Land use classification based on decision tree in karst rocky desertification areas

Kurst rocky desertification is a process of land degradation. The spatial data of desertification areas´ land use is mainly through the interpretation of satellite images to get. Supervised classification and unsupervised classification are traditional interpretation methods, but their classification precision are low. And the result of desertification data automatic extracted by them also can´t make us satisfactory. Now, a new image interpretation method ,decision tree classification can be employed. In this study, we use the ASTER image data, DEM data and lithologic data, and extract normalized difference vegetation index, ration vegetation index from image data, the slope from DEM. By using lithologic data and extracted data, we set decision tree classification rules and construct a decision tree. Then on the ENVI software support, we get decision tree classification images. By comparing the reference data and surveying field, we access the category classification accuracy and kappa coefficient. Finally the results showed that using decision tree approaching to land use classification in karst rocky desertification areas, we can get better land classification results and rocky desertification information. The results also prove automatic extracting the rocky data in karst rocky desertification is feasible .