Research And Implementation Of Remote Sensing Image Management System For Spatial-temporal Information Mining

With the development of remote sensing technology, the number of remote senseing image grows at a very fast speed. People's capability of collecting and storing images has greatly exceed the capability of analyzing and achieving information from the images, and this contributes to the development of remote sensing image mining technology. Remote sensing image mining technology is not only the application of image mining technology in the field of remote sensing, but also is an important branch of spatial data mining technology. Due to the large amount of remote sensing images, storing and querying these images are very important and indispensable steps. This thesis designs and implements a remote sensing image management system for spatial-temporal information mining, which depends on the "Spatial-temporal Information Mining" project.This thsis introduces ArcGIS, the related technology and development tools at first. Then design and implementation of this system are described in detail. This thesis analyzes the overall demand of this system, designs the whole thinking and presents the framework of the system. It implements the basic functions of remote sensing image management, including importing spatial data, managing relational data, data query and visualization. This thesis uses C# as the main programming language, and.NET Framework to build the whole framework of this system. It implements the graphical user interface, which is friendly and easy to operate. Besides, this thesis uses SQL Server 2005 to manage data, and implements data storage and query functions with database technology. Spatial data are inserted into relational database by ArcSDE interface, which manages the geographic database. It also uses ArcGIS Engine to import shapefile into spatial database.For the convenience of users to query and view the data mining results, this thesis designs and implements the result visualization module to solve the problem of association rules'display by using JAVA GUI technology. It also supports the query and view of association rules for users, including querying rules' ID, rules'length and support. Last, it displays the rules which meet the query to users.In addition, for the problem of data query's efficiency, this thesis implements the optimization of query time by using Rtree index. It designs data query module based on the Counting algorithm, and compares the influence on the query time by using Rtree index with SQL query.Finally, the system is tested comprehensively, and future work is presented.