双摄像机结构光三维测量系统自标定方法  

Self-calibration Method of Structured Light 3D Measurement System with Double Cameras

作　　者：范剑英[1] 王洋[1] 武利伟 王安[1] 倪岳松 

机构地区：[1]哈尔滨理工大学测控技术与仪器黑龙江省高校重点实验室,黑龙江哈尔滨150080

出　　处：《哈尔滨理工大学学报》2013年第6期10-15,共6页Journal of Harbin University of Science and Technology

基　　金：黑龙江省教育厅科学技术研究项目(12521091)

摘　　要：针对双摄像机与投影仪结构光三维测量系统的标定问题,提出了极线约束条件下的自标定方法.依据极线约束理论,通过标定物上特征点与对应双摄像机的两个像点确定极线方程,建立了两个摄像机之间的平移和旋转关系矩阵.将多编码周期三角形灰度分布图案投射到标定物,获取特征点在投影仪上对应点的坐标,采用改进8点法建立投影仪与双摄像机关系矩阵的两个估计矩阵,利用对极距离为目标函数进行估计矩阵优化,得到双摄像机和投影仪的参数矩阵.采用该系统对平面和人脸进行了三维测量实验,结果表明平面测量的最大绝对误差不大于1.7 mm,相对误差不大于0.18%,表明了该三维测量系统自标定方法的有效性.For the problem existing during the calibration between two cameras and a projector, a new self-cali- bration method based on epipolar constraint was proposed. Firstly, according to epipolar constraint theory, epipolar equation was framed by using the feature points on calibration board and their projections on the image planes of two cameras, and the rotation and translation matrixes between cameras were calculated. Then we project multiple cod- ing period fringe images to calibration object to get corresponding coordinate point projected by feature points. Fi- nally, parameter matrix of the relationship between projector and cameras were calculated by the joint efforts. One is the two estimation matrixes generated by improved 8 points method, the other is optimization of estimation matrix to objective function by pair and epipolar line. The maximum absolute error is 1.7 mm, and the maximum relative error is less than 0. 18%. Experimental results show that the proposed calibration method can effectively calibrate the structured light 3D measurement system.

关 键 词：结构光三维测量 极线约束 多编码周期相移 自标定 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP317[自动化与计算机技术—计算机科学与技术]