AIVR Focus
What is it
AIVR Focus is One Big Circle’s line-scanning system, which captures high-definition imagery of track conditions using cameras attached to the train’s underbody.
The system automatically detects and analyses key components such as fishplates, welds, joints, switches and crossings (S&C), fasteners and railhead contamination, delivering precise geospatial positioning for remote inspection.
It makes this data available for viewing in an intuitive interface that enables users to fully explore the footage and perform measurements as well as analysing for extra insight.
The AIVR Focus underbody camera and lighting array is mounted beneath a train to capture continuous line-scan imagery of the track.
Why it matters
AIVR Focus enables track engineers to virtually verify and inspect assets, to gain further insight or arrange for maintenance work to repair any defects.
It reduces the need for track workers to conduct manual visual inspections (BVI) during night possessions or poor conditions, reducing safety risks and access constraints.
Machine learning flags defects early, enabling predictive maintenance that prevents failures, optimises schedules and cuts operational disruptions.
When: key dates
One Big Circle introduced AIVR Focus in 2023 as part of their expanding AIVR product suite, building on earlier forward-facing and thermal systems deployed since 2021. Development leveraged in-house AI expertise and rail partnerships, achieving commercial maturity by 2025 with integrations such as track geometry (AGS) and low-adhesion monitoring. The system is now deployed UK-wide on in service and measurement trains, notably Class 153 VIUs and Network Rail NMT.
Line-scan track imagery captured by AIVR Focus is reviewed on the AIVR Platform, with automatic detections flagged for review.
How it works
Underbody line-scan cameras are mounted on dedicated measurement vehicles (such as the Network Rail New Measurement Train and VIU Class 153s) and in-service passenger trains (for example, South Eastern Class 375s and LNER Class 800s). The sensors are configured in a five or seven-camera setup, with some perpendicular to the track (for overhead images) and others angled toward the gauge (inner) face of both tracks.
Controlled by AIVR Tempus, the camera captures one high resolution image line per 1mm of train movement, or up to 56,000 times per second when the train is travelling at 125mph.
Imagery streams via 4G/5G to the AIVR cloud platform, where Machine Learning models process data against Network Rail line references for sub-metre alignment. Via AIVR’s browser dashboard, engineers are able to look at all the track not just the identified issues to gain context and compare with historical views. They can also access annotated views and receive alerts of detected changes. Additional geometry and ballast data feeds help users to build a comprehensive picture of asset health.