GPR-based detection and classification of railway infrastructure using near-surface features
Noll M, Kohnert S, Caldero P (2026)
Publication Type: Journal article
Publication year: 2026
Journal
DOI: 10.1017/S1759078726103298
Abstract
This paper introduces a methodology for detecting and classifying railway infrastructure elements using ground-penetrating radar (GPR). Seven trips, covering dollar sign 57.4 dollar sign 57.4km across a German track network were recorded. The network includes ballasted tracks, switches, bridges with dominant metal elements, and overpasses. GPR data from a multi-channel sensor array is transformed through a preprocessing pipeline to generate spatial B-scans. Statistical and structural features, including energy, depth, and skewness, are extracted via an overlapping sliding window. The detection algorithm operates as a streaming process, combining correlation-based pattern matching with feature-based thresholds in a state-machine architecture. The evaluation results demonstrate overall good performance for infrastructure detection and classification (dollar sign 98 dollar sign 98-dollar sign 100 percent sign dollar sign 100%), with only minor misclassifications occurring for switch orientation estimation in scenarios involving closely spaced or physically interconnected switches (dollar sign 82 dollar sign 82-dollar sign 100 percent sign dollar sign 100%).
Authors with CRIS profile
Involved external institutions
Germany (DE)How to cite
APA:
Noll, M., Kohnert, S., & Caldero, P. (2026). GPR-based detection and classification of railway infrastructure using near-surface features. International Journal of Microwave and Wireless Technologies. https://doi.org/10.1017/S1759078726103298
MLA:
Noll, Maximilian, Sören Kohnert, and Pau Caldero. "GPR-based detection and classification of railway infrastructure using near-surface features." International Journal of Microwave and Wireless Technologies (2026).
BibTeX: Download