DOI QR코드

DOI QR Code

딥러닝 기반 픽셀 단위 콘크리트 벽체 균열 검출 방법

Deep Learning-based Pixel-level Concrete Wall Crack Detection Method

  • 투고 : 2022.12.14
  • 심사 : 2023.03.16
  • 발행 : 2023.04.20

초록

콘크리트는 압축력에 잘 저항하고 내구성이 우수하여 널리 사용되는 재료이다. 하지만 구조물은 시공 단계에서 주변 환경, 사용되는 재료의 특성에 따라 완공된 후 표면의 균열, 구조물의 침하 등 다양한 하자가 발생하거나 시간이 지남에 따라 콘크리트 구조물 표면에 결함이 발생한다. 그대로 방치하면 구조물에 심각한 손상을 초래하기 때문에 안전 점검을 통해 검사해야 한다. 하지만 전문 검사원들이 직접 조사하기에 비용이 높고 육안으로 판단하는 외관 검사법을 사용한다. 고층 건물일수록 상세한 검사가 힘들다. 본 연구는 노후화로 인해 콘크리트 표면에 발생하는 결함 중 균열을 탐지하는 딥러닝 기반 시맨틱 세그먼테이션 모형과 해당 모형의 특징 추출과 일반화 성능을 높이기 위한 이미지 어그멘테이션 기법을 개발하였다. 이를 위해 공개 데이터셋과 자체 데이터셋을 결합하여 시맨틱 세그먼테이션용 데이터셋을 구축하고 대표적인 딥러닝 기반 시맨틱 세그먼테이션 모형들을 비교실험하였다. 콘크리트 내벽을 중점으로 학습한 모형의 균열 추출 성능은 81.4%이며, 개발한 이미지 어그멘테이션을 적용한 결과 3%의 성능향상을 확인하였다. 향후 고층 건물과 같이 접근성이 어려운 지점을 드론을 통해 콘크리트 외벽에서 균열을 검출할 수 있는 시스템을 개발함으로써 실질적으로 활용할 수 있기를 기대한다.

Concrete is a widely used material due to its excellent compressive strength and durability. However, depending on the surrounding environment and the characteristics of the materials used in the construction, various defects may occur, such as cracks on the surface and subsidence of the structure. The detects on the surface of the concrete structure occur after completion or over time. Neglecting these cracks may lead to severe structural damage, necessitating regular safety inspections. Traditional visual inspections of concrete walls are labor-intensive and expensive. This research presents a deep learning-based semantic segmentation model designed to detect cracks in concrete walls. The model addresses surface defects that arise from aging, and an image augmentation technique is employed to enhance feature extraction and generalization performance. A dataset for semantic segmentation was created by combining publicly available and self-generated datasets, and notable semantic segmentation models were evaluated and tested. The model, specifically trained for concrete wall fracture detection, achieved an extraction performance of 81.4%. Moreover, a 3% performance improvement was observed when applying the developed augmentation technique.

키워드

과제정보

This work was supported by the National Research Foundation of Korea(NRF) grant funded by the Korea government (MSIT)(No. 2022R1F1A107008211).

참고문헌

  1. Oh SG. A View on water-laekage and contermeasure of concrete structure. Magazine of the Korea Concrete Institute. 2002 Nov;14(6):14-9.  https://doi.org/10.22636/MKCI.2002.14.6.14
  2. Understanding of Concrete [Internet]. Seoul (Korea): Korea Cement Association. Available from: http://www.cement.or.kr/about_2015/cont8.asp?sm=1_4_0 
  3. Kang HK, Hong SG. Causes and mitigation of concrete cracking. Magazine of the Korea Concrete Institute. 2008 Sep;20(5):61-8.  https://doi.org/10.22636/MKCI.2008.20.5.61
  4. Choo JH, Park SW, Kim HT, Jee KH, Yoon TG. Analysis and cause of occurrence of lining cracks on NATM tunnel based on the precise inspection for safety and diagnosis-Part I. Journal of Korean Tunnelling and Underground Space Association. 2011 May;13(3):199-214. https://doi.org/10.9711/KTAJ.2011.13.3.199 
  5. Kim SM, Sohn JM, Kim DS. A method for concrete crack detection using U-Net based image inpainting technique. Journal of the Korea Society of Computer and Information. 2020 Oct;25(10):35-42. https://doi.org/10.9708/jksci.2020.25.10.035 
  6. Sohn JM, Kim DS, Hwang HB. Improvement of learning concrete crack detection model by weighted loss function. Journal of the Korea Society of Computer and Information. 2020 Oct;25(10):15-22. https://doi.org/10.9708/jksci.2020.25.10.015 
  7. Kim GY, Kim DH, Kim DJ. Deep learning-based exterior wall crack detection method for building safety diagnosis. Journal of the Korean Institute of Plant Engineering. 2021 Dec;26(4):31-43. 
  8. Lee SM, Kim KY, Kim DJ. Development of robust semantic segmentation modeling on various wall cracks. Proceedings of the Korean Society of Computer Information Conference. 2022 Jul 14-16; Jeju, Korea. Yongin (Korea): Korea Society of Computer Information; 2022. p. 49-52. 
  9. Dung CV, Anh LD. Autonomous concrete crack detection using deep fully convolutional neural network. Automation in Construction. 2019 Mar;99:52-8. https://doi.org/10.1016/j.autcon.2018.11.028 
  10. Lee SI, Yang GM, Lee YM, Lee JH, Jeong YJ, Lee JG, Choi W. Recognition and visualization of crack on concrete wall using deep learning and transfer learning. Journal of the Korean Society of Agricultural Engineers. 2019 May;61(3):55-65. https://doi.org/10.5389/KSAE.2019.61.3.055 
  11. Szegedy C, Vanhoucke V, Ioffe S, Shlens J, Wojna Z. Rethinking the inception architecture for computer vision. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015 Jun 27-30; Las Vegas, NV; Manhattan (NY): IEEE; 2015. p. 2818-26. https://doi.org/10.1109/CVPR.2016.308 
  12. Seol DH, Oh JH, Kim HJ. Comparison of deep learning-based cnn models for crack detection. Journal of the Architectural Institute of Korea Structure & Construction. 2020 Mar;36(3):113-20. https://doi.org/10.5659/JAIK_SC.2020.36.3.113 
  13. Jung SY, Lee SK, Park CI, Cho SY, Yu JH. A method for detecting concrete cracks using deep-learning and image processing. Journal of the Architectural Institute of Korea Structure & Construction. 2019 Nov;35(11):163-70. https://doi.org/10.5659/JAIK_SC.2019.35.11.163 
  14. Redmon J, Farhadi A. YOLO9000: better, faster, stronger. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2017 July 21-26; Honolulu, HI. Manhattan (NY): IEEE; 2017. p. 7263-71. https://doi.org/10.1109/CVPR.2017.690 
  15. Lee MH, Nam KW, Lee CW. Crack detection on the road in aerial image using mask R-CNN. Journal of the Korea Industrial Information Systems Research. 2019 Jun;24(3):23-9. https://doi.org/10.9723/jksiis.2019.24.3.023 
  16. Yamane T, Chun P. Crack detection from a concrete surface image based on semantic segmentation using deep learning. Journal of Advanced Concrete Technology. 2020 Sep;18(9):493-504. https://doi.org/10.3151/jact.18.493 
  17. Kim SM, Sohn JM, Kim DS. A method for concrete crack detection using U-Net based image inpainting technique. Journal of The Korea Society of Computer and Information. 2020 Oct;25(10):35-42. https://doi.org/10.9708/jksci.2020.25.10.035 
  18. Ha JW, Park KW, Kim MS. A development of road crack detection system using deep learning-based segmentation and object detection. The Journal of Society for e-Business Studies. 2021 Feb;26(1):93-106. https://doi.org/10.7838/jsebs.2021.26.1.093 
  19. Panella F, Aldo L, Jan B. Semantic segmentation of cracks: Data challenges and architecture. Automation in Construction. 2022 Mar;135:104110. https://doi.org/10.1016/j.autcon.2021.104110 
  20. Krizhevsky A, Sutskever I, Hinton GE. Imagenet classification with deep convolutional neural networks. Communications of the ACM. 2017 Jun;60(6):84-90. https://doi.org/10.1145/3065386 
  21. Garcia-Garcia A, Orts-Escolano S, Oprea S, Villena-Martinez V, Garcia-Rodriguez J. A review on deep learning techniques applied to semantic segmentation. arXiv preprint. 2017 Apr;arXiv:1704.06857. https://doi.org/10.48550/arXiv.1704.06857 
  22. Hafiz AM, Bhat GM. A survey on instance segmentation: state of the art. International Journal of Multimedia Information Retrieval. 2020 Jul;9(3):171-189. https://doi.org/10.1007/s13735-020-00195-x 
  23. Long J, Shelhamer E, Darrell T. Fully convolutional networks for semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2015 Jun 7-12; Boston, MA. Manhattan (NY): IEEE; 2015. p. 3431-40. https://doi.org/10.1109/CVPR.2015.7298965 
  24. Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical Image Computing and Computer-assisted Intervention. 2015 Oct 5-9; Munich, Germany, Manhattan (NY): Springer; 2015. p. 234-241. https://doi.org/10.1007/978-3-319-24574-4_28 
  25. Arkin E, Yadikar N, Muhtar Y, Ubul K. A Survey of Object Detection Based on CNN and Transformer. 2021 IEEE 2nd International Conference on Pattern Recognition and Machine Learning (PRML). 2021 July 16-18; Chengdu, China. Manhattan (NY): IEEE; 2021. p. 99-108. https://doi.org/10.1109/PRML52754.2021.9520732 
  26. Chen LC, Papandreou G, Kokkinos I, Murphy K, Yuille AL. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2018 Apr;40(4):834-48. https://doi.org/10.1109/TPAMI.2017.2699184 
  27. Chen LC, Papandreou G, Schroff F, Adam H. Rethinking atrous convolution for semantic image segmentation. arXiv preprint. 2017 Dec;arXiv:1706.05587. https://doi.org/10.48550/arXiv.1706.05587 
  28. Chen LC, Zhu Y, Papandreou G, Schroff F, Adam H. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proceedings of the European Conference on Computer Vision (ECCV). 2018 Sep 8-14; Munich, Germany. Manhattan (NY): Springer; 2018. p. 801-8. https://doi.org/10.1007/978-3-030-01234-2_49 
  29. Hassan SI, Dang L, Im SH, Min KB, Nam JY, Moon HJ. Damage detection and classification system for sewer inspection using convolutional neural networks based on deep learning. Journal of the Korea Institute of Information and Communication Engineering. 2018 Mar;22(3):451-7. http://doi.org/10.6109/jkiice.2018.22.3.451 
  30. Paik SH, Choi DC, Kim YK, Jung SW, Kim DN. Implementation of the drones with deep-learning crack detection analysis for inspection of bridg. Journal of Korea Institute of Information Technology. 2021 Mar;19(3):45-52. http://dx.doi.org/10.14801/jkiit.2021.19.3.45 
  31. Ratner AJ, Ehrenberg H, Hussain Z, Dunnmon J, Re C. Learning to compose domain-specific transformations for data augmentation. Advances in Neural Information Processing Systems. 2017 Sep;30:3236-46. https://doi.org/10.48550/arXiv.1709.01643 
  32. Shorten C, Khoshgoftaar TM. A survey on image data augmentation for deep learning. Journal of Big Data. 2019 Jul;6(1):1-48. https://doi.org/10.1186/s40537-019-0197-0 
  33. Bochkovskiy A, Wang CY, Liao HYM. Yolov4: Optimal speed and accuracy of object detection. arXiv preprint. 2020 Apr;arXiv:2004.10934. https://doi.org/10.48550/arXiv.2004.10934 
  34. Hao W, Zhili S. Improved mosaic: algorithms for more complex images. Journal of Physics: Conference Series. 2020 Nov; 1684(1):012094. https://doi.org/10.1088/1742-6596/1684/1/012094