An atmospheric correction algorithm, defined as the solution of a statistical inference problem, has been developed to process satellite ocean-color data into water reflectance. The definition of the inversion algorithm relies on an estimate of the distribution of the uncertainties on the top-of-atmosphere (TOA) reflectance, corrected for molecular effects. This distribution is estimated from an in-situ match-up dataset. The forward operator is discretized using a radiative transfer code, and the theoretical solution is approximated numerically. SeaWiFS spectral bands and Case-1 waters are considered in the simulations. The inverse problem is signicantly ill posed, i.e., quite different water reflectance spectra may correspond closely to the observed TOA reflectance spectrum. In view of this, the solution is approximated in a Bayesian framework by the conditional expectation of the water reflectance given the TOA reflectance. Satellite estimates of marine reflectance agree with in situ measurements. The mean squared differences (in ×10-5) are 2.16 at 412 nm, 1.12 at 443 nm, 0.77 at 490 nm, 0.53 at 510 nm, 0.46 at 555 nm, and 0.03 at 670 nm, and the mean absolute relative difference is 19.7%. Application to SeaWiFS imagery shows a substantial noise reduction in the spatial elds of water reflectance compared with the corresponding SeaDAS-derived fields. The methodology allows the construction of uncertainties on the retrieved water reflectance, without shape restrictions, a perspective for future work.
[1]
H. Engl,et al.
Regularization of Inverse Problems
,
1996
.
[2]
Bruno Pelletier,et al.
Fields of non-linear regression models for atmospheric correction of satellite ocean-color imagery
,
2007
.
[3]
Bruno Pelletier,et al.
Remote sensing of phytoplankton chlorophyll-a concentration by use of ridge function fields.
,
2006,
Applied optics.
[4]
Roland Doerffer,et al.
Atmospheric correction algorithm for MERIS above case‐2 waters
,
2007
.
[5]
P. J. Werdell,et al.
An improved in-situ bio-optical data set for ocean color algorithm development and satellite data product validation
,
2005
.
[6]
Robert Frouin,et al.
A general ocean color atmospheric correction scheme based on principal components analysis: Part I. Performance on Case 1 and Case 2 waters
,
2007,
SPIE Optical Engineering + Applications.