三维地质体建模可视计算及并行化的研究与应用
|
摘要
可视计算可概括为以计算机图形形式展现数据的计算过程,涉及到计算机图形学、计算机视觉、人工智能、程序设计以及软件工程等学科。可视计算可以帮助人们从海量数据中解放出来,并发现隐藏在数据背后的规律和本质,以便更好地认识和利用这些数据。随着城市建设与野外工程的不断发展,地质勘探相关部门已经积累了大量的地质资料信息,采用合适的数据结构产生具有地质特征的数字模型,并运用科学可视化技术对地质体进行三维再现和交互操作,可以帮助地质专业人员对勘探数据做出正确解释和分析,从而提高地质分析的准确率和可靠性。
地质体模型的绘制一般采用面绘制技术和体绘制技术。面绘制技术可以产生比较清晰的等值面图像,适用于绘制表面特征明显的地层。但是该技术不能反映整个数据场内部细节。体绘制技术可以很好地弥补这个不足,能产生三维数据场的整体图像。但它涉及的数据量较多且计算量较大,因而绘制时间较长,随着并行处理体系结构的完善与并行处理算法的成熟,再加上图像处理具有内在的并行性,并行处理技术成为提高体绘制速度的主要研究方向。
本文首先研究了三维地质体的建模体系与建模方法,从地质应用的需求出发,以地质勘探数据为研究对象,研究地层数据的组织、空间建模与数学表达,在综合分析和比较目前比较流行的建模方法(面模型、体模型和面体结合模型)的基础上,提出了一种适合工程地质的地层数据模型。该模型基于面向对象的设计思想,并综合利用了不规则三角网(TIN)和广义三棱柱(GTP)模型。
在研究和总结体绘制的发展历程与关键技术基础之上,本文着重研究了体绘制中的光线投射算法,结合多核处理器机群系统,提出并实现了一种基于多层次并行编程模型的并行光线投射体绘制算法,并成功地将该算法应用于三维城市浅层地质模型,取得了良好的可视化效果。本文在研究过程中,分别对MPI环境和多层次并行编程MPI+OpenMP环境下的光线投射算法进行了不同计算规模的性能比较实验。实验和分析表明,多层次并行光线投射体绘制算法加快了体绘制的速度,同时结果还表明MPI+OpenMP多层次并行编程模型性能高于纯MPI编程模型的性能。此外,本文对VTK并行机制进行了详细的研究,采用实际数据对任务并行模型进程验证并取得了良好的加速效果。
基于地质体建模算法与可视化算法,本文设计并实现了一套三维地质体可视化系统(Geological Visualization System),包括数据预处理、数据建模、三维地质体显示等功能。该系统是在开源平台QT与开源可视化工具包VTK的基础上,采用面向对象的设计思路,借助Visual Studio 2005开发环境设计和实现。系统提供了三维显示和常用的用户交互操作,包括模型图像旋转、缩放、剖切、动态演化和等值面提取等。本文应用该系统对理论模型数据、实际钻孔数据和实际地震勘探数据三套资料进行了处理和分析,并取得了良好应用的效果。
Visual computing can be concluded as the computing procedure which expresses data in computer graphical way, it involves multiple disciplines such as computer graphics, computer vision, artificial intelligence, programming design and software engineering. Visual computing can help people relieve from sea level data and discover the pattern behind it. As the development of the city construction and field engineering, a lot of geological information has been accumulated. Adopting proper data structure to built geological digital model and using scientific visualization technology to represent and interact with the model, geological profession can make precise explanation and analysis about the data, so as to improve the accuracy and reliability of geological analysis.
Surface rendering and volume rendering are common techniques when rendering geological model. Surface rendering technique can help produce clear isosurface image, it is suitable for rendering the stratum which has obvious surface feature. But this technique could not reflect the inner detail of the whole data field. On the contrary, volume rendering can fix up this defect. This method is time consuming because of the large scale data and computation. As the development of parallel process architecture and parallel algorithm, plusing the inner parallelism of imaging processing, parallel visualization has became an important research area of improving the speed of volume rendering.
This thesis studied modeling architecture and modeling method of 3D geological data. On the background of the need of geological application and geological exploration data, this thesis delved deeper about the organization of stratum data, the mathematic expression of space modeling. After analyzing and comparing recent popular modeling like surface model, volume model and surface-volume model, a stratum model which fits engineering geology has been proposed. This model is based on Object-oriented-design. It makes full use of Triangulated Irregular Net (TIN) and General Tri-prism (GTP) model.
On the basis of researching and concluding the development procedure of volume rendering and its key technology, combined with multi-core cluster, a parallel ray casting volume rendering algorithm based on multi-hierarchical parallel programming model has been proposed in this thesis. This multi-hierarchical parallel algorithm has been used in 3D city shallow geological model successfully. According to the comparative experiments and the analysis of different computation scale on ray casting algorithm between the pure MPI programming model and the multi-hierarchical parallel MPI + OpenMP programming model, the results show that the multi-hierarchical parallel ray casting algorithm accelerates the volume rendering speed, the results also show the multi-hierarchical parallel MPI + OpenMP programming model has better performance than pure MPI programming model. In addition, this thesis researched the parallel mechanism supported by VTK and adopted real data to test the task parallism method and get good accelerating effect.
Based on geological modeling and visualization algorithms, this thesis designed and realized the Geological Visualization System which includes four parts (data processing, data modeling, 3D geological visualization, user interactive). To develop this system, this thesis used open source platform QT and Visualization Tool Kit, integrate Visual Studio 2005 and adopt Object-oriented method. This system can provide display and user interactive function such as rotation, zoom in/out, clip, color bar, slice animation and isosurface extracting etc. This thesis processed and analyzed the theoretical model data, real bore-hole data and real seismic exploration data and get good results.
地质体模型的绘制一般采用面绘制技术和体绘制技术。面绘制技术可以产生比较清晰的等值面图像,适用于绘制表面特征明显的地层。但是该技术不能反映整个数据场内部细节。体绘制技术可以很好地弥补这个不足,能产生三维数据场的整体图像。但它涉及的数据量较多且计算量较大,因而绘制时间较长,随着并行处理体系结构的完善与并行处理算法的成熟,再加上图像处理具有内在的并行性,并行处理技术成为提高体绘制速度的主要研究方向。
本文首先研究了三维地质体的建模体系与建模方法,从地质应用的需求出发,以地质勘探数据为研究对象,研究地层数据的组织、空间建模与数学表达,在综合分析和比较目前比较流行的建模方法(面模型、体模型和面体结合模型)的基础上,提出了一种适合工程地质的地层数据模型。该模型基于面向对象的设计思想,并综合利用了不规则三角网(TIN)和广义三棱柱(GTP)模型。
在研究和总结体绘制的发展历程与关键技术基础之上,本文着重研究了体绘制中的光线投射算法,结合多核处理器机群系统,提出并实现了一种基于多层次并行编程模型的并行光线投射体绘制算法,并成功地将该算法应用于三维城市浅层地质模型,取得了良好的可视化效果。本文在研究过程中,分别对MPI环境和多层次并行编程MPI+OpenMP环境下的光线投射算法进行了不同计算规模的性能比较实验。实验和分析表明,多层次并行光线投射体绘制算法加快了体绘制的速度,同时结果还表明MPI+OpenMP多层次并行编程模型性能高于纯MPI编程模型的性能。此外,本文对VTK并行机制进行了详细的研究,采用实际数据对任务并行模型进程验证并取得了良好的加速效果。
基于地质体建模算法与可视化算法,本文设计并实现了一套三维地质体可视化系统(Geological Visualization System),包括数据预处理、数据建模、三维地质体显示等功能。该系统是在开源平台QT与开源可视化工具包VTK的基础上,采用面向对象的设计思路,借助Visual Studio 2005开发环境设计和实现。系统提供了三维显示和常用的用户交互操作,包括模型图像旋转、缩放、剖切、动态演化和等值面提取等。本文应用该系统对理论模型数据、实际钻孔数据和实际地震勘探数据三套资料进行了处理和分析,并取得了良好应用的效果。
Visual computing can be concluded as the computing procedure which expresses data in computer graphical way, it involves multiple disciplines such as computer graphics, computer vision, artificial intelligence, programming design and software engineering. Visual computing can help people relieve from sea level data and discover the pattern behind it. As the development of the city construction and field engineering, a lot of geological information has been accumulated. Adopting proper data structure to built geological digital model and using scientific visualization technology to represent and interact with the model, geological profession can make precise explanation and analysis about the data, so as to improve the accuracy and reliability of geological analysis.
Surface rendering and volume rendering are common techniques when rendering geological model. Surface rendering technique can help produce clear isosurface image, it is suitable for rendering the stratum which has obvious surface feature. But this technique could not reflect the inner detail of the whole data field. On the contrary, volume rendering can fix up this defect. This method is time consuming because of the large scale data and computation. As the development of parallel process architecture and parallel algorithm, plusing the inner parallelism of imaging processing, parallel visualization has became an important research area of improving the speed of volume rendering.
This thesis studied modeling architecture and modeling method of 3D geological data. On the background of the need of geological application and geological exploration data, this thesis delved deeper about the organization of stratum data, the mathematic expression of space modeling. After analyzing and comparing recent popular modeling like surface model, volume model and surface-volume model, a stratum model which fits engineering geology has been proposed. This model is based on Object-oriented-design. It makes full use of Triangulated Irregular Net (TIN) and General Tri-prism (GTP) model.
On the basis of researching and concluding the development procedure of volume rendering and its key technology, combined with multi-core cluster, a parallel ray casting volume rendering algorithm based on multi-hierarchical parallel programming model has been proposed in this thesis. This multi-hierarchical parallel algorithm has been used in 3D city shallow geological model successfully. According to the comparative experiments and the analysis of different computation scale on ray casting algorithm between the pure MPI programming model and the multi-hierarchical parallel MPI + OpenMP programming model, the results show that the multi-hierarchical parallel ray casting algorithm accelerates the volume rendering speed, the results also show the multi-hierarchical parallel MPI + OpenMP programming model has better performance than pure MPI programming model. In addition, this thesis researched the parallel mechanism supported by VTK and adopted real data to test the task parallism method and get good accelerating effect.
Based on geological modeling and visualization algorithms, this thesis designed and realized the Geological Visualization System which includes four parts (data processing, data modeling, 3D geological visualization, user interactive). To develop this system, this thesis used open source platform QT and Visualization Tool Kit, integrate Visual Studio 2005 and adopt Object-oriented method. This system can provide display and user interactive function such as rotation, zoom in/out, clip, color bar, slice animation and isosurface extracting etc. This thesis processed and analyzed the theoretical model data, real bore-hole data and real seismic exploration data and get good results.
引文
[1]唐泽圣.三维数据可视化[M].北京:清华大学出版社,1999.
[2]陈树铭,王满春,刘慧杰等.工程地质三维数字化及计算机三维实现[J].中国:土木工程学会第十届年会,2002.10.
[3]乌尽伦,刘瑜等.地理信息系统一原理、方法和应用[M].北京:科学出版社,2001.
[4] Chris Hardina. Modeling Geoscience Data in A Multisensory Virtual Environment. Computing in Science & Engineering. 2004,(l):89-92.
[5] Houlding S.W. 3D Geoscience Modeling Computer Techniques for Geologic characterization[M]. Berlin: Springer– Verlag,1994.
[6]齐安文,吴立新,候恩科等.三维地学模拟述评及其矿上应用关键问题.中国矿业. 2001,10(5):61-64.
[7]李智毅,唐辉明.岩土工程勘察[M].武汉:中国地质大学出版社,2005.
[8]中国测绘学会地图学与GIS专业委员会.地图学与地理信息系统工程的现状与趋势[J].测绘通报. 1997,(6):26~31.
[9]陈昌彦,张菊明,杜永廉等.边坡工程地质信息的三维可视化及其在三峡船闸边坡工程中的应用[J].岩土工程学报.1998.20(4):1-6.
[10]柴贺军,黄地龙,黄润秋.岩体结构面计算机三维扩展模型研究[J].地质灾害与环境保护. 1999,6(02):73-76.
[11]黄地龙,柴贺军,黄润秋.岩体结构可视化软件系统研究[J].成都理工学院学报, 2001,28(4):416-420.
[12]柴贺军.大型地质工程岩体结构三维可视化建模及其工程应用研究[D].四川大学, 2001, 10:l-10.
[13] TITAN4.0地学综合咨询系统产品发布会资料汇编.北京东方泰坦科技有限公司, 2001,5.
[14] W. E. Lorensen and H. E. Cline, Marching Cubes: A High Resolution 3D Surface construction Algorithm[J]. Computer Graphics Proceedings, Aug, 1997, 199-208.
[15] H. E. Cline and W. E. Lorensen, Two Algorithms for Three-dimensional Reconstruction of Tomograms[J], Medical Physics, 15(3), 1988,320-327.
[16] L. Westover, Footprint Evaluation for Volume Rendering[J], Computer Graphics, 24(4), 1990, 367-376.
[17] P. Lacroute and M. Levoy, Fast Volume Rendering Using a Shear-Warp Factorization of the Viewing Transformation[J], Computer Graphics, 25(4), 1994, 285-288.
[18] J. Wilhelms and V. Gelder, A Coherent Projection Approach for Direct Volume Rendering[J], Computer Graphics, 25(4), 1990, 41-47.
[19] M. Levoy, Display of Surfaces from Volume Data[J], IEEE Computer Graphics and Application, 8(3), 1988, 29-37.
[20] K. Ma, J. S.Painter, C. D. Hansen, Parallel Volume Rendering Using Binary-Swap Composition[J], IEEE Computer Graphics and Application, July, 1994, 59-68.
[21]颜辉武,祝国瑞.地下水的体视化研究[M].武汉:武汉大学出版社. 2004
[22]钟登华,李明超.水利水电工程地质三维建模与分析理论及实践[M].北京:中国水利水电出版社. 2006.
[23]杨东来,张永波,王新春等.地质体三维建模方法与技术指南[M].北京:地质出版社,2007.
[24]韩国建,郭达志,金学林.矿体信息的八叉树存储和检索技术[J].测绘学报,1992, 21(l), 13-17.
[25]张煜,白世伟.一种基于三棱柱体体元的三维地层建模方法及应用[J].中国图像图形学报. 2001,6(3),285-290.
[26]吴立新,史文中, Christopher Gold. 3DGIS与3DGMS中的空间构模技术[J],地理与地理信息科学, 2003, 19(l):5-11.
[27]龚键雅,夏宗国.矢量与栅格集成的三维数据模型[J].武汉测绘科技大学学报,1997, 22(l),7-15.
[28]李清泉,李德仁.三维空间数据模型集成的概念框架研究[J].测绘学报,1998,27(4), 325-330.
[29]徐长青,许志闻,郭晓新.计算机图形学[M].北京:机械工业出版社,2004.
[30]贾志刚.精通OpenGL[M].北京:电子工业出版社,1998.
[31] www.vtk.org. VTK 5.2.0 Documentation[M]. USA,2008.
[32] VTK User Guide[M]. (http://kitware.com/Porduest/VTKguide.html)
[33] William J Schroeder, Kenneth M Martin, William E Lorensen. The design and implement of An 0bject 0rienetd Toolkit For 3D Graphics and Visua1ization[M].
[34]武晓波,王世新,肖春生. Delaunay三角网的生成算法研究[J].测绘学报,1999, 28(1):28-35.
[35] http://baike.baidu.com/view/1691145.htm. 2010.5.
[36] Lewis B A. and Robinson J S. Triangulation of Planar Regions with Applications[J], The Computer Journal, 1978,21(4):324~332.
[37] Wu L X, Topological relations embodied in a generalized tri-prism (GTP) model for a 3D geoscience modeling system[J], Computer & Geosciences, 2004, 405~418.
[38]施木俊,熊毅明,甄云鹏.基于工程勘察钻孔数据的三维地层模型的自动构建[J].城市勘测,2006,5:62~65.
[39] www.CadCaeCam.com, 2010.3.
[40]陈国良编.并行计算--结构算法编程[M].北京:高等教育出版社,2003.8.
[41]罗省贤,何大可.基于MPI的网络并行计算环境及应用[M].成都:西南交通大学出版社,2001.
[42] Michael J. Parallel Programming in C with MPI and OpenMP影印版[M].北京:清华大学出版社,2005.
[43] James Ahrens, Charles Law, Will Schroeder, Kenneth M Martin. A Parallel Approach for Efficiently Visualizing Extremely Large Time-Varying Databases. Los Alamos National Laboratory -- Technical Report.
[44]韩伟杰,张文.基于VTK的并行可视化服务器设计与实现[J].超级计算通讯, 2006, 4(2).
[45] http://education.siggraph.org/resources/cgsource/instructional-materials/archives/ volume-visualization-data-sets. 2010.5.
[2]陈树铭,王满春,刘慧杰等.工程地质三维数字化及计算机三维实现[J].中国:土木工程学会第十届年会,2002.10.
[3]乌尽伦,刘瑜等.地理信息系统一原理、方法和应用[M].北京:科学出版社,2001.
[4] Chris Hardina. Modeling Geoscience Data in A Multisensory Virtual Environment. Computing in Science & Engineering. 2004,(l):89-92.
[5] Houlding S.W. 3D Geoscience Modeling Computer Techniques for Geologic characterization[M]. Berlin: Springer– Verlag,1994.
[6]齐安文,吴立新,候恩科等.三维地学模拟述评及其矿上应用关键问题.中国矿业. 2001,10(5):61-64.
[7]李智毅,唐辉明.岩土工程勘察[M].武汉:中国地质大学出版社,2005.
[8]中国测绘学会地图学与GIS专业委员会.地图学与地理信息系统工程的现状与趋势[J].测绘通报. 1997,(6):26~31.
[9]陈昌彦,张菊明,杜永廉等.边坡工程地质信息的三维可视化及其在三峡船闸边坡工程中的应用[J].岩土工程学报.1998.20(4):1-6.
[10]柴贺军,黄地龙,黄润秋.岩体结构面计算机三维扩展模型研究[J].地质灾害与环境保护. 1999,6(02):73-76.
[11]黄地龙,柴贺军,黄润秋.岩体结构可视化软件系统研究[J].成都理工学院学报, 2001,28(4):416-420.
[12]柴贺军.大型地质工程岩体结构三维可视化建模及其工程应用研究[D].四川大学, 2001, 10:l-10.
[13] TITAN4.0地学综合咨询系统产品发布会资料汇编.北京东方泰坦科技有限公司, 2001,5.
[14] W. E. Lorensen and H. E. Cline, Marching Cubes: A High Resolution 3D Surface construction Algorithm[J]. Computer Graphics Proceedings, Aug, 1997, 199-208.
[15] H. E. Cline and W. E. Lorensen, Two Algorithms for Three-dimensional Reconstruction of Tomograms[J], Medical Physics, 15(3), 1988,320-327.
[16] L. Westover, Footprint Evaluation for Volume Rendering[J], Computer Graphics, 24(4), 1990, 367-376.
[17] P. Lacroute and M. Levoy, Fast Volume Rendering Using a Shear-Warp Factorization of the Viewing Transformation[J], Computer Graphics, 25(4), 1994, 285-288.
[18] J. Wilhelms and V. Gelder, A Coherent Projection Approach for Direct Volume Rendering[J], Computer Graphics, 25(4), 1990, 41-47.
[19] M. Levoy, Display of Surfaces from Volume Data[J], IEEE Computer Graphics and Application, 8(3), 1988, 29-37.
[20] K. Ma, J. S.Painter, C. D. Hansen, Parallel Volume Rendering Using Binary-Swap Composition[J], IEEE Computer Graphics and Application, July, 1994, 59-68.
[21]颜辉武,祝国瑞.地下水的体视化研究[M].武汉:武汉大学出版社. 2004
[22]钟登华,李明超.水利水电工程地质三维建模与分析理论及实践[M].北京:中国水利水电出版社. 2006.
[23]杨东来,张永波,王新春等.地质体三维建模方法与技术指南[M].北京:地质出版社,2007.
[24]韩国建,郭达志,金学林.矿体信息的八叉树存储和检索技术[J].测绘学报,1992, 21(l), 13-17.
[25]张煜,白世伟.一种基于三棱柱体体元的三维地层建模方法及应用[J].中国图像图形学报. 2001,6(3),285-290.
[26]吴立新,史文中, Christopher Gold. 3DGIS与3DGMS中的空间构模技术[J],地理与地理信息科学, 2003, 19(l):5-11.
[27]龚键雅,夏宗国.矢量与栅格集成的三维数据模型[J].武汉测绘科技大学学报,1997, 22(l),7-15.
[28]李清泉,李德仁.三维空间数据模型集成的概念框架研究[J].测绘学报,1998,27(4), 325-330.
[29]徐长青,许志闻,郭晓新.计算机图形学[M].北京:机械工业出版社,2004.
[30]贾志刚.精通OpenGL[M].北京:电子工业出版社,1998.
[31] www.vtk.org. VTK 5.2.0 Documentation[M]. USA,2008.
[32] VTK User Guide[M]. (http://kitware.com/Porduest/VTKguide.html)
[33] William J Schroeder, Kenneth M Martin, William E Lorensen. The design and implement of An 0bject 0rienetd Toolkit For 3D Graphics and Visua1ization[M].
[34]武晓波,王世新,肖春生. Delaunay三角网的生成算法研究[J].测绘学报,1999, 28(1):28-35.
[35] http://baike.baidu.com/view/1691145.htm. 2010.5.
[36] Lewis B A. and Robinson J S. Triangulation of Planar Regions with Applications[J], The Computer Journal, 1978,21(4):324~332.
[37] Wu L X, Topological relations embodied in a generalized tri-prism (GTP) model for a 3D geoscience modeling system[J], Computer & Geosciences, 2004, 405~418.
[38]施木俊,熊毅明,甄云鹏.基于工程勘察钻孔数据的三维地层模型的自动构建[J].城市勘测,2006,5:62~65.
[39] www.CadCaeCam.com, 2010.3.
[40]陈国良编.并行计算--结构算法编程[M].北京:高等教育出版社,2003.8.
[41]罗省贤,何大可.基于MPI的网络并行计算环境及应用[M].成都:西南交通大学出版社,2001.
[42] Michael J. Parallel Programming in C with MPI and OpenMP影印版[M].北京:清华大学出版社,2005.
[43] James Ahrens, Charles Law, Will Schroeder, Kenneth M Martin. A Parallel Approach for Efficiently Visualizing Extremely Large Time-Varying Databases. Los Alamos National Laboratory -- Technical Report.
[44]韩伟杰,张文.基于VTK的并行可视化服务器设计与实现[J].超级计算通讯, 2006, 4(2).
[45] http://education.siggraph.org/resources/cgsource/instructional-materials/archives/ volume-visualization-data-sets. 2010.5.