대한토목학회논문집 (KSCE Journal of Civil and Environmental Engineering Research)

  • 제42권1호
  • /
  • Pages.117-126
  • /
  • 2022
  • /
  • 1015-6348(pISSN)
  • /
  • 2799-9629(eISSN)
대한토목학회 (Korean Society of Civil Engeneers)
권호 표지
DOI QR코드

DOI QR Code

교량 구조물 손상탐지를 위한 Open Set Recognition 기반 다중손상 인식 모델 개발

Development of Open Set Recognition-based Multiple Damage Recognition Model for Bridge Structure Damage Detection

  • 김영남 (성균관대학교 전자전기컴퓨터공학과, 차세대융합기술연구원) ;
  • 조준상 (한국도로공사 도로교통연구원) ;
  • 김준경 (차세대융합기술연구원) ;
  • 김문현 (성균관대학교 소프트웨어학과) ;
  • 김진평 (차세대융합기술연구원)
  • Kim, Young-Nam (Department of Electrical and Computer Engineering, Sungkyunkwan University, Advanced Institute of Convergence Technology) ;
  • Cho, Jun-Sang (Korea Expressway Corporation) ;
  • Kim, Jun-Kyeong (Advanced Institute of Convergence Technology) ;
  • Kim, Moon-Hyun (Sungkyunkwan University) ;
  • Kim, Jin-Pyung (Advanced Institute of Convergence Technology)
  • 투고 : 2021.07.21
  • 심사 : 2021.12.07
  • 발행 : 2022.02.01
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

현재 국내 교량 구조물은 지속적으로 증가 및 대형화되고 있으며 그에 따라 공용된 지 30년 이상 된 노후 교량도 꾸준히 늘어나고 있다. 교량 노후화 문제는 국내뿐 아니라 전 세계적으로도 심각한 사회 문제로 다루어지고 있으며, 기존 인력 위주의 점검 방식은 그 한계점을 드러내고 있다. 최근 들어 딥러닝 기반의 영상처리 알고리즘을 활용한 각종 교량 손상탐지 연구가 이루어지고 있지만 교량 손상 데이터 세트의 한계로 인하여 주로 균열 1종에 국한된 교량 손상탐지 연구가 대부분이고, 이 또한 Close set 분류모델 기반 탐지방식으로서 실제 교량 촬영 영상에 적용했을 시 배경이나 기타 객체 등 학습되지 않은 클래스의 입력 이미지들로 인하여 심각한 오인식 문제가 발생할 수 있다. 본 연구에서는 균열 포함 5종의 교량 손상을 정의 및 데이터 세트를 구축해서 딥러닝 모델로 학습시키고, OpenMax 알고리즘을 적용한 Open set 인식 기반 교량 다중손상 인식 모델을 개발했다. 그리고 학습되지 않은 이미지들을 포함하고 있는 Open set에 대한 분류 및 인식 성능평가를 수행한 후 그 결과를 분석했다.

Currently, the number of bridge structures in Korea is continuously increasing and enlarged, and the number of old bridges that have been in service for more than 30 years is also steadily increasing. Bridge aging is being treated as a serious social problem not only in Korea but also around the world, and the existing manpower-centered inspection method is revealing its limitations. Recently, various bridge damage detection studies using deep learning-based image processing algorithms have been conducted, but due to the limitations of the bridge damage data set, most of the bridge damage detection studies are mainly limited to one type of crack, which is also based on a close set classification model. As a detection method, when applied to an actual bridge image, a serious misrecognition problem may occur due to input images of an unknown class such as a background or other objects. In this study, five types of bridge damage including crack were defined and a data set was built, trained as a deep learning model, and an open set recognition-based bridge multiple damage recognition model applied with OpenMax algorithm was constructed. And after performing classification and recognition performance evaluation on the open set including untrained images, the results were analyzed.

키워드

과제정보

본 연구는 국토교통부/국토교통과학기술진흥원의 지원으로 수행되었음(22LTSM-B156015-03).

참고문헌

  1. Bendale, A. and Boult, T. E. (2015). "Towards open world recognition." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, Massachusetts, USA, pp. 1893-1902.
  2. Bendale, A. and Boult, T. E. (2016). "Towards open set deep networks." Proceedings of the IEEE Conference on Computer Vision and pattern Recognition, Las Vegas, Nevada, USA, pp. 1563-1572.
  3. Cha, Y. J., Choi, W. R. and Buyukozturk, O. (2017). "Deep learning-based crack damage detection using convolutional neural networks." Computer-Aided Civil and Infrastructure Engineering, Vol. 32, No. 5, pp. 361-378. https://doi.org/10.1111/mice.12263
  4. Fisher, R. A. and Tippett, L. H. C. (1928). "Limiting forms of the frequency distribution of the largest or smallest member of a sample." Mathematical Proceedings of the Cambridge Philosophical Society, Cambridge University Press, Vol. 24. No. 2, pp. 180-190.
  5. Jenkinson, A. F. (1955). "The frequency distribution of the annual maximum (or minimum) values of meteorological elements." Quarterly Journal of the Royal Meteorological Society, Vol. 81, No. 348, pp. 158-171. https://doi.org/10.1002/qj.49708134804
  6. Kim, S. M., Sohn, J. M. and Kim, D. S. (2020). "A method for concrete crack detection using U-Net based image inpainting technique." Journal of The Korea Society of Computer and Information, Vol. 25, No. 10, pp. 35-42. https://doi.org/10.9708/JKSCI.2020.25.10.035
  7. Korea Expressway Corporation Research Institute (KECRI) (2015). Prediction model for long-term maintenance costs of highway bridges.
  8. Korea Expressway Corporation Research Institute (KECRI) (2020). Image-based inspection techniques of bridge for seoul-sejong expressway.
  9. LeCun, Y., Bottou, L., Bengio, Y. and Haffner, P. (1998). "Gradient-based learning applied to document recognition." Proceedings of the IEEE, Vol. 86, No. 11, pp. 2278-2324. https://doi.org/10.1109/5.726791
  10. Liu, Y., Tang, Y., Zhang, L., Liu, L., Song, M., Gong, K., Peng, Y., Hou, J. and Jiang, T. (2020). "Hyperspectral open set classification with unknown classes rejection towards deep networks." International Journal of Remote Sensing, Vol. 41, No. 16, pp. 6355-6383. https://doi.org/10.1080/01431161.2020.1754492
  11. Neal, L., Olson, M., Fern, X., Wong, W. K. and Li, F. (2018). "Open set learning with counterfactual images." Proceedings of the European Conference on Computer Vision, ECCV, Munich, Germany, pp. 1-16.
  12. Nguyen, A., Yosinski, J. and Clune, J. (2015). "Deep neural networks are easily fooled: High confidence predictions for unrecognizable images." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Boston, Massachusetts, USA, pp. 427-436.
  13. Ozgenel, C. F. (2019). "Concrete crack images for classification." Mendeley Data, Version 2, DOI: 10.17632/5y9wdsg2zt.2.
  14. Park, J. M., Kim, H. S., Shin, D. H., Park, M. S. and Kim, S. H. (2019). "A study on machine learning algorithm suitable for automatic crack detection in wall-climbing robot." Kips Transactions on Software and Data Engineering, Vol. 8, No. 11, pp. 449-456. https://doi.org/10.3745/KTSDE.2019.8.11.449
  15. Scheirer, W. J., Rocha, A. R., Sapkota, A. and Boult, T. E. (2012). "Toward open set recognition." IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 35, No. 7, pp. 1757-1772. https://doi.org/10.1109/TPAMI.2012.256
  16. Scheirer, W. J., Rocha, A., Micheals, R. J. and Boult, T. E. (2011). "Meta-recognition: The theory and practice of recognition score analysis." IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 33, No. 8, pp. 1689-1695. https://doi.org/10.1109/TPAMI.2011.54
  17. Scherreik, M. D. and Rigling, B. D. (2016). "Open set recognition for automatic target classification with rejection." IEEE Transactions on Aerospace and Electronic Systems, Vol. 52, No. 2, pp. 632-642. https://doi.org/10.1109/TAES.2015.150027
  18. Seol, D. H., Oh, J. H. and Kim, H. J. (2020). "Comparison of deep learning-based CNN models for crack detection." Journal of The Architectural Institute of Korea Structure & Construction, Vol. 36, No. 3, pp. 113-120.
  19. Simonyan, K. and Zisserman, A. (2014). "Very deep convolutional networks for large-scale image recognition." arXiv preprint arXiv:1409.1556.
  20. Sun, X., Li, X., Ren, K. and Song, J. (2019). "Solving the defect in application of compact abating probability to convolutional neural network based open set recognition." 2019 IEEE 31st International Conference on Tools with Artificial Intelligence, ICTAI, Portland, USA.