D 2.1 - Benchmarking for AVRIS and definition for solutions for track assets

The automatic visual inspection is a process involving the use of inspection trains with onboard systems digital video, to capture information of the track and its surroundings and an analytical part using pattern recognition tools, in real-time and during post-processing.

It can be used mainly for maintenance and/or safety processes.

These identification and classification tools may be used based on Inspection Activity, normally neural networks, either through EML or deep learning.

The automatic visual inspection of the S&Cs is a specific part of the surveillance of the track and of its surroundings. Its main objective is to verify that the safety of rail traffic is assured.

The European Strategy “Challenge 2050” considers that the use of technology in railway infrastructure should have a targeted investment in the assets of this system, where with the trains, system automation and well-motivated staff, shall all provide a picture of a reliable and high-capacity railway system.

Also, the Rail Technical Strategy Europe (RTSE) mentions that both the “Internet of Things” (IoT) – a smart network of inter-communicating assets in combination with the use of advanced monitoring, should give as results of the analytics of the adequate asset management tools a comparison of maintenance and/or replacement strategies. This comparison should be considering track and infrastructure based on traffic levels and whole life evaluation, and thus, aiming at the optimisation of the strategy and overall cost approach.

However, the construction of an updated inventory of assets, deeper that the required Registry of Infrastructure by the European Union Agency for Railways, remains still as a challenge to complete for the needed implementation for both monitoring and analytics on condition.