Introducing a model for evaluating concrete structure performance using deep convolutional neural network

Sang Hyo Lee, Ha Young Kim, Hyun Kyu Shin, Youjin Jang, Yong Han Ahn

Research output: Contribution to journalArticle

Abstract

A material or environmental factor can decrease the durability of a concrete structure. If the decreased durability is not caused by environmentally induced deterioration, it may signify a problem with the material itself, which is a serious issue affecting the entire structure. The evaluation and prediction of the durability of a concrete structure is considered a critical research area. Existing evaluation and prediction methods using non-destructive inspection or core sampling have limitations. Some studies applied image processing techniques to overcome these shortcomings; nonetheless, a problem remains in analyzing the obtained images. To solve this problem, a performance evaluation system model for concrete structures is proposed. The model is based on a deep convolutional neural network (DCNN). This model is intended to establish a foundation for solving difficult problems in the construction industry. The proposed concrete performance evaluation model defines a concrete surface image as a parameter of input data and uses a DCNN-based machine learning algorithm to produce the performance evaluation results as output data. This model consists of a database construction phase, a DCNN-based algorithm design phase, and an algorithm implementation phase. The complete development of this model is expected to resolve many problems of structure maintenance systems, which are the products of socio-environmental changes in the Republic of Korea. Furthermore, various data collected during the whole life cycle of a structure can be utilized to apply the proposed deep learning technique to various construction areas.

Original languageEnglish
Pages (from-to)285-295
Number of pages11
JournalInternational Journal of Sustainable Building Technology and Urban Development
Volume8
Issue number3
DOIs
Publication statusPublished - 2017 Jan 1

Fingerprint

Concrete construction
Neural networks
Durability
Concretes
Construction industry
Learning algorithms
Deterioration
Learning systems
Life cycle
Image processing
Inspection
Sampling

All Science Journal Classification (ASJC) codes

  • Building and Construction

Cite this

@article{94d63f12a56e40cb89e92518c2be2b23,
title = "Introducing a model for evaluating concrete structure performance using deep convolutional neural network",
abstract = "A material or environmental factor can decrease the durability of a concrete structure. If the decreased durability is not caused by environmentally induced deterioration, it may signify a problem with the material itself, which is a serious issue affecting the entire structure. The evaluation and prediction of the durability of a concrete structure is considered a critical research area. Existing evaluation and prediction methods using non-destructive inspection or core sampling have limitations. Some studies applied image processing techniques to overcome these shortcomings; nonetheless, a problem remains in analyzing the obtained images. To solve this problem, a performance evaluation system model for concrete structures is proposed. The model is based on a deep convolutional neural network (DCNN). This model is intended to establish a foundation for solving difficult problems in the construction industry. The proposed concrete performance evaluation model defines a concrete surface image as a parameter of input data and uses a DCNN-based machine learning algorithm to produce the performance evaluation results as output data. This model consists of a database construction phase, a DCNN-based algorithm design phase, and an algorithm implementation phase. The complete development of this model is expected to resolve many problems of structure maintenance systems, which are the products of socio-environmental changes in the Republic of Korea. Furthermore, various data collected during the whole life cycle of a structure can be utilized to apply the proposed deep learning technique to various construction areas.",
author = "Lee, {Sang Hyo} and Kim, {Ha Young} and Shin, {Hyun Kyu} and Youjin Jang and Ahn, {Yong Han}",
year = "2017",
month = "1",
day = "1",
doi = "10.12972/susb.20170027",
language = "English",
volume = "8",
pages = "285--295",
journal = "International Journal of Sustainable Building Technology and Urban Development",
issn = "2093-761X",
publisher = "Taylor and Francis Ltd.",
number = "3",

}

Introducing a model for evaluating concrete structure performance using deep convolutional neural network. / Lee, Sang Hyo; Kim, Ha Young; Shin, Hyun Kyu; Jang, Youjin; Ahn, Yong Han.

In: International Journal of Sustainable Building Technology and Urban Development, Vol. 8, No. 3, 01.01.2017, p. 285-295.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Introducing a model for evaluating concrete structure performance using deep convolutional neural network

AU - Lee, Sang Hyo

AU - Kim, Ha Young

AU - Shin, Hyun Kyu

AU - Jang, Youjin

AU - Ahn, Yong Han

PY - 2017/1/1

Y1 - 2017/1/1

N2 - A material or environmental factor can decrease the durability of a concrete structure. If the decreased durability is not caused by environmentally induced deterioration, it may signify a problem with the material itself, which is a serious issue affecting the entire structure. The evaluation and prediction of the durability of a concrete structure is considered a critical research area. Existing evaluation and prediction methods using non-destructive inspection or core sampling have limitations. Some studies applied image processing techniques to overcome these shortcomings; nonetheless, a problem remains in analyzing the obtained images. To solve this problem, a performance evaluation system model for concrete structures is proposed. The model is based on a deep convolutional neural network (DCNN). This model is intended to establish a foundation for solving difficult problems in the construction industry. The proposed concrete performance evaluation model defines a concrete surface image as a parameter of input data and uses a DCNN-based machine learning algorithm to produce the performance evaluation results as output data. This model consists of a database construction phase, a DCNN-based algorithm design phase, and an algorithm implementation phase. The complete development of this model is expected to resolve many problems of structure maintenance systems, which are the products of socio-environmental changes in the Republic of Korea. Furthermore, various data collected during the whole life cycle of a structure can be utilized to apply the proposed deep learning technique to various construction areas.

AB - A material or environmental factor can decrease the durability of a concrete structure. If the decreased durability is not caused by environmentally induced deterioration, it may signify a problem with the material itself, which is a serious issue affecting the entire structure. The evaluation and prediction of the durability of a concrete structure is considered a critical research area. Existing evaluation and prediction methods using non-destructive inspection or core sampling have limitations. Some studies applied image processing techniques to overcome these shortcomings; nonetheless, a problem remains in analyzing the obtained images. To solve this problem, a performance evaluation system model for concrete structures is proposed. The model is based on a deep convolutional neural network (DCNN). This model is intended to establish a foundation for solving difficult problems in the construction industry. The proposed concrete performance evaluation model defines a concrete surface image as a parameter of input data and uses a DCNN-based machine learning algorithm to produce the performance evaluation results as output data. This model consists of a database construction phase, a DCNN-based algorithm design phase, and an algorithm implementation phase. The complete development of this model is expected to resolve many problems of structure maintenance systems, which are the products of socio-environmental changes in the Republic of Korea. Furthermore, various data collected during the whole life cycle of a structure can be utilized to apply the proposed deep learning technique to various construction areas.

UR - http://www.scopus.com/inward/record.url?scp=85065577157&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065577157&partnerID=8YFLogxK

U2 - 10.12972/susb.20170027

DO - 10.12972/susb.20170027

M3 - Article

AN - SCOPUS:85065577157

VL - 8

SP - 285

EP - 295

JO - International Journal of Sustainable Building Technology and Urban Development

JF - International Journal of Sustainable Building Technology and Urban Development

SN - 2093-761X

IS - 3

ER -