Extending the Capabilities of Muography as an NDT-CE Tool using Machine Learning

The civil engineering industry faces increasing demand for innovative non-destructive evaluation methods, particularly for critical infrastructure such as bridges. Existing techniques such as ground-penetrating radar and ultrasonic echo measurements suffer from limited resolution, imaging artefacts, and reduced feature detection capabilities, especially at greater depths. Muography is an emerging non-invasive technique that constructs three-dimensional density maps by detecting the interactions of naturally occurring cosmic-ray muons within the scanned volume. Due to their high momenta, cosmic-ray muons can penetrate to depths where other techniques falter, and since the source is natural, there are no radiation safety risks involved. However, the reliance on a natural source constrains the muon flux, resulting in prolonged acquisition times and potentially noisy image reconstructions. Framing the problem as an image processing task, we have used data obtained from Geant4 Monte-Carlo simulations to perform upsampling and segmentation to optimise imaging times and improve data interpretability of concrete block volumes. This includes objects such as tendon ducts, rebar grids, and air voids with future work aiming to automate defect classification such as honeycombing, duct grouting voids and duct cable corrosion. A discussion on the design and operation of the current simulations and machine learning models for upsampling and segmentation will be provided.