Convolutional neural networks performance evaluation applied to automated pavement crack detection
DOI:
https://doi.org/10.14295/transportes.v28i5.2283Keywords:
Convolutional neural network. Pavement management systems. Automated pavement crack detection. Computing vision.Abstract
This research aims to analyze the performance of Convolutional Neural Network (CNN) as an automated tool applied to pavement surface crack detection. A group of pictures from different segments of chip seal pavement, acquired from photographic recording systems mounted on specific vehicles, was evaluated. An open-source machine learning library PyTorch available in the Python script language was applied to evaluate the images. The influence of three techniques used to process the pictures and the complexity of neural networks on the crack identification performance are discussed as well. The accuracy, precision, recall, and F1 score metrics were used to assess the neural network performance. The results show a good performance of the selected algorithm for pavement crack detection based on the observed metrics. Furthermore, it was found that the complexity of the neural network is an important factor that should be considered during the image classification process.
Downloads
References
Arpit, D.; S. Jastrzębski; N. Ballas; D. Krueger; E. Bengio; M. S. Kanwal; T. Maharaj; A. Fischer; A. Courville; Y. Bengio and S. Lacoste-Julien (2017). A Closer Look at Memorization in Deep Networks. Proceedings of the 34th International Conference on Machine Learning, Sydney, Australia, PMLR 70, 2017. ArXiv - a repository of electronic preprints, 1–10. arXiv:1706.05394
Bradski, G. (2000). The OpenCV Library. Dr. Dobb’s Journal of Software Tools. Available at: http://drdobbs.com/opensource/184404319
Dalla Rosa, F.; N. G. Gharaibeh; E. G. Fernando and A. Wimsatt (2016). Quality Assurance for Automated and Semi-Automated Pavement Condition Surveys. International Conference on Transportation and Development 2016. p. 192–201. doi:10.1061/9780784479926.018
Dung, C. V. and L. D. Anh (2019). Autonomous concrete crack detection using deep fully convolutional neural network. Automation in Construction, v.99, p. 52–58. doi:10.1016/j.autcon.2018.11.028
Fan, Z.; S. Member; Y. Wu; J. Lu and W. Li (2018). Automatic Pavement Crack Detection Based on Structured Prediction with the Convolutional Neural Network. ArXiv - a repository of electronic preprints, p. 1–9. arXiv:1802.02208
Haykin, S. (2009). Neural networks and learning machines. (3rd ed). Pearson, Ontario.
Khan, S.; H. Rahmani; S. A. A. Shah and M. Bennamoun (2018). A Guide to Convolutional Neural Networks for Computer Vision. Synthesis Lectures on Computer Vision, v.8 n.1, p. 1–207.DOI: 10.2200/s00822ed1v01y201712cov015
Koch, C.; K. Georgieva; V. Kasireddy; B. Akinci and P. Fieguth (2015). A review on computer vision based defect detection and condition assessment of concrete and asphalt civil infrastructure. Advanced Engineering Informatics, v.29, n.2, p. 196–210. DOI: 10.1016/j.aei.2015.01.008
Li, S.; Y. Cao and H. Cai. (2017). Automatic Pavement-Crack Detection and Segmentation Based on Steerable Matched Filtering and an Active Contour Model. Journal of Computing in Civil Engineering, v.31, n.5, doi:10.1061/(ASCE)CP.1943-5487.0000695
Ong, G. P.; S. Noureldin and K. Sinha (2011). Technical report: Automated Pavement Condition Data Collection Quality Control, Quality Assurance, and Reliability. doi:10.5703/1288284314288
Osman, M. K.; M. H. M. Noor; A. Ibrahim; N. M. Tahir; N. M. Yusof and N. Z. Abidin (2019). Deep convolution neural network for crack detection on asphalt pavement. International Conference on Nanomaterials: Science, Engineering and Technology (ICoNSET) 2019. v. 1349, doi:10.1088/1742-6596/1349/1/012020
Paszke, A.; S. Gross; F. Massa; A. Lerer; J. Bradbury; G. Chanan; T. Killeen; Z. Lin; N. Gimelshein; L. Antiga; A. Desmaison; A. Kopf; E. Yang; Z. DeVito; M. Raison; A. Tejani; S. Chilamkurthy; B. Steiner; L. Fang; J. Bai and S. Chintala (2019). PyTorch: An Imperative Style, High-Performance Deep Learning Library. H. Wallach, H. Larochelle, A. Beygelzimer, F. dtextquotesingle Alché-Buc, E. Fox, & R. Garnett (Eds), Advances in Neural Information Processing Systems 32 (p. 8024–8035). Available at: http://papers.neurips.cc/paper/9015-pytorch-an-imperative-style-high-performance-deep-learning-library.pdf
Pianucci, M. N.; C. S. Pitombo and A. L. Cunha (2019) Previsão da demanda por viagens domiciliares através de método sequencial baseado em população sintética e redes neurais artificiais. v. 27, n.4, p. 1–23. doi:10.14295/transportes.v27i4.1406
Pierce, L. M. and N. D. Weitzel (2019). Automated Pavement Condition Surveys. Automated Pavement Condition Surveys. Transportation Research Board, Washington, D.C. doi:10.17226/25513
Silva, W. R. L. and D. S. Lucena (2018). Concrete Cracks Detection Based on Deep Learning Image Classification. Proceedings, v.2, n.8. doi:10.3390/icem18-05387
Sun, Y.; E. Salari, and E. Chou (2009). Automated pavement distress detection using advanced image processing techniques. Proceedings of 2009 IEEE International Conference on Electro/Information Technology. EIT 2009. p. 373–377. doi:10.1109/EIT.2009.5189645
Zhang, A.; K. C. P. Wang; B. Li; E. Yang; X. Dai; Y. Peng; Y. Fei; Y. Liu; J. Q. Li and C. Chen (2017). Automated Pixel-Level Pavement Crack Detection on 3D Asphalt Surfaces Using a Deep-Learning Network. Computer-Aided Civil and Infrastructure Engineering, v.32, n.10, p. 805–819. doi:10.1111/mice.12297
Zhang, L.; F. Yang; Y. D. Zhang and Y. J. Zhu (2016). Road crack detection using deep convolutional neural network. Proceedings - International Conference on Image Processing (ICIP). p. 3708–3712. doi:10.1109/ICIP.2016.7533052
Downloads
Published
How to Cite
Issue
Section
License
Authors who submit papers for publication by TRANSPORTES agree to the following terms:
- The authors retain the copyright and grant Transportes the right of first publication of the manuscript, without any financial charge, and waive any other remuneration for its publication by ANPET.
- Upon publication by Transportes, the manuscript is automatically licensed under the Creative Commons License CC BY 4.0 license. This license permits the work to be shared with proper attribution to the authors and its original publication in this journal.
- Authors are authorized to enter into additional separate contracts for the non-exclusive distribution of the version of the manuscript published in this journal (e.g., publishing in an institutional repository or as a book chapter), with recognition of the initial publication in this journal, provided that such a contract does not imply an endorsement of the content of the manuscript or the new medium by ANPET.
- Authors are permitted and encouraged to publish and distribute their work online (e.g., in institutional repositories or on their personal websites) after the editorial process is complete. As Transportes provides open access to all published issues, authors are encouraged to use links to the DOI of their article in these cases.
- Authors guarantee that they have obtained the necessary authorization from their employers for the transfer of rights under this agreement, if these employers hold any copyright over the manuscript. Additionally, authors assume all responsibility for any copyright infringements by these employers, releasing ANPET and Transportes from any responsibility in this regard.
- Authors assume full responsibility for the content of the manuscript, including the necessary and appropriate authorizations for the disclosure of collected data and obtained results, releasing ANPET and Transportes from any responsibility in this regard.
