Identificação e Classificação de Patologias em Pavimentos Rodoviários Flexíveis e Rígidos Utilizando YOLOv8

• Danillo Marcus Farias Marinho do Monte IME
• Ronaldo Ribeiro Goldschmidt IME
• Marcos dos Santos IME

DOI: https://doi.org/10.5753/webmedia.2025.16131

Resumo

Road pavement infrastructure plays a fundamental role in mobility and economic development, with Brazil having over 213,500 km of paved roads. However, conventional inspection methods for pavement pathologies are predominantly manual, subjective, costly, and pose risks to evaluators. Recent advances in Deep Learning, particularly Convolutional Neural Networks (CNNs), have emerged as promising alternatives for automatic defect identification through computer vision. A systematic literature review identified 209 studies published between 2019-2024, revealing that most existing work focuses exclusively on flexible pavements, leaving gaps in rigid pavement pathology detection. Hence, this study proposes a hybrid model based on the YOLOv8 neural network for automatic identification and classification of pathologies in both flexible and rigid pavements. The proposed model incorporates a BACKbone + CSPNet architecture that reduces computational costs without compromising performance. The methodology involved collecting approximately 4,000 images using a customized approach with Go-Pro cameras equipped with GPS across different Brazilian regions under various climatic conditions, followed by manual annotation of 1,000 images for training. The model achieved promising results with mAP50 of approximately 90%, precision up to 96%, and mean Pixel Accuracy of 88%. An extra validation was conducted on 65 road segments of 1 km each across three Brazilian regions, where 51 segments with various pathologies were successfully identified and classified.

Palavras-chave: Computer Vision, YOLOv8, Road Pavements, Pathology Detection, Convolutional Neural Networks

Referências

S. Amani, M. Reza, and K. Johnson. 2024. Impact of Pavement Condition on Road Safety: A Comprehensive Analysis. Transportation Research Part C: Emerging Technologies 145 (2024), 103–118. DOI: 10.1016/j.trc.2024.103118

Deeksha Arya, Hiroshi Maeda, Sanjay Kumar Ghosh, Durga Toshniwal, Alexander Mraz, Takehiro Kashiyama, and Yoshihide Sekimoto. 2021. Deep Learningbased Road Damage Detection and Classification for Multiple Countries. In Automation in Construction, Vol. 132. 103935. DOI: 10.1016/j.autcon.2021.103935

L. B. Bernucci, L. M. G. Motta, J. A. P. Ceratti, and J. B. Soares. 2018. Pavimentação Asfáltica: Formação Básica para Engenheiros (3 ed.). PETROBRAS/ABEDA, Rio de Janeiro.

DNIT. 2003. DNIT 005/2003-TER: Defeitos nos Pavimentos Flexíveis e Semi-rígidos - Terminologia. Technical Report. DNIT, Rio de Janeiro.

DNIT. 2003. DNIT 008/2003-PRO: Levantamento Visual Contínuo para Avaliação da Superfície de Pavimentos Flexíveis e Semi-rígidos - Procedimento. Technical Report. DNIT, Rio de Janeiro.

DNIT. 2003. DNIT 010/2003-PRO: Levantamento Visual Contínuo para Avaliação da Superfície de Pavimentos Rígidos - Procedimento. Technical Report. DNIT, Rio de Janeiro.

DNIT. 2023. Anuário Estatístico dos Transportes Terrestres 2023.

Markus Eisenbach, Ronny Stricker, Daniel Seichter, Karl Amende, Klaus Debes, Maximilian Sesselmann, Dirk Ebersbach, Ulrike Stoeckert, and Horst-Michael Gross. 2017. How to Get Pavement Distress Detection Ready for Deep Learning? A Systematic Approach. In International Joint Conference on Neural Networks. Anchorage, 2039–2047. DOI: 10.1109/IJCNN.2017.7966101

Mark Everingham, Luc Van Gool, Christopher K. I. Williams, John Winn, and Andrew Zisserman. 2010. The Pascal Visual Object Classes (VOC) Challenge. International Journal of Computer Vision 88, 2 (2010), 303–338. DOI: 10.1007/s11263-009-0275-4

Tom Fawcett. 2006. An introduction to ROC analysis. Pattern Recognition Letters 27, 8 (2006), 861–874. DOI: 10.1016/j.patrec.2005.10.010

Ian Goodfellow, Yoshua Bengio, and Aaron Courville. 2016. Deep Learning. MIT Press, Cambridge.

Kasthurirangan Gopalakrishnan, Siddhartha Kumar Khaitan, Alok Choudhary, and Ankit Agrawal. 2017. Deep Convolutional Neural Networks with transfer learning for computer vision-based data-driven pavement distress detection. Construction and Building Materials 157 (2017), 322–330. DOI: 10.1016/j.conbuildmat.2017.09.110

Jiaqi Huyan,Wei Li, Susan Tighe, Zhengchao Xu, and Jianlong Zhai. 2020. CrackUNet: A novel deep convolutional neural network for pixelwise pavement crack detection. Structural Control and Health Monitoring 27, 8 (2020), e2551. DOI: 10.1002/stc.2551

Sergey Ioffe and Christian Szegedy. 2015. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. Proceedings of the 32nd International Conference on Machine Learning (2015), 448–456.

Yann LeCun, Yoshua Bengio, and Geoffrey Hinton. 2015. Deep Learning. Nature 521, 7553 (2015), 436–444. DOI: 10.1038/nature14539

H. Maeda, Y. Sekimoto, T. Seto, T. Kashiyama, and H. Omata. 2018. Road Damage Detection Using Deep Neural Networks with Images Captured Through a Smartphone. Computer-Aided Civil and Infrastructure Engineering 33, 6 (2018), 498–513. DOI: 10.1111/mice.12387

Hamed Majidifard, Yaw Adu-Gyamfi, and William G. Buttlar. 2020. Deep machine learning approach to develop a new asphalt pavement condition index. Construction and Building Materials 247 (2020), 118513. DOI: 10.1016/j.conbuildmat.2020.118513

R. Mkwata and E. Chong. 2022. Effect of pavement surface conditions on road traffic accident - a review. E3S Web of Conferences 347 (2022), 01017.

Danillo Marcus Farias Marinho do Monte. 2025. Identificação e Classificação de Patologias em Pavimentos Rodoviários: Uma Abordagem Baseada em Visão Computacional. Dissertação (Mestrado em Sistemas e Computação). Instituto Militar de Engenharia, Rio de Janeiro.

E. Onler and N. Köycü. 2024. Wheat powdery mildew detection with yolov8 object detection model. Applied Sciences 14, 16 (2024), 7052. DOI: 10.3390/app14167052

Rafael Padilla, Sergio L. Netto, and Eduardo A. B. da Silva. 2020. A Survey on Performance Metrics for Object-Detection Algorithms. International Conference on Systems, Signals and Image Processing (2020), 237–242. DOI: 10.1109/IWSSIP48289.2020.9145130

Joseph Redmon, Santosh Divvala, Ross Girshick, and Ali Farhadi. 2016. You Only Look Once: Unified, Real-Time Object Detection. In IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, 779–788. DOI: 10.1109/CVPR.2016.91

P. Sadeghi and A. Golroo. 2024. Investigating the impact of pavement condition and weather characteristics on road accidents (review paper). International Journal of Crashworthiness 29, 6 (2024), 973–989.

W. R. L. Silva and D. S. Lucena. 2020. Automated Road Crack Detection Using Deep Learning. Engineering Applications of Artificial Intelligence 89 (2020), 103–112. DOI: 10.1016/j.engappai.2019.103112

Ultralytics. 2023. YOLOv8: A New State-of-the-Art Computer Vision Model. [link].

K. C. P. Wang, A. Zhang, J. Q. Li, Y. Fang, C. Chen, and B. Li. 2017. Deep Learning for Asphalt Pavement Cracking Recognition Using Convolutional Neural Network. In Airfield and Highway Pavements Conference. Philadelphia, 166–177. DOI: 10.1061/9780784480922.015

E. J. Yoder and M. W. Witczak. 1975. Principles of Pavement Design (2 ed.). John Wiley & Sons, New York.

L. Zhang, F. Yang, Y. D. Zhang, and Y. J. Zhu. 2016. Road Crack Detection Using Deep Convolutional Neural Network. In IEEE International Conference on Image Processing. Phoenix, 3708–3712. DOI: 10.1109/ICIP.2016.7533052