摘要
由于成像过程中存在种种非线性因素,如径向畸变与切向畸变,摄像机的最终成像模型为非线性模型。为了更好的实现物点与像点间的非线性映射关系,神经网络技术被引入了摄像机标定工作中。本文重点对基于神经网络技术的摄像机标定方法进行了深入的研究。
目前利用神经网络进行摄像机标定的研究,只能实现单个位置上或小范围空间的标定精度要求。当标定空间扩大到一定范围,标定精度和标定速度的冲突便不可调节。本文分别介绍了摄像机标定与神经网络这两个学科各自具有的特点。在此基础上,本文重新定位了二者间的结合点,提出一种新的标定方法,使神经网络专注于非线性因素的计算。
本文从摄像机针孔模型出发,推导了摄像机的成像原理;证明物点分别沿世界坐标系三个坐标轴(XW轴、YW轴和ZW轴)方向运动时存在截然不同的成像规律,提出一种新的并行标定方法,对物点的三维信息(XW、YW和ZW)的标定工作进行并行处理;提出了一种新的归一化算法,进一步提高了XW和YW的标定精度。
对于文中提出的标定方法,本文进行了摄像机标定实验以及相应的实验分析。通过误差分析,本文指出了提高标定精度的可能途径。实验结果表明:本文的标定方法可以在保证标定精度与标定速度的同时扩大摄像机标定空间;本文提出的归一化算法能够进一步提高XW和YW的标定精度;ZW标定是整个大范围标定的关键,其重构标准差大于XW和YW的重构标准差。提高ZW轴标定精度,将是下一步工作的重点。
As the existence of the many unwanted factors, such as radial distortion and decentering distortion, the model of camera imaging is actually a nonlinear one. In order to successfully realize this kind of nonlinear mapping relationship between the 3D object points and their corresponding 2D image points, neural networks were and are still used. This dissertation laid emphasis on camera calibration methods based on neural network technique.
Current researches on camera calibration using neural network can only keep their precision within a small space, and there has not been such a method that can solve the conflict between calibration precision and speed as the space extends. This dissertation first introduced the characteristics of each fields, camera calibration and neural network, respectively, found another way of combining the two and therefore a new calibration method, and made it possible for the neural network to focus on nonlinear mapping.
Using the pinhole model, this dissertation proved that there are different imaging rules along the three axes (XW axis, YW axis and ZW axis) in the world coordinate, and a brand new calibration method was created to calibrate the camera along these three directions respectively, namely parallel calibration. Besides, a new normalization algorithm was proposed and better calibration precisions were gained along both XW axis and YW axis.
Experiments and analysis were done to testify calibration methods proposed in this dissertation. After the analysis, we offered ways that may improve calibration precision. Experiment results show that parallel calibration method belongs with large scale calibration since it has extended the calibration scale while keeping both calibration precision and speed. Also, there comes the conclusion that calibration along ZW axis is critical to the whole large scale calibration since the standard deviation of reconstruction along ZW axis is much higher than those along XW axis and YW axis. Therefore, a calibration method that can improve calibration precision along ZW will be our next destination.
引文
[1]马颂德,张正友.计算机视觉——计算理论与算法基础[M].北京:科学出版社, 1998.
[2]张广军.机器视觉[M].北京:科学出版社, 2005.
[3]李华,胡占义.摄像机标定问题研究[D].中国科学院硕士论文, 2001.
[4] C. S. Fraser, M. R. Shortis, G. Ganci. Multi-Sensor System Self-Calibration[C]. Proceedings of the International Society for Optical Engineering Conference on Videometrices IV, Philadelphia, Pennsylvania, USA, 1995, Vol.2598:2~18.
[5]刘俏.基于改进遗传算法的摄像机标定技术研究[D].大连理工大学硕士学位论文, 2007.
[6] Y. I. Abdel-Aziz, H. M. Karara. Direct Linear Transformation into Object Space Coordinates in Close-Range Photogrammetry[C]. The Symposium on Close-Range Photogrammetry, Urbana, Illinois, USA, 1971:1~18.
[7]葛宝臻,李晓洁,邱实.基于共面点直接线性变换的摄像机畸变校正[J].中国激光, 2010, 37(2): 488~494.
[8] B. Hallert. Notes on Calibration of Cameras and Photographs in Photogrammetry. Photogrametria, 1968(23):163~178.
[9] W. Faig. Calibration of Close-Range Photogrammetric System: Mathematical Formulation. Photogrammetric Eng[J]. Remote Sensing, 1975, 41(12):1479~1486.
[10] R. Y. Tsai. An Efficient and Accurate Camera Calibration Technique for 3D Machine Vision[C]. Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Miami Beach, FL, USA, 1986:364~374.
[11] R. Y. Tsai. A Versatile Camera Calibration Technique for High-Accuracy 3D Machine Vision Metrology Using Off-the-Shelf TV Cameras and Lenses[J]. IEEE Journal of Robotics and Automation, 1987, RA-3(4):323~340.
[12] Reg G. Willson. Modeling and Calibration of Automated Zoom Lenses[J]. PhD Dessertation, Carnegie Mellon University, 1994.
[13] D. C. Brown. A Strategy for Multi-Camera On-the-Job Self-Calibration[J]. Institute Fur Photogrammetric Stuttgart, Festschrift, Ackermann, zum 60. Geburtstag, Tokyo, Iapan, 1989:9~21.
[14] J. Wen, G. Schweitzer. Hybrid Calibration of CCD cameras using artificial neural nets[C]. Proc. Int Joint Conf. Neural Networks. New York, USA, 1991:337~342
[15] M. Lynch, C. Dagli. Backpropagation neural network for stereoscopic vision calibration[C]. Proc. SPIE. Int. Soc. Opt. Eng., 1992:289~298
[16]袁野,仲崇权,欧宗瑛.基于BP网络的足球机器人视觉系统的标定[J].机器人, 2001, 23(7): 752~755.
[17]李国友,徐鹏,李慧光,马静霞.基于优化RBF-DDA神经网络的摄像机标定[J].计算机工程, 2007, 33(24): 244~246.
[18]刘胜,傅荟璇,王宇超.基于LS-SVM的立体视觉摄像机标定[J].光电工程, 2008, 35(10): 21~25.
[19] J. Weng, P. Cohen, M. Herniou. Camera Calibration with Distortion Models and Accuracy Evaluation[C]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 1992, 14(10):965~980.
[20] O. D. Faugeras, G. Toscani. Camera Calibration for 3D Computer Vision[C]. Proceedings of the International Workshop on Machine Vision and Machine Intelligence, Atlantic, NJ, USA, 1987:240~247.
[21] J. Weng, P. Cohen, M. Herniou. Calibration of Stereo Cameras Using a Non-Linear Distortion Model[C]. Proceedings of International Conference on Pattern Recognition, 1990:246~253.
[22]邱茂林,马颂德,李毅.计算机视觉中摄像机标定综述[J].自动化学报, 2000, 26(1):43~55.
[23] Z. Y. Zhang. A Flexible New Technique for Camera Calibration[C]. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2000, 22(11):1330~1334.
[24] G. Q. Wei, S. D. Ma. Implicit and Explicit Camera Calibration: Theory and Experiment[C]. IEEE Trans. Pattern Recognition and Machine Intelligence, 1994, 16(5): 469~480.
[25] R. I. Hartley. Estimation of Relative Camera Position for Uncalibrated Camera[C]. Proc. of the ECCV’92. Italy, 1992: 379~387.
[26] S. D. Ma. A Self-calibration Technique for Active Vision System[C]. IEEE Trans. Robotics and Automation, 1996, 12(1):114~120.
[27]朱嘉,李醒飞,徐颖欣.摄像机的一种主动视觉标定方法[J].光学学报, 2010, 30(5): 1297~1303.
[28] O. D. Faugeras, S. Maybank. Motion from point matches: multiplicity of solutions[J]. Int. J. Computer Vision, 1990, 4:225~246.
[29] Y. Ma,B. Vidal,J. Zosecka, et a1.Kruppa Equation Revisited:Its Renormalizationand Degeneracy[C].In: Proc. On ECCV’00. 2000: 561~577.
[30] W. McCulloch, W. Pitts. A Logical Calculus of the Ideas Immanent in Nervous Activity[J]. Bulletin of Mathematical Biophysics, 1943, 5:115~133.
[31] J. J. Hopfield, D. W. Tank. Neural Networks and Physical Systems with Emergent Collective Computational Abilities[C]. Proceedings of the National Academy of Sciences, 1982, 79(8):2554~2558.
[32] D. E. Rumelhart and J. L. McClelland (Eds.). Parallel Distributed Processing[M]. USA: MIT Press, 1986.
[33]张德丰. MATLAB神经网络应用设计[M].北京:机械工业出版社, 2009.
[34] T. H. Martin, B. D. Howard, H. B. Mark. Neural Network Design[M]. USA: PWS Publishing Company, 1996.
[35] K. Funahashi. On the Approximate Realization of Continuous Mapping by Neural Networks[J]. Neural Networks, 1989, 2(3): 183~192.
[36] JUN J, KIM C. Robust Camera Calibration Using Neural Network[C]. Proceedings of the IEEE Region 10 conference, 1999: 694~697.
[37]刘宏建,罗毅,刘允才.可变精度的神经网络摄像机标定法[J].光学精密工程, 2004, 12(4): 443~448.
[38] Chen D, Zhang G. A New Sub-pixel Detector for X-corners in Camera Calibration Targets[C]. The 13th International Conference in Central Europe on Computer Graphics, Visualization and Computer Vision, 2005: 97~100.
[39] Shin’ichi Tamura, Masahiko Tateishi. Capabilities of a Four-layered Feedforward Neural Network: Four Layers versus Three[C]. IEEE Transactions on Neural Networks, 1997, 8(2): 251~255
[40]袁野,欧宗瑛,田中旭.应用神经网络隐式视觉模型进行立体视觉的三维重建[J].计算机辅助设计与图形学学报, 2003, 15(3): 293~296.