Research on New Classification Methods of Remote Sensing of Mass Ingredient without Vegetation of Hei Shan Gorge in Yellow River Basin

Method of normalized spectrum was presented for problem-saving of spectral complexity and separating capacity, which were used to differ prtrous mountain from exposed soil and desert. Using the method, normalized spectral index (NSI) was established; then, the preous mountain index (RMI) was created; finally, we established desert-exposed soil difference model (DS-Def). The above results indicated that the precision is higher than traditional classification. But the method is too complex to extract the information quickly, so we selected above sensitive factors as new bands to classify mass ingredient in non-vegetation area using supervise classification. The result indicted that the method is relatively simple and effective.