Journal of the Architectural Institute of Korea (대한건축학회논문집)

Architectural Institute of Korea (대한건축학회)
권호 표지
DOI QR코드

DOI QR Code

Analyzing Deep Learning-based Techniques in Concrete Crack Detection Technology

딥러닝 기반 콘크리트 균열 검출 기술에 관한 연구

  • Kim, Ki-Woong (Dept. of Architectural Engineering, Daejin University) ;
  • Yoo, Moo-Young (Dept. of Architectural Engineering, Daejin University)
  • Received : 2023.10.24
  • Accepted : 2024.02.20
  • Published : 2024.03.30
⟨ Previous Next ⟩

Abstract

When buildings deteriorate, they may develop defects like surface cracks and structural subsidence. If left unaddressed, these issues can significantly weaken the structure, potentially leading to collapse accidents. Detecting cracks promptly is crucial to prevent such outcomes. With the advancements in artificial intelligence, researchers are exploring deep learning techniques to identify microscopic cracks, replacing traditional manual methods. As AI technology progresses, diverse AI models have emerged, enhancing the reliability of crack detection data for field inspections. This study focuses on leveraging the Yolo model, known for its superior performance and faster data acquisition compared to other AI models. By incorporating object detection methods used by CNN, the study aims to enhance the detection performance of the model by considering various variables across different AI models and detection techniques.

Keywords

Acknowledgement

본 연구는 국토교통부 국토교통 DNA플러스 융합기술대학원 육성사업의 연구비 지원에 의해 수행되었습니다. 과제번호: RS-2023-00250434

References

  1. Construction site inspection manual (2013). Daejeon: Construction Management Office, Daejeon Regional Land Management Agency. 
  2. Do, Y. S. (2020). "Prediction of concrete cracks using machine learning." Master's thesis in Korea Hanyang University Graduate School of Engineering, Seoul. 
  3. Do, Y, S. (2020). "Predicting concrete cracks using machine learning." Graduate School of Engineering, Hanyang University, Seoul. 
  4. Ham, S, W., Bae, S. H., Lee, I. P., Lee, G. P., & Kim, D. G. (2022). "How to evaluate the deep learning model of concrete lining crack division." Journal of the Korean Tunnel Underground Space Association 24.6 : 513-524. 
  5. Kang, H. G., & Hong, S. G. (2008). "causes and mitigations of concrete cracks." Journal of the Concrete Society 20.5, 61-68. 
  6. Kang, K. S., & Ryu, H. G. (2023). "Deep learning - based method of detecting cracks in pixel unit concrete wall." Journal of the Korean Society of Architecture and Construction 23.2, 197-207. 
  7. Kim, B. H., Kim, G. S., Jin, S. M., & Jo, S. J. (2019). "A Comparative Study on Performance of Deep Learning Models for Vision-based Concrete Crack Detection according to Model Types." Journal of the Korean Society of Safety 34.6, 50-57.  https://doi.org/10.14346/JKOSOS.2019.34.6.50
  8. Kim, J. H., Kim, K. Y., & Kim, D. J. (2023). "Domain adaptive crack detection system based on deep learning for building inspection." Korea Telecommunications Association(The Journal of the KICS), 48.5, 567-580. 
  9. Kim, J. S. (2023). "Comparison of the concrete crack detection performance of segmentation and object detection models." Journal of the Korean Society of Industry and Technology 24.4, 31-41.  https://doi.org/10.31159/ksmrt.2023.33.2.31
  10. Kim, S. D. (2020). "A Study on the Detection of Fire in a Ship Engine Room Using the YOLO Algorithm." Korea Master's thesis Mokpo National Maritime University, South Jeolla Province. 
  11. Kim, S. H. (2022). "A Study on the Improvement of Prosthetic Image Detection by U-NET Model." Graduate School of Yeungnam University, Gyeongsangbuk-do. 
  12. Kim, S. H. (2022). "Study on improving orthoimage detection accuracy using U-NET model." Domestic doctoral thesis, Yeungnam University Graduate School. 
  13. Penalty point management standards for construction work, etc. (related to Article 87 (5)) 
  14. Road pavement crack detection using deep learning with synthetic data - Scientific Figure on ResearchGate. Available (2023). from: https://www.researchgate.net/figure/VGG16-U-Net-architecture_fig2_348661034 [accessed 12 Jul] 
  15. Seol, D. H. (2020). "Comparison of deep learning based convolution neural network model for concrete crack detection." Graduate School of Kyungpook National University. 
  16. Shin, S. Y., Lee, S. H., & Han, H. H. (2021). "A study on Residual U-Net for Semantic Segmentation based on deep learning." Digital Convergence Study 19.6: 251-258. 
  17. Tsang, S. H. (2018). Review: Inception-v4 - Evolved From GoogLeNet, Merged with ResNet Idea (Image Classification), towards data science.