Machine Learning & AI
With a high volume of rich, consistent, and accessible video data captured by the full scope of our AIVR systems, our integrated Machine Learning enhances railway safety analysis, automatically detecting assets and their condition. We work closely with rail industry experts and technology partners to ensure our datasets provide vital insights for predictive rail maintenance and easier defect detection.
View automatically detected assets on the AIVR Platform
Accessing data in the AIVR Platform is enhanced through automatic asset detection features, developed by our in-house Machine Learning team.
Users can easily navigate to detected assets or receive alerts based on their condition along with their associated locational positioning on the AIVR Platform.
AIVR’s Machine Learning models can be employed across the full breadth of data on the system to detect:
- Track Components
- Thermal Hotspots
- Trackside Tensioners
- Troughing
- Graffiti
- Lineside Assets e.g. Signals, Scrap Rail, Speed Boards
- Wet Beds, Voiding and Low Ballast
- Limited Clearances
- Corona Discharge
And more!
Examples of Machine Learning & Artificial Intelligence in AIVR:
Thermal Hotspots
Thermal footage is recorded alongside Forward-Facing Video, where the readings and images are viewable concurrently in the AIVR Platform. Machine Learning models automatically pinpoint frames within the footage which exceed a pre-set temperature, allowing users to review both FFV and Thermal data to identify whether intervention is required.
Track Componentry
Line-scanning data captured by the AIVR Focus system provides an incredibly comprehensive capture of the track using underbody cameras. Machine Learning is applied to the high-definition imagery to automatically detect componentry such as Fish Plates, Fasteners, Switches and Crossings, Welds, and further points of interest. This advanced track component monitoring enables detailed assessment and issue identification for further investigation.
Graffiti
Thousands of images of lineside graffiti were annotated in rapid time to create a dataset to train a Machine Learning model. This automated graffiti detection enables targeted and effective responses to areas of vandalism and trespass, enhancing railway security measures.
Signals
Using repetitive data capture from any AIVR device in any location, signals and their ID plates are being detected and mapped. This process supports dynamic rail signal management, building a dynamic asset map and enabling engineers to navigate to any signal by its unique ID. Similar AI is being applied to detect and map OLE structure IDs.
Scrap Rail
Scrap Rail, as well as other lineside items such as ballast bags, are being detected using AI. As part of the continuous work on Track Worker Safety, hazards in the cess and four foot are identified and located on the rail map. This continuously evolving AI helps create a safer working environment and the ability to locate, remove or redeploy assets such as scrap rail.
Corona Discharge
UV Corona cameras, installed onto a measurement fleet or in-service train, enable automated data capture with real-time data transmission to capture the electricity OLE. Machine Learning models identifying occurrences of Corona within the UV data support advanced OLE performance monitoring and predictive maintenance.
Learn more about our OLE Project
Balance Weights
As part of wider OLE maintenance practices, we have used AIVR Machine Learning to automatically identify and measure balance weights along the trackside to support predictive monitoring of Overhead Wires.
Learn more about our OLE Project
Virtual Signal Sighting
Augmented Reality is used to support requirements for Design and Infrastructure Sighting, facilitating virtual inspection and planning in railway environments.
What our partners say about us
Ramboll
"We feel that we have plenty to offer on this sustainability journey and AiVR is certainly part of that ability to reduce our carbon footprint."
QTS
"Its been a great tool for me to use and has saved on having to carry out site visits all of the time and with more and more videos being added there is good choice most of the time when looking at a specific area."
Jamie Mcllvaney, Isolations Manager
QTS
"It really is a massive benefit having such a vast coverage of cab footage without the hassle of having to arrange track access and often drive hundreds of miles in the process."
Barry McLeish, Operations Director
Network Rail
"Very good equipment to have to give us a heads-up before the patrols are due as some of these faults may not have been seen for a few months until the patrols are completed. I can now plan ahead and get the faults checked and the work completed before it affects the running of any passenger train."
Section Supervisor CRE
Network Rail
"A fantastic example of use of trainborne technology to improve performance & intervene before failure, and in such a critical area; well done to the project team working on thermal imaging."
Infrastructure Director
Senior Planning Manager
"AIVR by One Big Circle has enabled our planning and site teams to gain a better understanding of site conditions and layout before attending the location due to the availability of regularly updated high quality footage."
Construction Manager
"Great solution that’s enabled early planning of work sites in a controlled space prior to going trackside - reducing exposure to the railway environment and improving site knowledge."
Principal Civil Engineer
"We find the AIVR system to be a useful and user friendly tool for both increasing our pre-site survey spatial awareness around proposed assets and obtaining a base understanding of the lie of the land for our many long linear projects."
Network Data Manager
"AIVR is a good tool for Access Point work, such as identification of track entrance Stairs and Platform ends. I have used AIVR to support already collected data with further verification confirmation."
DigiSig Rail Group
"We have found the AIVR software really useful and a great tool to assist us with our Signalling Design projects!"
Russell Simpson
GGP Consult
"GGP find the AIVR system to be a useful and user friendly tool for both increasing our pre-site survey spatial awareness around proposed assets and obtaining a base understanding of the lie of the land for our many long linear projects."
Euan Allison, Principal Civil Engineer
Kent Group
"We think it is a great tool and really helps us with our planning – we are just scratching the surface of using it “properly,” I think so expect to get more value as we go."
Robert Kent, Operations Director
Kier
Anthony Brooks, Senior Planning Manager
VolkerRail
"AIVR has helped me with signal sighting activities on a number of schemes, where alternative footage was not readily available."
Story Plant
"….this is an Ideal tool, points out everything in the site that could be a hazard, and can be used to aid the site briefings too."
Signal Sighter
"Ease of use and an extensive video library made for rapid deployment of this tool to the project teams. Gives an up-to-date window view into project environments."
Siemens
Shovlin Plant Hire
"AIVR has become an essential component of our POS planning process."
Gaby
Download
Want to show your team the capabilities of AIVR Machine Learning & AI? Download the product PDF below and share it with important stakeholders.