Scaling Up the Structural Assessment of Ageing Bridges with UAVs

The risk of bridge collapse is real and rising. Thousands of bridges in Europe and North America have outlived the loads, traffic and climate they were designed for, and most of them are never monitored. Over the past decade, geomatics has given structural engineers new tools to assess these structures: laser scanning and photogrammetry produce point clouds that can be segmented automatically into structural members, converted into BIM and finite-element models, and calibrated with vibration tests. Yet inaccessible structures still escape us. Even after dense terrestrial and mobile scanning campaigns, inner members, joints and under-deck areas remain hidden, and the geometry is rarely reliable enough to drive a structural model. To understand how a damaged steel truss bridge really behaves, we had to take it to the laboratory, where a full-scale test revealed latent resistance mechanisms that no field survey could have captured, and where hundreds of simulated damage scenarios trained AI models for damage detection and early warning. The methods exist; what is missing is the data, at the scale the problem demands. Bringing the structural assessment of ageing structures into everyday practice still faces challenges of access, repeatability, quantified damage, non-contact measurement of the structural response and cost. This talk argues that UAVs are the missing link. It reviews some of the most relevant proposals published in recent years by the drone community and discusses how, combined with the methods developed in the laboratory, they could allow structural assessment to be deployed massively on real bridges, as well as the questions of accuracy, regulation and trust that remain open before that becomes routine.