Anomaly Detection in Railway Bridges Using InSAR and In-Situ Monitoring Data

Monitoring aging bridges is essential for transport network safety. Traditional Structural Health Monitoring (SHM) relies on contact sensors, which face high costs and sparse spatial coverage. Satellite-based Interferometric Synthetic Aperture Radar (InSAR) offers a promising complementary technology, providing spatially distributed displacement measurements over vast areas. This thesis develops a multi-source integration framework to evaluate InSAR's operational reliability for bridge monitoring. The methodology fuses satellite observations, in-situ sensor data, and environmental variables, using the permanently monitored "Viadotto Km 191,400_152" railway bridge as a case study. Persistent Scatterers (PS) from Ascending and Descending Sentinel-1 EGMS datasets are spatially matched to the bridge using a Haversine-distance procedure. Satellite Line-of-Sight (LOS) data is geometrically decomposed into vertical and east-west displacements for direct comparison with in-situ sensors. ERA5-Land variables are incorporated to account for meteorological influences. To distinguish structural anomalies from environmental kinematics, a multivariate statistical approach utilizing the Mahalanobis distance is implemented. This formulates an Environmental Compensation Damage Index (ECDI), isolating abnormal structural behavior from natural, thermally induced deformations. Results show that despite higher noise than contact sensors, InSAR successfully captures dominant low-frequency structural trends, particularly seasonal thermal cycles. Findings confirm that environmentally compensated InSAR data effectively complements traditional SHM, enabling scalable, cost-efficient, and sensor-light infrastructure monitoring.