基于时空图的交通流量统计和交通状态检测

提出了一种新的基于时空图的交通流量统计和交通状态检测方法。首先,通过人机交互的方法设定检测线,并利用检测线计算时空图;然后,对时空图进行边缘提取、图像分割等处理,利用时空图上车辆的边缘、形状和占道率等信息,计算出一段时间内的交通流量。此外,还通过时空图的边缘信息的差异,将当前时间段的交通状态分为通畅、拥挤和堵塞三种不同的情况。实验结果表明,在摄像机安装位置合适的情况下,该方法统计交通流量的误差低于8%,判断交通状态的误差为0,具有很好的商业实用性。

Traffic flow estimation and traffic state detection based on the space-time map

We propose a novel traffic flow estimation and traffic state detection approach. Firstly, the detection line used to compute the space-time map is set via man-machine interaction, and the processes such as edge detection, image segmentation and so on are then performed in the space-time map. The traffic flow for a period of time is estimated by using the information of edge, shape and the rate of road occupation. In addition, the current traffic state is divided into three levels: clear, crowd and jam by using the difference between edge information in the space-time map. Experimental results show that the proposed approach can effectively compute the traffic flow and detect the traffic state, and it is practical for commerce.