基于不规则三角网的DSM快速生成方法设计与实现

发布时间：2019-01-14 15:31

【摘要】：如今，地理信息系统地发展日益成熟，机载激光雷达技术的发展一日千里，人们对于三维数字城市地图的需求越来越迫切，如何从海量数据中快速获取到高分辨率的数字表面模型成为众多学者关注的问题。数字表面模型常规的表示方法有两种：格网法和不规则三角网模型。格网法效率高但失真较大，不规则三角网模型可以较好地描述真实地形但算法复杂度较高。针对如何使用不规则三角网快速生成数字表面模型的问题，本文进行了一系列的探讨和实验。首先总结了经典的不规则三角网构网方法——分而治之算法、逐点插入算法和三角网生长算法的优缺点，然后针对这三种算法构网效率低的问题，面向快速处理大量数据生成DSM的应用，提出了一种新的不规则三角网构网流程。本文算法主要包括建立虚拟网格、网格插值、生成初始三角网、冗余数据处理和密集数据处理五个步骤。通过建立虚拟网格可以对三维离散数据点进行快速定位，有效的网格插值有助于生成高分辨率的DSM，构造初始三角网的时间复杂度为O (N)，大大提高了不规则三角网的构网效率，，冗余数据处理和密集数据处理是对初始三角网的优化，可以一定程度上解决网格插值带来的信息冗余和网格定位带来的信息损失，因此能够有效提高不规则三角网的质量。同时，本文采用OpenGL作为三维显示工具，使用VC++编写软件，对所提出的算法进行了实验和分析，经验证，该算法时间效率较高，具有一定的实用性。
[Abstract]:Nowadays, with the development of geographic information system (GIS) and the rapid development of airborne lidar technology, the demand for 3D digital city map is becoming more and more urgent. How to quickly obtain high resolution digital surface models from massive data has become a problem that many scholars pay attention to. There are two kinds of conventional representation methods of digital surface model: grid method and irregular triangular mesh model. The grid method has high efficiency and high distortion. The irregular triangular network model can describe the real terrain well but the complexity of the algorithm is high. In order to solve the problem of how to use irregular triangulation to quickly generate digital surface model, a series of experiments are carried out in this paper. This paper first summarizes the advantages and disadvantages of the classical irregular triangulation network-divide-and-conquer algorithm, point-by-point insertion algorithm and triangulation algorithm, and then aims at the low efficiency of these three algorithms. In this paper, a new irregular triangulation network flow is proposed for fast processing of large amount of data to generate DSM. The algorithm consists of five steps: creating virtual grid, grid interpolation, generating initial triangulation, redundant data processing and dense data processing. By establishing virtual mesh, 3D discrete data points can be located quickly. Effective mesh interpolation is helpful to generate high-resolution DSM, to construct the initial triangulation. The time complexity of constructing the initial triangulation network is O (N),. The efficiency of irregular triangulation is greatly improved. Redundant data processing and dense data processing are the optimization of initial triangulation, which can solve the information redundancy caused by grid interpolation and the information loss caused by grid location to some extent. Therefore, the quality of irregular triangulation can be improved effectively. At the same time, OpenGL is used as a 3D display tool, and the software is written by VC. The proposed algorithm is tested and analyzed. It is proved that the algorithm is efficient in time and practical.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：P208;TP391.41

【参考文献】