[4Q3-J-13] AI application: sensing of artifacts

[4Q3-J-13-02] Image-based AI system for evaluating surface-treated-concrete

〇Michio Imai¹, Manato Nakamura¹, Shuji Matsumoto¹, Sohei Arisaka¹, Kenichiro Tsuyuki¹, Takekazu Matsukawa¹, Shuji Yanai¹, Chaninthorn Amornsawaddirak (1. Kajima Corporation)

Keywords:concrete, construction quality, image processing

In civil engineering structures under construction, quality evaluation of the surface-treated-concrete, which is a joint between fixed and fresh concrete, is critical for structural integrity. Although the authors had already developed an evaluation method based on a digital camera histogram of the targeted concrete, the method has a difficulty with dealing with a wide range of conditions. In this study, an AI-based method has been applied to this image processing application, and then advantages of deep learning has been successfully confirmed.