Abstract
Haze, an atmospheric condition that degrades many remote sensing images, is part of the atmospheric correction. In this study, a new method is proposed to reduce the effect of haze on hyperspectral imagery over inland water bodies using the spectral unmixing approach. If haze is present in the image, spectral unmixing can result an end member corresponding to haze, and a corresponding haze abundance map (HAM) can be generated. The spectrally and spatially varying haze contribution to reflectance spectra, referred to as hyperspectral haze reflectance cube, is estimated using the HAM and the reflectance spectra extracted from haze affected areas of the hyperspectral image. Actual reflectance from water bodies can thus be obtained by subtracting this cube from the observed hyperspectral reflectance cube. Two hyperspectral images acquired with Airborne Visible Infrared Imaging Spectrometer-Next Generation (AVIRIS-NG) sensor over the Ujani reservoir of Maharashtra, India, are used to develop and validate the proposed hyperspectral haze correction method. We demonstrate the successful removal of haze effects by validating with field measured reflectance spectra, visual, spectral and statistical interpretation of the haze corrected AVIRIS-NG imagery. The proposed methodology can be readily implemented for existing and future hyperspectral imagery from NASA’s Plankton, Aerosol, Cloud and ocean Ecosystem (PACE) mission.
A new approach for haze correction in the hyperspectral image over water bodies.
Haze treated as spatially and spectrally varying atmospheric condition.
Developed hyperspectral haze reflectance abundance surface.
Spectral and statistical validation of the proposed haze correction approach.
Highlights
Disclosure statement
The authors have no potential conflicts of interest.
Data availability statement
The AVIRIS-NG data used in this study are publicly available at avirisng.jpl.nasa.gov/