Knowledge Discovery And Application Based On Landuse Spatlal Database

A plenty of survey data of land-use and multi-source & multi-temporal remote sensing data have been accumulated in China , some of which have been used to establish the spatial database. Since there is no potential relationship in the database, the knowledge can not be quickly excavated when facing the mass data. It`s hard to find the relationships and rules of data and forecast the trends according to the existing data, the fact that lacking of methods of excavating knowledge from mass data leads to the appearance of phenomenon〝people are flooding by data, while still hungering for knowledge〞.Discovering knowledge from spatial database which based on sub-compartment by combining the land use database and multi-source & multi-temporal remote sensing data is still a problem waited to be deeply researched..Because of diversity and complexity of GIS data, it has been an important direction and subject in the area of spatial data to use common methods to get knowledge and excavate the relationships and rules existing in geosciences knowledge from mass of geosciences data. Therefore, under the background of this condition, the paper find geosciences knowledge and remote sensing knowledge of kinds of land types from different topography and scales based on spatial database, then use these knowledge into the judgment of sub-compartment property.The paper chooses two study area: mountain area and hilly. In mountain, discovering remote sensing knowledge from TM image, and finding spectral information of various land types from TM image. In the hilly, discovering multitemporal spectra information of dry land and paddy from MODIS images,and descryibing their change process of spectra. Discovering the altitude and slope of various land types based on geosciences knowledge of land-use database. The paper carries out the discovery of spectral knowledge and multidimensional characteristics knowledge from ALOS. Final,the paper take hilly as an example to establish the database, then apply it to the judgment of the property of sub- compartment and the result can be checked by using the land-use types. The experiment indicates that the property of sub- compartment can be judged by using the knowledge database, the first level judgment precision can be 88.11%, the second level approaches 89.75%, and the third level achieves 92.62%; the total precision is 93.03%.