Carbonate Rocks Lithological Discrimination Using Multi-source Remote Sensing Data in Southwestern China

The difficulty in geological mapping by using remote sensing is lithological discrimination, especially in southwest China karst area, where carbonate rocks with deep soil and flourishing vegetation cover the surface. In this paper, the use of multi-source remote sensing data, such as TM, SPOT and ASTER, for lithological discrimination was evaluated. The results indicate that the accuracy of lithological discrimination using single RS data not high, e.g., 69.36% for ASTER, 64.37% for TM and 54.41% for SPOT, but when more sorts of RS data were used for classification, higher accuracy was obtained. Except for spectral information, the inclusion of variogram texture images in image classification may considerably improve the classification accuracy. In the case of using 4 SPOT spectral bands and its 4 texture images, 6 TM spectral bands, 14 ASTER spectral bands and its 3 texture images extracted from its 3 VNIR spectral bands for classification, the final overall classification accuracy is 82.01%.