Atmospheric correction for waterbody images acquired by spaceborne hyperspectral sensor CHRIS

CHRIS (Compact High Resolution Imaging Spectrometer) is an exploring hyperspectral remote sensor, carried on board PROBA-1 satellite. This satellite was launched in October, 2001 by the European Space Agency. With the properties of high spatial resolution, multi-angle observation and hyperspectral imagery, CHRIS offers a precious data resource for monitoring water quality. To monitor water quality by remote sensing technique quantitatively, a crucial step is to accomplish atmospheric correction accurately to gain above-water reflectance. Compared with remote sensing images over land, water reflectance is a weak signal, which requires more accurate atmospheric correction. 6S (Second Simulation of Satellite Signal in the Solar Spectrum) and MODTRAN (Moderate resolution TRANsmittance code) are two widely-used radiative transfer models. Most softwares utilize these two models to perform atmospheric correction of remote sensing images. This paper apply two sorts of softwares, including the 6S-based REMS (Remote Sensing Environmental Monitoring System) and MODTRAN-based ACORN (Atmospheric CORrection Now) to remove atmosphere effects over three CHRIS images acquired under different view angles over Meiliang Bay, Taihu Lake. We compared the reflectance spectra extracted from the atmospherically corrected images with in-situ measurement. The result shows that the water reflectance corrected by REMS and ACORN are quite close to the measured data. The correlation coefficients over the whole spectral range reach 90%. The angle properties of in-situ measured data are analyzed, and the angle properties of corrected images are more obvious than those of in-situ observation. The difference values of reflectance under the three view angles are bigger in green light than in red and near infrared light, agreeing with the measurement. The angle properties of reflectance corrected by ACORN better correspond with those of measured reflectance.