智能交通信号控制中的动态背景更新的改进方法

An Improved Method of Updating the Background Dynamically Used in ITS

提出一种改进的动态背景更新方法,用帧差法进行车辆检测,结合边缘检测对于边缘的敏感性,可以有效降低阴影对检测结果影响的优点,消除帧差法中阴影的影响。并且考虑光线的变化,对每帧图像的平均灰度值的过滤来减少光线影响带来的误判。实验表明,这种方法可以排除背景的干扰,有效检测车辆。

In this paper, an improved method of updating the background dynamically is proposed on the basis of the method of the difference between two frames. The difference between the current frame and the background is used in detecting the present and moving of the vehicle with the edge detection to weaken the impact of the shadow. In addition the average gray of each frame is considered to avoid the wrong recognition. Experiment indicates that this method can effectively detect the vehicle and weaken the impact by the background.