摘要
三维重建是计算机视觉技术的主要内容之一,研究了由两幅或多幅二维图像恢复出被拍摄物体的深度信息的方法。其中基于两幅图像的双目视觉技术是一个研究热点。双目立体视觉模仿人眼与人类视觉的立体感知过程,从两个视点观察同一景物,以获取不同视角下的图像,通过三角测量原理计算图像像素间的位置偏差,以获取景物的深度信息。双目立体视觉技术在虚拟现实、机器视觉、多媒体教学、数字娱乐、产品外观设计、雕刻与建筑等领域都有着广泛的应用。
本文在分析研究大量相关技术和文献的基础上,对立体视觉领域中的摄像机标定、立体匹配、三维重建这三个关键技术进行了研究。主要研究如下:
1.自制标定模板,用普通数码相机拍摄两幅不同角度的带有标定模板的手机照片。利用Tsai两步法对照相机进行标定,求出了照相机的内外参数。
2.对所拍摄的照片进行滤波、锐化等图像处理,采用基于特征的匹配方法,从两幅图像中分别提取特征点,然后利用极线约束等匹配准则对一幅图像中的特征点在另一幅图像中寻找匹配点。
3.在已获得照相机的内外参数和完成左右图像的特征点匹配的基础上,根据立体视觉原理实现了对特征点的三维重建,给出了特征点的空间坐标。最后在计算机上通过OpenGL编程将这些点显示出来,实现了手机的三维重建。
One of the main component of the computer vision is 3D reconstruction.It is to restore the depth information of an object by two or more than two images of two-dimensional.The technology of binocular stereo vision based on two images is a hot point.The binocular stereo vision has a strong resemblance to the visual perception procedure of mankind. We can get different images with two different view points and then calculate the position deviation of images pixels based on triangular measure.Finally, we can attain the depth information of the object.The binocular stereo vision widely used in many fields such as virtual reality, machine vision, multimedia education,digital entertainment, appearance design of industrial products, sculpture and architecture etc.
Based on the analysis and research of much technology and literatures concerned,this paper studies the basic problems of stereo visual domain such as camera calibration,stereo matching and the 3D reconstruction and so on.The corresponding work are given as follows:
1 .Do the calibration template by ourself then shoot two different angles images of the cell phone with a calibration template. We use the Tsai's two-step method for calibration and obtained the internal and external parameters.
2.Process these photos by image smoothing, image sharpening and other image pre-processing.We use the operator to obtain the feature points of both images.We search this image's matching point in other image based on the epipolar line restriction and so on.based on the matching method of feature.
3.Based on the results of the cameras calibration and feature points matching,we reconstructed the character points according to the principle of the stereo vision ,and get the 3D coordinates of the character points. At last we use OpenGL program in the computer to display the disperse points and implement the three-dimensional reconstruction of the cell phone.
引文
[1]刘传才.图像理解与计算机视觉.[M].厦门:厦门大学出版社,2002
[2]Rangachar Kasturi.Machine Vision[M].China Machine Press,2003
[3]李伟.立体视觉若干基本问题的研究:[硕士学位论文][D].昆明理工大学,2005.12
[4]王新宇.基于计算机立体视觉的三维重建:[硕士学位论文][D].中南大学,2004.5
[5]佘彦杰.基于多幅图像序列的三维重建:[硕士学位论文][D].吉林大学,2006.4
[6]霍长娟,张桦,邓熹.几种摄像机标定方法的比较[J].天津理工大学学报.Vol.23,NO5,pp.75-77,2007(10)
[7]张备伟.摄像机标定方法研究:[硕士学位论文][D].安徽:安徽大学,2001.5
[8]王明昕,赵东标,邵泽明.CCD摄像机内外参数标定技术研究[J].机械电子.pp12-14,2004(3)
[9]刘元坤,苏显渝,吴庆阳.基于傅里叶条纹分析的多摄像机标定方法[J].光子学报.Vol.36,No.9,pp.1734-1737,2007(9)
[10]盖绍彦,达飞鹏.一种新的相位法三维轮廓测量系统模型及其标定方法研究[J].自动化学报.Vol.33,No.9,pp.902-910,2007(9)
[11]段成钢,李秀智.一种基于三消隐点的摄像机自标定算法[J].计测技术.Vol.27,No.4,pp.20-25,2007
[12]Hakan Bacakoglu and Mohamed S.Kamel,Senior Member,IEEE.A Three-Step Camera Calibration Method.IEEE Transansactions on instrumentation and measurement,VOL.46,NO.5,Octobier 1997
[13]于洪川,吴福朝,袁波,韦穗.基于主动视觉的摄像机自标定方法[J].机器人.Vol.21,No1,pp1-7,1999(1)
[14]SD.Faugeras,O.T.Luong,and S.J.Maybank."Camera self-calibration:theory and experiments".European Conference on Computer Vision,volume 588,1992,pp.321
[15]Ma S D.A self-calibration technique for active vision systems.IEEE Trans Robotics and Automation,1996,12(1):114-120
[16]Zhang Z Y.A flexible new technique for camera calibration[J].IEEE Transactions on pattern Analysis and Machine Intelligence,2000,22(11):1330-1334
[17]Tsai R Y.An efficient and accurate camera calibration technique for 3D machine vision[A].IEEE Computer Society Conference onComputer Vision and Pattern Recognition[C].Miami:IEEE Comput Soc Press,1986,:364-374
[18]马颂德,张正友.计算机视觉[M].北京:科学出版社,1998
[19]Weng J Y,Cohen P,Herniou M.Camera calibration with dis-tortion model and accuracy evaluation.IEEE Trans Pattern Analysis and Machine Intelligence,1992,14(10):965-980
[20]J.Weng,P.Cohen and M.Herniou."Camera Calibration with Distortion Models and Accuracy Evaluation."[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,vol.14,no.10,October1992,pp 965-980
[21]郑南宁.计算机视觉与模式识别[M].北京:国防工业出版社,1998
[22]马连峰,张秋菊,李挺.一种基于自制平面模版的摄像机标定方法[J].中国制造业信息化Vol.36,NO.17,pp.45-48,2007(9)
[23]陈书海,傅录祥.实用数字图像处理[M].北京:科学出版社.2005
[24]朱虹.数字图像处理基础.北京:科学出版社,2005
[25]霍宏涛.数字图像处理[M].机械工业出版社.2004
[26]刘丹.计算机图像处理的数学和算法基础[M].北京:国防工业出版社.2005
[27]高和蓓,胡学龙,周磊.一种基于感兴趣区域的图像检索算法[J].微计算机应用.Vol.28,No.3,2007(3),pp.225-228
[28]Harris C,Stephens M.A Combined Comer and Edge Detector[C].Proceedings of the 4th Alvey Vision Conference.Manchester,England.1988,:147-151.
[29]Ruey-Feng Chang,Chii-Jen Chen,and Chen-Hao Liao.Region-based Image Retriveal Using Egeflow Segmentation and Region Adjacency Graph.IEEE,2004:1883-1886
[30]徐志祥,吴波,田一翔,田礼乔.基于角点特征的近景影像匹配算法[J].测绘信息与工程.Vol.31,No.4,pp.40-41,2006
[31]孙龙祥,程义民,王以孝,孙启彬.深度图像分析[M].电子工业出版社.1996
[32]Chun-Hua Zhuang,Zhen-Yu Li,Pu Wang and Chong-Zheng Sun.An Area and Feature-based Matching Method for 3D Reconstruction[C].IEEE,2007
[33]R.Hartley.A Linear Method for Reconstruction from Lines and Points.In proceedings of International Conference on Computer Vision,1995,882-887
[34]李秀智,张广军.三目视觉图像的极线校正方法[J].光电工程.Vol.34,No.10,pp.50-54,2007(10)
[35]张金苓,王琰.基于特征约束的图像点匹配算法[J].沈阳理工大学学报.Vol.25,No.6,pp.14-16,2006(12)
[36]舒远,谈正,丁礼儒.基于高精度匹配点的对极几何估计[J].工程图学学报.2005
[37]范秋凤,杨国胜,马晓燕,王应军.基于对应点匹配的物体深度信息测量[J].计算机应用.Vo.l 26 No.8,pp.1873-1875,2006(8)
[38]Donald Hearn M.Pauline Baker.计算机图形学(第三版)[M].北京:电子工业出版社.2005
[39]William H.Press Saul A.Teukolsky William T.Vetterling Brian P.Flannery.C数值算法(第二版)[M].北京:电子工业出版社.2004
[40]何渝.计算机常用数值算法与程序(C++版)[M].人民邮电出版社.2003
[41]Richard S.Wright,Jr.Benjamin Lipchak.OpenGL超级宝典(第三版)[M].人民邮电出版社.2005
[42]Lee Adams.三维图形与动画程序设计[M].中国广播电视出版社.1992
[43]神会存.基于手持单相机序列图像的三维重建[J].计算机工程与应用.2006.23.
[44]汪伟,李郝林.基于数码相机图像的三维重构[J].上海理工大学学报.Vol.27,No.5,pp.429-432,2005(5)
[45]于君明,黄小波.基于数码相机立体像对的物体3维信息快速获取方法[J].测绘学院学报.Vol.19No.4,pp.276-279,2002(12)
[46]高爽,肖扬.简化的图像三维重建算法[J].北方交通大学学报,2004.428(2):30-34
[47]高月芳,罗飞,曹建忠.由单幅二维灰度图像重构物体表面形状[J].计算机应用研究.Vol.24,No.8,pp.195-197,2007(8)
[48]吴凤和,张晓峰,施法中.基于单幅图像数据的三维重构方法研究[J].中国机械工程.Vol.18,No.17,pp.2071-2075,2007(9)
[49]邓俊峰.大范围的建筑三维重构的方法研究:[硕士学位论文][D].武汉:华中科技大学,2004.5
[50]史德林.基于场景的植物三维重建技术研究:[硕士学位论文][D].江苏:江苏大学,2006,5