Radargrammetric digital surface models generation from high resolution satellite SAR imagery: Methodology and case studies

The goal of this paper is to investigate the potential of high resolution SAR satellite imagery for DSMs generation using the radargrammetric technique. This study is methodological, devoted to illustrate both the fundamental advantages of this approach and also its drawbacks. As for photogrammetry, the achievable accuracy level of a radargrammetric generated DSM is strictly related both to the image orientation and to the image matching procedure. A rigorous orientation model based only on metadata information and an innovative matching strategy have been developed, so that a complete suite for the DSMs generation through radargrammetry has been embedded in SISAR, a scientific software developed at the Geodesy and Geomatic Division of the University of Rome “La Sapienza”. Here we discuss the results coming from two COSMO-SkyMed SpotLight stereo pairs (ascending and descending) acquired over the area of Como (Northern Italy), characterized by a mixed land cover (flat urban area, steep forested mountain slopes). Three DSMs (ascending, descending and merged) have been generated and compared with a LiDAR DSM; the accuracy of the merged product is around 7m, better than the accuracy of the ascending and descending DSMs (around 8–10 m).