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.
出处
《公路交通科技》
CAS
CSCD
北大核心
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)
关键词
运动分割
速度场
视频图像处理
交通监测系统
Segmentation of moving objects
Vdocity field
Video image processing
Traffic monitoring system
