ABSTRACT
Transportation infrastructure have a considerable impact on the economic, social and political growth of nowadays society. In particular, highway bridges, due to their inherent vulnerability, are affected by aging and degradation processes over time, which lead to a loss of the expected performances. Periodic inspection is therefore essential to ensure their safety and serviceability. In common practice, defects assessment is performed through visual inspections carried out by engineers that come close to the bridges to perform the inspection. However, these activities are laborious, require huge cost and time and have significant safety implications. Moreover, there might be a wide disparity in the quality of visual inspections depending on the experience, knowledge, and competence of inspectors. To cope with these issues, the use of drones, can allow performing visual inspections, maximizing damage identification quality and minimizing costs as well as user safety. This work illustrates an innovative and powerful methodology to perform completely remote visual inspections using drones. In particular, a tool to perform remote inspections using the digital twin of a bridge is presented, combining state-of-the art 3D reconstruction methods with AI technologies for defects detection and high-resolution 3D rendering. The tool empowers inspectors with a Web application which improves and speeds up the process of defects detection, verification and tracking through time. We demonstrate significant improvements upon precision and recall of defects detection with respect to on-site inspections even in the most challenging cases. Significant experience from different Italian bridges is shown, confirming the opportunities and challenges of performing remote inspections applying the developed methodology.
