Algorithms to Improve Spectral Discrimination from Indian Hyperspectral Sensors Data

With the inclusion of a hyperspectral imager (HySI) sensor on the Indian Mini Satellite (IMS-1) Earth observation mission and subsequently near identical hyperspectral sensor on the Chandrayaan-1 lunar mission, 64-band hyperspectral data from both these missions have provided the user community rich information to explore new algorithms to exploit sensorspecific parameters and to interpret and/or classify the features in multi-resolution frame. In this article, methods to improve spectral uniqueness present in the HySI by analysing adjacent bands' spectral overlaps, by implementing spectral deconvolution and reconstruction techniques are presented. Similarly, the use of multi-resolution approach for fast searching of standard spectral library end-members for better discrimination of hyperspectral pixel data are also discussed along with applications in Earth and lunar surface hyperspectral image interpretation. These spectral analyses techniques are useful in discriminating subtle differences in spectral signatures that help study the origin of secondary craters and gullies/ landslides on the lunar surface.