FIG Peer Review Journal


Optical Non-Contact Railway Track Measurement with Static Terrestrial Laser Scanning to Better Than 1.5mm RMS (7493)

Anita Soni, Stuart Robson and Barry Gleeson (United Kingdom)
Miss Anita Soni
Research Engineer
University College London
Dept of CEGE, Chadwick Building
University College London
United Kingdom
Corresponding author Miss Anita Soni (email: anita.soni.10[at], tel.: +44 7921219059)

[ abstract ] [ paper ] [ handouts ]

Published on the web 2015-03-31
Received 2014-10-20 / Accepted 2015-02-07
This paper is one of selection of papers published for the FIG Working Week 2015 in Sofia, Bulgaria and has undergone the FIG Peer Review Process.

FIG Working Week 2015
ISBN 978-87-92853-35-6 ISSN 2307-4086


The railway industry requires track to be monitored for a variety of reasons, particularly when any type of physical works take place within the vicinity of the asset (e.g. demolition, construction and redevelopment works). Terrestrial laser scanning (TLS) has considerable potential as a survey method for rail measurement due to its non-contact nature and independence from physical targeting at track level. The consensus from recently published work using static terrestrial laser scanning is that rail measurements to the order of 3mm RMS are routinely possible. Such measures are appropriate for extracting the gauge, cant and twist parameters required by the rail industry, however engineering specifications designed to ensure safe and comfortable running of the trains ideally require measurements of better quality. This paper utilises standard design rail profiles from the UK industry to optimise the way in which TLS point cloud data are fitted to the rail geometry. The work is based on the use of off the shelf phase-based TLS systems each capable of delivering single point measurements of the order of 5mm to cooperative surfaces. The paper describes a workflow which focuses the fitting process onto discrete planar rail elements derived from the design rail geometry. The planar fitting process is improved through understanding how data from these scanners respond to rail surfaces. Of particular importance is the removal of noisy data from the shiny running surfaces. Results from a sequence of multi-station TLS surveys of the same set of double tracks taken from platform level highlight the capability to obtain fits to the rail model of better than 1.5mm RMS. Whilst fitting can be carried out on a single side of a rail, the paper highlights the challenge of obtaining an accurate TLS registration necessary to extract both sides of each rail to the same level of accuracy. This configuration is proven over inter-TLS instrument separations of the order of 30m and demonstrates the TLS network coverage necessary to achieve such results even in the presence of an occluding electric third rail.
Keywords: Laser scanning; Engineering survey; Deformation measurement; Railway track; Geometry