Detection of Railway Track Anomalies Using Interferometric Time Series of TerraSAR-X Satellite Radar Data

Abstract

In this study, we investigate a new approach to detect railway track anomalies based on surface displacements retrieved from interferometric time series of satellite radar data. The condition of the ballast substructure is critical for the longevity of railway tracks. Early detection of problematic track sections allows for interventions to extend service life. In chord-based measurements, as routinely performed with survey trains, undulations in relative height along the track are used as a proxy for detecting track anomalies. These train-based methods for track condition assessment are costly and are limited by the availability of track recording vehicles. We developed and assessed a similar approach using spaceborne synthetic aperture radar (SAR) data to derive variations of surface deformations as a proxy for track anomalies. Utilizing time series of TerraSAR-X observations, we conducted a persistent scatterer interferometry (PSI) analysis to estimate surface deformations along railway tracks. We developed a new methodology for analyzing the obtained deformation rates and their longitudinal variability over approximately 50 km of railway tracks of the Swiss Federal Railways network. We analyzed various features and assessed their potential to identify track anomalies. Comparing our results with the track classification based on the chord-based measurements, we found a correlation of our derived PSI-based features to problematic track sections using the chord-based classifications as a reference. Our proposed approach of a satellite-based track anomaly detection system provides a useful and potentially cost-effective additional source of information to monitor railway tracks.

Keywords

• Persistent scatterer interferometry (PSI)
• railway monitoring
• synthetic aperture radar (SAR)
• TerraSAR-X