4DM completes project to develop algorithms for near – real time flood extent and depth from RADARSAT-2 and LiDAR

4DM, in collaboration with C-Core, recently completed a project on the development of algorithms and software tools to support near real-time mapping of flooding, commissioned by Natural Resources Canada’s Emergency Geomatics Services (EGS). The objectives of the project were to (1) improve the accuracy of flood extent mapping by complementing an existing EGS open-water detection algorithm with specialized approaches for vegetated and urban areas and a GIS-based conflation post-processing, and (2) enable calculation of flood depth within the inundated area.

Within the scope of this project 4DM has developed an algorithm and a software toolset for performing conflation post-processing of remotely-sensed flood extent areas (open-water only, as well as combined with specialized vegetated/urban algorithms) to ensure the hydrologic validity of the resulting flood extent, and for determination of flood depth. 4DM also conducted research for the urban flood detection algorithm.

Preliminary results from the Richelieu and Red rivers indicate that the conflation post-processor significantly improves the accuracy of flood extent mapping in comparison to the existing open-water algorithm in areas where the flow pattern is primarily 1-D (well defined valleys) and high-resolution DEM is available. The approach currently does not perform as well in flat terrain where flow patterns are more 2-D in nature and/or DEM quality is lower.