视频监视中运动目标的检测与跟踪算法

Moving Object Detection and Tracking Algorithm for Video Monitoring

提出一种视频监视中运动目标的精确检测、提取以及跟踪算法。该算法采用基于自适应背景图像估计与当前多帧图像的混合差实现快速精确地检测和提取目标 ,使用扩展的Kalman滤波器预测运动目标下一时刻可能处于的区域 ,缩小了目标跟踪时的搜索范围。充分利用运动目标检测的结果 ,提高了目标的匹配效率及跟踪速度。

A fast moving object detection and tracking algorithm for video monitoring is proposed. An accurate object detection and extriaction algorithm based on the difference between adaptive background estimation and several adjacent frames is used. An extended Kalman filter is used to predict the moving object's position so as to track moving objects in realtime at the next moment, thus reducing the search range. The results of moving object detection are utilized to improve the object matching efficiency. The experiment result shows that the proposed algorithm is efficient. \;