Automated Concrete Cube Failure Mode Classification Using Deep Learning And Engineering-Guided Decision Logic
DOI:
https://doi.org/10.29284/fkkb3m03Keywords:
Concrete cube failure mode classification, Deep learning–based inspection, YOLOv8 object detection, Multi-view image analysis, Engineering-guided decision rules, Construction quality control.Abstract
Compression strength of concrete on cubes is a traditional building quality control check. Nevertheless, the evaluation of cube failure mechanisms is still mainly based on visual inspection (manual), which is subjective and varies among different individuals with a high degree of misclassification, resulting in material waste, higher costs, and unnecessary environmental pollution. Although there has been study on crack detection for the surface of concrete, automatic categorization of engineering-valid cube failure modes is still scanty. This paper presents an automated fault mode classification framework which combines deep learning-based object detection with rule-based engineering decision logic. A hybrid YOLOv8 and convolutional neural network-based architecture is proposed to detect objects and cracks at multiple scales. The system provides a six-face assessment of each concrete cube and adopts morphological image processing to improve the rapid detection and estimation of rep crack boundaries along with noise reduction. Categorization of failure mode is done by decision rule on rule-based engineering based on the conventional method of testing. Additionally, a dataset including 100 concrete cube samples (and 600 images) was extracted and annotated by professionals. Severe class imbalance was dealt with by creating a balanced dataset via augmenting and synthesizing data. The experimental results show that the classification accuracy of all works are 94.3% on the original data and 96.8% on the balanced data, whilst cube detection has precision = 97.8%, recall = 96.5%, mAP@0. 5 of 0.987. The system significantly shortens time of inspection by about 85% compared to manual assessment with high reliability. Results show that coupling deep learning with model-based decision logic yields robust scalable and objective failure mode classification paving the way for automated quality control in sustainable construction applications.
References
1. Scrivener, K.L., John, V.M., & Gartner, E.M. (2018). Eco-efficient cements: Potential economically viable solutions for a low-CO₂ cement-based materials industry. Cement and Concrete Research, 114, 2–26.
2. Mehta, P.K., &Monteiro, P.J.M. (2019). Concrete: Microstructure, Properties, and Materials (4th ed.). McGraw-Hill Education.
3. Neville, A.M. (2020). Properties of Concrete (6th ed.). Pearson Education.
4. Sadouskaya, A., Leanovich, S., Budrevich, N.A., &Sadovskaya, E.A. (2021). Compressive strength of steel fibre reinforced concrete obtained by testing cubes and cylinders. MATEC Web of Conferences, 350, 00008.
5. Halstead, P.E. (1969). The significance of concrete cube tests. Magazine of Concrete Research, 21(69), 187–194.
6. Mukhopadhyay, T., Naskar, S., Dey, S., &Chakrabarti, A. (2019). Condition assessment and strengthening of aged structures. Practice Periodical on Structural Design and Construction, 24(3), 04019013.
7. Kim, J., Cho, Y., Lee, J., & Kim, Y.H. (2022). Defect detection and characterization in concrete based on FEM and ultrasonic techniques. Materials, 15(22), 8160.
8. Shahrokhinasab, E., Hosseinzadeh, N., Monirabbasi, A., &Torkaman, S. (2020). Performance evaluation of image-based crack detection systems in concrete structures. Journal of Infrastructure Systems, 26(4), 04020039.
9. Dorafshan, S., Maguire, M., & Hoffer, N. (2018). Challenges in automated crack detection of concrete structures. Computer-Aided Civil and Infrastructure Engineering, 33(8), 642–660.
10. Andrew, R.M. (2019). Global CO₂ emissions from cement production. Earth System Science Data, 11, 1675–1710.
11. Miller, S.A., Horvath, A., &Monteiro, P.J.M. (2021). Readily implementable techniques can cut annual CO₂ emissions from the global cement industry by at least 20%. Environmental Science & Technology, 52(9), 5560–5569.
12. Malhotra, V.M., &Carino, N.J. (2019). Handbook on Nondestructive Testing of Concrete (2nd ed.). CRC Press.
13. Lorenzi, A., Tisbierek, F., & Dal Bó, R. (2020). Ultrasonic testing for damage assessment of concrete structures. Construction and Building Materials, 247, 118517.
14. Taffese, W.Z., &Sistonen, E. (2017). Machine learning for durability and service-life assessment of reinforced concrete structures. Automation in Construction, 77, 1–14.
15. Abdel-Qader, I., Abudayyeh, O., & Kelly, M.E. (2020). Analysis of edge-detection techniques for crack identification in concrete. Journal of Computing in Civil Engineering, 34(2), 04019045.
16. Savio, A.A.D., Torres, A.L., Salas, D., Olivera, M.A.V., & Ibarra, G.T.U. (2023). Detection and evaluation of construction cracks through image analysis using computer vision. Applied Sciences, 13(17), 9662.
17. Cha, Y.J., Choi, W., &Büyüköztürk, O. (2017). Deep learning-based crack damage detection using convolutional neural networks. Computer-Aided Civil and Infrastructure Engineering, 32(5), 361–378.
18. Xu, Y., Bao, Y., Chen, J., Zuo, W., & Li, H. (2019). Surface crack detection using deep learning-based fully convolutional networks. Computer-Aided Civil and Infrastructure Engineering, 34(7), 567–582.
19. Li, S., Zhao, X., Zhou, G., & Du, K. (2021). Deep learning-based crack detection for concrete structures: A review. Automation in Construction, 125, 103606.
20. Zhang, L., Yang, F., Zhang, Y.D., & Zhu, Y.J. (2020). Road crack detection using deep convolutional neural networks. IEEE Transactions on Intelligent Transportation Systems, 21(4), 1525–1535.
21. Zou, Q., Cao, Y., Li, Q., Mao, Q., & Wang, S. (2021). CrackTree: Automatic crack detection from pavement images. Neurocomputing, 408, 161–171.
22. Lin, T.Y., Dollár, P., Girshick, R., He, K., Hariharan, B., &Belongie, S. (2020). Feature pyramid networks for object detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44(5), 2733–2747.
23. Bochkovskiy, A., Wang, C.Y., & Liao, H.Y.M. (2020). YOLOv4: Optimal speed and accuracy of object detection. IEEE Conference on Computer Vision and Pattern Recognition Workshops.
24. Redmon, J., &Farhadi, A. (2018). YOLOv3: An incremental improvement. arXiv preprint → (used only conceptually; implementation basis)
25. Downey, A., D’Alessandro, A., Ubertini, F., &Laflamme, S. (2018). Automated crack detection in conductive smart-concrete structures. Measurement Science and Technology, 29(3), 035107.
26. Deng, L., Wang, J., Xu, S., & Zhang, Y. (2022). Dataset challenges and solutions for concrete crack detection using deep learning. Automation in Construction, 136, 104158.
27. Shorten, C., &Khoshgoftaar, T.M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6, 60.
28. Frid-Adar, M., et al. (2018). GAN-based synthetic medical image augmentation. Neurocomputing, 321, 321–331.
29. BS EN 12390-3:2019. Testing hardened concrete – Compressive strength of test specimens. British Standards Institution.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
