Advancements in Satellite SAR Methods for Monitoring Civil Infrastructure

Synthetic Aperture Radar (SAR) imaging enables large-scale, remote monitoring of civil infrastructure. SAR allows for monitoring without installation or maintenance, unlike traditional contact-type sensors. Whilst SAR interferometry (InSAR) is widely used for land monitoring, its application to bridge monitoring is still constrained by important limitations of the technique. Moreover, the use of SAR imagery for vibration-based structural health monitoring (VBSHM) remains largely unexplored in the current literature. This thesis investigates two complementary SAR-based approaches for bridge monitoring: spatio-temporal InSAR (ST-InSAR) for measuring long-term displacement and micro-Doppler SAR (MDSAR) for measuring structural vibrations. The first part of the thesis discusses the issues of conventional InSAR for bridge applications, including phase ambiguity and the use of deformation models that do not accurately represent structural behaviour. These factors degrade temporal coherence and reduce the number of reliable measurement points on bridges. A new physics-based ST-InSAR framework is introduced that accounts for the spatial correlation of persistent scatterers (PSs) on the bridge within the InSAR processing pipeline, thereby resolving InSAR issues. By applying physical constraints obtained from feasible structural deformation fields, the interpretation of InSAR phase data is enhanced. ST-InSAR is then applied to the Colle Isarco viaduct case study, where PSs with near-zero temporal coherence achieve values of 0.70–0.80, and displacement errors are reduced to levels comparable to those of topographic surveying. The second part of the thesis investigates vibration measurements from single-spaceborne SAR acquisitions, using MDSAR to estimate the line-of-sight (LOS) velocity time histories. On-site experiments on isolated targets demonstrate the successful measurement of controlled vibrations, even thoughm they are affected by significant noise. The identification of dominant frequencies is successful for velocities as low as 1 mm/s, and the accuracy of the frequency estimation depends solely on the SAR acquisition duration. Typically achievable resolutions are 0.03–0.06 Hz for acquisition durations of 20–30 seconds. Experimental results obtained on a real bridge, specifically the South Portland Street suspension bridge, are consistent with those found in the isolated target tests. Dominant frequencies are identified for independent SAR acquisitions. However, full modal extraction, including mode shapes and damping ratios, is not currently achievable with available technology but is expected to become possible with future developments in satellite SAR systems.