Fusion strategies for high resolution urban DEM

With the development of telecommunication and the extensive use of GIS for urban planning, the need for high quality DEM is growing. Computation of urban DEM is concerned with getting a high resolution product in a wide area at the lowest cost, which are rather demanding requirements. Data fusion seems to be one way to obtain such high level data. In this paper, two approaches for the fusion process are proposed. The first one is called top down strategy and consists in the refinement of a part of the DEM of interest for the aimed application, using multi-sensor and multi-resolution data. The second one is called bottom up approach and consists in the computation of a DEM, also using multi-sensor and multi-resolution information. The data are fused in a competitive way, taking into account the desired application. Parallel with these strategies applicable to any representation of the DEM, we choose to use the Delaunay triangulation. Using this flexible modeling, we preserve the data quality and facilitate the fusion by allowing easy deletion or addition of points. It also permits the integration of constraints to structure the DEM. An application of the fusion process will consist in the computation of an urban DEM at high resolution