TY - GEN
T1 - Analysis of Machine Learning for Detect Concrete Crack Depths Using Infrared Thermography Technique
AU - Jang, Arum
AU - Kim, Jihyung
AU - Park, Min Jae
AU - Ju, Young K.
AU - Kim, Sung Jig
N1 - Funding Information:
This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. NRF - 2020R1A2C3005687, NRF -2021R1A5A1032433, and NRF-2020R1A2C1013287) .
Publisher Copyright:
© 2022 IABSE Symposium Prague, 2022: Challenges for Existing and Oncoming Structures - Report. All rights reserved.
PY - 2022
Y1 - 2022
N2 - Recently, much research with high-tech technology is being conducted in building inspection. In previous studies, thermography technology quickly and accurately inspected the concrete crack defects, and several machine learning models can reliably predict the crack depths. In this study, the most proper model would be proposed according to the concrete crack by evaluating the adaptability of the seven machine learning models. The models also predicted the crack depths, and the data were applied to each machine learning considering concrete temperature and external parameters. In machine learning, less critical features were ignored by filtering existing data to find useful features related to crack depths. Machine learning models are evaluated, and the structures of the models were investigated to determine the feature importance and part dependence. Those enabled us to decide the most proper machine learning according to the cracks.
AB - Recently, much research with high-tech technology is being conducted in building inspection. In previous studies, thermography technology quickly and accurately inspected the concrete crack defects, and several machine learning models can reliably predict the crack depths. In this study, the most proper model would be proposed according to the concrete crack by evaluating the adaptability of the seven machine learning models. The models also predicted the crack depths, and the data were applied to each machine learning considering concrete temperature and external parameters. In machine learning, less critical features were ignored by filtering existing data to find useful features related to crack depths. Machine learning models are evaluated, and the structures of the models were investigated to determine the feature importance and part dependence. Those enabled us to decide the most proper machine learning according to the cracks.
KW - Building inspection
KW - Crack depth prediction
KW - Infrared thermography technique
KW - Machine learning
KW - Thermal camera
UR - http://www.scopus.com/inward/record.url?scp=85133522304&partnerID=8YFLogxK
M3 - Conference contribution
AN - SCOPUS:85133522304
T3 - IABSE Symposium Prague, 2022: Challenges for Existing and Oncoming Structures - Report
SP - 758
EP - 765
BT - IABSE Symposium Prague, 2022
PB - International Association for Bridge and Structural Engineering (IABSE)
T2 - IABSE Symposium Prague 2022: Challenges for Existing and Oncoming Structures
Y2 - 25 May 2022 through 27 May 2022
ER -