Underwater photogrammetry: Full Motion Video and integration of machine learning algorithms - a case study applied to MSFD seabed habitat monitoring

The Marine Strategy Framework Directive (MSFD) aims to effectively protect the marine environment across European Seas, adopting measures based on ecological ndicators, methodological standards, and monitoring programs. A specific monitoring program status is carried out by Italy within Descriptor 1 - Biodiversity (D1) aiming to characterize their environmental status for the following benthic habitats: i) Seagrass (Posidonia oceanica); ii) Reefs (Coralligenous and Cold-water corals, CWCs); and iii) Rhodoliths beds. This study presents the application of a multi-resolution and multi-scale approach for the MSFD monitoring program. A method based on machine learning algorithms is tested to improve and speed up the mapping and monitoring procedures. To this end, the Full Motion Video (FMV) technique was integrated using underwater hotogrammetry, DEM (Digital Elevation Model), highresolution multibeam bathymetry (MBES), multibeam backscatter data and ROV imaging.