Article Article
Railroad Crosstie Deflection Measurement via Ultrasonic Airborne Sonar and Computer Vision Techniques

ABSTRACT A smart tie-tracking technology is proposed to measure the deflections of railroad crossties by means of non-contact ultrasonic testing in sonar mode and computer vision techniques. The sensing layout consists of an array of air-coupled capacitive transducers (in pulse-echo mode) and a high frame-rate camera, rigidly connected to the main frame of train car. The acquisition system is programmed such that the synchronized waveforms and images are collected and saved as train car moves. In the processing stage, a machine learning-based image classification approach is developed to discriminate tie/ballast images and demarcate the crossties’ boundaries. The relative deflections of the identified crossties are eventually computed by tracking the arrival time of the reflected waves from the surfaces flagged as tie. Further inspection of the deflection profiles can reveal crossties with potential poor ballast support condition. The proposed ‘tie sonar’ system was prototyped and used to reconstruct the deflection profile of the crossties scanned during a series of test runs at the Rail Defect Testing Facility of UC San Diego as well as the BNSF yard in San Diego, CA.

DOI: 10.32548/RS.2022.011

References

(1) Yu, H., 2016, “Estimating Deterioration in the Concrete Tie-Ballast Interface based on Vertical Tie Deflection Profile: A Numerical Study,” Proceedings of ASME Joint Rail Conference, April 2016, Columbia, SC, USA, JRC2016-5783, pp 1-9.

(2) Marquis, B., LeBlanc, J., Yu, H. and Jeong, D., 2014, CSX Derailment on Metro-North Tracks in Bronx, NY, July 18, 2013, Report to National Transportation Safety Board.

(3) Bay, H., Ess, A., Tuytelaars, T. and Van Gool. K., 2008, “Speeded-Up Robust Features (SURF),” Computer Vision and Image Understanding, 110(3), pp 346-359.

(4) Zhang, Y., Jin, R. and Zhou, Z. H., 2010, “Understanding Bag-of-Words Model: A Statistical Framework,” International Journal of Machine Learning and Cybernetics, 1, pp 43-52.

Metrics
Usage Shares
Total Views
125 Page Views
Total Shares
0 Tweets
125
0 PDF Downloads
0
0 Facebook Shares
Total Usage
125