【作者】 陈建华；

【导师】 苗放；

【作者基本信息】 成都理工大学 ， 地球探测与信息技术， 2008， 博士

【摘要】 空间数据管理是地理信息科学的重要研究内容,也是空间信息系统相关厂商重点研发的部分。空间数据管理的高效与否,直接关乎空间信息系统的成败、关乎空间数据的有效利用、关乎空间信息技术的发展与应用,意义重要。GML作为Internet环境下空间信息的全新、简洁、统一的编码标准,为空间数据建模、栅/矢一体化GIS体系研究、WebGIS研发、空间数据共享与互操作研究等提供了新的方法和途径。作为基于XML的地理标记语言,鉴于GML的文本性、自描述性、可扩展性、内容与表现分离性、开放性、互操作等特性,使得GML被广泛关注、研究与实践,大量甚至海量的GML空间数据正不断产生。如何对其进行有效管理,从而实现GML空间数据的高效存储、索引、查询、传输与交换等,是GML相关研究者、GML空间数据使用者所必须面对的重要问题;也是当前和未来以GML为核心的空间信息系统、GML空间信息集成系统、GML空间信息共享与互操作体系等能否有效实现、实施、运行所必须面对的关键问题。针对GML空间数据的管理,存在基于关系型数据库管理和基于原生模式XML数据库管理两大类;由于GML数据的层次嵌套模型特点,采用关系模型对文档型GML空间数据进行存储管理并不理想。由于GML数据XML格式的特点,采用原生XML数据库对GML空间数据进行管理具有与生俱来的优势,但是由于GML数据的空间特性,原生XML数据库无法直接对GML数据进行存储管理,必须对其进行扩展研究。而从国内外研究进展来看,GML空间数据管理机制的研究,尤其是原生XML数据库环境下GML空间数据管理机制的研究还处于非系统化、非全面化、非应用化的部分内容研究、初始研究阶段。论文着力对原生模式环境下GML空间数据的存储、索引、查询、传输与交换,以及与应用的集成等进行了较为全面的研究。主要表现为:(1)简述XML与地理空间数据,分析GML特征、GML核心模式与GML应用模式,根据GML规范与建模思想,给出一个二维校园平面建模示例。(2)研究原生GML空间数据库系统架构体系,包括:概念、内容、体系结构等;同时研究给出GML空间数据原生模式存储模型,其中涉及GML实例文档与GML应用模式的相关性与独立性。(3)研究GML空间数据原生模式存储索引机制与算法。结合XML索引编码方案、XML查询的结构连接算法与传统空间数据索引算法,提出基于Ex-Dewey前缀编码与R树的GML空间数据索引机制。(4)研究原生GML空间数据查询机制与查询语言,集成原生GML空间数据索引机制,深化研究并拓展基于XQuery的GML空间数据查询语言。(5)研究GML空间数据的高效压缩传输方案、GML空间数据与KML、SVG格式数据的交换方法。(6)研究原生GML空间数据库系统与数字地球原型系统的应用集成。提出一种原生GGEarth空间数据应用服务系统架构体系;设计其原型系统以实现以原生GML空间数据库为标准数据中心,以数字地球平台原型系统客户端为前端的数据显示、查询、更新、分析为一体的空间数据应用服务体系。论文主要研究成果与创新点表现为:(1)提出了一种原生GML空间数据库系统架构体系。给出了原生GML空间数据库系统定义、系统架构与研究内容。同时研究确立了以元素为粒度的原生GML空间数据存储模型。(2)提出了一种针对XML/GML非空间数据索引构建的Ex-Dewey前缀编码方案。Ex-Dewey前缀编码方案在消除Dewey编码方案中节点的插入、删除影响其后趋同胞节点及其子孙节点编码变更等缺点的同时,提出了节点插入、删除对已有节点编码值串无影响的更新策略,并保留了Dewey前缀编码的优点。同时提出了一种针对Ex-Dewey编码方案的XML/GML非空间数据查询的结构连接算法—ED-XQ-SJ。给出了算法思想、描述与验证。算法能够根据XML/GML非空间数据索引信息直接判断潜在先辈、子孙节点集中节点间的先辈—子孙(或父—子)关系,而无需访问实际存储的节点。算法在降低复杂度的同时,也明显减少了I/O开销。(3)提出了一种基于Ex-Dewey前缀编码与R树的GML空间数据索引机制。GML实例文档中的GML空间数据采用R树构建索引,而非空间数据则基于Ex-Dewey前缀编码方案进行索引构建,二者独立又紧密关联,共同实现GML空间数据的索引。(4)深化研究并拓展了基于XQuery的GML空间数据查询语言,并命名其为:GXQuery。在前人研究的基础上,全面从扩展XQuery的原因、GML空间数据查询语言的特征、XQuery空间扩展的内容、GXQuery查询语言系统结构、GXQuery查询语言实现方法、GXQuery查询语言查询实例等方面进行了深化研究,拓展了GML空间数据查询语言的组成、内容与应用。(5)提出了一种原生GGEarth空间数据应用服务系统架构体系。给出了研究内容和关键实现技术;并开发了GGEarth实验系统,对5·12汶川地震时序分布和数字九寨虚拟游览进行了实验运行。更多还原

【Abstract】 Spatial Data Management is very important research content in Geoinformatics, and is primarily research and development part to spatial infromation systems’ vendors. It is efficient or not directly concerns to rise and decline of spatial information systems, concerns to the efficient use of spatial data, concerns to the development and application of spatial information technology.GML is a kind of novel, succinct, unified encoding international standard for spatial information under the Internet enviroment, which provides new methods and approaches for spatial data modeling, raster and vector unified GIS research, WebGIS research and development, spatial data sharing and interoperability research. As based on XML, GML has characteristics of textuality, self-description, extensibility, separability of content and presentation, openness, interoperability, which makes GML is extensively concerned, studied and practised, and meanwhile, lots of even mass GML spatial data are being produced. How to efficiently manage these GML data and then efficiently implement storage, index, query, transport and transformations to GML spatial data is a key problem to GML researchers and GML spatial data users, and also is a key problem to efficiently achieve, implement and carry out for GML spatial infromation systems, GML spatial infromation integration systems, GML spatial information sharing and interoperability systems. To GML spatial data management, there have two main categories: based on RDBMS and based on native XML Database. As the feature of level embeding model of GML data, using RDBMS to manage GML data is not perfect. And because of the feature of XML like data format, using native XML Database to manage GML spatial data has the vantages of birth. But the spatiality of GML data, navtive XML Database can not directly manage them, it must be extended. While seeing the research progress home and abroad, GML spatial data management, especially based on native XML Database is still in stage of scattered, partial and initiatory.This thesis mainly concerns the research on GML spatial data’s storage, index, query, transport and transformations based on native mode, and its integration with application. As follows. (1) The general view of XML and geo-spatial data is given, the GML feature, GML schema and GML application schema are analysed. And according to GML specifications and modeling concepts, a two-dimensional campus modeling example is given. (2) The achitecture of Native GML Spatial Database System is researched, including concept, content and system achitecture. And the native mode storage model of GML spatial data is also researched, involving the correlativity and independence between GML instance data and GML application schema. (3) The indexing mechanism and algorithm of GML spatial data storage are researched. Combined with XML indexing encoding scheme, XML query structure join algorithm and traditional spatial data indexing algorithm, a kind of GML spatial data indexing mechanism based on Ex-Dewey prefix encoding and R-tree is proposed. (4) Native mode GML spatial data querying mechanism and query language are researched. Integrated with native mode GML spatial data indexing mechanism, GML spatial data query language based on XQuery is deeply researched and extended. (5) Efficient compression transport scheme of GML spatial data and transformations methods between GML data and KML data, or SVG format data are researched. (6) The integrating application of Native GML Spatial Database System and digital earth prototype system is researched. A kind of achitecture of native GGEarth spatial data application and service system is proposed. Its prototype system is designed in order to implement a new kind of spatial data application and service system involving functions with data displaying, querying, updating and analysing, which uses Native GML Database as the standard data center and the digital earth platform prototype system client as its front-end.The main research findings and innovations of this thesis are as follows. (1) A kind of achitecture of Native GML Spatial Database System is proposed. Its concept, system achitecture and content are detailedly given. And also element is ascertained as the storage granularity of native GML spatial data storage model. (2) A kind of Ex-Dewey prefix encoding scheme aiming at XML or GML non-spatial data indexing construction is proposed. Ex-Dewey encoding scheme eliminates Dewey encoding scheme’s shortcoming of node’s inserting and deleting effecting its following siblings’ and corresponding descendants’ nodes encoding, and proposes new updating strategy of node’s inserting and deleting no effecting others’ encoding value. It also keeps the advantages of Dewey encoding scheme. And a corresponding structure join algorithm - ED-XQ-SJ for XML or GML non-spatial data querying is proposed. The algorithm’s idea, description and verification are given. The algorithm can directly determine the ancestors-descendants or parents-children relationship between potential ancestor node set and descendant node set, and do not access the real storage nodes. It reduces complexity, and obvious decreases the I/O overhead as well. (3) A kind of GML spatial data indexing mechanism based on Ex-Dewey prefix encoding and R-tree is proposed. It uses R-tree to construct spatial index for GML spatial data from GML instance document, and uses Ex-Dewey prefix encoding scheme to construct index for GML non-spatial data, they are independent but closely correlative, and together accomplish the indexing of GML data. (4) GML spatial data query language based on XQuery is deeply researched and extended, and named GXQuery. Based on research of former researchers, this thesis deeply researches GXQuery from reasons of extending XQuery, features of GML spatial data query language, spatial extending content of XQuery, system achitecture of GXQuery, implementation methods of GXQuery and query examples of GXQuery. It extends GML spatial data query language’s constitution, content and application. (5) A kind of native GGEarth spatial data application and service system achitecture is proposed. Its main research content and key implementing techniques are given. And a GGEarth experimental system is developed, which has made experimental running for ’5·12’ Wenchuan Earthquake time sequence distribution and digital Jiuzhai virtual touring.更多还原