Vision-based railway inspection system
Project lead
Research topic Perception and Learning, Design and Manipulation, Modelling and Control
Industry application Assistive Robotics, Construction
Railways are an important part of the Australian transport network, with over 40,000 km of railway track currently in use across the country. Australia also uses the world’s heaviest heavy haul trains, with axle loads of up to 40 tonnes and train lengths of 2.5km or more. Inspecting and maintaining such as vast rail network is challenging, with many inspection and maintenance systems relying on costly manual operations.
This project aims to use vision-based methods for modelling and measuring railway tracks, and enhancing the accuracy of the current reconstruct results by optimising current algorithms in the MVS pipeline.