The winner of the IAG Video Contest for the period October-December 2025 is Rosa Colacicco (Italy)!

Despite major advances in Earth Observation (EO), real-time observation of river flood propagation remains limited. Satellites provide valuable but infrequent snapshots of inundation, while surface flow kinematics, the key drivers of flood energy, sediment transport, and channel reshaping, are often inferred from models rather than directly observed.
RivAIr bridges this gap through a drone-based edge computing platform for real-time hydrodynamic observation. Designed as a fully autonomous sensing node integrable within drone-in-a-box infrastructure, RivAIr uses Convolutional Neural Networks and optical flow algorithms to map river water extent and surface velocity fields on board and in real time. These measurements can reveal sudden geomorphological changes that shape flood hazard levels, eliminating ground intervention or post-processing.
Through rapid, repeated overpasses, RivAIr captures continuous data on the evolving geometry and dynamics of inundation, linking in-situ processes and satellite EO for improved flood forecasting and hazard assessment.