一种提高视频车速检测精度的算法分析和实现  

Analysis and realization of an algorithm to improve the accuracy of video-based vehicle speed measurement

作　　者：孙宁[1] 张重德[1] 

机构地区：[1]合肥工业大学信息化建设与发展中心,安徽合肥230009

出　　处：《合肥工业大学学报（自然科学版）》2014年第12期1462-1467,1527,共7页Journal of Hefei University of Technology：Natural Science

基　　金：合肥工业大学科学研究发展基金资助项目(2013HGXJ0434)

摘　　要：通过对基于虚拟线的机动车视频测速方法原理和精度不理想的分析,文章提出了一种提高测速精度的方案。该方案在车道上定义了一个矩形检测区,由于透视现象该检测区在视频图像上呈现为梯形,采用摄像机定标方法可将二维图像上梯形区域中任意一点的图像坐标较准确地映射为检测区内对应点的地面坐标。通过在车辆经过检测区的视频图像中选取适当的图像帧,可以获得车辆的位移及对应的时间间隔,从而获得较高精度的速度测量结果。T he principle and the insufficient accuracy of the video‐based vehicle speed measurement using two virtual lines are analyzed ,and a scheme to improve the speed measurement accuracy is proposed .A detection area is defined as a rectangle on a road lane formed with two lane lines and two traverse lines ,but displayed as a trapezoid because of perspective phenomenon on the images captured by video camera .By using camera linear calibration transformation ,the two‐dimensional image coordinates of any point within the trapezoid can be accurately mapped to the ground coordinates of corresponding point in the rectangle .By selecting adequate image frames captured during the running of the vehicle across the detection area ,the displacements of the vehicle and associate time intervals can be obtained by the proposed algorithm and thus the vehicle speed can be achieved with better accuracy .

关 键 词：测速精度 检测区 摄像机定标 图像帧 

分 类 号：U491.2[交通运输工程—交通运输规划与管理]