Deep Learning Based Framework for Automated Road Crack Detection and Severity Assessment under Different Weather Conditions
Keywords:
Crack geometry analysis, Deep learning, Dry and wet surface conditions, Road crack detectionAbstract
Crack detection and assessment of roads and concrete are vital for infrastructure and safety management. This paper proposed an end-to-end deep learning based framework for automated crack detection, classification, and severity analysis under different surface conditions. The simulation implemented and tested four architectures, namely CNN, BiGRU, BiLSTM, and a hybrid BiGRU–BiLSTM–CNN model. Using a large dataset of 5,200 crack and non-crack photos collected from the METU campus and road photos taken using a smartphone, we trained and assessed our approach. Regression and classification metrics such as RMSE, MAE, R2, Cohen's Kappa, F1-score, and AUC were also used in the study to evaluate model performance. The CNN outperformed recurrent and hybrid models in capturing complex spatial crack information, according to the experimental data (R² = 0.92, F1-score = 0.97, AUC = 0.99). The suggested framework included geometric crack analysis and 3D heat map visualization in addition to probabilistic crack classification to measure relative depth and crack shape. The result offers an approach to real-time road condition evaluation using conventional cameras that is clear, dependable, and reasonably priced.
Downloads
References
Milling, A., Martin, H., and Mwasha, A. (2023). Design, construction, and in-service causes of premature pavement deterioration: a fuzzy Delphi application. Journal of Transportation Engineering, Part B: Pavements, 149(1), 05022004.
Rincon, L. F., Moscoso, Y. M., Hamami, A. E. A., Matos, J. C., and Bastidas-Arteaga, E. (2024). Degradation models and maintenance strategies for reinforced concrete structures in coastal environments under climate change: a review. Buildings, 14(3), 562.
Septiyani, Y. N. (2024). The impact of load traffic of road deterioration in urban areas: case study Jalan KH Abdul Halim Majalengka. LEADER: Civil Engineering and Architecture Journal, 2(4), 911-919.
Bhosale, T., Attar, A., Warang, P., and Patil, R. (2022). Object detection for autonomous guided vehicle. ASEAN Journal of Science and Engineering, 2(3), 209-216.
Pohan, M. A. R., Utama, J., and Herdiana, B. (2024). Novel motion planning strategy with fuzzy logic for improving safety in autonomous vehicles in response to risky road user behaviors. ASEAN Journal of Science and Engineering, 4(3), 471-484.
Maske, M. (2022). Review of applications of ground penetrating radar as an NDT tool. ASEAN Journal of Science and Engineering, 2(2), 115-128.
Merkle, D., Frey, C., and Reiterer, A. (2021). Fusion of ground penetrating radar and laser scanning for infrastructure mapping. Journal of Applied Geodesy, 15(1), 31-45.
Rangole, A., De, S., Kuchekar, N., and Raj, A. B. (2024). A comprehensive review of ground penetrating radar: Techniques, applications and future directions. International Journal of Engineering Research and Reviews, 12(3), 30-53.
Shayea, G. G., Zabil, M. H. M., Habeeb, M. A., Khaleel, Y. L., and Albahri, A. (2025). Strategies for protection against adversarial attacks in AI models: An in-depth review. Journal of Intelligent Systems, 34(1), 20240277.
Yang, Y., Pan, Z., Sun, J., Welch, J., and Klionsky, D. J. (2024). Autophagy and machine learning: Unanswered questions. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease, 1870(6), 167263.
Chen, Y., Kang, J., Feng, L., Yuan, L., Liang, J., Zhao, Z., and Wu, B. (2024). Deep learning-based frequency-multiplexing composite-fringe projection profilometry technique for one-shot 3D shape measurement. Measurement, 233, 114640.
Hadj-Attou, A., Kabir, Y., and Ykhlef, F. (2023). Hybrid deep learning models for road surface condition monitoring. Measurement, 220, 113267.
Huang, Z., Xu, G., Zhang, X., Zang, B., and Yu, H. (2025). Three-dimensional ground-penetrating radar-based feature point tensor voting for semi-rigid base asphalt pavement crack detection. Developments in the Built Environment, 21, 100591.
Afzal, A. (2024). A hybrid approach combining computer vision and machine learning for enhanced image analysis. Academia Nexus Journal, 3(3), 1-15.
Choudhary, G., and Sethi, D. (2023). From conventional approach to machine learning and deep learning approach: an experimental and comprehensive review of image fusion techniques. Archives of Computational Methods in Engineering, 30(2), 1267-1304.
Sajid, M., Malik, K. R., Almogren, A., Malik, T. S., Khan, A. H., Tanveer, J., and Rehman, A. U. (2024). Enhancing intrusion detection: a hybrid machine and deep learning approach. Journal of Cloud Computing, 13(1), 123.
Zhuang, H., Cheng, Y., Zhou, M., and Yang, Z. (2025). Deep learning for surface crack detection in civil engineering: A comprehensive review. Measurement, 116908.
Thohari, A. N. A., Karima, A., Santoso, K., and Rahmawati, R. (2024). Crack detection in building through deep learning feature extraction and machine learning approch. Journal of Applied Informatics and Computing, 8(1), 1-6.
Khan, S., Jan, A., and Seo, S. (2023). Structural crack detection using deep learning: an in-depth review. Korean Journal of Remote Sensing, 39(4), 371-393.
El-Din Hemdan, E., and Al-Atroush, M. (2025). A review study of intelligent road crack detection: Algorithms and systems. International Journal of Pavement Research and Technology, 2025, 1-31.
Manoni, L., Orcioni, S., and Conti, M. (2024). Recent advancements in deep learning techniques for road condition monitoring: A comprehensive review. IEEE Access, 12, 154271-154293.
Gupta, P., and Dixit, M. (2022). Image-based crack detection approaches: a comprehensive survey. Multimedia Tools and Applications, 81(28), 40181-40229.
Birgani, S. A., Zadeh, S. S., Davari, D. D., and Ostovar, A. (2024). Deep Learning Applications for Analysing Concrete Surface Cracks. International Journal of Applied Data Science in Engineering and Health, 1(2), 69-84.
Rashid, T., Mokji, M. M., and Rasheed, M. (2025). Cracked concrete surface classification in low-resolution images using a convolutional neural network. Journal of Optics, 54(5), 3671-3683.
Vargas, I. G. (2024). Deep Learning Approaches for Defect Segmentation on Composite Materials using Infrared Thermography. https://repositorio.ufu.br/handle/123456789/44621
Yang, S., Shi, H., Yin, J., Xu, X., Yao, J., and Liu, S. (2025). PLI and CNN-BiLSTM: An enhanced data augmentation and deep learning approach for defect recognition in spiral welded pipe based on ultrasonic guided waves. IEEE Sensors Journal, 25(19), 35916-35929.
Fan, Z., Y. Wu, J. Lu, and W. Li. (2018). Automatic pavement crack detection based on structured prediction with the convolutional neural network. arXiv preprint arXiv:1802.02208. DOI: https://doi.org/10.48550/arXiv.1802.02208.
Ali, L., et al. (2021). Performance evaluation of deep CNN-based crack detection and localization techniques for concrete structures. Sensors. 21(5): p. 1688. DOI: https://doi.org/10.3390/s21051688.
Matarneh, S., et al. (2024). Evaluation and optimisation of pre-trained CNN models for asphalt pavement crack detection and classification. Automation in Construction. 160: p. 105297. DOI: https://doi.org/10.1016/j.autcon.2024.105297.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
