Hyperspectral remote sensing for land degradation mapping in China

Abstract

Land degradation is a major environmental problem internationally. Soil degradation is one of the key factors of land degradation, which is related to susceptibility to erosion, soil suitability, and soil characteristics especially at regional scale. It is important and meaningful to evaluate objectively land degradation at regional scale. The study is to present the classification approaches for land degradation by Degraded Soil Line Index (DSLI) and object-oriented method by determination of land degradation spectral response units (DSRU) compared to the spectral angle mapping (SAM) method using Hyperion image data for mapping land degradation. The method was tested in a study area located in Hengshan county in ShaanXi province, China, where is in the agriculture-pasture mixed area in Loess Plateau in China with complex physical geographical situation. The results showed that the three methods of SAM, DSLI and DSRU have the ability to map land degradation and degraded soil classes, and the performance of the methods of DSLI and SAM is different and DSLI is prior to SAM for land degradation mapping in the study area. Moreover, the results also showed that the object-oriented analysis method based on DSRU approach is valid for extraction of land degradation information and clearly shows the degraded land classes with an overall accuracy of 0.88 and Kappa coefficients of 0.86.