Reliable measurement of small displacements in railway infrastructure is critical for improving condition monitoring and supporting data-driven maintenance decisions. However, traditional measurement techniques typically require track access, intrusive instrumentation, or localized sensing systems that limit spatial coverage and operational feasibility on high-traffic corridors. This project proposes the development and validation of a UAV-based measurement framework capable of extracting millimeter-scale displacement information from repeated aerial imagery without the use of ground control points.
The proposed approach integrates high-resolution unmanned aerial vehicle (UAV) imagery with an optimization-based geometric correction procedure that leverages internal geometric invariants within railway track components, such as tie-plate dimensions and rail spacing. These invariants are used to correct residual anisotropic scaling and perspective distortions across image acquisitions, enabling consistent coordinate reconstruction across survey dates. The framework will be designed to support scalable monitoring of linear infrastructure by minimizing field installation requirements while maintaining metric accuracy suitable for engineering analysis.
To demonstrate the feasibility and practical value of the method, the framework will be applied to a case study quantifying longitudinal rail–crosstie relative displacement near railroad structures across several heavy axle load freight corridors. Statistical analysis will be used to examine how anchoring condition, traffic directionality, track geometry, and cumulative loading influence observed movement amplitudes.
The primary outcome of this work will be a validated methodology for extracting engineering grade measurements from UAV imagery in railway environments. The results will demonstrate the potential for UAV-based sensing to support safe, scalable, and repeatable infrastructure monitoring, with applications extending beyond rail creep measurement to broader track condition assessment and structural monitoring.