Comparison of efficient techniques of hyper-spectral image preprocessing for mineralogy and vegetation studies

Abstract

In the present study, EO-1 hyperion data is used for processing. Out of 242 bands, 163 bands are taken in the calibrating condition. To extract the vegetation and mineralogy from hyperion imagery needs various preprocessing steps such as bad bands removal, destriping, radiometric calibration and reflectance generation. Vertical destriping process is performed with local destriping algorithm. Radiance of the imagery generates in the BIL format at the scale factor of 0.1. Log residuals, flat field correction, IARR, QUAC and FLAASH atmospheric correction methods are used to remove error from bands and generate the reflectance. Comparative analysis is also performed on various atmospheric corrections. Results showed that FLAASH is the efficient atmospheric correction method compared to other methods.