基于虚拟窗口的自适应车流量检测应用研究

基于视频交通检测中实时性和准确性要求,研究了三帧差分、背景差分及动态阈值等算法,提出增加虚拟窗口的方法,使整个车流量检测算法只针对有效序列图像信息进行处理.该方法通过排除差异大图像求均值,快速初始化虚拟窗口背景模型,在检测窗口内使用三帧差分和动态阈值准确定位运动车辆目标,实时更新背景后,根据背景差分法得到的窗口图像信息统计车流情况.实验结果表明,该方法可以有效应用于基于视频的实时车流量检测中.

Research and application of adaptive traffic flow detection based on virtual window

In view of the real-time and accuracy requirements in traffic detection based on video,the three frame differencing algorithm,background subtraction algorithm,dynamic threshold algorithm are studied,and a method of introducing virtual window in the whole algorithm is presented to make the traffic flow detection dealing only with effective image information of the video.This method firstly removes some images which have large differences with others in a certain successive video frames,then averages the remaining images and quickly initializes background model for the first virtual detection window.Accurately locates the moving targets in virtual window by using three frame differencing and dynamic threshold methods.Statistics traffic flow on the basis of window image information made by the difference method after updating background image in real-time.Experimental results show that the method is effective to use in the traffic flow detection based on video.