Digital Surface Model Extraction and Re nement through Image Segmentation - Application to the ISPRS Benchmark Stereo Dataset

This paper proposes a methodology for the geometric refinement of an arbitrary DSM using high or very high resolution satellite scenes through an advanced hierarchical image segmentation and shows the results obtained on a dataset over Catalonia, Spain, provided by ISPRS WGI/4 through the project “Benchmarking and quality analysis of DEM generated from high and very high resolution optical stereo satellite data”. During this experiment a WorldView-1 quasi-nadir scene was used for the enhancement of the DSM generated by a Cartosat-1 stereopair. The original and final DSMs are compared to the Lidar DSM on the same area for quality analysis.