多帧速度场估计方法及在交通视频监测中的应用  

Multi-frame Estimation Based on Velocity Field and Its Application in Traffic Video Monitoring

作　　者：房桂英[1] 纪筱鹏[1] 李莉[1] 邹大鹏 

机构地区：[1]中国海洋大学,山东青岛266071 [2]山东省莱阳市开发区,山东青岛265200

出　　处：《公路交通科技》2006年第4期132-134,共3页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金资助项目(60374031);北京大学视觉与听觉信息处理实验室基金资助项目(0306)

摘　　要：基于速度场的运动估计方法是进行运动物体分割的有效途径,提出了一种视频图像处理中基于多帧速度场的运动物体分割方法,针对简单速度场运动物体分割方法的不足,提出对多帧速度场图像提取的运动物体信息相融合的方法。试验结果证明:用此方法提取的运动物体核心点更完备,提高了运动物体分割的准确性。该方法在智能交通、视频监控系统中都有广泛应用。Motion estimation based on velocity field is an effective method for moving objects segmentation.A new segmentation of moving objects algorithm based on velocity field is presented in this paper. To improve the simple segmentation method for moving objects, we propose a method to interfuse the information of multi-velocity fields. The new method is testified to obtain more information of the moving objects and improve the quality of the segmentation. This method can be used in intelligent traffic system and video monitoring system widely.

关 键 词：运动分割 速度场 视频图像处理 交通监测系统 

分 类 号：TN91[电子电信—通信与信息系统]