End-to-End Physics-Informed Representation Learning from and for Satellite Ocean Remote Sensing Data : Applications to Satellite Altimetry and sea Surface Currents
|Author(s)||Fablet Ronan1, Amar Mohamed Mahmoud1, Febvre Quentin1, Beauchamp Maxime1, Chapron Bertrand2|
|Affiliation(s)||1 : IMT Atlantique, UMR CNRS Lab-STICC, Brest, France.
2 : Ifremer, UMR CNRS LOPS, Brest, France
|Meeting||XXIVth ISPRS Congress - Intenational Society for Photogrammetry and Remote Sensing Congress. 31 August - 02 September 2020, Nice, France (virtual event) , 05 - 09 July 2021, Nice, France (virtual event) , 06 - 11 June 2022, Nice, France|
|Source||ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences (2194-9050) (Copernicus GmbH), 2021-06-17 , Vol. V-3-2021 , P. 295-302|
|Keyword(s)||Space oceanography, sea surface dynamics, satellite altimetry, SWOT mission, end-to-end learning, inverse problems, data assimilation, space-time interpolation, short-term forecasting, adaptive sampling.|
This paper addresses physics-informed deep learning schemes for satellite ocean remote sensing data. Such observation datasets are characterized by the irregular space-time sampling of the ocean surface due to sensors’ characteristics and satellite orbits. With a focus on satellite altimetry, we show that end-to-end learning schemes based on variational formulations provide new means to explore and exploit such observation datasets. Through Observing System Simulation Experiments (OSSE) using numerical ocean simulations and real nadir and wide-swath altimeter sampling patterns, we demonstrate their relevance w.r.t. state-of-the-art and operational methods for space-time interpolation and short-term forecasting issues. We also stress and discuss how they could contribute to the design and calibration of ocean observing systems.