All Issue

2024 Vol.11, Issue 4 Preview Page

Original Article

Application of Deep Learning Model to Detect Cracked Pixels in Images with Semantic Segmentation

이미지의 균열 픽셀 탐지를 위한 딥러닝 모델의 적용과 의미론적 분할 활용

Jaehyun Shin1
,
Dong Sop Rhee2*
신 재현1
,
이 동섭2*
1Assistant Professor, Department of Civil and Environmental Engineering, Gachon University, Seongnam 13120, Korea
2Research Fellow, Department of Hydro Science and Engineering Research, Korea Institute of Civil Engineering and Building Technology, Goyang 10223, Korea
1가천대학교 토목환경공학과 조교수
2한국건설기술연구원 수자원하천연구본부 연구위원
*Corresponding Author
31 December 2024. pp. 177-185
PDF XML Citation
Abstract

To apply damage detection methods through deep learning of crack images that could be used in water resources facilities, an image processing algorithm that could detect crack damage image data was collected, utilized, and trained in this study. Using the U-Net structure based on convolution among deep learning techniques, we were able to classify crack damage areas within the image through classification for each pixel through learning and feature extraction of structural crack damage images. During this process, the activation map was visualized to check the model's response to crack features to determine which part of the image the model was focusing on. Crack detection results showed high accuracy and precision compared to the true value through image mask processing, confirming that this technique could be used for crack detection. The model learned in this study demonstrated detection capabilities of crack detection in acquired images as well as application to images of water resources facilities. It is expected to be used for monitoring safety diagnosis and evaluating water resources structures in the future.

Keywords
Crack
Deep learning
Image
Semantic segmentation
U-Net

본 연구에서는 수자원시설에서 활용될 수 있는 검출 기술로 균열 이미지의 딥러닝 학습을 통한 손상탐지 기법의 적용을 위하여 균열 손상 이미지 데이터를 수집, 활용, 학습하여 이를 탐지할 수 있는 이미지처리 알고리즘의 적용을 수행하였다. 딥러닝 기법 중 합성곱을 기반으로 한 U-Net 구조를 활용하여 구조물 균열손상 이미지들의 학습과 특징 추출을 통하여 각 픽셀 별 분류를 통해 이미지 내에서 균열손상 부위를 분류해 낼 수 있었다. 이 과정 중 활성화 맵을 시각화하여 균열 특징에 대한 모델의 반응을 확인하여 모델이 이미지의 어느 부분에 주목하고 있는지를 확인하였다. 균열 검출 결과 이미지 마스크 처리를 통한 참 값과 비교하여 높은 정확도와 정밀도가 나타나 본 기법이 균열 탐지에 적용성이 있음을 확인하였다. 본 연구에서 학습된 모델은 취득한 이미지에서의 균열 검출을 비롯하여 수자원 시설물 이미지에 대한 적용에도 검출 능력이 드러났으며, 향후 수자원 구조물의 안전 진단 및 평가를 위한 모니터링에서 활용될 것으로 기대된다.

키워드
균열
딥러닝
이미지
의미론적 분할
U-Net
References
1

Bao, Y., Chen, Z., Wei, S., Xu, Y., Tang, Z., and Li, H. 2019. The state of the art of data science and engineering in structural health monitoring. Engineering 5(2): 234-242.

10.1016/j.eng.2018.11.027
2

Cha, Y.J., Ali, R., Lewis, J., and Büyükӧztürk, O. 2024. Deep learning-based structural health monitoring. Automation in Construction 161: 105328.

10.1016/j.autcon.2024.105328
3

Fan, X., Cao, P., Shi, P., Wang, J., Xin, Y., and Huang, W. 2021. A nested unet with attention mechanism for road crack image segmentation. In 2021 IEEE 6th International Conference on Signal and Image Processing (ICSIP), pp. 189-193.

10.1109/ICSIP52628.2021.9688782
4

Feng, C., Zhang, H., Wang, H., Wang, S., and Li, Y. 2020. Automatic pixel-level crack detection on dam surface using deep convolutional network. Sensors 20(7): 2069.

10.3390/s2007206932272652PMC7180706
5

Fukushima, K. 1980. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biological cybernetics 36(4): 193-202.

10.1007/BF003442517370364
6

Huang, B., Zhao, S., and Kang F. 2023. Image-based automatic multiple-damage detection of concrete dams using region-based convolutional neural networks. Journal of Civil Structural Health Monitoring 13: 413-429.

10.1007/s13349-022-00650-9
7

Krizhevsky, A., Sutskever, I., and Hinton, G.E. 2012. ImageNet classification with deep convolutional neural networks. In Advances in Neural Information Processing Systems, pp. 1097-1105.

8

LeCun, Y., Bengio, Y., and Hinton, G. 2015. Deep learning. Nature 521(7553): 436-444.

10.1038/nature1453926017442
9

LeCun, Y., Bottou, L., Bengio, Y., and Haffner, P. 1998. Gradient-based learning applied to document recognition. Proceedings of the IEEE 86(11): 2278-2324.

10.1109/5.726791
10

Lin, T.Y., Dollár, P., Girshick, R., He, K., Hariharan, B., and Belongie, S. 2017. Feature pyramid networks for object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 2117-2125.

10.1109/CVPR.2017.106PMC5744014
11

Matarneh, S., Elghaish, F., Rahimian, F.P., Abdellatef, E., and Abrishami, S. 2024. Evaluation and optimization of pre-trained CNN models for asphalt pavement crack detection and classification. Automation in Construction 160: 105297.

10.1016/j.autcon.2024.105297
12

Ozgenel, Ç.F. and Sorguc A.G. 2018. Performance comparison of pretrained convolutional neural networks on crack detection in buildings. In Isarc. Proceedings of the International Symposium on Automation and Robotics In Construction 35: 1-8.

10.22260/ISARC2018/0094
13

Ronneberger, O., Fischer, P., and Brox, T. 2015. U-net: Convolutional networks for biomedical image segmentation. In Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015: 18th International Conference, Springer International Publishing, Munich, Germany, pp. 234-241.

10.1007/978-3-319-24574-4_28
14

Tomaszkiewicz, K. and Owerko, T., 2023. A pre-failure narrow concrete cracks dataset for engineering structures damage classification and segmentation, Scientific Data 10(1): 925

10.1038/s41597-023-02839-z38129453PMC10739794
15

Yang, F., Zhang, L., Yu, S., Prokhorov, D., Mei, X., and Ling, H. 2019. Feature pyramid and hierarchical boosting network for pavement crack detection. IEEE Transactions on Intelligent Transportation Systems 21(4): 1525-1535.

10.1109/TITS.2019.2910595
16

Zhang, L., Yang, F., Zhang, Y.D., and Zhu, Y.J. 2016. Road crack detection using deep convolutional neural network. In 2016 IEEE International Conference on Image Processing (ICIP), pp. 3708-3712.

10.1109/ICIP.2016.7533052
Information
  • Publisher :Korean Society of Ecology and Infrastructure Engineering
  • Publisher(Ko) :응용생태공학회
  • Journal Title :Ecology and Resilient Infrastructure
  • Journal Title(Ko) :응용생태공학회 논문집
  • Volume : 11
  • No :4
  • Pages :177-185
  • Received Date : 2024-12-06
  • Revised Date : 2024-12-17
  • Accepted Date : 2024-12-18
  • DOI :https://doi.org/10.17820/eri.2024.11.4.177
Journal Informaiton Ecology and Resilient Infrastructure Ecology and Resilient Infrastructure
Online Submission
  • NRF
  • KISTI Current Status
  • KISTI Cited-by
  • crosscheck
  • orcid
  • open access
  • ccl
Journal Informaiton Journal Informaiton - close