【作者】 石海潜；

【导师】 吴秋云；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2011， 硕士

【摘要】 随着空间数据采集技术的快速发展,人类获取并累积了大量的空间数据,这些数据资源在网络上呈现分布式存储以及异构的特征,使多源空间数据的集成成为迫切需要解决的问题。目前数据集成的各种解决方案在实际应用中仍存在一些不足,集中表现在:一是数据整合加工的时间长,数据访问的时效性差;二是系统耦合性强,接入新数据源时配置复杂,可扩展性差;三是系统依赖或受限于具体的技术,开放性差。针对上述不足,论文研究了一种网络环境下对分布式异构空间数据进行松散耦合数据集成的方法,主要思路是,基于面向服务架构(SOA,ServiceOrientedArchitecture),增强分布式数据源的访问接口规范性;通过对不同空间数据源内部数据的高效组织和快速预处理,提高单源数据实时访问效率,进而提高整体系统集成访问效率。论文主要工作有:1、提出了一个基于SOA的多源空间数据集成访问框架,该框架是松耦合的,具有良好的可扩展性,能够保证各数据源的自治性,且接入配置灵活、简单。2、采用了一种面向快速可视化的空间数据剖分组织方法以及基于视点可见区域的数据动态调度策略,通过构建了一个统一的可视化框架,提高了空间数据的访问效率。3、提出了面向服务发布的遥感影像数据金字塔索引并行构建方法,该方法采用MPI并行编程技术,在集群环境下实现数据的并行处理,加快了空间数据的预处理速率。4、设计实现了两种空间数据服务发布与集成访问方法,分别为网络地图服务WMS和瓦片地图服务TMS,这两种方法均是面向网络、符合SOA架构的。5、设计实现了基于SOA的多源空间数据集成访问原型系统,通过原型系统验证了本文所提出的若干关键技术与解决方案的可行性。更多还原

【Abstract】 With the rapid development of spatial data acquisition technology, human beingsacquire and accumulate large amount of spatial data resources, which is distributed andheterogeneous on network. This makes the multi-source spatial data integratingtechnology becoming an urgent problem that need to be solved.While in practical appliance, there are still some deficiencies in the solutions ofdata integration at present. Firstly, data processing takes long time, making theefficiency of data accessing is bad. Secondly, the system’s expansibility is bad becauseit is strong coupling. Connecting a new data source is complex. Thirdly, the system’sopenness is bad because it is dependent or limited by special technique.According to the deficiencies discussed above，this paper studied a data integrationmethod in the form of loose coupling which is used to access the distributed andheterogeneous data in the network environment. The main idea is: Enhance thenormative of the distributed data source accessing interface based on SOA (ServiceOriented Architec). With efficient organization and rapid preprocessing of the spatialdata from different data source, improve the accessing efficiency when accessing tosingle data source in real time, and then furthermore improve the integrating accessefficiency of the whole system. The main jobs of this paper are:1. Propose a SOA-based multi-source spatial data integrating access framework.The framework is loose coupling and has good expansibility. This ensure the autonomyof the data resources. Besides the connection configuration is flexible and simple.2. Use a data subdivision and organization method that orients rapid visualization,a data dynamic scheduling method based on the viewpoint of viewable region. Bybuilding a unified visualization framework, the accessing efficiency of spatial data isimproved.3. Propose a remote image pyramid index parallel building method. The methodprocesses data in the cluster environment using MPI parallel programming technics. Itspeeds up the spatial data preprocessing rate.4. Design and implement two methods which are used to publish and integratingaccess the spatial data service. The methods are called WMS (Web Map Service) andTMS (Tile Map Service) respectively. They are network-oriented and in line with SOA.5. Design and implement a SOA-based multi-source spatial data integrating accessprototype system. The prototype system validates the feasibility of the key technologiesand solutions proposed in the paper.更多还原