RULE BASED CLASSIFICATION APPROACH FOR MAPPING LAND DEGRADATION

摘要

Land degradation is one of the major ecological problems faced by the world today. Remote sensing, now-a-days plays an important role to map degraded lands with regard to their extension, type, and severity. Traditional classification methods are mostly based on the properties of spectral reflectance. Thus, classification of features having similar reflecting properties or using remotely sensed data which have insufficient number of bands is often challenged because of some difficulties such as mixed classes. For land degradation mapping, it is observed that getting accurate result is not easy because various types of degraded areas appear only slightly different in terms of their spectral information. To overcome this challenge, a classification approach using multi-inputs as indices was implemented in this study. This study focuses on mapping land degradation using two methods, namely, supervised classification and rule based mapping approach. Compared with supervised classification for the area under study, rule based classified map using different set of indices is found more suitable for the extraction of degraded landforms.