Innovative strategies to preserve the Italian engineering heritage: the historical tunnels.

Italy, a nerve center for Western culture, holds the largest number of artistic and cultural assets declared World Heritage by UNESCO. From the Romans to the present day, an ever-growing infrastructure system, rich in tunnels, bridges and viaducts, has been the expression of a high engineering expertise. For the management of the aforementioned complex infrastructure heritage, the development of automated control and maintenance plans is one of the issues on which the engineering and research community focuses its resources and efforts. In this study, an approach is proposed to automate the process of classifying defects in tunnels using deep learning techniques to protect and maintain the concrete tunnel lining. The acquisition of images from non-destructive monitoring techniques, such as Ground Penetrating Radar, within a supervised learning process allows the creation of an effective tool for the automatic detection of severe defects such as cracks, anomalies, and voids. The obtained results provided for a high degree of accuracy in identifying the tunnels’ structural condition. The use of the developed strategy, based on machine learning and non-invasive inspection techniques, is costeffective for infrastructure managers. Such a procedure reduces both the number of invasive interventions on the tunnel lining and the time and cost associated with employing specialized technicians.