Karst rocky desertification is the most serious problems of land degradation in karst regions

southwest China. Remote sensing technique is the promising method to assess and monitor the degree and extent of karst rocky desertification at large scale. In this study

based on field spectral reflectance measurements

the traditional vegetation indices （VIs） and linear spectral unmixing （LSU） are assessed to extract the key indicators of karst rocky desertification. Karst rocky desertification synthesis index （KRDSI） has been developed with the unique of spectral features observed in non-vegetation land cover types. The results show that VIs could be used to extract the fractional cover of green vegetation

and they are not sensitive to soil background. Both VIs and LSU can efficiently extract the fractional cover of non-green vegetation. Compared with LSU

KRDSI shows more consistent results with the field measurement of non-vegetation land cover fractions. This study indicates that evaluation indicators of karst rocky desertification can be extracted from the Hyperion image with the combination of vegetation indices and KRDSI values.