The Nautilos project continues to push the boundaries of ocean observation technology! A new research publication, Flying Robots Teach Floating Robots: A Machine Learning Approach for Marine Habitat Mapping Based on Combined Datasets, presents an innovative methodology for mapping marine habitats by integrating aerial and underwater data.
Developed under Nautilos for the task “Controlled scenario testing of sensors, buoy, lander, and ASV joint operations,” led by HCMR, this study combines Unmanned Aerial Vehicles (UAVs) and Autonomous Surface Vehicles (ASVs) to improve habitat classification, particularly for Posidonia oceanica meadows in the Mediterranean.
UAVs provide high-resolution optical data in shallow waters, while ASVs equipped with multibeam echosounders (MBES) capture detailed acoustic information at greater depths. By using machine learning, researchers trained the ASV’s classifier with insights from UAV imagery, achieving over 85% accuracy in habitat mapping.
This breakthrough helps overcome depth limitations in traditional aerial mapping, offering a more comprehensive and accurate approach to marine habitat assessment. It marks a significant step forward in utilizing advanced robotics and artificial intelligence for ocean monitoring.
Read the research publication here