Research And Application Of GML Spatial Database Prototype System Based On Db4o

As a result of the application and researching of GML is more extensive, vast amounts of spatial GML data are producing, the number of spatial data which exist in XML format encoding of document is large, a lot of troubles is bringing to data managers and users, so people want to find a suggestion which could make better of the storage and managemen of spatial GML data in database. This is the effective guarantee to make realize of Spatial GML Data Sharing and Interoperability, it is also a good way to realize the spatial data centralized management and maintenance at this stage, but currently do not have a real sense of spatial database or Spatial Database Engine can provide a effective storage and management for GML spatial data.Based on the characteristics of spatial GML data and the background which is talked, a relevant research has been carried out about the storage of GML spatial data on land use and index query in this paper. The main contents include the following:(1) According to the XML features which GML dacuments have, combining with the adnantages of VTD-XML, a GML spatial data parse engine that coubled be used for land-use spatial data parse has been realized based on C#.NET, the engine can parse GML document which contains large amounts of data fastly and accurately;(2) A comparion of the current main storage management model is done, and proposes a suggestion of import spatial data into GML object database and has a research to the open source object database engine db4o, also use C#.NET language to develope a prototype system of object GML spatial database, what the main idea is to import object-oriented data of land-use into the db4o database, the process of objectification algorithm improved mapping method of ODMG and a mapping rules of mapping GML schema to object database is proposed, the rules greatly reduces the number of objects and improve the performance of the system;(3) Today huge amounts of GML data are very common, to query data content of interest from vast amounts of GML data, spatial data indexing is particularly important. For the disadvantage of B-tree index which is provided by db4o, this paper study and implement the classic R-Tree spatial index which is widely used in Oracle Spatial, PostGIS and ArcSDE for PostgreSQL/Informix. Finally, create a index in the object database which stores the spatial database, and store the index into object database, relative to the general spatial database index, the index of the object spatial database is stored directly as the form of object, improve the efficiency of the index file system access, then analyzed and implemented the query processing and optimization of spatial data, and a GML land Query system wiich used C#.NET language is been achieved. Through the analysis of spatial index performance, verify the feasibility of the spatial index and query speed efficiency.