Atmospheric correction algorithm over coastal and inland waters based on the red and NIR bands: application to Landsat-8/OLI and VNREDSat-1/NAOMI observations
Ngoc, D.D.; Loisel, H.; Duforet-Gaurier, L.; Jamet, C.; Vantrepotte, V.; Goyens, C.; Xuan, H.C.; Minh, N.N.; Van, T.N. (2019). Atmospheric correction algorithm over coastal and inland waters based on the red and NIR bands: application to Landsat-8/OLI and VNREDSat-1/NAOMI observations. Optics Express 27(22): 31676-31697. https://dx.doi.org/10.1364/OE.27.031676 In: Optics Express. Optical Society of America: Washington. e-ISSN 1094-4087, more |   |
| Authors | | Top | - Ngoc, D.D.
- Loisel, H.
- Duforet-Gaurier, L.
| - Jamet, C.
- Vantrepotte, V.
- Goyens, C., more
| - Xuan, H.C.
- Minh, N.N.
- Van, T.N.
|
| Abstract | Water pixel extraction and correction of the atmospheric signal represent prerequisite steps prior to applying algorithms dedicated to the assessment of water quality of natural surface water bodies. The recent multiplication of medium spatial resolution sensors (10–60 m) provides the required constellation to monitoring bio-optical and biogeochemical parameters of surface waters at the relevant spatial-temporal scales. Here we present a new approach to identify water pixels and to extract the atmospheric contribution to the top of atmosphere signal measured by the NAOMI sensor on board the first Vietnamese satellite, VNREDSat-1. After verifying the TOA calibration of NAOMI through a vicarious calibration exercise, we adapt a recent water pixel extraction algorithm (WiPE) to NAOMI, and develop a new atmospheric correction algorithm (referred to as red-NIR) based on the use of the red and NIR bands (the only bands available for that purpose on NAOMI) and spectral relationships. The evaluation of red-NIR with a match-up data set gathering remote sensing reflectance, Rrs, measurements performed at the AERONET-OC stations in moderately turbid waters indicates excellent performance in the blue and green part of the spectrum (similar to the performances reached by the SeaDAS NIR-SWIR algorithms) and lower accuracy in the red. Intercomparison of simultaneous images collected by NAOMI and OLI over a more turbid water body shows an excellent agreement between the NAOMI-Rrs estimated by the present processing, and the OLI-Rrs estimated from the ACOLITE algorithm. This approach will allow sensors that do not have SWIR bands, such as SPOT-6 and -7, to be processed, making their data exploitation available for long-term temporal analyses. |
|
