情報化施工(2)

[VI-1086] Labor-saving trial of road surface damage detection using images"Advanced quality control in pavement repair work using video"

〇kenji watanabe¹, Yukinobu Yanagisawa¹, Takaaki Yokoyama², Yogo Kurokawa³, Kiyotaka Suda⁴, Norio Kani⁵ (1.APPLIED TECHNOLOGY, 2.Ritsumeikan University, 3.IKEE, 4.environment climate techno, 5.Kani Construction)

Keywords:Pavement repair work, Road surface damage, Open data, Image recognition, Deep learning, Automatic detection

A learning model for image recognition by deep learning was constructed, and automatic damage detection from road surface images was attempted.

In the construction of the learning model, the road damage open data was used, the accuracy of automatic detection was increased by adding teacher data, and the degree of labor saving in the detection work was evaluated.