Passive and active reflectors for InSAR applications in urban areas

In the framework of interferometric applications devoted to ground instabilities, artificial reflectors, both passive corner reflectors and active transponders, are usually designed and exploited to calibrate the interferometric measurements as well as to provide displacement measurements where natural coherent targets are missing within the area of interest. Passive devices have drawbacks related to their large size (specifically for long SAR wavelengths) and weight that make deployment problematic in difficult areas, as well as to possible loss of phase signal coherence due to geometric deformations and material degradation. Compact active transponders have been proposed as alternatives to corner reflectors, showing promising results but also some issues related to their relatively high cost and signal stability. The aim of this study is to assess the reliability and limits of corner reflectors and active transponders for supporting multi-temporal SAR interferometry (MTInSAR) in displacement measurements, with a specific focus on urban areas. An experimental calibration site consisting of two corner reflectors and three active transponders was established on the roof of the Physics Department of the Polytechnic University of Bari (PhyBA) (Southern Italy). The site is operated by Geophysical Applications Processing (GAP), in collaboration with two research institutions (CNR and PoliBA). The corner reflectors are trihedral structures with internal edge lengths of 69.5 cm (CR0) and 1.05 m (CR1). The active reflectors include a C-band Electronic Corner Reflector (ECR-C) developed by MetaSensing and two Active Radar Calibrators, a C-band unit (ARC-C) and an X-band unit (ARC-X), both developed by the Remote Sensing Group of PhyBA. The ARC devices, designed in the early 1990s for NASA/JPL airborne campaigns and the SIR-C/X-SAR mission, use military-grade Gallium Arsenide Field-Effect Transistor amplifiers and horn antennas. In this setup, ECR-C and ARC-C are tuned for the ESA Sentinel-1 SAR mission, while ARC-X is used for the ASI COSMO-SkyMED constellation. ECR-C can switch electronically between ascending and descending passes, whereas ARC-C and ARC-X require mechanical reorientation. This calibration site provides a controlled environment to test passive and active reflectors under real urban conditions and by simulating displacements at different rates and directions. Data acquired up to 2026 from Sentinel-1 and COSMO-SkyMED along both ascending and descending passes were processed through the SPINUA MTInSAR algorithm to derive time series of both SAR amplitude and displacement values. These estimations were then analysed for assessing the quality of backscattering and interferometric phase with respect to reflector size, geometry, polarisation, and distance from other reflectors and structures. Results show that measured backscatter is consistent with theoretical expectations. Reflector size proved to be a key factor: CR0, being smaller, was noisier and more affected by nearby structures, while CR1 provided more stable responses. Active reflectors offered strong amplitude signals, though their phase stability was sometimes lower than expected. Performance was also influenced by interactions among artificial reflectors as well as with other structures present on the building roof. Specific experiments were designed and performed to investigate the possibility of monitoring subpixel differential displacements by exploiting both the polarisation diversity and the time diversity induced by a delay line. Since January 2024, ARC-C has been operated in VH polarisation, while CR0, which responds in VV polarisation, has been positioned at about 1 m from ARC-C and occasionally moved upward up to 1 cm. By processing both VH and VV Sentinel-1 datasets, it was possible to identify the two reflectors and correctly measure the line-of-sight component of the induced CR0 displacements. Similar results were obtained by setting ARC-C in VV polarisation and by delaying its response through a 30 m long coaxial cable calibrated for C-band. The delayed ARC-C signals were imaged a few pixels away from the CR0 in the VV Sentinel-1 data, thus allowing us to analyse separately the two targets and thus distinguish their behaviour in time. These configurations allowed us to test the measurement of the differential displacement between two point-like targets located within the same Sentinel-1 resolution cell, opening opportunities for monitoring closely spaced scatterers and for exploiting the VH polarisation, which has so far been underused in MTInSAR applications. Moreover, since in urban settings the precise geocoding of the structures becomes pivotal, we assessed the performance of geocoding of Sentinel-1 data in relation to the increase of the Sentinel-1A orbital tube from mid-April 2024, which improved the precision of target elevation estimation through MTInSAR. Further work focused on innovative reflector geometries and low-cost transponders. Specifically, a hexagon-periodic corner reflector with a side length of 46 cm was designed to achieve a more compact and less visually intrusive structure, suitable for architecturally sensitive environments. A prototype was built and tested after 3D electromagnetic simulations. The reflector showed good performance at C-band, but it exhibited a highly directive response. Despite this drawback, the design remains promising for applications where aesthetics is a primary concern. In addition, a prototypal low-cost transponder was designed and built up by PoliBA. Experimental tests have so far been performed using both Sentinel-1 C-band data and COSMO-SkyMED X-band data. There is, however, a growing interest in extending these investigations to the L-band, where transponders could provide a major advantage compared to passive corner reflectors in terms of physical size. Moreover, L-band is expected to ease the preservation of transponder phase stability due to the lower frequency with respect to C/X bands. In the future, the proposed calibration site may become an open experimental facility, providing data available for testing and validating interferometric techniques.

This work was supported in part by ESA under the project “Quantum computing for ground motion measurement.”, ESA/RFQ/3-17708/22/I-DT-lr. The authors also wish to acknowledge all researchers and professionals involved in research activities related to the Experimental SAR Calibration Site of Bari (Italy) under self-financed projects.