Land use/cover is one of the most important factors effecting Soil and Water Loss (SWL) which causes sediment yield. Employment of remote sensing in SWL and land use/cover is more and more popular The classification precision of remotely sensed images is hard to be improved further. I4ence, it is necessary to append geographical data for multiple source data classification. Rule-based expert System is considered as an effective approach for multiple source data classification. Difficulties in capturing expertise and regional limitation greatly discount the method. Geographic information System (GIS) spatial database contains much knowledge useful for classification. This paper presents a method of deriving knowledge with and from GIS. and expresses them as rules in classification through a case study on land use/cover in the Xiamen Island, Fujian. It is shown that the result from the proposed method is more accurate than the Maximum Likelihood Classification.
展开▼