Translator Disclaimer
Presentation + Paper
12 September 2021 Monitoring of bridges by MT-InSAR and unsupervised machine learning clustering techniques
Author Affiliations +
Continuous monitoring of critical infrastructures is crucial to prevent catastrophic events such as collapse of viaducts and prioritising maintenance interventions. However, developing effective monitoring approaches must rely on the collection of a variety of information, such as the time series of structural deformations. In this context, various ground-based non-destructive testing (NDT) methods have been used in monitoring the structural integrity of transport infrastructures. However, these require routine and systematic application at the network level over long periods of time to build up a solid database of information, involving many efforts from stakeholders and asset owners in the sector. To this effect, satellite-based remote sensing techniques, such as the Multi-temporal Interferometric Synthetic Aperture Radar (MT-InSAR), have gained momentum due to the provision of accurate cumulative structural displacements in bridges. Although the application of the InSAR monitoring technique is established, this is limited to the high required time for the interpretation of the data with high spatial and temporal density. This research aims to demonstrate the viability of the MT-InSAR techniques for the structural assessment of bridges and the monitoring of damage by structural subsidence, using high-resolution SAR datasets, integrated with complementary Ground-Based (GB) information. To this purpose, high-resolution SAR dataset of the COSMO-SkyMed (CSK) mission provided by the Italian Space Agency (ASI), were acquired and processed in the framework of the ASI-Open Call approved Project “MoTiB” (ID 742). In particular, a Persistent Scatterer Interferometry (PSI) analysis is applied to identify and monitoring the structural displacements at the Rochester Bridge, in Rochester, Kent, UK. In order to explore the viability of Machine Learning algorithms in detecting critical situations in the monitoring phases, an Unsupervised ML Clustering approach, which generates homogeneous and well-separated clusters, is implemented. Each PS data-point is located to specific cluster groups, based on the deformation-trend and the values of displacements of the historical time-series. This research paves the way for the development of a novel interpretation approach relying on the integration between remote-sensing technologies and on-site surveys to improve upon current maintenance strategies for bridges and transport assets.
Conference Presentation
© (2021) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Valerio Gagliardi, Fabio Tosti, Luca Bianchini Ciampoli, Fabrizio D'Amico, Amir M. Alani, Maria Libera Battagliere, and Andrea Benedetto "Monitoring of bridges by MT-InSAR and unsupervised machine learning clustering techniques", Proc. SPIE 11863, Earth Resources and Environmental Remote Sensing/GIS Applications XII, 118630I (12 September 2021);


Synthetic aperture radar

Machine learning

Remote sensing



X band

Back to Top