CHARACTERIZATION OF GLACIATED AREA USING HYPERSPECTRAL IMAGING

Abstract

NASA’s Earth Observing - 1 (EO-1) Hyperion satellite, launched in 2000, served as a stage to obtain hyperspectral images from the space. Hyperspectral imaging sensors provide valuable information to aid to the identification of geological and topographical features and other targets on the earth surface. The contiguous bands of narrow bandwidth helps in generating laboratory equivalent spectral signature and imaging techniques provides wider coverage area at a time. This advantage of hyperspectral imaging can be used in mapping mineral abundances over any area, thereby reducing tedious field data collection. The present study takes Hyperion images to characterize glaciated area in Satopanth Glacier, Uttarakhand, India using Hyperion data and obtain spectral features of glacial rocks. Pre-processing of the data becomes mandatory for correct identification of the areas before applying atmospheric correction models and consists first of removing bad bands and noise prone bands which reduces the number of bands to 180 from 242 followed by destriping of bands using local techniques. After that FLAASH atmospheric correction model is applied which retrives a reflectance image which is further used for characterizing the study area. Characterization of the area under study involved the use of MNF transform, followed by Pixel Purity identification. This was followed by end member collection which was used to characterize the study area. By comparing spectral signatures generated using hyperspectral with the USGS library spectral data and with the field maps, it has been observed that hyperspectral data is in agreement with ground based and library data. Further it is recommended that spectral unmixing study may be carried to separate pixels and characterize different features and minerals on earth surface