Railway Infrastructure Classification and Instability Identification Using Sentinel-1 SAR and Laser Scanning Data

Area of Science:

• Geodesy and Remote Sensing
• Civil Engineering
• Infrastructure Monitoring

Background:

• Satellite radar interferometry (InSAR) is valuable for monitoring linear infrastructure like railways.
• Limitations in Sentinel-1 SAR imagery resolution and geolocation hinder precise scatterer identification and instability detection.
• Existing methods struggle to accurately associate radar scatterers with specific railway components and identify unstable segments from single-geometry Line of Sight (LOS) data.

Purpose of the Study:

• To improve the 3-D geolocation accuracy of Sentinel-1 derived persistent scatterers (PS) for railway infrastructure.
• To classify identified PS into distinct railway components (rails, embankments, surroundings).
• To develop a method for identifying unstable railway segments using decomposed LOS deformation, focusing on vertical rail settlement.

Main Methods:

• A two-step method was employed to enhance the 3-D geolocation of Sentinel-1 PS, incorporating laser scanning data.
• Classified railway infrastructure components based on improved PS geolocation.
• Decomposed LOS deformation to derive near-vertical rail settlement for instability detection.
• Utilized 170 Sentinel-1a/b ascending datasets from January 2017 to December 2019.

Main Results:

• Achieved 98% association of PS with real objects at a 25% significance level after geolocation improvement.
• PS settlement measurements showed good agreement with in-situ Rail Infrastructure aLignment Acquisition (RILA) data.
• The average standard deviation of PS settlement measurements was 6.16 mm, indicating reliable precision.

Conclusions:

• The enhanced PS geolocation and classification method significantly improves the reliability of InSAR for railway structural health monitoring.
• The technique effectively identifies unstable railway segments by analyzing decomposed vertical rail settlement.
• The findings demonstrate the potential of InSAR for precise and large-scale monitoring of railway infrastructure health.