节点文献
真三维地理信息系统中海量数据处理技术的应用研究
【作者】 莫映;
【导师】 周文业;
【作者基本信息】 首都师范大学 , 计算机应用技术, 2004, 硕士
【摘要】 随着“数字地球”概念的提出和不断推广,“地理信息数字化”已成为计算机在传统学科领域的一项重要应用。但是,传统地理信息系统的数据通常被组织在二维平面中。为了更接近实际地重构和描绘地理数据,地理信息系统从二维向三维方向发展已经成为不可逆转的趋势。 与此同时,三维可视化作为科学可视化技术的一支,以其形象直观和易于理解的特点来表现各类庞杂纷繁的数据,因而受到人们普遍的欢迎。它与地理信息系统的结合形成了所谓的“真三维地理信息系统”。 GIS数据往往包含大量信息,一般都会达到几百兆甚至几个京的大小,因此被称为“海量数据”。另外,GIS数据本身的复杂性也很突出。同时,在某些对建模要求较高的应用领域,对三维地理信息系统提出了较高的实时性和交互性要求。 鉴于三维地理信息系统及GIS数据的上述特点,以目前计算机的硬件水平,还无法将所有数据一次性载入内存。由此,在现有条件下,根据实际需要,如何适时地对GIS数据进行有效地简化,便成为一个急需解决的关键问题。 另外,以往的地理信息系统多基于单机环境实现,与之相应的显示优化技术也是基于单机的。但GIS数据的特性,使得在现有技术条件下的单机系统的性能无法很好的满足应用需求,因此,考虑如何在分布环境下,通过多台计算机协同处理,共同完成GIS数据的复杂运算和调度处理,以减轻单机环境系统资源紧张,CPU满负荷运转的瓶颈问题,也是一项很有现实意义的研究内容。 针对上述两个方面,笔者在论文中作了初步的有益的尝试和探讨,并取得了一定的成绩。在GIS数据简化方面,本文通过采用一种基于四叉树结构的动态层次细节技术,根据视觉特点和场景本身的复杂度,对地形模型进行了简化,并对算法做了大量优化,从而一定程度的缓解了内存资源紧张的问题,并提高了显示速度。 在分布环境数据调度方面,本文借鉴了现有分布式体系架构,并结合了三维地理信息系统自身的特点,定义了一个系统逻辑模型。并在此基础上,采用分块策略,结合预先处理的流水线作业模式,给出了一个数据调度策略,从而保证了地形计算与绘制的并行进行。
【Abstract】 With the spreading of the concept of "Digital Earth", the geography information digitalization has become an important application in the field of traditional science, such as geography, which computer technology is used. To describe the data set more intuitional and practical, the trend of GIS from 2D to 3D is undoubted. And it is a hotspot nowadays.Meanwhile, 3D visualization is popular because of its live presentation and easy-to-understand. How to integrate 3D visualization technology into the legacy geography system? It is a significant theme.Generally speaking, there will be a very large quantity of data in normal geography system especially spatial data as well as image data. In addition, most of GIS data are complicated especially in some certain of professional applications.On the other hand, we can’t load the whole data set into the memory because of the present hardware limit. To resolve the bottleneck, we should concern about how to schedule data effectively, how to simplify data according to the current viewpoint and environment context, how to render the scene rapidly by rendering algorithm, such as Level of Detail.At the same time, most of GIS applications are based on stand-alone system. The corresponding optimized algorithms are also stand-alone based. But with the complexity and the great capacity of GIS data, as well as the real time and interactive feature of geography system, the performance of stand-alone system is not meet the application requirements very well. To resolve the problem, a reasonable solution is to transplant the system to a distributed parallel environment.In this paper, we mainly focused on Level of Detail algorithm and distributed data schedule that is mentioned above.
【Key words】 3D visualization; Large scale data set processing; Level of Detail; Dynamic data schedule strategy; Distributed technology;
- 【网络出版投稿人】 首都师范大学 【网络出版年期】2004年 04期
- 【分类号】TP399
- 【被引频次】5
- 【下载频次】489