Abstract:

This paper is about Middle Inner Mongolia region, which is a typical region of desertification. The structure of the decision tree and the corresponding position of each target are determined according to the desertification land classification. According to the spectrum-reflecting characteristic of targets and the integrated feature of TM and ETM⁺ image data, proper identifying and extracting methods are applied. So the potential of the remote sensing data is maximized. From the multi-band and multi-time TM and ETM⁺ remote sensing data, the spectrum-reflecting properties, the Soil-Adjusted Vegetation Index(SAVI), Soil Brightness Index(TC1), and Moisture Index(MI)are extracted. And these indexes are used to distinguish and classify the different land types. The result shows the decision tree and multi-layer method is efficient to decrease the intrusion of the extra information in the pixel identification and make the targets definite. Through Image Algebra Detection method and Visual Interpretation method, the changes of the desertification from 1990 to 2000 are collected, and the evolvement rule of desertification in the study area is analyzed. Finally, with the technology of GIS, the analysis of the ecological environment quality changes and climate factors changes reveals that the main factors to influence the environmental changes are precipitation and evaporation(desiccation).

Key words: desertification land, decision tree, multi-layer extraction technology, evolvement analysis, environment quality, remote sensing investigation, Inner Mongolia