多目立体视觉中图像匹配技术研究
|
摘要
本文研究的主要内容是多目立体视觉的图像特征点匹配问题。图像特征点匹配指的是找出物体从不同角度拍摄后在不同图像上投影点的对应。为了提高图像匹配算法的精度和可靠性,首先将一套明显区别于背景物体的标记点粘贴到物体上对其进行拍摄,然后对拍摄图像进行处理,实现不同图像间同名点的匹配。为了使得特征点匹配以及三维重建的结果更加精确,本文对边缘检测和特征点匹配问题进行了研究。主要的研究内容及创新点如下：1)、提出了一种改进的OFMM算法。由于OFMM算法运算量大,对OFMM算法的模板进行了改进,使边缘判定条件发生了改变。2)、提出了一种新的标记点自动检测算法。为了使边缘检测算法达到高效、快速、准确的运行效果,首先使用Canny算法对图像进行初始边缘定位,然后根据标记点的特征对无效边缘进行过滤,其次使用改进后的OFMM算法对图像边缘进行进一步修正,再次对标记点使用改进的算法分类,最后确定标记点的亚像素中心坐标。3)、提出一种基于相机姿态的匹配算法。首先根据已经匹配的编码元获取各幅图像的相机姿态,然后在相机姿态已知的情况下,根据极线约束实现非编码元的匹配,最后为了增加匹配点对并提高匹配的准确性,使用对称性距离约束对初匹配结果进行精匹配。
Image matching in multi-stereo vision is studied in this thesis. To improve the precision and reliability of matching of corresponding points between different images taken from different orientation by one single camera, a set of refrence points is applied as an assistant utility to measure the object. The main contents and contributions of this thesis are as follows:(1) An improved algorithm is put forward based on the OFMM. Owning to the great computational complexity of the OFMM algorithm, Canny operator is used to detect initial edge locations before using this algorithm. The templates of OFMM are improved, the condition have changed when we judge which pixel is edge. (2) A novel algorithm of automatic reference point detection is put forward and implemented. Firstly, the Canny operator is utilized to detect initial edge locations in pixel. Secondly, useless edges are filtered by using inherent attributes of the ellipse, such as length, roundness, convex-concave and enclosure. Thirdly, accurate edges in sub-pixel are obtained by using the improved orthogonal Fourier-Mellin moment operator. Fourthly, reference points are classified by using the improved method. At last, sub-pixel center coordinates are identified. (3) A matching algorithm based on camera poses is put forward. Firstly, poses of the camera are obtained based on coded targets which are matched. Secondly, uncoded targets are matched by using epipolar constraint based on camera poses obtained. At last, refined matching is performed based on initial matching by means of symmetrical distance constraint to increase number of corresponding points and to improve the accuracy of the matching between corresponding points.
Image matching in multi-stereo vision is studied in this thesis. To improve the precision and reliability of matching of corresponding points between different images taken from different orientation by one single camera, a set of refrence points is applied as an assistant utility to measure the object. The main contents and contributions of this thesis are as follows:(1) An improved algorithm is put forward based on the OFMM. Owning to the great computational complexity of the OFMM algorithm, Canny operator is used to detect initial edge locations before using this algorithm. The templates of OFMM are improved, the condition have changed when we judge which pixel is edge. (2) A novel algorithm of automatic reference point detection is put forward and implemented. Firstly, the Canny operator is utilized to detect initial edge locations in pixel. Secondly, useless edges are filtered by using inherent attributes of the ellipse, such as length, roundness, convex-concave and enclosure. Thirdly, accurate edges in sub-pixel are obtained by using the improved orthogonal Fourier-Mellin moment operator. Fourthly, reference points are classified by using the improved method. At last, sub-pixel center coordinates are identified. (3) A matching algorithm based on camera poses is put forward. Firstly, poses of the camera are obtained based on coded targets which are matched. Secondly, uncoded targets are matched by using epipolar constraint based on camera poses obtained. At last, refined matching is performed based on initial matching by means of symmetrical distance constraint to increase number of corresponding points and to improve the accuracy of the matching between corresponding points.
引文
[1]Linda G. Shapiro, George C. Srockman著,赵清杰,钱芳,蔡利栋译.计算机视觉[M].北京：机械工业出版社,2005.
[2]Zafer Iscan, Zumray Dokur, Tarner OImez. Tumor detection by using Zernike moments on segmented magnetic resonance brain images [EB/OL]. http://www.sciencedirect.com. 2009-8-20.
[3]A. Sluzek. On moment-based local operators for detecting image patterns[J]. Image and Vision Computing.2005,23(3):287-298.
[4]Zafer Iscan, Ayhan Yuksel, Zumray Dokur, Mehmet Korurek, Tamer Olmez. Medical image segmentation with transform and moment based features and incremental supervised neural network[J]. Digital Signal Processing,2009,19(5):890-901.
[5]Cao W P, Che R S, Ye D. An illumination-independent edge detection and fuzzy enhancement algorithm based on wavelet transform for non-uniform weak illumination images[J]. Pattern Recognition Letters,2008,29(3):192-199.
[6]周玲.基于多帧数字图像的三维重建关键技术研究[D].南京：南京航空航天大学,2006.
[7]马国红,朱政强等.基于图像的焊件深度信息提取[A].第七届中国机器人焊接学术与技术交流会议[C]：长春.2008,12,08：245-253.
[8]张光富.基于合成视觉的3D重建技术研究[D].浙江：浙江大学.2008.
[9]Wan,Y L,Cao Y D,Li D,et al.Finite element method edge detection algorithm based on multiscale fields[J]. Journal of Computational Information Systems,2007,3(4):1417-1424.
[10]Chen B,He L,Liu P.A morphological edge detector for gray-level image thresholding[A]. Proceedings of the 2nd International Conference on Image Analysis and Recognition[C],2005, 3656:659-666.
[11]Jiang J A,Chuang C L,Lu Y L,et al.Mathematical-morphology-based edge detectors for detection of thin edges in low-contrast regions[J]. Institution of Engineering and Technology Image Processing,2007,1(3):269-277.
[12]Liu T,Luo X QPeng C L,et al.Improved morphological edge detection algorithm for ultrasound heart ventricular wall image in the motion of its rotation[A]. Proceedings of the 1st International Conference on Bioinformatics and Biomedical Engineering[C],2007,960-963.
[13]陈君,戚飞虎,一种新的基于特征点的立体匹配算法[J].中国图象图形学报,2005,10(11)：1411-1414.
[14]荆丽秋.双目视觉系统标定与匹配的研究与实现[D].哈尔滨：哈尔滨工业大学,2009.
[15]乔警卫,胡少兴.三维重建中特征点提取与匹配算法研究[A].第八届全国虚拟现实与可视化学术会议(CCVRV' 08) [C]:福州.2008,09：400-403.
[16]易成涛,王孝通,徐晓刚.基于极线约束的角点匹配快速算法[A].第八届全国虚拟现实与可视化学术会议(CCVRV' 08) [C]:福州.2008,09：371-374.
[17]Evans A N,Liu X U.A morphological gradient approach to color edge detection [J]. IEEE Transactions on Image Processing,2006,15(6):1454-1463.
[18]Ferencik M, Lisauskas J B, Cury R C,et al. Improved vessel morphology measurements in contrast-enhanced multi-detector computed tomography coronary angiography with non-linear post-processing[J]. European Journal of Radiology,2006,57(3):380-383.
[19]王红梅,张科,李言俊.图像匹配研究进展[J].计算机工程与应用.2004(19)：42-46.
[20]张强.图像匹配算法研究[D].西安：西安电子科技大学,2006.
[21]罗海燕.多时相图像匹配技术研究[D].南京：南京理工大学,2008.
[22]Liu F, Deng J, Cui P Y,et al. A method for lunar obstacle detection based on multi-scale morphology transformation[J]. Proceedings of the IEEE International Conference on Mechatronics and Automation,2007,2736-2740.
[23]连静.图像边缘特征提取算法研究及应用[D].吉林：吉林大学,2008.
[24]S. D. Cochran, G. Medioni.3-d surface description from binocular stereo[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1992,14(10):981-994.
[25]R. J. Woodham. Photometric Method for Determining Shape from Shading[J]. Image Understanding,1984:97-126.
[26]J. Fan and L. B. Wolff. Surface Curvature and Shape Reconstruction from Unknown Multiple Illumination and Integrability[J]. Computer Vision and Image Understanding,1997,65(2): 347-359.
[27]J. Davis, X Chen. A Laser Range Scanner Designed for Minimum Calibration Complexity [A]. Proc.3nd 3DIM[C],2001:91-98.
[28]T. Stahs, F.Wahl. Fast and robust range data acquisition in a low-cost environment[A]. SPIE, 1395, Close-Range Photogrammetry Meets Machine Vision[C],1990:496-503.
[29]C. S. Chen, Y. P. Hung, C. C. Chiang, et al.Range data acquisition using color structured lighting and stereo vision[J]. Image and Vision Computing,1997,15(6):445-456.
[30]郑建冬,张丽艳,杜小宇.近景摄影测量中带有隐式图像校正的精确三维目标定位[J].中国航空学报.2009,6：649-657.
[31]赵恒,李辉,费向东.自动化几何校正中相继位姿自动校正方法[J].微计算机信息.2010,26(1)：172-216.
[32]周振德,胡立生,顾军.自平衡电动车定姿系统设计[J].控制工程.2009,16：45-60.
[33]Canny J.A computational approach to edge detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(1):679-697.
[34]Canny J.Finding edges and lines in images.MIT Artificial Intelligence Lab, Cambridge, MA, Technology Report,1983.
[35]T. J. Bin, Ao Lei, Cui Jiwen等.Subpixel edge location based on orthogonal Fourier-Mellin moments[J]. Image and Vision Computing,2008(26):563-569.
[36]Yunlong Sheng, Lixin Shen. Orthogonal Fourier-Mellin moments for invariant pattern recognition[J]. JOSA A.1994,11(6):1748-1757.
[37]Chao Kan, Mandyam D. Srinath. Invariant character recognition with Zernike and orthogonal Fourier-Mellin moments[J]. Pattern Recognition.2002(35):143-154.
[38]Keith Forbes, Anthon Voigt, Ndimi Bodika. An Inexpensive, Automatic and Accurate Camera Calibration Method[A]. In:Proceedings of the Thirteenh Annual Symposium of the Pattern Recognition Association of South Africa (PRASA 2002) [C], November 2002.
[39]C.T. Schneider. DPA-WIN-A PC Based Digital Photogrammetric Station for Fast and Flexible On-Site Measurement. In:International Achieves of Photogrammetry and Remote Sensing, Vienna,1996, XXXI(B):530-533.
[40]Heuvel van den, F.A. Kroon R.J.G.A. Digital Close-range Photogrammetry Using Artifical Tragets. In:Int. Archives of Photogrammetry and Remote Sensing, Washingtion D.C.,1992, 29(B5):222-229.
[41]Sung Joon Ahn, Wolfgang Rauh. Circular Coded Target for Automation of Optical 3D-Measurement and Camera Calibration[J]. International Journal of Pattern Recognition and Artificial Intelligence,2001,15(6):905-919.
[42]张广军.机器视觉[M].北京：科学出版社,2005.
[43]马颂德,张正友.计算机视觉[M].北京：科学出版社,2003.
[44]Richard Hartley, Andrew Zisserman著.韦穗,杨尚骏,章权兵,胡茂林译.计算机视觉中的多视图几何[M].安徽：安徽大学出版社,2002.
[45]管业鹏,童林夙,尹涵春.基于数码相机的三维物体空间几何位置的摄影测量[J].电子学报,2002,30(6)：849-852.
[46]于起峰,孙祥一,权铁汉等.用标定和亚象素技术实现三维运动目标的高精度测量[J].宇航学报,1999,20(3)：38-42
[47]Sung Joon Ahn, Wolfgang Rauh, Matthias Rechnagel. Ellipse Fitting and Parameter Assement of Circular Object Targets for Robot Vision. In:Proceedings of the 1999 IEEE/RSJ International Conference on Intelligent Robots and Systems, Kyongju Korea, October 17-21, 1999:525-530
[48]M.R.Shortis, T.A.Clarke, T.Short. A Comparison of Some Techniques for the Subpixel Location of Discrete Target Images[A]. Videometrics Ⅲ, SPIE[C]. Boston, USA,1994. 2350:239-250.
[49]钟慧湘,基本矩阵计算方法的研究[D].吉林：吉林大学,2005.
[50]戚世贵,戚素绢.一种基于图像特征点的图像匹配算法[J].理论与方法.2008,27(1)：3-15.
[51]黄俊敏.图像匹配算法研究及其在几何尺寸测量中的应用[D].湖北：湖北工业大学,2009.
[52]赵亮亮.双目立体视觉中的图像匹配技术研究[D].南京：南京航空航天大学,2007.
[53]张维忠.基于多幅图像的空间点定位与曲线结构三维重建[D].南京：南京航空航天大学,2007.
[54]刘建伟,梁晋,梁新合等.大尺寸工业视觉测量系统[J].光学精密工程.2010,18(1)：126-134.
[55]Ito Y, Sato T, Yamashita N, et al. Impulse noise detector using mathematical morphology. Proceedings of IEEE International Symposium on Circuits and Systems,2006,4261-4264.
[56]David G. Lowe. Distinctive Image Features from Scale-Invariant Keypoints. Distinctive Image Features from Scale-Invariant Interest Points[J]. International Journal of Computer Vision, 2004,60(2):91-110.
[57]赵恒,李辉,费向东.自动化几何校正中相机位姿自动校正方法[J].微计算机信息.2010,26(1)：172-216.
[58]张维中,杨厚俊,张丽艳,油世明,王静.基于多幅图像的同名曲线亚像素匹配算法.北京邮电大学学报,2008,31(4)：66-69.
[2]Zafer Iscan, Zumray Dokur, Tarner OImez. Tumor detection by using Zernike moments on segmented magnetic resonance brain images [EB/OL]. http://www.sciencedirect.com. 2009-8-20.
[3]A. Sluzek. On moment-based local operators for detecting image patterns[J]. Image and Vision Computing.2005,23(3):287-298.
[4]Zafer Iscan, Ayhan Yuksel, Zumray Dokur, Mehmet Korurek, Tamer Olmez. Medical image segmentation with transform and moment based features and incremental supervised neural network[J]. Digital Signal Processing,2009,19(5):890-901.
[5]Cao W P, Che R S, Ye D. An illumination-independent edge detection and fuzzy enhancement algorithm based on wavelet transform for non-uniform weak illumination images[J]. Pattern Recognition Letters,2008,29(3):192-199.
[6]周玲.基于多帧数字图像的三维重建关键技术研究[D].南京：南京航空航天大学,2006.
[7]马国红,朱政强等.基于图像的焊件深度信息提取[A].第七届中国机器人焊接学术与技术交流会议[C]：长春.2008,12,08：245-253.
[8]张光富.基于合成视觉的3D重建技术研究[D].浙江：浙江大学.2008.
[9]Wan,Y L,Cao Y D,Li D,et al.Finite element method edge detection algorithm based on multiscale fields[J]. Journal of Computational Information Systems,2007,3(4):1417-1424.
[10]Chen B,He L,Liu P.A morphological edge detector for gray-level image thresholding[A]. Proceedings of the 2nd International Conference on Image Analysis and Recognition[C],2005, 3656:659-666.
[11]Jiang J A,Chuang C L,Lu Y L,et al.Mathematical-morphology-based edge detectors for detection of thin edges in low-contrast regions[J]. Institution of Engineering and Technology Image Processing,2007,1(3):269-277.
[12]Liu T,Luo X QPeng C L,et al.Improved morphological edge detection algorithm for ultrasound heart ventricular wall image in the motion of its rotation[A]. Proceedings of the 1st International Conference on Bioinformatics and Biomedical Engineering[C],2007,960-963.
[13]陈君,戚飞虎,一种新的基于特征点的立体匹配算法[J].中国图象图形学报,2005,10(11)：1411-1414.
[14]荆丽秋.双目视觉系统标定与匹配的研究与实现[D].哈尔滨：哈尔滨工业大学,2009.
[15]乔警卫,胡少兴.三维重建中特征点提取与匹配算法研究[A].第八届全国虚拟现实与可视化学术会议(CCVRV' 08) [C]:福州.2008,09：400-403.
[16]易成涛,王孝通,徐晓刚.基于极线约束的角点匹配快速算法[A].第八届全国虚拟现实与可视化学术会议(CCVRV' 08) [C]:福州.2008,09：371-374.
[17]Evans A N,Liu X U.A morphological gradient approach to color edge detection [J]. IEEE Transactions on Image Processing,2006,15(6):1454-1463.
[18]Ferencik M, Lisauskas J B, Cury R C,et al. Improved vessel morphology measurements in contrast-enhanced multi-detector computed tomography coronary angiography with non-linear post-processing[J]. European Journal of Radiology,2006,57(3):380-383.
[19]王红梅,张科,李言俊.图像匹配研究进展[J].计算机工程与应用.2004(19)：42-46.
[20]张强.图像匹配算法研究[D].西安：西安电子科技大学,2006.
[21]罗海燕.多时相图像匹配技术研究[D].南京：南京理工大学,2008.
[22]Liu F, Deng J, Cui P Y,et al. A method for lunar obstacle detection based on multi-scale morphology transformation[J]. Proceedings of the IEEE International Conference on Mechatronics and Automation,2007,2736-2740.
[23]连静.图像边缘特征提取算法研究及应用[D].吉林：吉林大学,2008.
[24]S. D. Cochran, G. Medioni.3-d surface description from binocular stereo[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,1992,14(10):981-994.
[25]R. J. Woodham. Photometric Method for Determining Shape from Shading[J]. Image Understanding,1984:97-126.
[26]J. Fan and L. B. Wolff. Surface Curvature and Shape Reconstruction from Unknown Multiple Illumination and Integrability[J]. Computer Vision and Image Understanding,1997,65(2): 347-359.
[27]J. Davis, X Chen. A Laser Range Scanner Designed for Minimum Calibration Complexity [A]. Proc.3nd 3DIM[C],2001:91-98.
[28]T. Stahs, F.Wahl. Fast and robust range data acquisition in a low-cost environment[A]. SPIE, 1395, Close-Range Photogrammetry Meets Machine Vision[C],1990:496-503.
[29]C. S. Chen, Y. P. Hung, C. C. Chiang, et al.Range data acquisition using color structured lighting and stereo vision[J]. Image and Vision Computing,1997,15(6):445-456.
[30]郑建冬,张丽艳,杜小宇.近景摄影测量中带有隐式图像校正的精确三维目标定位[J].中国航空学报.2009,6：649-657.
[31]赵恒,李辉,费向东.自动化几何校正中相继位姿自动校正方法[J].微计算机信息.2010,26(1)：172-216.
[32]周振德,胡立生,顾军.自平衡电动车定姿系统设计[J].控制工程.2009,16：45-60.
[33]Canny J.A computational approach to edge detection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1986,8(1):679-697.
[34]Canny J.Finding edges and lines in images.MIT Artificial Intelligence Lab, Cambridge, MA, Technology Report,1983.
[35]T. J. Bin, Ao Lei, Cui Jiwen等.Subpixel edge location based on orthogonal Fourier-Mellin moments[J]. Image and Vision Computing,2008(26):563-569.
[36]Yunlong Sheng, Lixin Shen. Orthogonal Fourier-Mellin moments for invariant pattern recognition[J]. JOSA A.1994,11(6):1748-1757.
[37]Chao Kan, Mandyam D. Srinath. Invariant character recognition with Zernike and orthogonal Fourier-Mellin moments[J]. Pattern Recognition.2002(35):143-154.
[38]Keith Forbes, Anthon Voigt, Ndimi Bodika. An Inexpensive, Automatic and Accurate Camera Calibration Method[A]. In:Proceedings of the Thirteenh Annual Symposium of the Pattern Recognition Association of South Africa (PRASA 2002) [C], November 2002.
[39]C.T. Schneider. DPA-WIN-A PC Based Digital Photogrammetric Station for Fast and Flexible On-Site Measurement. In:International Achieves of Photogrammetry and Remote Sensing, Vienna,1996, XXXI(B):530-533.
[40]Heuvel van den, F.A. Kroon R.J.G.A. Digital Close-range Photogrammetry Using Artifical Tragets. In:Int. Archives of Photogrammetry and Remote Sensing, Washingtion D.C.,1992, 29(B5):222-229.
[41]Sung Joon Ahn, Wolfgang Rauh. Circular Coded Target for Automation of Optical 3D-Measurement and Camera Calibration[J]. International Journal of Pattern Recognition and Artificial Intelligence,2001,15(6):905-919.
[42]张广军.机器视觉[M].北京：科学出版社,2005.
[43]马颂德,张正友.计算机视觉[M].北京：科学出版社,2003.
[44]Richard Hartley, Andrew Zisserman著.韦穗,杨尚骏,章权兵,胡茂林译.计算机视觉中的多视图几何[M].安徽：安徽大学出版社,2002.
[45]管业鹏,童林夙,尹涵春.基于数码相机的三维物体空间几何位置的摄影测量[J].电子学报,2002,30(6)：849-852.
[46]于起峰,孙祥一,权铁汉等.用标定和亚象素技术实现三维运动目标的高精度测量[J].宇航学报,1999,20(3)：38-42
[47]Sung Joon Ahn, Wolfgang Rauh, Matthias Rechnagel. Ellipse Fitting and Parameter Assement of Circular Object Targets for Robot Vision. In:Proceedings of the 1999 IEEE/RSJ International Conference on Intelligent Robots and Systems, Kyongju Korea, October 17-21, 1999:525-530
[48]M.R.Shortis, T.A.Clarke, T.Short. A Comparison of Some Techniques for the Subpixel Location of Discrete Target Images[A]. Videometrics Ⅲ, SPIE[C]. Boston, USA,1994. 2350:239-250.
[49]钟慧湘,基本矩阵计算方法的研究[D].吉林：吉林大学,2005.
[50]戚世贵,戚素绢.一种基于图像特征点的图像匹配算法[J].理论与方法.2008,27(1)：3-15.
[51]黄俊敏.图像匹配算法研究及其在几何尺寸测量中的应用[D].湖北：湖北工业大学,2009.
[52]赵亮亮.双目立体视觉中的图像匹配技术研究[D].南京：南京航空航天大学,2007.
[53]张维忠.基于多幅图像的空间点定位与曲线结构三维重建[D].南京：南京航空航天大学,2007.
[54]刘建伟,梁晋,梁新合等.大尺寸工业视觉测量系统[J].光学精密工程.2010,18(1)：126-134.
[55]Ito Y, Sato T, Yamashita N, et al. Impulse noise detector using mathematical morphology. Proceedings of IEEE International Symposium on Circuits and Systems,2006,4261-4264.
[56]David G. Lowe. Distinctive Image Features from Scale-Invariant Keypoints. Distinctive Image Features from Scale-Invariant Interest Points[J]. International Journal of Computer Vision, 2004,60(2):91-110.
[57]赵恒,李辉,费向东.自动化几何校正中相机位姿自动校正方法[J].微计算机信息.2010,26(1)：172-216.
[58]张维中,杨厚俊,张丽艳,油世明,王静.基于多幅图像的同名曲线亚像素匹配算法.北京邮电大学学报,2008,31(4)：66-69.