RCL-YOLO: A Reparameterized Network for Strip Steel Surface Defect Detection
DOI:
https://doi.org/10.54097/924jb628Keywords:
Strip Steel Surface Defect Detection, YOLOv8n, RepStem, C2f-RMBC, Lighten Cross-AttentionAbstract
Strip steel surface defect detection is essential for intelligent quality control in modern steel manufacturing. Defects such as crazing, inclusion, patches, pitted surface, rolled-in scale, and scratches usually have weak texture, low contrast, irregular morphology, and large scale variation, which makes accurate real-time detection difficult. This paper proposes RCL-YOLO,a lightweight strip steel surface defect detector based on YOLOv8n, to deal with these issues. A RepStem-based downsampling module is introduced to strengthen shallow local feature extraction and blueuce the loss of fine defect details during spatial resolution blueuction. C2f-RMBC (Reparameterized Mobile Bottleneck Convolution C2f) is then built to improve local structural modeling for weak and irregular defect patterns while keeping an efficient reparameterized inference structure. In addition, a Lighten Cross-Attention Neck is used to improve the interaction between high-resolution texture features and low-resolution semantic features, which helps multi-scale defect representation and blueuces background interference. Experiments are conducted on the NEU-DET surface defect dataset. The results show that RCL-YOLO achieves an mAP@0.5 of 78.20%, surpassing the YOLOv8n baseline by 2.54 percentage points. RCL-YOLO uses only 2.84M parameters and 7.6 GFLOPs, and its end-to-end inference speed reaches 129.05 FPS. Taken together, these results indicate that the proposed method maintains a good balance between detection accuracy,a lightweight structure, and real-time performance, which makes it suitable for industrial strip steel surface defect inspection.
Downloads
References
[1] Wang, J., Chen, T., Xu, X., Zhao, L., Yuan, D., Du, Y., Guo, X., & Chen, N. (2025). An improved YOLOv8 model for strip steel surface defect detection. Applied Sciences, 15(1), 52. https: //doi.org/10.3390/app15010052.
[2] Kou, X., Liu, S., Cheng, K., & Qian, Y. (2021). Development of a YOLO-V3-based model for detecting defects on steel strip surface. Measurement, 182, 109454. https://doi. org/ 10. 1016/ j. measurement.2021.109454.
[3] Zheng, X., Zheng, S., Kong, Y., & Chen, J. (2021). Recent advances in surface defect inspection of industrial products using deep learning techniques. The International Journal of Advanced Manufacturing Technology, 113, 35–58. https://doi. org/ 10.1007/s00170-021-06582-2.
[4] Wen, X., Shan, J., He, Y., & Song, K. (2023). Steel surface defect recognition: A survey. Coatings, 13(1), 17. https:// doi. org/ 10.3390/coatings13010017.
[5] Girshick, R., Donahue, J., Darrell, T., & Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 580–587). https:// doi. org/10.1109/CVPR.2014.81.
[6] Girshick, R. (2015). Fast R-CNN. In Proceedings of the IEEE international conference on computer vision (pp. 1440–1448). https://doi.org/10.1109/ICCV.2015.167.
[7] Ren, S., He, K., Girshick, R., & Sun, J. (2017). Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39(6), 1137–1149. https://doi. org/10. 1109/ TPAMI. 2016.2577403.
[8] Cai, Z., & Vasconcelos, N. (2018). Cascade R-CNN: Delving into high quality object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 6154–6162). https://doi.org/ 10.1109/ CVPR. 2018.00644.
[9] Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C. Y., & Berg, A. C. (2016). SSD: Single shot MultiBox detector. In Proceedings of the European conference on computer vision (pp. 21–37). https://doi.org/10.1007/978-3-319-46448-0_2.
[10] Tan, M., Pang, R., & Le, Q. V. (2020). EfficientDet: Scalable and efficient object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 10781–10790). https://doi.org/10. 1109/CVPR42600. 2020. 01080.
[11] Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779–788). https://doi.org/ 10.1109/ CVPR. 2016.90.
[12] Li, C., Li, L., Jiang, H., Weng, K., Geng, Y., Li, L., et al. (2022). YOLOv6: A single-stage object detection framework for industrial applications. arXiv preprint arXiv:2209.02976.
[13] Wang, C. Y., Bochkovskiy, A., & Liao, H. Y. M. (2023). YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 7464–7475). https://doi.org/10.1109/CVPR52729.2023.00717.
[14] Jocher, G., Chaurasia, J., Qiu, A., & Stoken, K. (2023). Ultralytics YOLOv8 [Computer software]. GitHub. https:// github. com/ultralytics/ultralytics.
[15] Lin, T. Y., Dollár, P., Girshick, R., He, K., Hariharan, B., & Belongie, S. (2017). Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 2117–2125). https:// doi. org/ 10.1109/CVPR.2017.374.
[16] Liu, S., Qi, L., Qin, H., Shi, J., & Jia, J. (2018). Path aggregation network for instance segmentation. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 8759–8768). https://doi.org/10. 1109/ CVPR. 2018. 00913.
[17] Song, K., & Yan, Y. (2013). A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects. Applied Surface Science, 285, 858–864. https:// doi. org/ 10.1016/j.apsusc.2013.08.132.
[18] Vasu, P. K. A., Gabriel, J., Zhu, J., Tuzel, O., & Ranjan, A. (2023). FastViT: A fast hybrid vision transformer using structural reparameterization. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 5785–5795). https://doi.org/10.1109/ICCV51070.2023.00528.
[19] Ding, X., Zhang, X., Ma, N., Han, J., Ding, G., & Sun, J. (2021). RepVGG: Making VGG-style ConvNets great again. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 13733–13742). https://doi.org/10. 1109/ CVPR46437.2021.01350.
[20] Wang, Y., Li, Y., Wang, G., & Liu, X. (2024). PlainUSR: Chasing faster ConvNet for efficient super-resolution. arXiv preprint arXiv:2409.13435.
[21] Howard, A., Sandler, M., Chu, G., Chen, L. C., Chen, B., Tan, M., et al. (2019). Searching for MobileNetV3. In Proceedings of the IEEE/CVF international conference on computer vision (pp. 1314–1324). https://doi.org/10.1109/ICCV.2019.01404.
[22] Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., et al. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30.
[23] Woo, S., Park, J., Lee, J. Y., & Kweon, I. S. (2018). CBAM: Convolutional block attention module. In Proceedings of the European conference on computer vision (pp. 3–19). https:// doi. org/10.1007/978-3-030-01234-2_1.
[24] Yan, Q., Feng, Y., Zhang, C., Pang, G., Shi, K., Wu, P., et al. (2025). HVI: A new color space for low-light image enhancement. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition.
[25] Wang, C. Y., Yeh, I. H., & Liao, H. Y. M. (2024). YOLOv9: Learning what you want to learn using programmable gradient information. arXiv preprint arXiv:2402.13616.
[26] Wang, A., Chen, H., Liu, L., Chen, K., Lin, Z., Han, J., et al. (2024). YOLOv10: Real-time end-to-end object detection. arXiv preprint arXiv:2405.14458.
[27] Jocher, G., Chaurasia, J., Qiu, A., & Stoken, K. (2024). Ultralytics YOLOv11 [Computer software]. GitHub. https:// github. com/ultralytics/ultralytics.
[28] Tian, Y., Ye, Q., & Doermann, D. (2025). YOLOv12: Attention-centric real-time object detectors. arXiv preprint arXiv:2502.12524.
[29] Jiang, J., Su, Y., & Song, J. (2026). A YOLOv8-based network for surface defect detection in strip steel. Journal of Computer and Communications, 14, 106–129. https://doi.org/ 10. 4236/ jcc. 2026.141009.
[30] Wu, Y., Chen, R., Li, Z., Ye, M., & Dai, M. (2024). SDD-YOLO: A lightweight, high-generalization methodology for real-time detection of strip surface defects. Metals, 14(6), 650. https://doi.org/10.3390/met14060650.
[31] Chu, Y., Yu, X., & Rong, X. (2024). A lightweight strip steel surface defect detection network based on improved YOLOv8. Sensors, 24(19), 6495. https://doi.org/10.3390/s24196495.
[32] Zhao, Y., Lv, W., Xu, S., Wei, J., Wang, G., Dang, Q., Liu, Y., & Chen, J. (2024). DETRs beat YOLOs on real-time object detection. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (pp. 16965–16974). https://doi.org/10.1109/CVPR52729.2024.01617.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Frontiers in Computing and Intelligent Systems
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
