一种视频运动目标轨迹的实时探测方法

Researches on Real-time Detecting Track of Moving Objects of Video

本文分析了视频系列图像的特点 ,在运动目标的连续帧相减探测方法的基础上 ,进行了运动轨迹自动探测理论分析和实验研究 ,提出了局部运动轨迹合成目标整体运动轨迹的方法。在该方法中 ,第 1步 ,把图像相减 ,对结果进行对比度增强后 ,分成一定大小的子块 ;第 2步 ,求出每个子块的重心位置和重量 ,按一定阈值把背景区子块的重量置为零 ,基本上保留下运动目标区的子块 ;第 3步 ,按相邻关系把非零值子块组合成各自的子块群 ;最后 ,求出各个子块群的重心位置 ,就求出了各个运动目标的位置。

Some analyses and experiments of detecting moving track of video objects based on successive frame subtraction are proposed. A method of synthesizing the object-moving track by part moving tracks is suggested. First, the difference image between current frame and previous frame is processed in contrast enhancement, and is divided into blocks; second, weight of every block image is calculated and forced to zero if the weight less than a threshold, so these blocks of moving objects are remained; third, all adjacent blocks are grouped; finally, the weight position of very groups are calculated, which is the position of moving objects. The researching results have proved the speediness and validity of the detecting method.