Fluvial remote sensing of river bathymetry is crucial for characterizing the topography of the riverbed and monitor changes in habitat at large scales. Hyperspectral data enables bathymetric retrieval through optical models. On the Ain River (France), multiple hyperspectral aerial campaigns with different sensors were conducted and processed to create bathymetric maps of the river for different flow conditions. In particular, a continuous bathymetric map was produced for a 20 km reach of the river with a median error of 20 centimetres for depths up to 2.5 metres. Despite the uncertainties of the models tested, the results are more robust spatially and over a wider range of depth and flow conditions than optical models based on traditional colour imagery.