The real-time visual measurement system of rail creeping of high-speed railway

Jingbo Xu,Xiaomeng Cui,Wenbo Ma

doi:10.1177/0020294019901303

Abstract

Changes in temperature and stress will lead to the rail creeping of high-speed railway, which becomes a hidden danger in the operation of trains. This paper studies a real-time visual measurement system for creeping displacement monitoring. The bilateral line extraction to determine the target location overcomes the influence of ambient light on image grayscale. The dynamic region of interest setting method is produced to lock and track the target. The self-calibration technology makes the system suitable for field application. The remote transmission of monitoring data is realized through narrow band internet of things (NB-IOT). These methods solve the problems in practical application. The monitoring system provides a reliable guarantee for the safe and stable operation of high-speed railway.

Highlights

Rail creeping, known as railway line creeping, is a kind of longitudinal creeping phenomenon along the railway
In view of the foregoing situation, this paper studies a rail displacement monitoring device based on machine vision
Bilateral threshold segmentation is performed in the region of interest (ROI) setting region, and the lines on both sides are extracted in the monitored target ‘‘square.’’ Through the compensation calculation, we can obtain the pixel at the center of the ‘‘square’’ and monitor its position change, that is, just the rail displacement

Summary

Introduction

Known as railway line creeping, is a kind of longitudinal creeping phenomenon along the railway. The whole image processing mainly includes image acquisition, grayscale transform, image filtering, region of interest (ROI) setting, threshold segmentation, line extraction and parameter calculation. Bilateral threshold segmentation is performed in the ROI setting region, and the lines on both sides are extracted in the monitored target ‘‘square.’’ Through the compensation calculation, we can obtain the pixel at the center of the ‘‘square’’ and monitor its position change, that is, just the rail displacement.

Results

Conclusion