基于行车记录仪视频的事故碰撞前场景车距计算
详细信息 查看官网全文
- 英文论文题名：The Calculation of the Distance of In-front Car in Pre-crash Scene Based on VDR Videos
- 论文作者：王琴 ; 陈伟 ; 陈君毅 ; 王宏雁
- 英文论文作者：Wang Qin ; Chen Wei ; Chen Junyi ; Wang Hongyan ; School of Automotive Studies ; Tongji University ; Chinese and German College ; Tongji University
- 年：2016
- 作者机构：同济大学汽车学院;同济大学中德学院;
- 论文关键词：行车记录仪 ; 摄像头标定 ; 前车车距 ; 交通事故深度研究
- 英文论文关键词：VDR ; camera calibration ; distance between vehicles ; in-depth research of traffic accident
- 会议召开时间：2016-10-26
- 会议录名称：2016中国汽车工程学会年会论文集
- 英文会议录名称：Proceedings of 2016 SAE-China Congress
- 语种：中文
- 分类号：U463.6;U491.31;TP391.41
- 学会代码：QCGC
- 会议名称：2016中国汽车工程学会年会暨展览会
- 会议地点：中国上海
- 主办单位：中国汽车工程学会
- 学会名称：中国汽车工程学会
- 页数：4
- 文件大小：376k
- 原文格式：O
- 会议级别：全国
摘要
本文基于事故车辆行车记录仪视频数据,提出一种用于交通事故深度研究的前车车距测量方法。利用matlab标定工具箱得到摄像头的内部参数,去除图像畸变。选取事故发生前车辆行驶过的某一特征场景,利用三维扫描仪获取该场景中地面参考点的坐标,建立世界坐标系和像素坐标系之间的变换关系。结合车道消失点与摄像头光心之间的几何关系,计算事故过程中的前车车距。实验结果表明,当行车记录仪与障碍物之间的距离小于25m时,误差范围在5%以内。相对于传统摄像头事前标定来计算车距,本文提出了一种适用于事故研究的事后标定算法,对交通事故深度调查具有重要意义。
A new method for calculating the distance between vehicles in pre-crash scene,which could be applied in the in-depth research of traffic accidents,was proposed based on VDR(Video Drive Recorder) videos.The images of video were undistorted based on the intrinsic parameters of the camera obtained by Matlab calibration Toolbox.The image pixel coordinate system was transformed into the world coordinate system by measuring the coordinates of reference points on the ground with 3D scanner.Finally based on the geometrical relationship between the vanishing point of lane lines and the optic center of camera,the distance between vehicles in Pre-crash scene was calculated.The results of the experiment show that the error range of measurement is below 5% within 25 m.Compared to traditional method of distance calculation in previous researches,camera calibration can be conducted after crash in this new method,which is of great significance to the in- depth research of traffic accidents.
A new method for calculating the distance between vehicles in pre-crash scene,which could be applied in the in-depth research of traffic accidents,was proposed based on VDR(Video Drive Recorder) videos.The images of video were undistorted based on the intrinsic parameters of the camera obtained by Matlab calibration Toolbox.The image pixel coordinate system was transformed into the world coordinate system by measuring the coordinates of reference points on the ground with 3D scanner.Finally based on the geometrical relationship between the vanishing point of lane lines and the optic center of camera,the distance between vehicles in Pre-crash scene was calculated.The results of the experiment show that the error range of measurement is below 5% within 25 m.Compared to traditional method of distance calculation in previous researches,camera calibration can be conducted after crash in this new method,which is of great significance to the in- depth research of traffic accidents.
引文
[1]Z.Y Zhang.A Flexible Technique for Camera Calibration.Technical Report MSR-TR-2001-120,Microsoft Research,December 1998
[2]于泓.摄像机标定算法研究[D].济南:山东大学,2006:15-27
[3]贾丹.摄像机现场标定算法研究[D].哈尔滨:哈尔滨工程大学,2007:14-47
[4]Stein G P,Mano O,Shashua A.Vision-based ACC with a single camera;bounds on range and range rate accuracy;Intelligent Vehicles Symposium,2003.Proceedings.IEEE,2003.
[5]余厚云,张为公,基于单目视觉传感器的车距测量与误差分析[J].传感器与微系统.2012,31(9):10-13
[6]郭磊,徐友春,李克强,等.基于单目视觉的实时测距方法研究[J].中国图像图形学报.2006,11(1):74-81
[7]王荣本,李斌,储江伟,等公路上基于车载单目机器视觉的前方车距测量方法的研究[J].公路交通科技.2001,18(6):94-98
[2]于泓.摄像机标定算法研究[D].济南:山东大学,2006:15-27
[3]贾丹.摄像机现场标定算法研究[D].哈尔滨:哈尔滨工程大学,2007:14-47
[4]Stein G P,Mano O,Shashua A.Vision-based ACC with a single camera;bounds on range and range rate accuracy;Intelligent Vehicles Symposium,2003.Proceedings.IEEE,2003.
[5]余厚云,张为公,基于单目视觉传感器的车距测量与误差分析[J].传感器与微系统.2012,31(9):10-13
[6]郭磊,徐友春,李克强,等.基于单目视觉的实时测距方法研究[J].中国图像图形学报.2006,11(1):74-81
[7]王荣本,李斌,储江伟,等公路上基于车载单目机器视觉的前方车距测量方法的研究[J].公路交通科技.2001,18(6):94-98