中文摘要：

基于视频图像处理方法,提出了一种复式伸缩窗来实时跟踪交叉口排队车辆队尾和队头的位置变化,从而准确描述交叉口车辆排队形成和消散过程。通过检测指定区域内车辆是否存在和是否运动,分别构建跟踪排队队尾和队头的队尾伸缩窗和队头伸缩窗。描述排队首尾伸缩变化的复式伸缩窗则由这两个伸缩窗相互协作所构成。根据跟踪队尾和队头的结果,车辆排队长度和停车延迟时间等重要参数就可以轻易得到。试验结果表明本文算法能实时准确地跟踪队尾和队头的位置,能适应不同天气环境和光照变化,其准确率达到92%以上,较好地满足车辆堵塞监控和交通信号灯控制的需要。

英文摘要：

Based on video image processing technology,an efficient vehicle queue and dissipation detection method was proposed.The method can track the position changes of vehicles queue tail and head at intersection in real time,and accurately describe the formation and dissipation of the queue.By vehicle presence detection and movement analysis on the monitored region,queue tail and head flexible windows were adopted to track the queue tail and head positions.These two cooperative flexible windows constitute a duplex flexible window to detect the position changes of the queue tail and head.The queue length and stop delay can be calculated by the tracking result of the queue tail and head.The experimental result shows that the accuracy rate of the presented algorithm is more than 92% by real-time tracking tail and head positions under various illumination and weather conditions.Therefore,the proposed algorithm can be used in traffic congestion monitoring and traffic signal controlling.