Determination of Railway Track Gauge with Image Processing

Abstract

With the increase in the use of rail transportation systems, rail controls have gained importance. Especially in recent years, rail damage detection methods based on image processing have been developed because it produces fast and accurate results instead of rail controls with traditional methods. In this study, the damages that may occur as a result of physical or environmental effects in the railway track span were examined. For this, the railway line was monitored with a drone moving at a fixed height. In order to obtain clearer images, vibrations in the recorded video were removed and frames were created. The created frames were preprocessed using MATLAB and the distance between the rails was measured and the possible expansion and contraction errors on the rails were examined. In the developed method, a healthier measurement was achieved by taking images from a certain height using the drone and improving the vibrations in the images. The developed method was tested on images obtained in sunny and normal daylight, and it was shown in the experimental results section that it produced high-fidelity results. © 2021 IEEE.