节点文献

三维体数据生成及三维缓冲区分析

Three-Dimensional Volume Data Generation and Three-Dimensional Buffer Analysis

【作者】 邱华

【导师】 陈学工;

【作者基本信息】 中南大学 , 计算机科学与技术, 2011, 硕士

【摘要】 在三维地理信息系统(GIS)中,三维数据模型与数据结构是研究的核心。从数据描述格式的角度划分,三维空间数据模型可以归纳为面模型和体模型两种。由于体模型可以把空间对象以离散的量化体素值进行表示,在表现三维空间物体内部特征上有很大的优势,利于对物体进行空间分析,所以它是空间信息系统的重要基础之一。本文以三维体数据为基础,主要研究了三维体数据的生成算法和基于体数据表示的三维缓冲区分析算法。在三维体数据生成方面,本文提出了一种针对三角网格模型的快速实体体素化算法。该算法的实现分为两个步骤:第一,表面体素化,即将三角网格模型中的三角面片细划为若干足够小的小三角面片,体素化各小三角面片顶点,将复杂的面体素化转换为简单的点体素化;第二,内部体素化,运用扩展的深度缓存原理得到体模型在三维空间的六个方向投影,利用这些投影进行内部体素的判定,并结合边标志填充算法有效减少了需进行判定的体素个数。对于精细复杂的大规模三角网格模型,本文算法能确实有效的生成逼近原模型的26-连通的体素模型,且具有高效的时间效率。在三维缓冲区分析方面,本文使用了三维带符号的欧式距离变换算法,但在高分辨率下体数据的数据量极大,甚至可能超出内存的装载容量,导致一般算法难以使用。因此,基于一般海量数据的分块处理方法,本文给出了针对三维缓冲区分析的海量数据处理策略。实验证明,该方法能有效的完成海量数据下的三维缓冲区分析。

【Abstract】 In the research of the three-dimensional Geographic Information System (GIS), three-dimensional data model and data structure is the core. The three-dimensional spatial data model includes the surface model and the volume model according to the data description format. Because of spatial objects can be rendered by the discrete voxel values, volume model has more advantages in showing inner characters of spatial object. It has higher efficiency in some spatial analysis. So it is one of important basis for spatial information system. The algorithm of three-dimensional volume data generation and the algorithm of three-dimensional buffer analysis for volume data model are mostly researched in this thesis.In the aspect of three-dimensional volume data generation, a rapid solid voxelization algorithm based on triangular mesh model is propoes in this thesis. The algorithm can be realized in two steps:the first step is surface voxelization, dividing the triangular faces of triangular mesh model into some other new minor triangle chip which is sufficiently small, then voxelizing these triangle vertexes, which can convert surface voxelization to points voxelization; the second step is internal voxelization, the internal voxel can be recognized by using the six directional projection of the voxel model which can be obtained by extending the Z-buffer algorithm, the number of voxel which needs recognizing can be reduced by combining with the edge marking fill algorithm. As for complicated and fine triangular mesh model, the algorithm not only can efficiently generate voxel model of 26-connected which is approached to original model, but also have high time efficiency.In the aspect of three-dimensional buffer analysis, the algorithm of three dimensional signed euclidean distance transform is used, but the data quantity in high resolution is large, and even may larger than loading capacity of memory, it leads the general algorithm cannot be used. Therefore, a massive data processing strategy of three-dimensional buffer is given in this thesis based on the partition strategy. The results of experiment show that the method can effectively complete the three-dimensional buffer analysis under the massive data.

  • 【网络出版投稿人】 中南大学
  • 【网络出版年期】2012年 01期
节点文献中: 

本文链接的文献网络图示:

本文的引文网络