Remote sensing data has provided synoptic information on the global ocean continuously for the past 4 decades at a spatial resolution between the meter and tens of kilometers. Its analysis is often done through very standard processing routines that, over the years, have been challenged by machine learning approaches.
Machine learning approaches have been widely used for decades over a broad range of applications such as computer vision, neurosciences, or geophysical fields. Over the last few years artificial intelligence methods, in particular deep learning, has become the dominant approach for much ongoing work in the field of machine learning. Beyond dealing with the huge quantity of remote sensing images of diverse nature, deep learning also enables a broad range of applications. Indeed, it has a great potential to explore new data-driven and learning-based methodologies and propose computationally efficient strategies able to benefit from the large amount of observational remote sensing, to finally improve the quality of oceanographic research approaches.
This Research Topic call is designed to gather a large range of uses of AI and deep learning methods applied to remote sensing in any field of ocean sciences. Original research articles covering recent research about the following topics (but not limited to) are invited for this special issue:
• Filling/dealing with spatio-temporal gaps in remote sensing datasets
• Transfer of learning approaches across satellite sensors, for instance for downscaling issues and/or sensors’ and geophysical noise removal
• Data fusion from multiple satellite sensors, for instance, to obtain high-spatial and high-spectral resolution images
• Prediction of new variables from remotely sensed data
• Reconstruction of underwater vertical profiles from surface observations
• Analysis and detection of temporal patterns and dynamical structures (e.g., frontal zones or mesoscale eddies)
• Classification issues for biogeochemical ocean processes (e.g., phytoplankton functional types, optical water types, biogeographical regions)
• Identification of connections between different ocean processes (e.g., physical vs. biogeochemical variables)
• Exploring new data-driven and learning-based methodologies for ocean-related studies from remote sensing data
• Past reconstruction, projection and forecasting of ocean essential variables time series (e.g., SST, salinity, sea-level rise, ocean color)
Authors can choose between two types of contributions containing novel and original research: 1) a full-length Original Research article (12 000 words and 15 Figures max) or 2) a Brief Research Report (4 000 words and 4 Figures max).
Dr. Ronan Fablet has two industrial collaborations:
- Research Grant from Microsoft (EU for Oceans AI award)
- Ongoing industrial partnerships with NavalGroup, Eodyn, CLS, OceanDataLab, OceanNext and Mercator Ocean Intl
All other Topic Editors declare no competing interests with regards to the Research Topic subject.