Monitoring of ground displacement-induced railway anomalies using PS-InSAR techniques
Rui Tao, Albert Lau, Mats Emil Mossefin, Gefei Kong, Steinar Nordal, Yutao Pan
Abstract
• Spatio-temporal approach reduces outliers and geocoding errors in PS-InSAR data. • PS-InSAR, measurement train data, and incident logs are compared for validation. • Site visits validate detected anomalies and identify their physical causes. • PS-InSAR technique shows relatively promising alignment with other data sources. Uneven railway displacement develops after repeated train loads and long-term degradation, affecting the riding experience and operation safety. Persistent scatterer interferometric synthetic aperture radar (PS-InSAR) technique monitors ground displacement by identification of PS points over satellite-collected SAR images. Direct application of PS-InSAR technique involves multi-source measurement errors. Therefore, a spatiotemporal processing approach for PS-InSAR data is proposed to mitigate displacement measurement outliers and geocoding errors of PS points, resulting in more robust detection of railway displacement anomalies. The processed PS-InSAR data is compared with measurement train data and inspection logs of a 52-kilometer railway section, showing promising correspondence with locations found by land-based observations and logs. Within an acceptable positional deviation, 4 out of 7 InSAR-detected anomalies correspond to reported incidents, while 2 of the remaining 3 anomalies were also captured by measurement trains. Site visits to the anomalous railway segments verify the effectiveness of the proposed PS-InSAR approach.