting the moving object accurately,and it also can improve reliability of object tracking.

Key words: Meanshift,Dynamic threshold,Five frame difference,Template updating