The performance of tunnels and underground systems rely heavily on having a clear understanding of the physical condition of the structure and its components over time.

As the need for resilient and sustainable infrastructure becomes increasingly pressing, maintenance strategies are shifting towards more proactive and predictive management. To enable this, digitalisation of assets and inspection is playing a crucial role in accelerating this shift, allowing cost-effective, rapid and high-quality data collection, better health and safety conditions and enhanced collaboration. 

Tunnel vision

Arup has developed Loupe 360, a tunnel inspection visualisation and analytics platform. This leverages the power of 360-degree imagery and computer vision, to enable rich visualisation and asset analysis, supported by machine learning defect detection, inventorying and condition rating analytics.

Loupe 360 is being used by asset owners, maintainers and decision makers to more reliably assess the condition and behaviour of their tunnel systems, in a user-friendly web environment, away from the in-tunnel environment, allowing clients to reach better informed decisions in intervention, planning and engagement. Its utility has been demonstrated across the asset life cycle, from as-built verification assurance through to asset life extension assessment, with typical application supporting the routine and detailed visual inspection. 

Automated tunnel inspections: how it works

The 360-imagery data capture can be acquired from static collection through to fully automated data capture at speed. Inspection is supported by Arup’s deep knowledge in specification and acquisition of high-quality datasets and physical hardware build. In combination with the Loupe 360 platform, Arup’s automation of the tunnel inspection imagery capture has demonstrated significant time efficiency and data quality achievements, as well as health, safety and welfare benefits, ultimately automating and digitising the conventional manually intensive visual inspection process.