摘要
CCD相机,又称数字相机,它是运用CCD光电成像技术发展起来的新一代相机。近年来,CCD相机逐步应用于近景摄影测量、低空摄影测量。摄影测量的成果精度能否满足工程要求,取决于相机的技术参数和数据处理的技术方法。因此相机技术参数的检校是保证成果质量能否满足精度要求的关键问题之一。
对普通非量测型数码相机的检校,一种常用的方法是直接线性变换(DLT)模型法,其基本思想是在空间设置足够数量的控制点,事先测定每个控制点的空间坐标位置,用相机对控制点进行拍摄,将像点坐标和空间坐标代入构像方程,就可计算出该相机的内外方位元素。当考虑镜头畸变误差时,就要加上系统误差改正数。从而实现相机主距,像主点坐标以及因光学畸变等因素引起的系统误差改正数的检校改正。经过大量专家学者多年的研究分析,DLT直接线性变换相机检校法已经发展成熟,检校成果点位精度已达到0.35像素。但是该方法需要建立高精度控制场,给摄影测量的野外工作带来不便。
基于该原因,本文提出了一种新的无需物方控制的相机自检校法,即基于共面条件的连续像对相对定向的方法,实现相机各参数值的检校。并与DLT方法的相机检校结果相比较,以检验各个检校参数的可靠性。实验结果表明,共面条件检校方法是可靠的。
CCD camera, has another name called digital camera. It is a new generation camera with the development of digital imaging technology. In recent years, CCD camera applies to close range photogrammetry and low altitude photogrammetry progressively. Whether the precision of photogrammetric results could meet the project request, depending on the technical parameter of the camera and data processing technical methods. So calibration of camera parameter is a key problem to guarantee result quality.
According to the calibration of the ordinary non-metric digital camera, one commonly used method is the Direct Linear Transformation (DLT). Its basic thought is that establishing enough control points in 3D space, determining the space coordinates for each control point beforehand, taking the photography of the control point with the camera, and then leading the picture coordinates and the spatial coordinates into the DLT conformation equation, calculating inner and external elements of this camera. When considering lens distortion error, the system error must be added. Thus realizing the calibration correction of the camera host , the coordinates of the principal point as well as the system error because of the lens distortion . After many years’research analysis by many experts, the DLT calibration law already developed maturely, and the position precision of calibration results is already achieving 0.35 pixels. But this method needs to establish the high accuracy controlling field, bringing great inconvenience to the photographic field operation.
Based on this reason, this article proposed one new camera calibration method without the object control,namely successive stereopairs relative orientation based on coplane condition, realizing the camera calibration to various parameters, and through the DLT model to examine the reliability of the camera calibration results. The experimental result indicated that, calibration method based on the coplane condition is reliable.
引文
[1] 王 冬.基于多片空间后方交会的 CCD 相机检校:[硕士学位论文].山东科技大学,2003
[2] 朱 庆,张庆珩.论 CCD 成像系统检校的 DLT 方法.铁路航测,1994(3):1~6
[3] Fauguras O D, Luong Q T, Maybank S J.Camera self_calibration: Theory and experiments [J]. Computer Vision_Eccv 92, Lecture Notes in Computer Science, 1992, 588. 321~334
[4] Maybank S J, Faugeras O D. A theory of self_calibration of a moving camera [J]. International Journal of ComputerVision , 1992, 8 (2): 123~151
[5] Hartley R.Euclidean reconstruction from Uncalibrated views [J].Applications of Invariance in Computer Vision,1994, 825: 237~256
[6] 冯文灏.近景摄影测量:物体外形与运动状态的摄影法测定.武汉:武汉大学出版社,2002
[7] 崔红霞,孙 杰,林宗坚,储美华.非量测数码相机的畸变差检测研究.北京:测绘科学2005,30(1):105~107
[8] 王之卓.摄影测量原理[M].北京:测绘出版社,1979
[9] 冯文灏.关于近景摄影机检校的几个问题[J].测绘科学, 2000, 10:1~3
[10] 陈新玺,李 浩,张曼祺.普通数码相机构像畸变差两种检校模型的比较.北京:北京测绘,2005(4):51~54
[11] 邱茂林,马颂德,李 毅.计算机视觉中摄相机定标综述[J] .自动化学报, 2000, 26 (1): 24
[12] Luong Q T, Faugeras O D.Self_Calibration of a Moving Camera from Point Correspondences and Fundamental Matrices[J]. International Journal of Computer Vision, 1997, 22 (3): 261~289
[13] Hartley R.Kruppa equations derived from the fundamental matrix [J].IEEE Transaction on Pattern Analysis and Machine Intelligence, 1997, 19 (2): 133~135.
[14] 雷 成,吴福朝,胡占义.Kruppa 方程与摄相机自检校[J].自动化学报, 2001, 27 (5): 9
[15] Zhengyou Zhang.A Flexible Camera Calibration By Viewing a Plane From Unknown Orientations [R].Technical Report, Microsoft Research, 1998
[16] 张勇军.基于序列图像的工业钣金件三维重建与视觉检测[D].武汉大学, 2002
[17] Qing Wu, Guangyou Xu, Lei Wang.A Three_stage System for Camera Calibration [Z].In SPIE Conference Proceeding: Multispectral Image Processing and Pattern Recognition,Wuhan, China 2001. 22~24
[18] 徐丽华,谢东海,江万寿.基于预检校的数码相机自检校研究.河南理工大学学报,2005,24(1):50~60
[19] 朱肇光,孙 护,崔炳光.摄影测量学.北京:测绘出版社,1995
[20] 张祖勋,张剑清.数字摄影测量学[M].武汉:武汉测绘科技大学出版社,1997
[21] 冯文灏.非地形摄影测量[M].北京: 测绘出版社, 1985
[22] 苗红杰,赵文吉,刘先林.数码相机检校和摄像测量的部分问题探讨.首都师范大学学报(自然科学版),2005,26(1):117~120
[23] 邓 钢.非量测影像数字化近景摄影测量的部分问题探讨.昆明:昆明理工大学学报,2002,27(4):6~11
[24] 杨化超,邓喀中,张书毕,周 鸣,赵 卿.附加约束条件的 DLT 用于非量测数码相机的检校.北京:测绘工程,2005,14(40):30~33
[25] 王庆有.CCD 应用技术.天津:天津大学出版社,2000
[26] 王文颖等.工程与近景摄影测量.北京:地质出版社,1994
[27] 程效军,罗 武.基于非量测数字相机的近景摄影测量.铁 路 航 测 2002(1): 9~11
[28] Fraser, C. S.Photogrammetric Measurement to One Pare in A Million.Phtogrammetric Engineering and Remote Sensing, 1992
[29] Fraser, C. S.A Correction Model for Variation of Distortion within the photogrammetric Field.Close-Range Phtogrammetry Meets Machine Vision, Zurich, Switjierland, 1990
[30] 李建松,冯文灏等.大型钢结构部件的高精度检测与放样系统.钢结构,2001
[31] Beyer, H. A.Geometric and Radiometric Analysis of a CCD-Camera Based Photogrammetric Close-Range System.Mittelliungen Nr.51, Institute for Geodesy and Photogrammetry, 1992
[32] Atkinsn A.and etc.Photogrammetric Capabilities of the KODAK DC 40, DCS 420 and DCS 460 Digital Cameras.The Photogrammetriy and Surveying.University of London, 1996
[33] Abdel-Aziz,Y.I. and Karara H.M..Direct Linear Transformation form Comparator Coordinates into Object-Space Coordinates in Close-Range Photogrammetry.American Society of Photogrammetry, Falls Church Virginia, USA,1971
[34]Gruen,A.Tracking Moving Objects with Digital Photogrammetry System.Photogrammetic Record, 1992
[35] Gruen,A.Proceedings on Fast Proceeding of Phtogrammetric data.Interlarken, Switzerland, June, 1987
[36] Scott, P. J.Measurement without Photographs.Photogrammetric Record, 1981,10(58)
[37] 李志林,朱庆.数字高程模型[M].武汉:武汉测绘科技大学出版社,2000
[38] 冯文灏.非地形摄影测量[M].北京: 测绘出版社, 1985
[39] 王宝丰.计算机视觉工业测量系统的建立与标定[D].郑州: 信息工程大学测绘学院, 2004
[40] Fraser C S. Digital camera self-calibration [J]. JPRS,1997, 4: 149~159.
[41] 孙杰,林宗坚,崔红霞.无人机低空遥感监测系统[J].遥感信息, 2003, (1): 49~50.
[42] 崔红霞,孙杰,林宗坚.无人机遥感设备的自动化控制系统[J].测绘科学, 2004, 29(1): 47~49.
[43] Hartley R I.In defence of the 8-point algorithm, in Proceedings of the 5th International Conference on ComputerVision [Z].Boston, MA, 1995, 1064-1070, IEEE Comput. Soc., Los Alamitos, CA, 199
[44] Devernay F, Faugeras O.Straight Lines have to be Straight.Machine Vision and Application, 2001(13):14~24
