基于双目立体视觉的图像匹配与三维重建
|
摘要
基于立体视觉技术的检测设备能够实现智能化、小型化、数字化、网络化和多功能化,具有非接触、精度高、在线检测、实时分析与控制、连续工作等特点,能够适应多种危险的应用场合,广泛应用于军事、医学、工业、农林业、航空航天、科学研究等领域。双目立体视觉是立体视觉的一个重要的分支。它直接模拟人类视觉处理景物的方式,可以在多种条件下灵活地测量景物的立体信息。
完整的双目立体视觉测量过程包括图像获取、摄像机标定、双目视觉系统标定、图像匹配和三维重建等主要阶段。本文对摄像机标定、图像匹配和三维重建进行了研究,其中重点对图像匹配的算法进行了深入研究。摄像机标定是几何校正和三维重建的基础,其精度往往决定着整个双目立体视觉系统的性能。本文研究了一种基于平面定标模板的标定方法,该方法的平面定标物便于在室内环境中使用,有效减少了随后立体匹配过程的计算复杂度。
本文选用局部特征作为匹配基元,运用一种具备较强鲁棒性、独特性和匹配速度的特征描述子——SIFT特征描述子对特征点进行描述,使用SIFT算法进行图像匹配。首先建立尺度空间,检测尺度空间极值点,然后精确定位特征点,同时去除不稳定的特征点,接着为每个特征点指定主方向参数,最后生成特征点描述符并进行匹配。
由于现实条件的复杂性,导致SIFT算法在运行时产生大量不必要的特征点,在时间和空间上均造成了大量的浪费,这不仅严重影响了算法的实时性,也导致了算法的精度的下降。为了提高图像匹配的速度和精度,本文在利用SIFT算法对两幅图像的特征点进行提取和描述后,采用一种改进的Kd-树算法对特征点进行检索,提高特征点匹配的效率；然后使用随机抽样一致性算法对匹配对进行提纯,得到更为精确的特征点匹配对。针对不同类型的立体图像进行匹配实验,实验结果验证了本文算法的有效性。本文还研究了双目立体视觉三维测量原理和双目立体视觉测量的数学模型,以及双目视觉系统的标定方法。
本文对双目立体视觉测量中的各项关键技术,包括摄像机标定、图像立体匹配和三维坐标计算等进行了方法研究,并最终通过软件编程完成了一个完整的三维重建过程,生成了三维云图,验证了本文研究的各项关键技术的正确性和可行性,为开发构建一个完整的双目立体视觉测量系统奠定了基础。
The testing equipment based on stereovision technology can be intelligent, miniaturized, digital, network and multi-functional.stereovision technology are of non-contact, high precision, online detection, real-time analysis and control, and continuous work. It can be used in many dangerous situations, and can be widely applied in military, medical, industrial, agriculture, forestry, aerospace and scientific research field. Binocular stereovision is an important branch in stereovision field. It simulates the vision process of humanbeing directly, and can measure the three-dimensional information of the object flexibly on many conditions.
The complete binocular stereovision measuring process can be divided into the following main phase:image acquisition, camera calibration, binocular vision system calibration, image matching and three-dimensional reconstruction. The processes of camera calibration, image matching and three-dimensional reconstruction are discussed in this paper, and the image matching algorithm is deeply studied. Camera calibration is the basis of geometric correction and three-dimensional reconstruction, its accuracy determines the performance of the whole binocular stereovision system. A calibration method based on plane calibration template is studied in this paper. The plane calibration object is suitable to be used in the indoor environment, and it can reduce the complexity of the following stero matching process effectively.
In this paper, the local feature is selected as the matching primitive, SIFT feature descriptor is adopted in the description of the feature points for its robustness, uniqueness and fast speed, and SIFT algorithm is applied in image matching. Firstly, the scale-space is created and the extreme points in the scale space are detected; then the feature points are located accurately and the unstable feature points are eliminated; after that, the main direction parameter of each feature point is specified; finally, the descriptor of the feature point is generated and matched.
Because of the complexity of the real environment, many unnecessary feature points are generated by SIFT algorithm. This results in the waste of time and space, it is not only seriously affects the real-time of the algorithm, but also leads to decrese of algorithm accuracy. In order to improve the speed and accuracy of image matching, after the feature points of the two images are extracted and descripted by SIFT algorithm, this paper adopts an improved Kd-tree algorithm to retrieve the feature points and improve the efficiency of the feature points matching; finally the matching pairs are purified by the RANSAC algorithm and more accurate feature points matching pairs are achieved. Matching experiments are conducted aiming at different kinds of three-dimensional images, the experiment results validate the efficiency of the algorithm Proposed in this paper this paper. Besides, this paper also discuss the theory and mathematical model of the binocular stereovision measuring, and the calibration method of the binocular stereovision system.
The critical technologies of the binocular stereovision measuring, including the camera calibration, three-dimensional image matching and 3D coordinate calculation are studied in this paper, then the complete three-dimensional reconstruction process is accomplished by the programing and the three-dimensional cloud picture is generated to validate the correctness and feasibility of the critical technologies, which settles a solid base for the development of a complete binocular stereovision measuring system.
完整的双目立体视觉测量过程包括图像获取、摄像机标定、双目视觉系统标定、图像匹配和三维重建等主要阶段。本文对摄像机标定、图像匹配和三维重建进行了研究,其中重点对图像匹配的算法进行了深入研究。摄像机标定是几何校正和三维重建的基础,其精度往往决定着整个双目立体视觉系统的性能。本文研究了一种基于平面定标模板的标定方法,该方法的平面定标物便于在室内环境中使用,有效减少了随后立体匹配过程的计算复杂度。
本文选用局部特征作为匹配基元,运用一种具备较强鲁棒性、独特性和匹配速度的特征描述子——SIFT特征描述子对特征点进行描述,使用SIFT算法进行图像匹配。首先建立尺度空间,检测尺度空间极值点,然后精确定位特征点,同时去除不稳定的特征点,接着为每个特征点指定主方向参数,最后生成特征点描述符并进行匹配。
由于现实条件的复杂性,导致SIFT算法在运行时产生大量不必要的特征点,在时间和空间上均造成了大量的浪费,这不仅严重影响了算法的实时性,也导致了算法的精度的下降。为了提高图像匹配的速度和精度,本文在利用SIFT算法对两幅图像的特征点进行提取和描述后,采用一种改进的Kd-树算法对特征点进行检索,提高特征点匹配的效率；然后使用随机抽样一致性算法对匹配对进行提纯,得到更为精确的特征点匹配对。针对不同类型的立体图像进行匹配实验,实验结果验证了本文算法的有效性。本文还研究了双目立体视觉三维测量原理和双目立体视觉测量的数学模型,以及双目视觉系统的标定方法。
本文对双目立体视觉测量中的各项关键技术,包括摄像机标定、图像立体匹配和三维坐标计算等进行了方法研究,并最终通过软件编程完成了一个完整的三维重建过程,生成了三维云图,验证了本文研究的各项关键技术的正确性和可行性,为开发构建一个完整的双目立体视觉测量系统奠定了基础。
The testing equipment based on stereovision technology can be intelligent, miniaturized, digital, network and multi-functional.stereovision technology are of non-contact, high precision, online detection, real-time analysis and control, and continuous work. It can be used in many dangerous situations, and can be widely applied in military, medical, industrial, agriculture, forestry, aerospace and scientific research field. Binocular stereovision is an important branch in stereovision field. It simulates the vision process of humanbeing directly, and can measure the three-dimensional information of the object flexibly on many conditions.
The complete binocular stereovision measuring process can be divided into the following main phase:image acquisition, camera calibration, binocular vision system calibration, image matching and three-dimensional reconstruction. The processes of camera calibration, image matching and three-dimensional reconstruction are discussed in this paper, and the image matching algorithm is deeply studied. Camera calibration is the basis of geometric correction and three-dimensional reconstruction, its accuracy determines the performance of the whole binocular stereovision system. A calibration method based on plane calibration template is studied in this paper. The plane calibration object is suitable to be used in the indoor environment, and it can reduce the complexity of the following stero matching process effectively.
In this paper, the local feature is selected as the matching primitive, SIFT feature descriptor is adopted in the description of the feature points for its robustness, uniqueness and fast speed, and SIFT algorithm is applied in image matching. Firstly, the scale-space is created and the extreme points in the scale space are detected; then the feature points are located accurately and the unstable feature points are eliminated; after that, the main direction parameter of each feature point is specified; finally, the descriptor of the feature point is generated and matched.
Because of the complexity of the real environment, many unnecessary feature points are generated by SIFT algorithm. This results in the waste of time and space, it is not only seriously affects the real-time of the algorithm, but also leads to decrese of algorithm accuracy. In order to improve the speed and accuracy of image matching, after the feature points of the two images are extracted and descripted by SIFT algorithm, this paper adopts an improved Kd-tree algorithm to retrieve the feature points and improve the efficiency of the feature points matching; finally the matching pairs are purified by the RANSAC algorithm and more accurate feature points matching pairs are achieved. Matching experiments are conducted aiming at different kinds of three-dimensional images, the experiment results validate the efficiency of the algorithm Proposed in this paper this paper. Besides, this paper also discuss the theory and mathematical model of the binocular stereovision measuring, and the calibration method of the binocular stereovision system.
The critical technologies of the binocular stereovision measuring, including the camera calibration, three-dimensional image matching and 3D coordinate calculation are studied in this paper, then the complete three-dimensional reconstruction process is accomplished by the programing and the three-dimensional cloud picture is generated to validate the correctness and feasibility of the critical technologies, which settles a solid base for the development of a complete binocular stereovision measuring system.
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]迟健男,王先梅,胡四泉,曹洋,等.视觉测量技术[M].北京：机械工业出版社,2011.
[3]贾云德.机器视觉[M].北京：科学出版社,2000.
[4]郑南宁.计算机视觉与模式识别[M].北京：国防工业出版社,1998.
[5]Vallerand S,Kanbara M,Yokoya N.Binocular vision-based augmented reality system with an increased registration depth using dynamic correction of feature positions [C]//International Conference on Virtual Reality,Los Angeles,USA,2003:271-272.
[6]Roberts L G.Machine perception of the three-dimensional solids[J].Optical and Electro-optical Information Proeessing,1965:159-197.
[7]Marr D.Vision:A computational investigation into the human representation and processing of visual information[M].San Francisco:W.H.Freeman and Company,1982.
[8]Okada K,Inaba M,Inoue H.Integration of real-time binocular stereo vision and whole body information for dynamic walking navigation of humanoid robot[C]//International Conference on Multi-sensor Fusion and Integration for Intelligent System,Tokyo,Japan, 2003:131-136.
[9]孙双花.视觉测量关键技术及在自动检测中的应用[D].天津：天津大学,2007.
[10]陈向伟.机械零件计算机视觉检测关键技术的研究[D].长春：吉林大学,2005.
[11]苗世迪.基于计算机视觉的机械零件尺寸测量研究[D].哈尔滨：哈尔滨理工大学,2006.
[12]沈满德.基于计算机视觉的破片参数精密测量技术研究[D].中国科学院研究生院,2008.
[13]李为民.大尺度范围内视觉测量技术研究[D].合肥：中国科学技术大学,2006.
[14]章毓晋.图象工程下册——图象理解与计算机视觉[M].北京：清华大学出版社,2000.
[15]高文,陈熙霖.计算机视觉——算法与系统原理[M].北京：清华大学出版社,1999.
[16]唐志豪.基于双目立体视觉的测量技术研究[D].苏州：江苏大学,2007.
[17]贺洪军.基于高速图像采集卡的图像处理[D].成都：电子科技大学,2005.
[18]黄侃.计算机立体视觉中的摄像机标定算法的研究与应用[D].沈阳：沈阳工业大学,2007.
[19]范志斌.基于图像局部不变性特征的机器人目标识别与定位[D].北京：北京交通大学,2011.
[20]刘瑶.遥操作机器人的双目视觉立体重建研究[D].北京：北京邮电大学,2009.
[21]程晓亮.基于双目立体视觉的三维重构研究[D].哈尔滨：哈尔滨工业大学,2006.
[22]刘盛夏.基于相位信息的立体图像匹配研究[D].上海：上海交通大学,2007.
[23]翟乃斌.基于计算机视觉的汽车整车尺寸测量系统的研究[D].长春：吉林大学,2007.
[24]王永强.基于约束算法的大尺寸三维视觉测量方法研究[D].北京：北京邮电大学,2007.
[25]夏永泉.计算机视觉中双目匹配相关技术研究[D].南京：南京理工大学,2007.
[26]Faugeras O D,Toscani G.The calibration problem for stereo[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition,Minmi Beach,Florida,June 22-26,1986:15-20.
[27]Tsai R Y.A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf cameras and lens[J].IEEE Transactions on Robotics and Automation,1987,3(4):323-344.
[28]Kim J S,Kim H W,Kweon I S.A camera calibration method using concentric circles for vision applications[C]//The 5th Asian Conference on Computer Vision,2002:515-520.
[29]Zhang Z.A flexible new technique for camera calibration[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(11):1330-1334.
[30]Basu A.Active calibration:alternative strategy and analysis[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,1993:495-500.
[31]Du F,Brady M,Self-calibration of the intrinsic parameters of camera for active vision system[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,1993:477-482.
[32]Ma S D.A Self-calibration technique for active vision system[J].IEEE Transaction on Robotics and Automation,1996,12(1):114-120.
[33]杨长江,孙凤梅,胡占义.基于二次曲线的纯旋转摄像机自标定[J].自动化学报,2001,27(3)：310-317.
[34]杨长江,孙凤梅,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.
[35]雷成,吴福朝,胡占义.一种新的基于主动视觉系统的摄像机自标定方法[J].计算机学报,2000,23(11)：1130-1139.
[36]李华,吴福朝,胡占义.一种新的线性摄像机自标定方法[J].计算机学报,2000,23(11)：1121-1129.
[37]吴福朝,李华,胡占义.基于主动视觉系统的摄像机自标定研究[J].自动化学报,2001,27(6)：752-762.
[38]胡占义,雷成,吴福朝.关于共面P4P问题的一点讨论[J].自动化学报,2001.27(6)：770-776.
[39]Hartley R.Self-calibration of stationary cameras[J].International Journal of Computer Vision,1997,229(1):2-5.
[40]Benallal M,Meunier J.Camera calibration with simple geometry[C]//The Processing of 2003 International Conference on Image and Signal,2003:165-170.
[41]Carvalho P C P,Szenberg F,Gattass M.Image-based modeling using a two-step camera calibration method[C]//Proceedings of International Symposium on Computer Graphics,Image Processing and Vision,1998:388-395.
[42]Wu Y H,Zhu H J,Hu Z Y,Wu F C.Camera calibration from the quasi-affine invariance of two parallel circles[C]//The 8th European Conference on Computer Vision (ECCV), 2004:190-202.
[43]Wu Y H,Hu Z Y.The invariant representations of a quadric cone and a twisted cubic[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003, 25(10):1329-1332.
[44]Yang A Y,Hong W,Ma Y.Structure and pose from single images of symmetric objects with applications to robot navigation[C]//Proceedings of IEEE International Conference on Robotics & Automation,Taipei,2003:1013-1020.
[45]Carvalho P C P.Szenberg F,Gattass M.Image-based modeling using a two-step camera calibration method[C]//Proceedings of International Symposium on Computer Graphics, Image Processing and Vision,1998:388-395.
[46]Wu Y H,Zhu H J,Hu Z Y,Wu F C.Camera calibration from the quasi-affine invariance of two parallel circles[C]//The 8th European Conference on Computer Vision (ECCV), 2004:190-202.
[47]Wu Y H,Hu Z Y.The invariant representations of a quadric cone and a twisted cubic[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003, 25(10):1329-1332.
[48]Yang A Y.Hong W,Ma Y.Structure and pose from single images of symmetric objects with applications to robot navigation[C]//Proceedings of IEEE International Conference on Robotics & Automation,Taipei,2003:1013-1020.
[49]Zhang Z Y.A Flexible New Technique for Camera Calibration,MS-RT-98-71[R].Microsoft Research,December 1998.
[50]肖志涛,国澄明,侯正信,于明.图像特征检测算法的分析和研究[J].中国图像图形学报,2004(12)：1414-1420.
[51]王永明,王贵锦.图像局部不变性特征与描述[M].北京：国防工业出版社,2010.
[52]Moravec H P.Towards Automatic Visual Obstacle Avoidance[C]//Proc.5th International Joint Conference on Artificial Intelligence,San Francisco,CA,USA,1977:584.
[53]Harris C,Stephens M.A Combined Corner and Edge Detector[C]//Proc.Alvey Vision Conf.Univ.Manchester,UK,1988:147-151.
[54]Lindeberg T.Edge detection and ridge detection with automatic scale selection[J]. International Journal of Computer Vision,30,2,1998:117-156.
[55]Mikolajczyk K,Schmid C.An affine invariant interest point detector[C]//In Proceedings of the 8th International Conference on Computer Vision,Vancouver,Canada,2002:128-142.
[56]Lindeberg T.Shape from texture from a multi-scale perspective[C]//In Proc.International Conference on Computer Vision,Germany,1993:683-691.
[57]Lowe D G.Object recognition from local scale-invariant features[C]//In International Conference on Computer Vision,Confu,Greece,1999:1150-1157.
[58]Bay H,Tuytelaars T,van Gool L.SURF:Speeded up robust features[C]//In 9th European Conference on Computer Vision,Graz,Austria,2006:346-359.
[59]Matas J,Chum O,Urban M,Pajdla T.Robust wide baseline stereo from maximally stable extremal regions[C]//In 13th British Machine Vision Conference University of Cardiff, 2002:761-767.
[60]Lindeberg T.Feature detection with automatic scale selection[J].International Journal of Computer Vision,1998,30(2):77-116.
[61]Lindeberg T.Detecting Salient Blob-Like Image Structures and Their Scale with a Scale-Space Primal Sketch:A Method for Focus-of-Attention[J].International Journal of Computer Vision,1993,11(3):283-318.
[62]黄令允.基于自适应阈值的SIFT算法研究及应用[D].大连：大连理工大学.2010.
[2]迟健男,王先梅,胡四泉,曹洋,等.视觉测量技术[M].北京：机械工业出版社,2011.
[3]贾云德.机器视觉[M].北京：科学出版社,2000.
[4]郑南宁.计算机视觉与模式识别[M].北京：国防工业出版社,1998.
[5]Vallerand S,Kanbara M,Yokoya N.Binocular vision-based augmented reality system with an increased registration depth using dynamic correction of feature positions [C]//International Conference on Virtual Reality,Los Angeles,USA,2003:271-272.
[6]Roberts L G.Machine perception of the three-dimensional solids[J].Optical and Electro-optical Information Proeessing,1965:159-197.
[7]Marr D.Vision:A computational investigation into the human representation and processing of visual information[M].San Francisco:W.H.Freeman and Company,1982.
[8]Okada K,Inaba M,Inoue H.Integration of real-time binocular stereo vision and whole body information for dynamic walking navigation of humanoid robot[C]//International Conference on Multi-sensor Fusion and Integration for Intelligent System,Tokyo,Japan, 2003:131-136.
[9]孙双花.视觉测量关键技术及在自动检测中的应用[D].天津：天津大学,2007.
[10]陈向伟.机械零件计算机视觉检测关键技术的研究[D].长春：吉林大学,2005.
[11]苗世迪.基于计算机视觉的机械零件尺寸测量研究[D].哈尔滨：哈尔滨理工大学,2006.
[12]沈满德.基于计算机视觉的破片参数精密测量技术研究[D].中国科学院研究生院,2008.
[13]李为民.大尺度范围内视觉测量技术研究[D].合肥：中国科学技术大学,2006.
[14]章毓晋.图象工程下册——图象理解与计算机视觉[M].北京：清华大学出版社,2000.
[15]高文,陈熙霖.计算机视觉——算法与系统原理[M].北京：清华大学出版社,1999.
[16]唐志豪.基于双目立体视觉的测量技术研究[D].苏州：江苏大学,2007.
[17]贺洪军.基于高速图像采集卡的图像处理[D].成都：电子科技大学,2005.
[18]黄侃.计算机立体视觉中的摄像机标定算法的研究与应用[D].沈阳：沈阳工业大学,2007.
[19]范志斌.基于图像局部不变性特征的机器人目标识别与定位[D].北京：北京交通大学,2011.
[20]刘瑶.遥操作机器人的双目视觉立体重建研究[D].北京：北京邮电大学,2009.
[21]程晓亮.基于双目立体视觉的三维重构研究[D].哈尔滨：哈尔滨工业大学,2006.
[22]刘盛夏.基于相位信息的立体图像匹配研究[D].上海：上海交通大学,2007.
[23]翟乃斌.基于计算机视觉的汽车整车尺寸测量系统的研究[D].长春：吉林大学,2007.
[24]王永强.基于约束算法的大尺寸三维视觉测量方法研究[D].北京：北京邮电大学,2007.
[25]夏永泉.计算机视觉中双目匹配相关技术研究[D].南京：南京理工大学,2007.
[26]Faugeras O D,Toscani G.The calibration problem for stereo[C]//IEEE Computer Society Conference on Computer Vision and Pattern Recognition,Minmi Beach,Florida,June 22-26,1986:15-20.
[27]Tsai R Y.A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf cameras and lens[J].IEEE Transactions on Robotics and Automation,1987,3(4):323-344.
[28]Kim J S,Kim H W,Kweon I S.A camera calibration method using concentric circles for vision applications[C]//The 5th Asian Conference on Computer Vision,2002:515-520.
[29]Zhang Z.A flexible new technique for camera calibration[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2000,22(11):1330-1334.
[30]Basu A.Active calibration:alternative strategy and analysis[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,1993:495-500.
[31]Du F,Brady M,Self-calibration of the intrinsic parameters of camera for active vision system[C]//Proceedings of IEEE Conference on Computer Vision and Pattern Recognition,1993:477-482.
[32]Ma S D.A Self-calibration technique for active vision system[J].IEEE Transaction on Robotics and Automation,1996,12(1):114-120.
[33]杨长江,孙凤梅,胡占义.基于二次曲线的纯旋转摄像机自标定[J].自动化学报,2001,27(3)：310-317.
[34]杨长江,孙凤梅,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.
[35]雷成,吴福朝,胡占义.一种新的基于主动视觉系统的摄像机自标定方法[J].计算机学报,2000,23(11)：1130-1139.
[36]李华,吴福朝,胡占义.一种新的线性摄像机自标定方法[J].计算机学报,2000,23(11)：1121-1129.
[37]吴福朝,李华,胡占义.基于主动视觉系统的摄像机自标定研究[J].自动化学报,2001,27(6)：752-762.
[38]胡占义,雷成,吴福朝.关于共面P4P问题的一点讨论[J].自动化学报,2001.27(6)：770-776.
[39]Hartley R.Self-calibration of stationary cameras[J].International Journal of Computer Vision,1997,229(1):2-5.
[40]Benallal M,Meunier J.Camera calibration with simple geometry[C]//The Processing of 2003 International Conference on Image and Signal,2003:165-170.
[41]Carvalho P C P,Szenberg F,Gattass M.Image-based modeling using a two-step camera calibration method[C]//Proceedings of International Symposium on Computer Graphics,Image Processing and Vision,1998:388-395.
[42]Wu Y H,Zhu H J,Hu Z Y,Wu F C.Camera calibration from the quasi-affine invariance of two parallel circles[C]//The 8th European Conference on Computer Vision (ECCV), 2004:190-202.
[43]Wu Y H,Hu Z Y.The invariant representations of a quadric cone and a twisted cubic[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003, 25(10):1329-1332.
[44]Yang A Y,Hong W,Ma Y.Structure and pose from single images of symmetric objects with applications to robot navigation[C]//Proceedings of IEEE International Conference on Robotics & Automation,Taipei,2003:1013-1020.
[45]Carvalho P C P.Szenberg F,Gattass M.Image-based modeling using a two-step camera calibration method[C]//Proceedings of International Symposium on Computer Graphics, Image Processing and Vision,1998:388-395.
[46]Wu Y H,Zhu H J,Hu Z Y,Wu F C.Camera calibration from the quasi-affine invariance of two parallel circles[C]//The 8th European Conference on Computer Vision (ECCV), 2004:190-202.
[47]Wu Y H,Hu Z Y.The invariant representations of a quadric cone and a twisted cubic[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2003, 25(10):1329-1332.
[48]Yang A Y.Hong W,Ma Y.Structure and pose from single images of symmetric objects with applications to robot navigation[C]//Proceedings of IEEE International Conference on Robotics & Automation,Taipei,2003:1013-1020.
[49]Zhang Z Y.A Flexible New Technique for Camera Calibration,MS-RT-98-71[R].Microsoft Research,December 1998.
[50]肖志涛,国澄明,侯正信,于明.图像特征检测算法的分析和研究[J].中国图像图形学报,2004(12)：1414-1420.
[51]王永明,王贵锦.图像局部不变性特征与描述[M].北京：国防工业出版社,2010.
[52]Moravec H P.Towards Automatic Visual Obstacle Avoidance[C]//Proc.5th International Joint Conference on Artificial Intelligence,San Francisco,CA,USA,1977:584.
[53]Harris C,Stephens M.A Combined Corner and Edge Detector[C]//Proc.Alvey Vision Conf.Univ.Manchester,UK,1988:147-151.
[54]Lindeberg T.Edge detection and ridge detection with automatic scale selection[J]. International Journal of Computer Vision,30,2,1998:117-156.
[55]Mikolajczyk K,Schmid C.An affine invariant interest point detector[C]//In Proceedings of the 8th International Conference on Computer Vision,Vancouver,Canada,2002:128-142.
[56]Lindeberg T.Shape from texture from a multi-scale perspective[C]//In Proc.International Conference on Computer Vision,Germany,1993:683-691.
[57]Lowe D G.Object recognition from local scale-invariant features[C]//In International Conference on Computer Vision,Confu,Greece,1999:1150-1157.
[58]Bay H,Tuytelaars T,van Gool L.SURF:Speeded up robust features[C]//In 9th European Conference on Computer Vision,Graz,Austria,2006:346-359.
[59]Matas J,Chum O,Urban M,Pajdla T.Robust wide baseline stereo from maximally stable extremal regions[C]//In 13th British Machine Vision Conference University of Cardiff, 2002:761-767.
[60]Lindeberg T.Feature detection with automatic scale selection[J].International Journal of Computer Vision,1998,30(2):77-116.
[61]Lindeberg T.Detecting Salient Blob-Like Image Structures and Their Scale with a Scale-Space Primal Sketch:A Method for Focus-of-Attention[J].International Journal of Computer Vision,1993,11(3):283-318.
[62]黄令允.基于自适应阈值的SIFT算法研究及应用[D].大连：大连理工大学.2010.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |