Spatio-temporal modelling of highway deformations from Sentinel-1 SAR images by Least Squares Collocation

Highways are a vital part of the global transportation network, reducing commute times and enhancing national connectivity. Given their long-standing construction, continuous geodetic monitoring is essential for maintenance and safety. Nowadays satellite-based methods are widely employed for this purpose, with Synthetic Aperture Radar (SAR) standing out due to its extensive spatial coverage and moderate temporal resolution. These advantages mitigate some limitations of in-situ instruments. This study focuses on a highway in Slovenia. It aims to process Sentinel-1 images with the Persistent Scatterer Interferometry (PSI) technique and successively perform displacement interpolation in time and space by considering both deterministic and stochastic modelling. The former consists in an estimation of biases and trends by Least Squares adjustment, while the latter relies on the Collocation approach with a suitable covariance modelling. Available GNSS receivers provide in-situ validation of the resulting spatio-temporal deformation model, thus ensuring the reliability of the proposed approach. The developed methodology provides promising results and offers the possibility of developing a scalable solution for global highway monitoring.