Spectral analysis of hyperion hyperspectral data for identification of mango (Mangifera indica) species on farmlands

Abstract

Fruit species like mango (Mangifera indica) is grown by farmers on their farms. The identification of Mango species using remote sensing requires its spectral signature. Hyperspectral data has great potential for discrimination of vegetation such as tree species. In the present study, EO-1 Hyperion hyperspectral data has been used for generation of spectral signature of Mango for the months of March, May and November. Spectral signatures were found different for three periods due to phenological stage of the trees. Hyperion image has been classified using developed signature by Spectral Angle Mapper method and reasonabley good accuracy (87%) was found in classification of Mango species. Thus developed spectral signatures for a tree species would help in identification of that species as well as accurate mapping on farmlands. Hence spectral library is imperative for agroforestry systems mapping as well as estimation of area under them.