一种摄像机复合标定方法及其在汽车碰撞实验中的应用  

A NOVEL COMPOSITE CALIBRATION METHOD OF VIDEO CAMERA AND ITS APPLICATION IN AUTOMOBILE CRASH TEST

作　　者：张星[1,2] 刘一松[1] 刘翔[3] 

机构地区：[1]江苏大学,江苏镇江212003 [2]江苏科技大学,江苏镇江212013 [3]上海工程技术大学,上海201620

出　　处：《计算机应用与软件》2010年第4期96-97,105,共3页Computer Applications and Software

基　　金：上海市教委优青专项基金(05XPYQ02);上海市教委基金(066NZ005)

摘　　要：提出了一种摄像机的复合标定方法:基于共面点分步标定摄像机的初始内外参数;基于最小二乘法求解径向畸变参数,进而迭代优化。标定点的识别中,中心点的识别采用了Hough直线检测及直线求交法,避免通常的Hough椭圆检测法带来的测量误差。应用于汽车碰撞过程的定性和定量分析,结果表明了其有效性与稳定性。A composite calibration method for video camera is presented in this paper,in which step-by-step calibration of primitive extrinsic and intrinsic parameters of the video camera is taken based on coplanar targets and the least squares method is taken for solving radial distortion parameters and then further iterative optimization is performed.Among the identification of calibration points,the algorithms of line detection based on Hough transform and of line intersection are taken for the identification of central position,so as to avoid the measurement error from usual oval detection methods base on Hough transform.Finally this method is applied to the qualitative and quantitative analysis of automotive crash course.The results show its obvious effectiveness and stability.

关 键 词：摄像机模型 摄像机标定 汽车碰撞实验 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]