基于边缘信息车流量检测方法的研究

摘要

For the problem of edge detection that the threshold is difficult to be determined which causes the inaccurate background edge information extraction, we present an improved edge information-based traffic flow detection method.It makes truncate process on the road edge image in advance to obtain an appropriate threshold, so that can accurately extract the edge information of the background; and in vehicle counting process, through combining the advantages of both the fixed window method and floating window method, we design a new floating window method to count vehicles.Experimental results show that the method has great improvement in the aspects of real time property, detection speed and correct rate, etc.,and can correctly and effectively detect the traffic flow in different environment.%针对边缘检测中阈值很难确定导致不能准确提取背景边缘信息的问题，提出一种改进的基于边缘信息车流量检测方法。该方法预先对路况边缘图像作削顶处理获取一个恰当的阈值，从而可以准确地提取背景边缘信息；且在车辆计数过程中，结合固定窗法和浮动窗法二者的优点，设计一种新的浮动窗法进行车辆计数。实验结果表明，该方法在实时性、检测速度、正确率等方面得到很大的改进，能够在不同环境下进行正确有效地车流量检测。