基于平面控制格网的相机标定算法研究  

Research on camera calibration algorithm using planar control grid

作　　者：刘超[1] 高井祥[1] 杨化超[1] 

机构地区：[1]江苏省资源环境信息工程重点实验室/中国矿业大学环境与测绘学院,江苏徐州221008

出　　处：《测绘科学》2009年第4期50-52,共3页Science of Surveying and Mapping

基　　金：国家自然科学基金(40774010/D0401)

摘　　要：提出了一种基于平面控制格网的相机标定方法。推导了一种简单实用的在平面场景中利用二维直接线性变换和共线方程分解相机外方位元素初值的算法;引入了基于共面条件的约束条件方程,从而建立了基于附加约束条件的自检校光束法平差用于相机的精确标定。实际图像数据实验结果表明该方法同单纯基于标定点的算法相比标定精度明显提高。主点和焦距均达到了子像素的标定精度。检查点的平面和深度方向相对精度优于1/27000和1/38000。可以满足高精度近景三维量测的要求。A new method using points and lines in planar scenes was put forward. First, a method used to determine the intrinsic parameters was discussed, and camera＇ s initial values of extrinsic parameters were deduced using 2D direct linear transformation and collinearity equations. Image coordinates of planar grid control points were detected using improved Hough transformation algorithm. Based on these, three types of error equation including coplanar condition, planar relative control and linear relative control were im- ported. Bundle adjustment with self-calibration and with constrained conditions was created and used to carry out camera calibration with high accuracy. Higher precision with sub-pixel has been obtained compared with the method only using point features, which can meet the precision requirements of close-range photogrammetry with high accuracy.

关 键 词：相机标定 直接线性变换 自检校光线束法平差 约束条件 

分 类 号：P232[天文地球—摄影测量与遥感]