Fast terrain modelling for hydrogeological risk mapping and emergency management: the contribution of high-resolution satellite SAR imagery

Abstract

Geomatic tools fast terrain modelling play a relevant role in hydrogeological risk mapping and emergency management. Given their complete independence from logistic constraints on the ground (as for airborne data collection), illumination (daylight), and weather (clouds) conditions, synthetic aperture radar (SAR) satellite systems may provide important contributions in terms of digital surface models (DSMs) and digital elevation models (DEMs). For this work we focused on the potential of high-resolution SAR satellite imagery for DSM generation using an interferometric (InSAR) technique and using a revitalized radargrammetric stereomapping approach. The goal of this work was just methodological. Our goal was to illustrate both the fundamental advantages and drawbacks of the radargrammetric approach with respect to the InSAR technique for DSM generation, and to outline their possible joint role in hydrogeological risk mapping and emergency management. Here, it is worth mentioning that radargrammetry procedures are independent of image coherence (unlike the interferometric approach) and phase unwrapping, as well as of parsimony (only a few images are necessary). Therefore, a short time is required for image collection (from tens of minutes to a few hours), thanks to the independence from illumination and weather. The most relevant obstacles of the technique are speckle and the lack of texture impact on image matching, as well as the well-known deformations of SAR imagery (layover and foreshortening), which may produce remarkable difficulties with complex morphologies and that must be accounted for during acquisition planning. Here, we discuss results obtained with InSAR and radargrammetry applied to a COSMO-SkyMed SpotLight triplet (two stereopairs suited for radargrammetry and InSAR, sharing one common image) acquired over suburbs of San Francisco (United States), which are characterized by mixed morphology and land cover. We mainly focused on urban areas and zones covered by bare soil and rocks. Image processing was performed using the well-known commercial software SARscape® for InSAR, and the radargrammetric suite implemented in SISAR, software developed at the Geodesy and Geomatic Division of the University of Rome “La Sapienza”. Global accuracies were approximately 5 m using both approaches. However, several differences in terrain morphology reconstruction were determined and are underlined and evaluated here, as well as a possible way to further enhance the results using the integration of InSAR and radargrammetry.