XIII Međunarodno naučno-stručno savetovanje Ocena stanja, održavanje i sanacija građevinskih objekata (str. 166-175)
АУТОР(И) / AUTHOR(S): Miloš Lukić,
Dejan Gavran, Sanja Fric, Vladan Ilić, Filip Trpčevski, Stefan Vranjevac, Nikola Milovanović
Download Full Pdf 
DOI: 10.46793/SGISXIII.16ML
САЖЕТАК / ABSTRACT:
The lifespan of a structure encompasses its initial construction, followed by its operational phase, and ultimately, its rehabilitation. To accurately determine the optimal timing for rehabilitation, it is essential to continuously monitor the current state of the structure. Applied to road infrastructure, this monitoring process entails the systematic collection of data on pavement damage. This paper provides an analysis of existing models for detecting pavement cracks and introduces a detection method leveraging artificial intelligence techniques, specifically deep neural networks, and transfer learning. During the selection process, various deep neural network architectures, including ResNET 50 and ResNET 34, were evaluated. Comparative analysis revealed that the ResNET 50 network achieved the highest accuracy (Acc) in crack detection.
КЉУЧНЕ РЕЧИ / KEYWORDS:
pavement cracks, flexible pavement, deep learning, transfer learning.
ЛИТЕРАТУРА / REFERENCES:
- Arbel and Hel-Or H. Texture-preserving shadow removal in color images containing curved surfaces. Proc. IEEE Conf. on Computer Vision and Pattern Recognition (CVPR07), 2007, page 1-8.
- Chambon and Moliard J.M. Automatic road pavement assessment with image processing: Review and comparison. International Journal of Geophysics, 2011, page 0-20.
- Finlayson G., Hordley S., Lu , and Drew M. On the removal of shadows from images. IEEE Trans. Pattern Anal. Machine Intell., 2006, 28(1) page 59-68.
- Finlayson G., Drew , and Lu C. Entropy minimization for shadow removal. Internat. J. Comput. Vision, 2009, 85(1), page 35-57.
- Lee B. J. and Lee D. A robust position invariant artificial neural network for digital pavement crack analysis. Proceedings of the TRB Annual Meeting, 2003.
- Lin H., Zhao J.-W., Chen Q.-S., Cai -R., and Zhou P. Eggshell crack detection based on acoustic impulse response and supervised pattern recognition. Czech Journal of Food Sciences- UZEI, 2009, vol. 27, page 393-402.
- Li , Hou X., Yang X., and Dong Y. Automation recognition of pavement surface distress based on support vector machine. Proceedings of the 2009 Second International Conference on Intelligent Networks and Intelligent Systems, 2009.
- Oliveira H. and Correia L. Supervised crack detection and classification in images of road pavement flexible surfaces. Recent Advances in Signal Processing, 2009.
- Raschka S. and Mirjalili Python Machine Learning: Machine Learning and Deep Learning with Python, Scikit-Learn, and TensorFlow, 2nd Edition. Packt Publishing, 2 edition, 2017.
- Zou Q., Cao Y., Li , Mao Q., and Wang S. Cracktree: Automatic crack detection from pavement images. Pattern Recognition Letters, 2011, 33(3), page 227-238.
- Wang , Li H., Wang K., He C., Bai M. Pavement crack detection on geodesic shadow removal with local oriented filter on LOF and improved level set. Construction and Building Materials, 2020, vol. 237.