Publications

Li, Q; Jiang, LL; Chen, YL; Wang, L; Wang, LX (2022). Evaluation of seven atmospheric correction algorithms for OLCI images over the coastal waters of Qinhuangdao in Bohai Sea. REGIONAL STUDIES IN MARINE SCIENCE, 56, 102711.

Abstract
Atmospheric correction (AC) over coastal waters is always critical in ocean color remote sensing. The performance of the AC algorithms determines the reliability of the retrieval of the water-leaving radiance and subsequent bio-optical parameters. In this study, the seven AC algorithms for the Ocean and Land Color Instrument (OLCI) were evaluated with an in situ dataset over the coastal waters of Qinhuangdao in the Bohai Sea, China: the Baseline Atmospheric Correction (BAC), the Case 2 Regional Coast Color Atmospheric Correction (C2RCC) and its alternative version (C2RCC-AltNets), the Management Unit Mathematical Models (MUMM), the polynomial-based algorithm applied to MERIS (POLYMER), the image correction for atmospheric effects (iCOR), and ACOLITE for Dark Spectrum Fitting (ACOLITE-DSF). For OLCI-derived remote sensing reflectance, R-rs(lambda), where lambda is light wavelength, on the respective match-ups, all algorithms except ACOLITE-DSF had higher accuracy in the wavelength range from 490 nm to 560 nm, with the absolute percentage difference (APD) < 43%. On the common match-ups of all algorithms, the performance of seven AC algorithms was better than that of the respective match-ups. POLYMER obtained the best performance in visible bands, with APD less than 24.04% and 17.13% for respective and common match-ups, respectively. In terms of blue/green band ratios, R-rs(490)/R-rs(560) obtained by all algorithms except C2RCC performed better than R-rs(442)/R-rs(560). POLYMER performed best in blue/green band ratios, with the APD of 10.38% and 5.51% for R-rs(442)/R-rs(560) and R-rs(490)/R-rs(560), respectively. Using a scoring scheme based on all statistical parameters, POLYMER scored the highest. ACOLITE-DSF represented the worst accuracy for all evaluations over our in situ dataset. (C) 2022 Elsevier B.V. All rights reserved.

DOI:
10.1016/j.rsma.2022.102711

ISSN: