A new research program “Mathematical Seafloor Geomorphology” has been launched to establish a method to create a high resolution seafloor topography from the existing low resolution data by super-resolution using machine learning technology, and to merge the bottom topography pattern by feature extraction estimated for disaster prevention, environment, organisms, resources etc., aiming to perform quantitative evaluation and analysis, and to investigate the relationship between each field and bottom topography. Contribution to The Nippon Foundation – GEBCO Seabed 2030 Project is expected and will be performed under the program.
Creation of high resolution and high accuracy seafloor topography map by super resolution using machine learning method such as deep learning and Bayesian inferenceScheme II
Extraction of image feature quantity using machine learningScheme III
Collation of the image features with the data for each field to detect the underlying phenomenon that forms the seafloor topographySpecial Scheme
Contribution to The Nippon Foundation-GEBCO Seabed 2030 Project