維持・管理／検査技術・診断 (8)

[VI-778] Quantification of the deep learning detection information in the dam management

*天方 匡純¹、安野 貴人¹、藤井 純一郎¹、嶋本 ゆり¹、大久保 順一¹ (1. 八千代エンジニヤリング)

Keywords:artificial intelligence, deep learning, maintenance management, dam, pop out

The social infrastructure facilities such as rivers and roads have missions to support many people’s life in our country, so they can't help being long and giant constructions. Dam is one of them. We need many resources in order to keep their structure’s function. But it’s difficult to keep original abilities of their structures in as same ways as now during populations are decreasing. So we have to apply new ICT in order to transform old ways by mainly human into rational and high productive ways by digital data. Our thesis offers the output that we extracted pop-out feature on the dam concrete through the deep learning network from photo images which we took with the drone and we quantified damage information.