Deep residual network training for reinforced concrete defects intelligent classifier.

Autor: Huang, Caiping, Zhai, Kai-kai, Xie, Xin, Tan, Jinjia
Předmět:
Zdroj: European Journal of Environmental & Civil Engineering; Dec2022, Vol. 26 Issue 15, p7540-7552, 13p
Abstrakt: As artificial intelligence was more and more widely used in civil engineering, traditional detection technologies were gradually being intelligent. In order to more accurately identify the apparent diseases of concrete, in the first, this paper collected pictures of four kinds of apparent diseases of concrete, including general diseases, weathering, rebar exposed, and cracks, and used image processing technology to expand the image set; secondly, the deep residual network model were established to get the classifiers of the four apparent diseases of concrete; finally, the transfer learning was used to optimize the deep residual network model to get the best classification result. The results showed that the concrete apparent disease classifier based on deep learning established in this paper can intelligently classify the images of the single disease of concrete. After the optimization of the transfer learning, the accuracy rate reached 91.3%, and the recognition accuracy of rebar exposed diseases had reached 97.6%, which could meet the needs of intelligent detection of concrete diseases in actual projects. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje