A model of intelligent interpreting soil erosion based on geographical knowledge

Soil erosion is a key issue of land degradation and a main phenomenon of environmental deterioration. RS (Remote Sensing) and GIS (Geography Information System) have made great progress in the research on soil erosion and soil and water conservation since the end of 1980s´. All methods applying RS and GIS to study soil erosion are usually divided into Visual interpretation, Spectral Analysis, Parameter Determination and Man-computer Interactive Interpretation. But the model of intelligent interpreting soil erosion based on the geographical knowledge (MIISE) is different from four groups of the methods above. MIISE is developed from Visual Interpretation and Man-computer Interactive Interpretation to survey soil erosion for large scale, by use of a new expert system of spatial feature. In the article, the first part introduces the principle MIISE based on. The second part shows how to employ computer to realize MIISE, and how to make up knowledge database and the structure of the experts system. The third part shows MIISE evaluation. MIISE was applied in the study area to interpret the soil erosion, and the result was compared with original soil erosion map surveyed by field method. Finally, it is concluded that the geographical knowledge is an effectual way to discover the relationship between the soil erosion and its influent factors, and MIISE is more accurate and efficient than usual method such as Visual interpretation, Spectral Analysis and Man-computer Interactive Interpretation.