Combining SAR imagery and topography for the automatic retrieval of floodwater depth maps
Résumé
This paper proposes and evaluate an innovative method for automatically retrieving floodwater depth maps by combining SAR-derived flood extent maps and topography data. The proposed algorithm can integrate as topography data either Digital Elevation Models (DEMs) or the Height Above Nearest Drainage (HAND) derived from these DEMs. The algorithm is applied using SAR images, DEMs and HAND maps of various spatial resolutions. We use as a test case the flood event that hit River Severn in the UK in July 2007. The maps obtained using our algorithms are evaluated using the results of shallow-water model simulations and aerial photographs. The results show that high level of accuracies (Root Mean Squared Deviations of Water Depth equal 0.52 cm) can be obtained if DEMs and SAR images of high resolution are used.