| With the progress of research on Lop Nur dried salt lake, more and more data including attribute and spatial need to be managed reasonably for application. This study uses a new kind of data model and database tool GEODATABASE to construct and develop the Lop Nur"Great Ear"area spatial database. This paper gives an introduction about spatial database system and how to construct GEODATABASE data model. This paper makes a classification of salt crust in Lop Nur lake areas by using of decision tree methods. Based on the design of storage model for vector database and multi-source remote sensing image database in logical and physical, This study has developed a spatial database which realize the storage and network share of the spatial information data of object areas.This study achieves the following results:Firstly, the spatial database of Lop Nur"Great Ear"area has been designed and realizationed based on GEODATABASE data model.Secondly, the relationships among multi-source remote sensing images have been established by using metadata table based on the classification of the spatial resolution of remote sensing image data, which are helpful for organizing and managing image data.Thirdly, the ETM+ images of Lop Nur salt crusts have been classified by using decision tree technology which gives an overall classification accuracy of 86.3% and the Kappa coefficient is 0.8420. The results show that the distribution of different types of salt crust is consistency with the"Great Ear"rings characteristics of remote sensing images in Lop Nur"Great Ear"dry salt lake area, the salt crusts intersect in the same gershgorim band of images, similar shapes of salt crusts show different hue due to the impact of humidity of surface on the remote sensing images.Finally, Lop Nur"Great Ear"spatial database system has been developed by using Visual Studio (C #). NET 2005 language, based on the platforms of ArcGIS Server and ArcSDE and SQL Server2000.
|
PDF Full Text Request