Holistic strategies for control of tunnel linings

Civil infrastructures, such as bridges and tunnels, play a crucial role in our society in the flow of people and goods. Automated assessment plans are needed to control the extensive infrastructure heritage, which is subject to ageing and aggressive external phenomena. The utilisation of deep learning techniques starting from an images database associated with degradation phenomena makes such automatization possible. A classification accuracy of 90% or more has been led. The paper illustrates the construction of an effective tool to automate the detection of several defects including cracks, anomalies, and voids and its promising integration within an overall tunnel control plan including coring, ultrasound, compression, and permeability tests. The results confirm the validity of the proposed methodology to obtain a full understanding of structural conditions.