基于图像的弩机三维重建及其机构原理研究
|
摘要
为了辅助弩机考古研究工作,本文采用了一种无接触式的基于图像的三维重建技术,通过数字化技术对弩机进行重建,为弩机的断代、分类以及机械性能的研究提供了一种新型的手段和方向。
论文首先介绍了几种重建的方法,分析了各种方法的优缺点,通过比较,针对本文弩机重建的要求以及文物的特殊性,采用了计算机视觉领域中基于图像的三维重建方法,分析了整个重建流程；基于对现有文献介绍的几种特征点检测算子及其优缺点分析,提出了一种新的特征点检测算法-等比提取法。综合了本文特征点检测的特点,利用灰度互相关信息和视差梯度测度及边缘约束条件进行了特征点的立体匹配,成功的把复杂的匹配问题简单化。其次,在分析了现有相机标定算法的优缺点上,采用基于透视变换模型的线性标定方法,利用双目立体视觉系统成像原理,根据不同位置拍摄的两幅图像照片中的对应点图像坐标,进行了左右像机的参数标定,该标定方法过程简便,精度基本满足弩机三维重建的要求。进而采用基于线性标定的三维重建方法对立体匹配点进行了三维重建,获得了其对应物点的三维坐标,实现了由图像获取弩机物体深度信息的目的。最后,利用所获得的一系列坐标在pro-e软件中实现了由点云数据建立弩机的三维实体模型的过程,并对弩机机构进行了静力学和运动学的仿真。
本文所研究基于图像的弩机三维重建理论及其算法,将计算机视觉方法引入考古研究工作领域,丰富了文物考古研究的理论和方法,倡导了科技考古新理念,具有较高的实际应用前景。
In this thesis, a non-contact-type image-based three-dimensional reconstruction technology is used to complement archaeological research work of crossbow.At the same time, the reconstruction of crossbow is carried out, which provides a new type of means and direction for the dating、classification and the mechanical properties research of crossbow.
Firstly, by comparing the advantages and disadvantages of several methods of reconstruction, the image-based three-dimensional reconstruction method in the computer vision field is used to analysis the entire reconstruction process for the reconstruction of the requirements of crossbow and the specificity of cultural relics. Based on the analysis of the advantages and disadvantages of several feature point detection operator introduced in existing literature, a new feature point detection algorithm called isocon extraction is raised. By using of gray-scale inter-related information and disparity gradient bound measure and marginal conditions, the three-dimensional feature points matching are carried out based on the combination of the characteristics of feature point detection in this thesis. And the complex matching problem is successfully simplified. Secondly, based on the analysis of the advantages and disadvantages of the existing camera calibration algorithm, the linear calibration method based on the perspective transformation model is used. Depending on the location of two images taken photos of the corresponding points in image coordinate, the parameters of camera calibration are carried out by using binocular stereo vision system imaging principle. This calibration process is simple and its accuracy can basically meet the requirements of three-dimensional reconstruction for crossbow. By using the method of three-dimensional based on linear calibration reconstruction, the three-dimensional reconstruction for matching points on the three-dimensional is carried out and its counterpart of the three-dimensional coordinates of points are got as well as the depth of information of crossbow by the image is achieved. Finally, by using these coordinates, the three-dimensional solid model of crossbow in Pro/E is established from point cloud data. At the same time, the kinematics and dynamics of imitation for crossbow is carried out.
The theory and its algorithm of image-based three-dimensional reconstruction in this thesis, which not only introduces the computer vision to the field of archaeological work but also enriches the cultural relics and archeology research theories and methods and advocacies the new concept of Archaeological Science and Technology, has higher practical application of the prospect.
论文首先介绍了几种重建的方法,分析了各种方法的优缺点,通过比较,针对本文弩机重建的要求以及文物的特殊性,采用了计算机视觉领域中基于图像的三维重建方法,分析了整个重建流程；基于对现有文献介绍的几种特征点检测算子及其优缺点分析,提出了一种新的特征点检测算法-等比提取法。综合了本文特征点检测的特点,利用灰度互相关信息和视差梯度测度及边缘约束条件进行了特征点的立体匹配,成功的把复杂的匹配问题简单化。其次,在分析了现有相机标定算法的优缺点上,采用基于透视变换模型的线性标定方法,利用双目立体视觉系统成像原理,根据不同位置拍摄的两幅图像照片中的对应点图像坐标,进行了左右像机的参数标定,该标定方法过程简便,精度基本满足弩机三维重建的要求。进而采用基于线性标定的三维重建方法对立体匹配点进行了三维重建,获得了其对应物点的三维坐标,实现了由图像获取弩机物体深度信息的目的。最后,利用所获得的一系列坐标在pro-e软件中实现了由点云数据建立弩机的三维实体模型的过程,并对弩机机构进行了静力学和运动学的仿真。
本文所研究基于图像的弩机三维重建理论及其算法,将计算机视觉方法引入考古研究工作领域,丰富了文物考古研究的理论和方法,倡导了科技考古新理念,具有较高的实际应用前景。
In this thesis, a non-contact-type image-based three-dimensional reconstruction technology is used to complement archaeological research work of crossbow.At the same time, the reconstruction of crossbow is carried out, which provides a new type of means and direction for the dating、classification and the mechanical properties research of crossbow.
Firstly, by comparing the advantages and disadvantages of several methods of reconstruction, the image-based three-dimensional reconstruction method in the computer vision field is used to analysis the entire reconstruction process for the reconstruction of the requirements of crossbow and the specificity of cultural relics. Based on the analysis of the advantages and disadvantages of several feature point detection operator introduced in existing literature, a new feature point detection algorithm called isocon extraction is raised. By using of gray-scale inter-related information and disparity gradient bound measure and marginal conditions, the three-dimensional feature points matching are carried out based on the combination of the characteristics of feature point detection in this thesis. And the complex matching problem is successfully simplified. Secondly, based on the analysis of the advantages and disadvantages of the existing camera calibration algorithm, the linear calibration method based on the perspective transformation model is used. Depending on the location of two images taken photos of the corresponding points in image coordinate, the parameters of camera calibration are carried out by using binocular stereo vision system imaging principle. This calibration process is simple and its accuracy can basically meet the requirements of three-dimensional reconstruction for crossbow. By using the method of three-dimensional based on linear calibration reconstruction, the three-dimensional reconstruction for matching points on the three-dimensional is carried out and its counterpart of the three-dimensional coordinates of points are got as well as the depth of information of crossbow by the image is achieved. Finally, by using these coordinates, the three-dimensional solid model of crossbow in Pro/E is established from point cloud data. At the same time, the kinematics and dynamics of imitation for crossbow is carried out.
The theory and its algorithm of image-based three-dimensional reconstruction in this thesis, which not only introduces the computer vision to the field of archaeological work but also enriches the cultural relics and archeology research theories and methods and advocacies the new concept of Archaeological Science and Technology, has higher practical application of the prospect.
引文
[1]郑南宁.计算机视觉与模式识别[M].北京：国防工业出版社,1998
[2]B. K. P. Horn.hape from Shading:A Method for Obtaining the Shape of a Smooth Opaque Object from One View. PhD thesis[D] MIT,1970
[3]M.Bichsel, A.P.Pentland, A simple algorithm for shape from shading[J]. IEEEP Proceedings of Computer Vision and Pattern Recognition.1992
[4]Ruo Zhang, Ping-Sing Tsai, James Cryer, Shape form Shading:A Surrey IEEE Transactions on Pattern Analysis and Machine Intelligence[J]. vol.2 l,no.8,pp:690-706,1999
[5]J.V. Stone. Shape from Texture:a computational analysis. PhD thesis University of Sussex[D].1990
[6]马颂德,张正友.计算机视觉-计算理论和算法基础[M].北京：科学出版社,1998
[7]章毓晋.图像处理与分析[M].北京：清华大学出版社,2000
[8]Joscph.H Intersection mappings for stereo vision[J]. SPIE,1991,1608:293-299
[9]王润生.图像理解[M].长沙：国防科技大学出版社,1995
[10]吴立德.计算机视觉[M].上海：复旦大学出版社,1993
[11]YShirai. Three Dimensional Computer Vision[J]. Symbolic Computation, spring-Verlag, Berlin, 1987
[12]B.F.Buxton, D.A.Castelow, M.J.Dell. Evaluation of stereo vision system for the Control of an autonomous guided vehicle[J]. SPIE,1993,1989:104-115
[13]李介谷.计算机视觉的理论和实践[M].上海：上海交通大学出版社,1991
[14]张健新.用于三维尺寸检测的双目视觉传感器[J].计量学报,1999,20(4)：95-100
[15]D. Marr, Vision, W. H. Freeman and Company, San Francisco.中译本：视觉计算理论[M],姚国正、刘磊、汪云九译,科学出版社,1988
[16]S. DePaoli, R. Zignalln, T. Skordas, H. Huot Soudain Evaluation of an Obstacle DetectionTechnique Based in Stereo-Vision[J]. SPIE,1993,1989:128-136
[17]M. Pilu. A Direct Method for Stereo Correspondence based on Singular Value Position[J].IEEE CVPR'92Puerto Rico, June1997
[18]Zhengyou Zhang. A New Multistage Approach to Motion and Structure Estimation by Gradually Enforcing Geometric Constraints. ACCV'98 Paper
[19]Longuet-Higgins, H. C. A computer algorithm for reconstructing a scene form two projectionnature[J].Vol.293,10, Sept.1981
[20]Yunde Zhong, Huaifeng Zhang. Control points Based Semi-Dense Matching[J]. ACCV2002:The 5th Asian Conference on Computer Vision, Melbourne, Australia
[21]陈明舟.主动光栅投影双目视觉传感器的研究[D].天津大学,2002
[22]王磊.三维重构技术的研究与应用[D].北京：清华大学,2002
[23]R. C. Jain、T.O.Binford、Ignorance、Myopia, Active Computer Vision System CVGIP Image Understanding[J].Vol.53, No.l, pp.112-117,1991
[24]R. MeClamtoek, Marr's Three Levels A Re-evalution Minds Mach[J]. Vol.1, No.2, PP.185-196, 1991
[25]赵小松,张宏伟,张国雄等.摄像机标定技术的研究[J].机械工程学报,2002,38(3)：149-151
[26]S. M. Smith and J. M. Brady, SUSAN-a new approach to low level image processing, International Journal of Computer Vision[J].,1997,23(1):45-78
[27]H. P. Moravec, Towards automatic visual obstacle avoidance. In Proc.of the 5th International Joint Conference on Artificial Intelligence, Cambridge,Massachusetts, USA, pp.584,1977
[28]C. Harris and M. Stephens. A Combined Corner and Edge Detector[J]. In Alvey Vision Conference page 147-151,1988
[29]Xinag-YuYu, HongSunJun Chen. Points Matching via Iterative Convex Hull Vertices pairing EE proceedings of the Fourth International Conference on Machine Learning and Cybemetics[J]. Guangzhou,18-21 August.Z005:5350-5354
[30]H. S. Balrd. Model-based image matching using location. Ph. D. thesis, Dept of Elec. Eng Compute Sci., Princeton Univ., NJ,1984
[31]S. Rnaade and A. Rosenefld, Point Pattern matching by relaxation, Pattern Recog[J].vol.12, pp 269-275,1980
[32]A. Goshtasby and GC. Stoekman Point Pattern matching using convex hull edgesIEEE Trans Systems Man Cybernet.SMC-15(5).1985:631-637
[33]Ranade. S, A. Rosenefld Point Pattern matching by relaxation[J].Pattern Recognition,1980 2(4):269-276
[34]Jezching Ton, Anil K. Jain Registering Landsat Images by Point Matching IEEE transactionson geoscience and remote sensing[J]. vol.27, no.51989:642-651
[35]Z. Zhnag, R. Deriche,O Faugeras. A Robust Technique for Matching Two Uncalibrated Image Through the Recovery of the Unknown Epipolar Geometry[J]. Artificial Intelligence,1995,78 (1-2):87-119
[36]张广军.机器视觉[M].科学出版社,2005
[37]R.Tsai. An efficient and accurate camera calibration technique for 3D machine vision[J]. In Proceedings of Computer Vision and Pattern Recognition,1986,364-374
[38]H.A.Martins, J. R. Birk, R. B. Kelley. Camera Models Based on Data from Two Calibration Planes[J]. Computer Graphics and Image Processing,1981,Vol.17:173-180
[39]Abdel-Asiz Y.I,Karara H.M. Direct linear transformation from comparator Coordinates into object space coordinates in close-range photogrametry[J]. In Proceedings Symposium on close-range photogrametry 1971,1-18
[40]Hartley. R. Estimation of relative camera positions for uncalibrated cameras 1992
[41]雷成,吴福朝.Kruppa方程与摄像机自标定[J].自动化学报,2001,27(5)：621-630
[42]杨长江,汪威,胡占义.一种基于主动视觉的相机内参数自标定方法[J].计算机学报,1998
[43]杨长江,苏凤梅,胡占义.基于平面二次曲线的摄像机标定[J].计算机学报,2000,23(5)：541-547
[44]王伟,吴成柯.估计基础矩阵的六点综合算法[J].中国科学：E辑技术科学,1997,27(2)：165-170
[45]顾晓东,王晓明.基于Huang-Faugeras约束的相机分步自标定方法[J].计算机辅助设计与图形学学报,2003,15(1)：34-37
[46]刘传才.图像理解与计算机视觉[M].厦门大学出版社,2002
[47]贾云得.机器视觉[M].科学出版社,2000
[48]刘传才.图像理解与计算机视觉[M].厦门大学出版社,2002
[49]李鹏,王军宁.摄像机标定方法综述[J].山西电子技术,2007,4：77-79.
[50]邱茂林,马颂德,李毅.计算机视觉中摄像机定标综述[J].2000,26(1)：43-55
[51]Faig W. Calibration of close-range photogrammetric systems:Mathematical formulation [J]. Phot-ogrammetricEngineering and Remote Sensing,1975,41 (12):1479-1486
[52]武广智.地图构建中基于视觉的物体识别与定位[D].杭州：浙江大学,2006：20,27-33
[53]S.D.Ma, Conies-Based stereo, Motion Estimation, and Pose Determination[J]. International Journal of Computer Vision, Vol.10, No.l,1993
[54]尚明妹,解凯.几种三维重建方法的比较[J].哈尔滨师范大学自然科学学报,2003,5
[55]高立萍.基于solidworks的弩机信息管理与运动仿真研究[D].西安理工大学硕士论文,2004
[56]孙恒,陈作模.机械原理[M].高等教育出版社.2002
[57]侯介人,程学华.秦陵弩和铜镞制造工艺的模拟实验研究[J].西北农业大学学报,Vol.23suppl.Dec,1995
[58]袁伟钢.双目视觉中特征点匹配算法研究[D].哈尔滨：哈尔滨工程大学,2007
[59]Xiaomin JI,Mei Xue, Wen Yan. A Method for Polyhedron Object Rapid Prototyping Based on Its Digital Images[J]. CAD/Graphics, Oct.2007
[60]闫雯.基于图像的三维建模算法与试验研究[D].西安：西安理工大学,2007
[61]王方.利用数码相机重构物体三维形体的方法研究[D].西安：西安理工大学,2008
[2]B. K. P. Horn.hape from Shading:A Method for Obtaining the Shape of a Smooth Opaque Object from One View. PhD thesis[D] MIT,1970
[3]M.Bichsel, A.P.Pentland, A simple algorithm for shape from shading[J]. IEEEP Proceedings of Computer Vision and Pattern Recognition.1992
[4]Ruo Zhang, Ping-Sing Tsai, James Cryer, Shape form Shading:A Surrey IEEE Transactions on Pattern Analysis and Machine Intelligence[J]. vol.2 l,no.8,pp:690-706,1999
[5]J.V. Stone. Shape from Texture:a computational analysis. PhD thesis University of Sussex[D].1990
[6]马颂德,张正友.计算机视觉-计算理论和算法基础[M].北京：科学出版社,1998
[7]章毓晋.图像处理与分析[M].北京：清华大学出版社,2000
[8]Joscph.H Intersection mappings for stereo vision[J]. SPIE,1991,1608:293-299
[9]王润生.图像理解[M].长沙：国防科技大学出版社,1995
[10]吴立德.计算机视觉[M].上海：复旦大学出版社,1993
[11]YShirai. Three Dimensional Computer Vision[J]. Symbolic Computation, spring-Verlag, Berlin, 1987
[12]B.F.Buxton, D.A.Castelow, M.J.Dell. Evaluation of stereo vision system for the Control of an autonomous guided vehicle[J]. SPIE,1993,1989:104-115
[13]李介谷.计算机视觉的理论和实践[M].上海：上海交通大学出版社,1991
[14]张健新.用于三维尺寸检测的双目视觉传感器[J].计量学报,1999,20(4)：95-100
[15]D. Marr, Vision, W. H. Freeman and Company, San Francisco.中译本：视觉计算理论[M],姚国正、刘磊、汪云九译,科学出版社,1988
[16]S. DePaoli, R. Zignalln, T. Skordas, H. Huot Soudain Evaluation of an Obstacle DetectionTechnique Based in Stereo-Vision[J]. SPIE,1993,1989:128-136
[17]M. Pilu. A Direct Method for Stereo Correspondence based on Singular Value Position[J].IEEE CVPR'92Puerto Rico, June1997
[18]Zhengyou Zhang. A New Multistage Approach to Motion and Structure Estimation by Gradually Enforcing Geometric Constraints. ACCV'98 Paper
[19]Longuet-Higgins, H. C. A computer algorithm for reconstructing a scene form two projectionnature[J].Vol.293,10, Sept.1981
[20]Yunde Zhong, Huaifeng Zhang. Control points Based Semi-Dense Matching[J]. ACCV2002:The 5th Asian Conference on Computer Vision, Melbourne, Australia
[21]陈明舟.主动光栅投影双目视觉传感器的研究[D].天津大学,2002
[22]王磊.三维重构技术的研究与应用[D].北京：清华大学,2002
[23]R. C. Jain、T.O.Binford、Ignorance、Myopia, Active Computer Vision System CVGIP Image Understanding[J].Vol.53, No.l, pp.112-117,1991
[24]R. MeClamtoek, Marr's Three Levels A Re-evalution Minds Mach[J]. Vol.1, No.2, PP.185-196, 1991
[25]赵小松,张宏伟,张国雄等.摄像机标定技术的研究[J].机械工程学报,2002,38(3)：149-151
[26]S. M. Smith and J. M. Brady, SUSAN-a new approach to low level image processing, International Journal of Computer Vision[J].,1997,23(1):45-78
[27]H. P. Moravec, Towards automatic visual obstacle avoidance. In Proc.of the 5th International Joint Conference on Artificial Intelligence, Cambridge,Massachusetts, USA, pp.584,1977
[28]C. Harris and M. Stephens. A Combined Corner and Edge Detector[J]. In Alvey Vision Conference page 147-151,1988
[29]Xinag-YuYu, HongSunJun Chen. Points Matching via Iterative Convex Hull Vertices pairing EE proceedings of the Fourth International Conference on Machine Learning and Cybemetics[J]. Guangzhou,18-21 August.Z005:5350-5354
[30]H. S. Balrd. Model-based image matching using location. Ph. D. thesis, Dept of Elec. Eng Compute Sci., Princeton Univ., NJ,1984
[31]S. Rnaade and A. Rosenefld, Point Pattern matching by relaxation, Pattern Recog[J].vol.12, pp 269-275,1980
[32]A. Goshtasby and GC. Stoekman Point Pattern matching using convex hull edgesIEEE Trans Systems Man Cybernet.SMC-15(5).1985:631-637
[33]Ranade. S, A. Rosenefld Point Pattern matching by relaxation[J].Pattern Recognition,1980 2(4):269-276
[34]Jezching Ton, Anil K. Jain Registering Landsat Images by Point Matching IEEE transactionson geoscience and remote sensing[J]. vol.27, no.51989:642-651
[35]Z. Zhnag, R. Deriche,O Faugeras. A Robust Technique for Matching Two Uncalibrated Image Through the Recovery of the Unknown Epipolar Geometry[J]. Artificial Intelligence,1995,78 (1-2):87-119
[36]张广军.机器视觉[M].科学出版社,2005
[37]R.Tsai. An efficient and accurate camera calibration technique for 3D machine vision[J]. In Proceedings of Computer Vision and Pattern Recognition,1986,364-374
[38]H.A.Martins, J. R. Birk, R. B. Kelley. Camera Models Based on Data from Two Calibration Planes[J]. Computer Graphics and Image Processing,1981,Vol.17:173-180
[39]Abdel-Asiz Y.I,Karara H.M. Direct linear transformation from comparator Coordinates into object space coordinates in close-range photogrametry[J]. In Proceedings Symposium on close-range photogrametry 1971,1-18
[40]Hartley. R. Estimation of relative camera positions for uncalibrated cameras 1992
[41]雷成,吴福朝.Kruppa方程与摄像机自标定[J].自动化学报,2001,27(5)：621-630
[42]杨长江,汪威,胡占义.一种基于主动视觉的相机内参数自标定方法[J].计算机学报,1998
[43]杨长江,苏凤梅,胡占义.基于平面二次曲线的摄像机标定[J].计算机学报,2000,23(5)：541-547
[44]王伟,吴成柯.估计基础矩阵的六点综合算法[J].中国科学：E辑技术科学,1997,27(2)：165-170
[45]顾晓东,王晓明.基于Huang-Faugeras约束的相机分步自标定方法[J].计算机辅助设计与图形学学报,2003,15(1)：34-37
[46]刘传才.图像理解与计算机视觉[M].厦门大学出版社,2002
[47]贾云得.机器视觉[M].科学出版社,2000
[48]刘传才.图像理解与计算机视觉[M].厦门大学出版社,2002
[49]李鹏,王军宁.摄像机标定方法综述[J].山西电子技术,2007,4：77-79.
[50]邱茂林,马颂德,李毅.计算机视觉中摄像机定标综述[J].2000,26(1)：43-55
[51]Faig W. Calibration of close-range photogrammetric systems:Mathematical formulation [J]. Phot-ogrammetricEngineering and Remote Sensing,1975,41 (12):1479-1486
[52]武广智.地图构建中基于视觉的物体识别与定位[D].杭州：浙江大学,2006：20,27-33
[53]S.D.Ma, Conies-Based stereo, Motion Estimation, and Pose Determination[J]. International Journal of Computer Vision, Vol.10, No.l,1993
[54]尚明妹,解凯.几种三维重建方法的比较[J].哈尔滨师范大学自然科学学报,2003,5
[55]高立萍.基于solidworks的弩机信息管理与运动仿真研究[D].西安理工大学硕士论文,2004
[56]孙恒,陈作模.机械原理[M].高等教育出版社.2002
[57]侯介人,程学华.秦陵弩和铜镞制造工艺的模拟实验研究[J].西北农业大学学报,Vol.23suppl.Dec,1995
[58]袁伟钢.双目视觉中特征点匹配算法研究[D].哈尔滨：哈尔滨工程大学,2007
[59]Xiaomin JI,Mei Xue, Wen Yan. A Method for Polyhedron Object Rapid Prototyping Based on Its Digital Images[J]. CAD/Graphics, Oct.2007
[60]闫雯.基于图像的三维建模算法与试验研究[D].西安：西安理工大学,2007
[61]王方.利用数码相机重构物体三维形体的方法研究[D].西安：西安理工大学,2008