ABSTRACT
At tunnel construction sites, disasters caused by tunnel face collapses are often serious once they occur. To avoid this situation, it is necessary to detect crack on tunnel face as soon as possible before the occurrence of bed rock cave-in, and secure time for evacuate from there. We are now developing a “tunnel face monitoring supportive system” that helps crack identification on tunnel surface using a camera with AI. This system enables to identify cracks accurately and quickly while avoiding work near dangerous points and ensuring safety. An objective and quantitative comparison of monitoring functions, such as crack identification by the camera-mounted AI, with human visual inspection confirmed that the average travel distance to detect cracks can be reduced by approximately 28% with the introduction of this system. It was found that this system reduced the time required to evacuate the tunnel and assisted in efficient crack monitoring.
