作者：李佳莹, 罗哉, 江文松, 郭斌

作者单位：中国计量大学计量测试工程学院，浙江 杭州 310018

关键词：相机标定;镜头畸变;交比不变性;分区域

摘要：

针对相机不同畸变程度的大视场区域很难同步等精度标定的问题，提出基于交比不变性的分区域相机畸变矫正方法。以相机主点为中心，按距离加权原则将相机视场区域分割为若干个区域；利用各分割视场区域中的多组共线4点坐标建立交比不变性畸变系数求解模型；据此实现分区域相机畸变矫正。此方法的测量误差的评价是通过比较直线偏移量来实现的。实验表明，两区域分割法和三区域分割法的直线偏移量误差比传统不分区方法分别降低14.7%和29.93%。此方法在传统方法的基础上进行改进，解决大视场区域非等精度标定。从实验结果可得到此方法优于传统方法。

Sub-regional camera distortion correction based on cross-ratio invariance
LI Jiaying, LUO Zai, JIANG Wensong, GUO Bin
College of Metrology & Measurement Engineering, China Jiliang University, Hangzhou 310018,China
Abstract: Since the distortion of a camera varies from large field of view (FOV) to FOV, a sub-regional camera distortion correction method was proposed based on cross-ratio invariance, in this paper, to calibrate the visual model with equal precision at the same time. The region of the FOV was divided into several parts by the distance weighting principle according to optical center of the camera. The distortion coefficient-solving model for each sub-region of FOV was built by using multiple sets of data with four points of a line based on the cross-ratio invariance principle. The sub-region of the FOV could be calibrated by the calculated distortion coefficients. The measurement error of the method was evaluated by the average linear offset(ALO). The experimental result showed that, comparing with the traditional non-partitioned algorithm, the error of ALO have a 14.7% reduction for the two-region algorithm and 29.93% reduction for three-region algorithm, respectively. This method was improved on the traditional method, and solved the unequal precision calibration in the large FOV. In all, the proposed method is superior to the traditional method.
Keywords: camera calibration;lens distortion;cross-ratio invariance;sub-region
2020, 46(8):125-130  收稿日期: 2019-11-19;收到修改稿日期: 2019-12-06
基金项目: 国家重点研发计划（2018YFF01012006，2017YFF0206306）；浙江省重点研发计划（2018C01063）；国家自然科学基金（51675499）
作者简介: 李佳莹（1993-），女，浙江绍兴市人，硕士研究生，专业方向为机器视觉及精密测量
参考文献
[1] 邱茂林, 马颂德, 李毅. 计算机视觉中摄像机定标综述[J]. 自动化学报, 2000, 26(1): 43-54
[2] FAIG W. Calibration of close-range photogrammetric systems: mathematical formulation[J]. Photogrammetric Engineering and Remote Sensing, 1975, 41(12): 1479-1486
[3] 同志学, 赵涛, 贺利乐, 等. 基于双目视觉的工程车定位与行驶速度检测[J]. 中国机械工程, 2018, 29(4): 423-428
[4] GIOI R G, MONASSE P, MOREL J M, et al. Lens distortion correction with a calibration harp[C]//In Proceeding of the 18th IEEE International Conference on Image Processing (ICIP), Brussels, Belgium, 2011: 617-620.
[5] ZHANG Z Y. A flexible new technique for camera calibration[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2000, 22(11): 1330-1334
[6] PARK J, BYUN S C, LEE B U. Lens distortion correction using ideal image coordinates[J]. IEEE Transaction on Consumer Electronics, 2009, 55(3): 987-991
[7] 贺俊吉, 张广军, 杨宪明. 基于交比不变性的镜头畸变参数标定方法[J]. 仪器仪表学报, 2014, 25(5): 597-599
[8] 方勇, 胡海彦, 江振治. 一种多面阵航测相机实验室标定系统及几何校正办法[J]. 测绘通报, 2017(S1): 1-4
[9] 梁灵飞. 基于交比不变和间距比不变的线阵相机畸变标定方法[J]. 科学技术与工程, 2014, 14(14): 248-251
[10] 杨彦景. 摄像机标定与畸变图像矫正算法的设计与实现[D]. 沈阳: 东北大学, 2008.
[11] 刘书桂, 姜珍珠, 董英华, 等. 采用移动光靶标的分区域相机标定[J]. 光学精密工程, 2014, 22(2): 259-265
[12] 刘明周, 蒋倩男, 葛茂根. 基于机器视觉的装配动作自动分割与识别[J]. 中国机械工程, 2017, 28(11): 1346-1354
[13] 郭辉, 李天子, 徐克科. 数码相机畸变差模型中不同参数组合对物方坐标精度的影响[J]. 工程勘察, 2012(9): 62-64
[14] TSAI R Y. A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off the-shelf TV cameras and lenses[J]. IEEE Journal on Robotics & Automation, 1987, RA-3(4): 323-344
[15] 陈辉, 吕乃光. 基于交比不变性的相机镜头畸变标定研究[J]. 北京机械工业学院学报, 2016, 21(1): 1-3
[16] 张广军. 机器视觉[M]. 北京: 科学出版社, 2005： 95-97.
[17] 刘晓磊, 张兴国, 吴衡, 等. 近景大视场分区域标定方法研究[J]. 计算机测量与控制, 2017, 25(11): 262-268
[18] 凌寒雨, 彭彬彬. 基于距离加权的分区域相机标定方法[J]. 中国电子科学研究院学报, 2019, 14(1): 67-74