Utilization of AI-Based Diagnostic Imaging for Advanced and Efficient Tunnel Maintenance

In tunnel inspection, close visual inspection and hammering sound test are usually carried out to check the deformation and deterioration of lining concrete. These inspections are carried out using elevated work vehicles to do the close visual check and hammering test. The deformation cracks and deterioration area found during the inspection are marked on the concrete lining by chalk. These marks are then recorded as sketches. Since these inspection works need to be done by a large number of technicians and are time-consuming works, it is necessary to improve the efficiency of the works and keep the quality of the sketching to avoid differences due to the sketching skills of technicians. In order to achieve these purposes, we developed two technologies. The first one is to obtain developed diagrams of lining surface automatically from point cloud data acquired by laser scanners. The second one is to extract cracks automatically from developed diagrams using AI (artificial intelligence). We are able to improve the efficiency and quality of inspection works with these technologies.