A Study on Hammering Test using Deep Learning
Tsubasa Fukumura, Hayato Aratame, Atsushi Ito, Masafumi Koike, Katsuhiko Hibino, Yoshihisa Kawamura · 2020
Technical infrastructure such as bridges and tunnels are critical in economic activities, but in recent years, aging of them has progressed, and demand for checking the soundness of them is expanding. There are several technologies to inspect the aging process of the structures to find cracks on a wall or flacking inside of concrete. Among the technologies to diagnose a structure, the hammering test is a traditional, easy, simple, and effective method. We developed an AI hammering checker by using the k-mean method. However, accuracy is not good in some tasks. So, we started to develop the new version of the AI hammering checker by using deep learning. We explain the outline of the new system in this paper. As a result, the accuracy was increased up to 93.9%.