In Japan, the deterioration of bridges constructed in the high economic growth period is progressing, and the maintenance of those bridges is a problem. Municipalities, as road administrators, are required to conduct close visual inspections of bridges (longer than 2m) once every five years, but municipalities face difficulties due to lack of financial and human resources. For this reason, the research of inspection and diagnosis technology is advanced for efficient inspection work. Especially, it is the new technology of ICT(information and communications technology), such as AI analysis of image data of bridge photographs.
In this study, we developed a bridge inspection support system that automatically detects cracks in concrete bridges from bridge photographs. This system uses AI of image processing by deep learning. By using AI, we will be able to detect cracks in a short time and inspect bridges more efficiently. However, it requires many photographs of huge amount of data for image analysis. And those images take time to upload to the system by mobile communication. Therefore, we verified the system operation using 5G mobile communication, which is characterized by high speed and large capacity.

Bridge Inspection System;5G;AI;Cloud;bridge;AI diagnosis