Radargrammetric generation of DEMs from high resolution satellite SAR imagery. A new tool for landslide hazard and vulnerability assessment

Spatial information acquisition and analysis tools play a fundamental role to supplying the information necessary to produce landslide inventories, which represent the foundations for quantifying landslide hazard and vulnerability. In this frame, fundamental data are Digital Surface Models (DSMs) and Digital Elevation Models (DEMs). The goal of this paper is just methodological, focused to illustrate the actual potentialities of DSMs generation from high resolution satellite Synthetic Aperture Radar (SAR) imagery with a radargrammetric stereo-mapping approach. The fundamental advantage of this approach is that it can work with just a couple of images (no matter for their coherence), which can be collected in a short time (half day to quite few days) thanks to the independence of satellite radar acquisition from weather (clouds), daylight and logistic constraints (as for airborne data collection). The suite for the DSMs generation through the radargrammetric approach has been implemented in SISAR (Software per Immagini Satellitari ad Alta Risoluzione), a scientific software developed at the Geodesy and Geomatic Institute of the University of Rome "La Sapienza". In order to demonstrate the radargrammetric mapping potentialities of high resolution SAR data, a test site was established in the area of Merano (Northern Italy), characterized by mixed morphology and land cover. The data available for the experiment were a COSMO-SkyMed SpotLight stereo pair and a LiDAR DEM, used as ground truth. An accuracy better than 3 m has been achieved in open areas and the implemented algorithm appears able to generate DSMs both over open and forested areas, where the accuracy is around 4 m. Therefore, radargrammetric generation of DSMs from high resolution satellite SAR imagery appears a valuable tool to supply topographic information for landslide inventories at different scales.