University of Twente Student Theses


Assessing the flangeway of railway switches using point clouds captured from a moving train

Ennin, Anna (2021) Assessing the flangeway of railway switches using point clouds captured from a moving train.

Link to full-text:
(only accessible for UT students and staff)
Abstract:Railway switches are the nodes of the track network and therefore must operate with high reliability. Switches are worn down by normal operations because they contain moving parts and experience tremendous forces from the trains they are directing. It is highly important to detect this wear before a train becomes derailed. However, visual assessment and manually measurement with tapes serve as the fundamental approach to obtain the parameters used to assess these switches. Mobile mapping systems have become a popular means for railway infrastructure companies to acquire data that are used for various track monitoring purposes. This research, therefore, presents the feasibility of obtaining one of the switch inspection parameters from a 3D point cloud captured with a mobile mapping system. One important assessment parameter of railway switches is the flangeway, which is the minimum distance between the stock and switch rail when the switch is opened. Finding this minimum distance rigorously requires continuous modelling for the rails. Hence, this study presents an algorithm that turns the discrete point clouds into continuous models to find the flangeway. First of all, the rails are extracted from the point clouds using segmentation. Direct measurement on the rail point clouds may result in incorrect determination of the flangeway. Therefore, we resort to determining the positions (running edge points) on the rails where the flangeway is measured from the point clouds using prior knowledge of the designed rail profile. Due to the different curvature of the rails, a B-spline curve is fitted to the running edge points to generate continuous models for the stock rail and switch rail. Finally, the models are used to compute rail-to-rail distances in a continuous manner, which allows for the extraction of the minimum distance (the flangeway). The prosed method has proven its ability to determine switch flangeway from point clouds and can be used for commercial purposes. Also, most of the methods are implemented in python and can easily be adapted by railway infrastructure management companies to determine the flangeway automatically.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Geoinformation Science and Earth Observation MSc (75014)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page