基于监控视频图像车辆测速的研究  

Research on Algorithm of Traffic Accident Vehicle Speed Based on Monitoring Video Image

作　　者：梁建术 贾汇沦 张琳琳 Liang Jianshu;Jia Huilun;Zhang Linlin(School of Mechanical Engineering,Hebei University of Science and Technology,Shijiazhuang City,Hebei Province 050018,China)

机构地区：[1]河北科技大学机械工程学院,河北省石家庄市050018

出　　处：《农业装备与车辆工程》2022年第6期87-89,110,共4页Agricultural Equipment & Vehicle Engineering

摘　　要：在监控摄像机参数未知的情况下,基于图像匹配技术和逆透视图变换标定法,提出了一种事故车辆测速的方法。该方法首先使用基于对应点的标定方法进行图像的标定;然后利用改进的归一化互相关算法对目标车辆进行匹配,得到目标车在每一帧中的像素坐标;最后,由所获取的目标车辆时间-位移曲线测得车速。试验结果表明,此方法操作简单方便、精度高且有较高的普适性。Based on the image matching technology and the inverted view transformation calibration method, a method for computing the speed of an accident vehicle is proposed under the condition that the parameters of the surveillance camera are unknown. First, the calibration method based on the corresponding point is used for the calibration of the image. Then the target vehicle is matched by the normalization interrelation algorithm, and the pixel coordinates of the target vehicle are obtained in each frame. Finally, the vehicle speed is measured by the obtained target vehicle time-displacement curve. It is indicated that the method is simple and convenient to operate,has high precision and high universality.

关 键 词：视频测速 归一化互相关匹配 对应点标定 车速鉴定 

分 类 号：U491.31[交通运输工程—交通运输规划与管理]