Crack characterisation in buildings utilizing deep learning techniques

• Autorzy: Jerushan J. , Jebadurai S. Vincent Sam , Staines J. Noble , Subin S. R. , Ebenezer V. , Ebenezer Shamila , Hemalatha G.

Identyfikatory

DOI

10.15199/48.2025.02.32

Warianty tytułu

PL

Charakterystyka pęknięć w budynkach z wykorzystaniem systemów głębokiego uczenia

• Języki publikacji: EN

Abstrakty

EN

A significant field of research is the use of deep learning algorithms to detect and characterize cracks in structures. Building cracks may cause catastrophic structural collapses that endanger people's lives and property. This issue can be helped by deep learning algorithms, which allow for the very accurate identification and categorization of various crack forms. The present study uses a data set of 5000 photos to examine how image pre-processing affects the effectiveness of Deep Learning crack detection. The outcomes demonstrated that the CNN model's ability to identify cracks in concrete buildings is unaffected by the use of a pretrained model with RGB weights. Pretrained VGG16 and the Keras Python library are used to create a CNN model. The SciKit Image Python package was employed to divide the original picture data set into five comparison sets. The created model performed better than 98% in terms of accuracy and F1 values.

PL

Istotnym obszarem badań jest wykorzystanie algorytmów głębokiego uczenia do wykrywania i charakteryzowania pęknięć w konstrukcjach. Pęknięcia w budynkach mogą powodować katastrofalne zawalenia konstrukcyjne, które zagrażają życiu i mieniu ludzi. Problem ten można rozwiązać za pomocą algorytmów głębokiego uczenia, które umożliwiają bardzo dokładną identyfikację i kategoryzację różnych form pęknięć. W tym badaniu wykorzystano zestaw danych 5000 zdjęć, aby zbadać, w jaki sposób wstępne przetwarzanie obrazu wpływa na skuteczność wykrywania pęknięć metodą głębokiego uczenia. Wyniki wykazały, że zdolność modelu CNN do identyfikowania pęknięć w betonowych budynkach nie jest naruszona przez użycie wstępnie wytrenowanego modelu z wagami RGB. Wstępnie wytrenowany VGG16 i biblioteka Keras Python są używane do tworzenia modelu CNN. Pakiet SciKit Image Python został użyty do podzielenia oryginalnego zestawu danych obrazu na pięć zestawów porównawczych. Utworzony model uzyskał wyniki lepsze niż 98% pod względem dokładności i wartości F1.

Słowa kluczowe

EN

characterisation; crack; deep learning; CNN; python

PL

charakterystyka; pęknięcie; głębokie uczenie; CNN; python

• Wydawca: Wydawnictwo SIGMA-NOT
• Czasopismo: Przegląd Elektrotechniczny
• Rocznik: 2025
• Tom: R. 101, nr 2
• Strony: 137--143
• Opis fizyczny: Bibliogr. 18 poz., rys.

Twórcy

autor

Jerushan J.

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0009-0005-8596-2560

autor

Jebadurai S. Vincent Sam

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0000-0002-9820-8103

autor

Staines J. Noble

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0009-0003-4372-3620

autor

Subin S. R.

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

autor

Ebenezer V.

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0000-0002-0801-6926

autor

Ebenezer Shamila

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0000-0002-5167-8968

autor

Hemalatha G.

• Department of Civil Engineering, Karunya Institute of Technology and Sciences, Coimbatore 641114, Tamil Nadu, India

• https://orcid.org/0000-0001-7067-3786

Bibliografia

• [1] R. Fan et al., “Road crack detection using deep convolutional neural network and adaptive thresholding,” IEEE Intelligent Vehicles Symposium, Proceedings, vol. 2019-June, pp. 474– 479, Jun. 2019, doi: 10.1109/IVS.2019.8814000.
• [2] L. Zhang, F. Yang, Y. Daniel Zhang, and Y. J. Zhu, “Road crack detection using deep convolutional neural network,” Proceedings - International Conference on Image Processing, ICIP, vol. 2016-August, pp. 3708–3712, Aug. 2016, doi: 10.1109/ICIP.2016.7533052.
• [3] W. Ouyang and B. Xu, “Pavement cracking measurements using 3D laser-scan images,” Meas Sci Technol, vol. 24, no. 10, p. 105204, Aug. 2013, doi: 10.1088/0957- 0233/24/10/105204.
• [4] D. Joshi, T. P. Singh, and G. Sharma, “Automatic surface crack detection using segmentation-based deep-learning approach,” Eng Fract Mech, vol. 268, p. 108467, Jun. 2022, doi: 10.1016/J.ENGFRACMECH.2022.108467.
• [5] H. N. Nguyen, T. Y. Kam, and P. Y. Cheng, “An Automatic Approach for Accurate Edge Detection of Concrete Crack Utilizing 2D Geometric Features of Crack,” J Signal Process Syst, vol. 77, no. 3, pp. 221–240, Oct. 2014, doi: 10.1007/S11265-013-0813-8/METRICS.
• [6] R. Ali, J. H. Chuah, M. S. A. Talip, N. Mokhtar, and M. A. Shoaib, “Structural crack detection using deep convolutional neural networks,” Autom Constr, vol. 133, p. 103989, Jan. 2022, doi: 10.1016/J.AUTCON.2021.103989.
• [7] G. X. Hu, B. L. Hu, Z. Yang, L. Huang, and P. Li, “Pavement Crack Detection Method Based on Deep Learning Models,” Wirel Commun Mob Comput, vol. 2021, no. 1, p. 5573590, Jan. 2021, doi: 10.1155/2021/5573590.
• [8] J. Zhong, J. Zhu, J. Huyan, T. Ma, and W. Zhang, “Multi-scale feature fusion network for pixel-level pavement distress detection,” Autom Constr, vol. 141, p. 104436, Sep. 2022, doi: 10.1016/J.AUTCON.2022.104436.
• [9] P. Wang et al., “An automatic building façade deterioration detection system using infrared-visible image fusion and deep learning,” Journal of Building Engineering, vol. 95, p. 110122, Oct. 2024, doi: 10.1016/J.JOBE.2024.110122.
• [10] V. P. Golding, Z. Gharineiat, H. S. Munawar, and F. Ullah, “Crack Detection in Concrete Structures Using Deep Learning,” Sustainability, vol. 14, no. 13, p. 8117, Jul. 2022, doi: 10.3390/su14138117.
• [11] V. Mandal, L. Uong, and Y. Adu-Gyamfi, “Automated Road Crack Detection Using Deep Convolutional Neural Networks,” in 2018 IEEE International Conference on Big Data (Big Data), IEEE, Dec. 2018, pp. 5212–5215. doi: 10.1109/BigData.2018.8622327.
• [12] K. C. Laxman, N. Tabassum, L. Ai, C. Cole, and P. Ziehl, “Automated crack detection and crack depth prediction for reinforced concrete structures using deep learning,” Constr Build Mater, vol. 370, p. 130709, Mar. 2023, doi: 10.1016/J.CONBUILDMAT.2023.130709.
• [13] K. Chen, G. Reichard, X. Xu, and A. Akanmu, “Automated crack segmentation in close-range building façade inspection images using deep learning techniques,” Journal of Building Engineering, vol. 43, p. 102913, Nov. 2021, doi: 10.1016/J.JOBE.2021.102913.
• [14] R. E. Philip, A. D. Andrushia, A. Nammalvar, B. G. A. Gurupatham, and K. Roy, “A Comparative Study on Crack Detection in Concrete Walls Using Transfer Learning Techniques,” Journal of Composites Science 2023, Vol. 7, Page 169, vol. 7, no. 4, p. 169, Apr. 2023, doi: 10.3390/JCS7040169.
• [15] J. J. Kim, A.-R. Kim, and S.-W. Lee, “Artificial Neural Network- Based Automated Crack Detection and Analysis for the Inspection of Concrete Structures,” Applied Sciences, vol. 10, no. 22, p. 8105, Nov. 2020, doi: 10.3390/app10228105.
• [16] R. Geirhos, C. Michaelis, F. A. Wichmann, P. Rubisch, M. Bethge, and W. Brendel, “ImageNet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness,” 7th International Conference on Learning Representations, ICLR 2019, Nov. 2018, Accessed: Jul. 21, 2024. [Online]. Available: https://arxiv.org/abs/1811.12231v3
• [17] L. A. Gatys, A. S. Ecker, and M. Bethge, “Image Style Transfer Using Convolutional Neural Networks,” Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, vol. 2016-December, pp. 2414–2423, Dec. 2016, doi: 10.1109/CVPR.2016.265.
• [18] K. Simonyan and A. Zisserman, “Very Deep Convolutional Networks for Large-Scale Image Recognition,” 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings, Sep. 2014, Accessed: Jul. 21, 2024. [Online]. Available: https://arxiv.org/abs/1409.1556v6