Research on intelligent recognition algorithm for subsurface road defects
Source abstract
Abstract
In recent years, frequent urban road collapses have led to casualties and severely compromised traffic safety. Currently, 3D Ground Penetrating Radar (GPR) technology is recognised as the most scientific and effective method for rapidly screening defects to prevent and mitigate the risk of road collapses. However, interpreting radar images still faces challenges such as heavy reliance on manual expertise, large data processing requirements, and low efficiency. Therefore, exploring AI-driven methods for the intelligent detection of underground road defects is of significant importance. This study enhances the original YOLOv5 algorithm by introducing a dedicated small-target detection layer and replacing the C3 backbone with the lightweight GhostNet architecture using Python platform, aiming to improve the precision and efficiency of detecting small-scale defects. Experimental results demonstrate that the optimised YOLOv5 achieves notable improvements in small-target detection, including increased precision, recall, accuracy, F1-score, and faster processing speed. In practical engineering applications, the proposed method achieves full detection of defect location with no missed cases. The rate of multiple identifications is approximately 22%, and the detection time is about 46% of the average time required for manual interpretation, meeting the requirements of real-world engineering applications.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Infrastructure Maintenance and Monitoring · Industrial Vision Systems and Defect Detection · Image Processing and 3D Reconstruction
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Guangdong Provincial Academy of Building Research Group Co. Ltd
Funding
Guangdong Provincial Academy of Building Research Group Co., Ltd
References (25)
Show source-provided references
1. Xia Y, Yang F, Qiao X, and Xu X. Ground penetrating radar system applied in the underground concealed object detection. In: Proceedings of the 15th International Conference on Ground Penetrating Radar [Internet]; Brussels: IEEE; 2014. p. 811–814. [cited 2025 May 12]. Available from: https://ieeexplore.ieee.org/document/6970538
2. 10.4236/jtts.2019.9201510.4236/jtts.2019.92015 ↗
3. 10.1016/j.rcim.2020.10196310.1016/j.rcim.2020.101963 ↗
4. 10.1016/j.neucom.2020.03.04810.1016/j.neucom.2020.03.048 ↗
5. Girshick R Donahue J Darrell T et al. Rich feature hierarchies for accurate object detection and semantic segmentation. 2014 [cited 2025 Apr 28]. Available from: http://arxiv.org/abs/1311.252410.1109/CVPR.2014.81 ↗
6. 10.1016/C2009-0-16511-X10.1016/C2009-0-16511-X ↗
7. 10.1016/j.ins.2020.02.06710.1016/j.ins.2020.02.067 ↗
8. 10.1016/j.joca.2020.03.00610.1016/j.joca.2020.03.006 ↗
9. Girshick R. Fast R-CNN [Internet]. 2015 [cited 2025 May 12]. Available from: http://arxiv.org/abs/1504.08083
10. Redmon J, Divvala S, Girshick R, and Farhadi A. You only look Once: unified, real-time object detection. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) [Internet]; Las Vegas, (NV), USA: IEEE; 2016. p. 779–788. Available from cited 2025 Apr 28. Available from http://ieeexplore.ieee.org/document/7780460/10.1109/CVPR.2016.91 ↗
11. Redmon J Farhadi A. YOLOv3: An Incremental Improvement [Internet]. 2018 [cited 2025 May 28]. Available from: http://arxiv.org/abs/1804.02767
12. Bochkovskiy A Wang C-Y Liao H-Y. YOLOv4: optimal speed and accuracy of object detection [Internet]. 2020 [cited 2025 Apr 28]. Available from: http://arxiv.org/abs/2004.10934
13. 10.3390/fire608029110.3390/fire6080291 ↗
14. 10.3390/s2022647610.3390/s20226476 ↗
15. 10.3390/rs1306108110.3390/rs13061081 ↗
16. 10.3390/rs1408189510.3390/rs14081895 ↗
17. Tan M Pang R Le QV. EfficientDet: scalable and efficient object detection [Internet]. 2020 [cited 2025 May 9]. Available from: http://arxiv.org/abs/1911.0907010.1109/CVPR42600.2020.01079 ↗
18. Han K Wang Y Tian Q et al. GhostNet: more features from cheap operations [Internet]. 2020 [cited 2025 Apr 26]. Available from: http://arxiv.org/abs/1911.11907
19. 10.1109/TPAMI.2015.238982410.1109/TPAMI.2015.2389824 ↗
20. He K, Zhang X, Ren S, et al. Deep residual learning for image recognition. In: 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) [Internet]; Las Vegas (NV) USA: IEEE; 2016. p. 770–778. Available from cited 2025 May 9]. Available from: http://ieeexplore.ieee.org/document/7780459/10.1109/CVPR.2016.90 ↗
21. Liu S, Qi L, Qin H, et al. Path Aggregation network for instance segmentation. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition [Internet]; Salt Lake City, UT. IEEE; 2018. p. 8759–8768. [cited 2025 May 9]. Available from: https://ieeexplore.ieee.org/document/8579011/10.1109/CVPR.2018.00913 ↗
22. 10.1016/j.conbuildmat.2020.12194910.1016/j.conbuildmat.2020.121949 ↗
23. 10.1007/s13349-024-00818-510.1007/s13349-024-00818-5 ↗
24. 10.1016/j.jappgeo.2024.10549110.1016/j.jappgeo.2024.105491 ↗
25. 10.48550/arXiv.2507.1108110.48550/arXiv.2507.11081 ↗