Atmospheric correction of HJ-1 CCD imagery over turbid lake waters

We have presented an atmospheric correction algorithm for HJ-1 CCD imagery over Lakes Taihu and Chaohu with highly turbid waters. The Rayleigh scattering radiance (L-r) is calculated using the hyperspectral L-r with a wavelength interval 1nm. The hyperspectral L-r is interpolated from L-r in the central wavelengths of MODIS bands, which are converted from the band response-averaged L-r calculated using the Rayleigh look up tables (LUTs) in SeaDAS6.1. The scattering radiance due to aerosol (L-a) is interpolated from L-a at MODIS band 869nm, which is derived from MODIS imagery using a shortwave infrared atmospheric correction scheme. The accuracy of the atmospheric correction algorithm is firstly evaluated by comparing the CCD measured remote sensing reflectance (R-rs) with MODIS measurements, which are validated by the in situ data. The CCD measured R-rs is further validated by the in situ data for a total of 30 observation stations within +/- 1h time window of satellite overpass and field measurements. The validation shows the mean relative errors about 0.341, 0.259, 0.293 and 0.803 at blue, green, red and near infrared bands. (C) 2014 Optical Society of America