2021 Volume 2 Issue J2 Pages 749-756
Reinforcement inspection of reinforced concrete structures in bridge construction has conventionally been performed by measuring the spacing using a tape measure and visually checking the number of reinforcement. In this research, we aim to automate and upgrade the conventional inefficient reinforcement inspection using image recognition AI technology. The newly developed AI technology automatically performs reinforcement inspection based on images taken by digital cameras using UAVs or measurement vehicles. The algorithm is based on object detection using deep learning, and recognizes the arrangement of reinforcing bars by learning the features of the intersection points where the bars intersect. We conducted a demonstration test using this technology in an actual bridge construction project, and confirmed that it can accurately measure the distance between rebars and automatically count the number of rebars.
