摘要
近年来，随着计算机在各个领域应用的不断扩大以及虚拟现实等技术的兴起，人们研究的对象已经从二维图像、图形升级到了三维数据的建模。三维数据建模可以广泛地应用于工业、国防、影视制作、游戏娱乐、文物保护、虚拟现实和可视化等诸多领域。本文针对我所自主研发的激光三维真彩扫描仪VividScan，重点对其后期数据处理，模型重构，模型简化，模型交互编辑等方面进行了深入的研究，设计开发了三维模型处理软件LSIVT。主要工作包括：
1．研究了目前国内外三维扫描技术发展的新动向和新技术，对部分三维数字化仪器的工作原理、工作流程和适用领域进行了分析和对比，指出了各自的优缺点。
2．针对我所自主开发研制的三维扫描仪获取的原始三维数据进行了二维简化，提出了一种新的基于面积误差的多边形逼近方法来提取曲线形状特征点。
3．实现了从原始的三维空间点云数据重新构造三维实体模型。在此基础上完成了对实体的纹理映射，从而获得生动的三维实体模型。
4．提出了一种基于体积误差的三角形折叠网格简化算法。实例应用证明这种算法能够在较好地保持三维模型基本形状的情况下，大大简化用于表示模型的三角面片的数目。
5．针对硬件扫描过程中出现的数据误差，本文实现了交互式模型编辑的功能。用户可以通过移动点和三角面片，删除、添加、拆分三角面片等操作完成对模型的修改，减少扫描误差对模型形状的影响，从而完善三维模型。
在此基础上，我们为VividScan研发了专用的三维模型处理软件LSIVT，并投入使用，这套软件满足了三维模型处理的基本需求，并提供了一些高级的功能，操作方便，适用性强。
3D Modelling has deprived 2D Image processing and 2D graphic of attraction and has become the focus of people's attention for its particular predominance and its wide applications in industry, national defence, movie making, entertainment, culture relic conservation, virtual reality, visualization technology and so on. In this paper, studies on scanned data pre-filtering, model reconstruction, model simplification and interactive model modification are carried out.
The significance of 3D technology and its developing status are presented in this paper. Advantages and disadvantages of some 3D digitizers and scanners available on market are listed according to our through research and comparison.
An improved polygon approximation algorithm is put forward, which can extract feature points on curves steadily and more accurately for a given tolerance. The initial scanned 3D data can be simplified greatly using this algorithm, and that helps to reconstruct the model more quickly.
The shortest-distance method is used to reconstruct 3D model and texture mapping is implemented to obtain more vivid model.
A new 3D mesh simplification method is put forward. Triangle is shrunken to a vertex and triangles around the shrunken one are modified based on the calculation of volume error. It can simplify the original mesh greatly and provides efficient error management. Result of experiment is given.
Interactive model modification is implemented to adjust the shape of model since error occurs while scanning. User can move the points and triangles to reshape the model. Adding,deleting and splitting triangles are also available to meet the need of model modification.
We developed a 3D Laser Scan System, including the hardware and software. As the software is concerned, we use Microsoft Visual C++6.0 and OpenGL. Initial data filtering and simplification, model reconstruction, mesh simplification and interactive
3D shape adjusting are the main functional parts to obtain a vivid model.
引文
[1] Ediz Saykol, Gurcan Gulesir, Ugur Gudukbay, Ozgur Ulusoy. A Kinematics-Based Method for Polygon Approximation[J], Advances in Information System (ADVIS'2002), LNCS Vol. 2457, 186～194.
[2] 樊宏斌．计算机辅助颅骨面貌复原技术的研究与实现[D]．西北大学硕士论文．西安：西北大学，2002．20～22
[3] 吴中海．一个适合特征计算的多边形逼近算法[J]．计算机学报，1997．20卷，1129-1132
[4] 章毓晋．图像处理和分析[M]．北京：清华大学出版，2000．220～222
[5] Hoppe H, 1996, Progressive Meshes[J]. Computer Graphics, Proceeding, Aug, 99～108
[6] Calvin A D, Taylor R H. 1996 Superfaces: Polygonal Mesh Simplification with Bounded Error[J]. IEEE Computer Graphics and Applications, May, 64～77
[7] Rossignac J, Borrel P. 1993. Multi-resolution 3D Approximation for Rendering Complex Scenes[J]. Proceedings, Conference on Geometric Modeling in Computer Graphics, Aug, 455～465
[8] Luebke D, Erikson C. 1997. View-dependent Simplification of Arbitrary Polygonal Environment[J]. Computer Graphics Proceedings, Aug., 199～208
[9] Schroeder W J, Zarge J A, Lorensen W E. 1992 Decimation of Triangle Meshes[J]. Computer Graphics, Proceedings, Aug., 65～70
[10] 张明敏，周昆，潘志庚．基于超包络的网格简化算法[J]．软件学报，1996．10卷，584～588
[11] 陶志良，潘志庚，石教英．基于能量评估的网格简化算法及其应用[J]．软件学报，1997．8卷，881～888
[12] 周昆，马小虎，潘志庚．基于重新划分三角形网格的简化方法[J]．软件学报，1998．9卷，405～408
[13] 陈洪亮，谭建荣．基于渐进物理属性的网格简化方法[J]．计算机学报，2000．23卷，285～292
[14] 吴亚东，刘玉树，高春晓．基于二次误差度量的网格简化算法[J]．北京理工大学学报，2000，20卷．607～612
[15] 蒋遂平，周明天，戴颖．基于法向的网格简化算法[J]．计算机学报，1999．22卷，1074～1079
[16] 王军安，魏生民．基于距离加权的二次误差测量的网格简化方法[J]．计算机辅助设计与图形学学报，2001．13卷，189～192
[17] 王海滨、阮秋琦．基于边折叠的多边形网格简化方法[J]．铁道学报，2000．22卷，52～55
[18] 和平鸽工作室．OpenGL高级编程与可视化系统开发(高级编程篇)[M]．北京：中国水利水电出版社
[19] 李捷．三维网格模型的简化及多分辨率表示[D]．清华大学博士论文．1998年7月
[20] 李捷，唐泽圣．三维伏在模型的实时连续多分辨率绘制．计算机学报，1998．21卷 481～491
[21] Keppel E. 1975. Approximating Complex Surfaces by Triangulation of Contour Lines. IBM Journals Res. Develop, January, 2～10
[22] Fuchs H, Kedem Z M, Uselton S P. 1977. Optimal Surface Reconstruction from Planar Contours. Communication of the ACM, 20(10), 693～702
[23] Zimmer Y, Tepper R, Akselrod S. 1997. An improved method to compute the convex hull of a shape in a binary image PR, 30,397～402
[24] 耿国华数据结构—C语言描述[M]．西安：西安电子科技大学出版社，2003．218～225
[25] http://www.InSpeck.com
[26] http://www.polhemus.com/Products.htm
[27] http://www.jyb.com.cn/cedudaily/r10/gaodeng/gaoke183.htm
[28] http://www.nwpu.edu.cn/xb/xb587.htm
[29] http://graphics.stanford.edu/projects/mich/
[30] J. Sklansky, Recognition of Convex Blobs, Pattern Recognition, Vol 2, 3～10, 1970
[31] U. Montanari. A Note on Minimal Length Polygonal Approximation to a Digital Contour, Comm. Of the ACM., Vol 13,41～47, 1970
[32] T. Pavlidis, Structural Pattern Recognition, Springer-verlag Berlin Heidelberg, 1977
[33] 李将云．图像处理种插值和缩放若干技术的研究[D]．浙江大学博士论文．2002
[34] 张献颖．计算机辅助文物复原关键技术研究[D]．西北大学硕士学位论文，2003
[35] Genapathy S, Dennehy T G. 1982. A New General Triangulation Method for Planar Contours. Computer Graphics, 16(3), 69～75