土木分野におけるAIの活用(5)

[CS15-39] Verification of image recognition AI system for rebar arrangement inspection

〇Takumi Kakiichi¹, Takashi Nakano¹, Shintaro Arai¹, Koichiro Tamura², Hiromi Nakagawa², Keita Sumiki² (1.JFE Engineering, 2.ACES)

Keywords:inspection of rebar arrangement, AI, DeepLearaning, drone, save labor

In the inspection of rebar arrangement of RC slab or composite slab, several days are required for the measurement because workers perform the measurement manually in the range determined by specifications, and the measurement value outside the range is unknown. In this paper, a new inspection system using image recognition AI system was applied and verified for steel plate-concrete composite slab to save labor in the measurement of rebar arrangement and to improve the inspection quality. As a result, it was confirmed that the measurement accuracy was within 5mm compared to conventional manual measurement, and the work time was reduced to 35% of the conventional work time.