PSI - Issue 84

Luciano Pavesi et al. / Procedia Structural Integrity 84 (2026) 583–590

587

of the processing chain, a DEM can be generated, with each pixel, whose spatial resolution depends on the specific satellite mission, assigned an elevation value. However, the accuracy of the resulting measurements depends on the availability of in-situ control points, which may be incorporated as constraints during the validation phase. Once all procedures have been completed, an accuracy assessment can be carried out to evaluate the quality of the derived elevations. At the current stage of the research, two COSMO-SkyMed Stripmap-mode SAR images (HH polarization, Ascending geometry) (Italian Space Agency, 2016; Tapete et al., 2021), acquired on 2025-01-08 and 2024-11-05, have been processed by InSAR. A CSK DEM with a spatial resolution of 3m has been generated and compared with the Shuttle Radar Topography Mission (SRTM) v3 at 3 arc-seconds (Farr et al., 2007), which has a spatial resolution of 90 m. The results are presented in the following Fig. 2, including the elevation profile along the A– A′ transect, which intersects the case study areas and the Arrone River (Italy).

Fig .2. Profile extracted along A- A’ section (a), and DEM comparison (b), SRTM3 v 3 DEM (90m) v ersus COSMO-SkyMed InSAR DEM (3m) - COSMO-SkyMed ® Products ©ASI, Italian Space Agency, 2025. All rights reserved Starting from high-resolution Digital Elevation Models derived from COSMO-SkyMed satellite imagery, RESCUE_SAT enables the identification of the locations of structures and infrastructure through the application of detection methodologies (Fig. 1b). The high spatial detail of the satellite-derived DEMs allows morphological discontinuities associated with artificial flow confinement to be distinguished from natural terrain features, supporting the systematic mapping of, e.g., levees across the river network. Levee detection techniques are based on the analysis of different combinations of morphological parameters, such as relative elevation, slope, aspect difference and profile curvature (Wing et al., 2019; Sasaki et al., 2023; Khanh et al., 2025) and on an image-processing algorithm, which reduces discontinuities in the levee horizontal alignment. The detailed procedure for levee detection is described in Khanh et al. (2025); it can be validated against other elevation sources (e.g., LiDAR-derived DEMs) where locally available. Fig. 3 shows its application to a reach of the Tiber River located downstream of the city of Rome. Specifically, Fig. 3 illustrates the morphological factors required for levee detection (evaluated by a high-resolution DEM) together with the final result (upper right and lower right panels of the figure). Notably, the method allows levee locations and elevation profile to be easily assessed using only a high resolution DEM, which represents a significant advantage with respect to the requirements of the ANSFISA guidelines; we recall that they require determining whether a channel is embanked in order to apply the appropriate assessment procedure. Similarly, the detection of transport infrastructures enables the localization of bridges and the characterization of their essential hydraulic attributes, including position, orientation and degree of interference with the river section.