Abstract
Steel is an indispensable material in modern industry, and its surface quality directly affects the performance and service life of products. To address problems of insufficient feature extraction capability, weak detection of small defects, and blurred target contours that lead to degraded edge information in steel surface defect detection, this paper proposes a novel edge-enhanced dual-branch steel surface defect recognition model, DEENet. First, a dual-encoder module based on CNN and Transformer is designed to extract image features and enhance the feature extraction capacity of the backbone network. Second, a Dual Channel Fusion module is introduced to perform cross-enhancement between the local features captured by the CNN and the global semantic features modeled by the Transformer, achieving feature complementarity and improving the detection accuracy for small defects. Finally, an edge enhancement module, C2f_EEM, is designed to highlight gradient differences between defective and normal regions through differential operations, thereby strengthening contour information and improving the model’s sensitivity to defect edges. Experimental results on the NEU-DET dataset show that, compared with other algorithms, DEENet achieves a superior mean Average Precision (mAP) of 81.4%, enabling more accurate detection of steel surface defects and providing valuable reference for defect inspection in real-world production scenarios.
Data availability
The publicly available dataset utilized in this research can be accessed via the following link: https://www.kaggle.com/datasets/kaustubhdikshit/neu-surface-defect-database.
References
Li, Q. Y., Li, W. G. & Tian, Z. Q. Surface defect recognition of hot-rolled strip based on self-distillation. China Metall. 34, 126–133. https://doi.org/10.13228/j.boyuan.issn1006-9356.20230527 (2024).
Zhang, C. J. et al. Morphological characteristics and formation causes of mountain-shaped surface cracks on Q355 hot-rolled plate. China Metall. 33, 55–63. https://doi.org/10.13228/j.boyuan.issn1006-9356.20220616 (2023).
Zhang, W. J. et al. Formation mechanism and process optimization of surface defects in hot-dip aluminum-zinc coated strip. Iron Steel 58, 87–95. https://doi.org/10.13228/j.boyuan.issn0449-749x.20220595 (2023).
Kang, Y. L. Progress and prospects of China’s rolling technology during the 13th Five-Year Plan. Iron Steel 56, 1–15. https://doi.org/10.13228/j.boyuan.issn0449-749x.20210324 (2021).
Wang, B., Hu, C., Wang, J., Wang, Y. Y. & Li, N. A novel eddy current testing method and its detection performance. China Metall. 31, 50–54. https://doi.org/10.13228/j.boyuan.issn1006-9356.20200350 (2021).
Chesnokova, A. A., Kalayeva, S. Z. & Ivanova, V. A. Development of a flaw detection material for the magnetic particle method. In J. Phys. Conf. Ser. 881, 012022 (2017).
Feng, X., Gao, X. & Luo, L. X-SDD: A new benchmark for hot rolled steel strip surface defects detection. Symmetry 13, 706 (2021).
Jani, M., Fayyad, J., Al-Younes, Y. & Najjaran, H. Model compression methods for YOLOv5: A review. arXiv 2307.11904 (2023).
Wang, C. Y., Bochkovskiy, A. & Liao, H. Y. M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. In: Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 7464–7475 (2023).
Varghese, R. & Sambath, M. Yolov8: A novel object detection algorithm with enhanced performance and robustness. In Proc. Int. Conf. Adv. Data Eng. Intell. Comput. Syst. (ADICS), 1–6 (2024).
Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. You only look once: Unified, real-time object detection. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit., 779–788 (2016).
Redmon, J. & Farhadi, A. YOLO9000: Better, faster, stronger. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit., 7263–7271 (2017).
Redmon, J. & Farhadi, A. Yolov3: An incremental improvement. arXiv 1804.02767 (2018).
Lu, J., Zhu, M., Ma, X. & Wu, K. Steel strip surface defect detection method based on improved YOLOv5s. Biomimetics 9, 28 (2024).
Lv, B. et al. Research on surface defect detection of strip steel based on improved YOLOv7. Sensors 24, 2667 (2024).
Zhang, W. K. & Liu, J. Steel surface defect detection based on improved YOLOv8s. J. Beijing Inf Sci. Technol. Univ. (Nat. Sci. Ed.) 38, 33–40. https://doi.org/10.16508/j.cnki.11-5866/n.2023.06.005 (2023).
Zhang, L., Wang, Z., Ma, Y. & Li, G. Steel surface defect detection algorithm based on improved YOLOv10. Sci. Rep. 15, 32827 (2025).
Ren, S., He, K., Girshick, R. & Sun, J. Faster R-CNN: Towards real-time object detection with region proposal networks. Adv. Neural Inf. Process. Syst. 28 (2015).
Dai, J., Li, Y., He, K. & Sun, J. R-FCN: Object detection via region-based fully convolutional networks. Adv. Neural Inf. Process. Syst. 29 (2016).
He, K., Gkioxari, G., Dollár, P. & Girshick, R. Mask R-CNN. In Proc. IEEE Int. Conf. Comput. Vis., 2961–2969 (2017).
Xia, B., Luo, H. & Shi, S. Improved Faster R-CNN based surface defect detection algorithm for plates. Comput. Intell. Neurosci. 2022, 3248722 (2022).
Liu, W., Weng, Y. S., Xiao, J. Q. & Xia, Y. Surface defect detection of strip steel using improved Mask R-CNN algorithm. Comput. Eng. Appl. 57, 235–242 (2021).
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C. Y. & Berg, A. C. SSD: Single shot multibox detector. In Eur. Conf. Comput. Vis., 21–37 (2016).
Liu, X. & Gao, J. Surface defect detection method of hot rolling strip based on improved SSD model. In Int. Conf. Database Syst. Adv. Appl., 209–222 (2021).
Wang, L., Liu, X., Ma, J., Su, W. & Li, H. Real-time steel surface defect detection with improved multi-scale YOLO-v5. Processes 11, 1357 (2023).
Song, X., Cao, S., Zhang, J. & Hou, Z. Steel surface defect detection algorithm based on YOLOv8. Electronics 13, 988 (2024).
Zhou, Y. & Zhao, Z. MPA-YOLO: Steel surface defect detection based on improved YOLOv8 framework. Pattern Recognit. 111, 111897 (2025).
He, L., Zheng, L. & Xiong, J. FMV-YOLO: A steel surface defect detection algorithm for real-world scenarios. Electronics 14, 1143 (2025).
Ayon, S. T. K., Siraj, F. M. & Uddin, J. Steel surface defect detection using learnable memory vision transformer. Comput. Mater. Contin. 82, 1 (2025).
Xu, W. Application of self-supervised learning in steel surface defect detection. J. Mater. Inform. 5, N-A (2025).
Wang, Q., Dong, H. & Huang, H. Swin-Transformer-YOLOv5 for lightweight hot-rolled steel strips surface defect detection algorithm. PLoS ONE 19, e0292082. https://doi.org/10.1371/journal.pone.0292082 (2024).
Liu, H. et al. CGTD-Net: Channel-wise global Transformer-based dual-branch network for industrial strip steel surface defect detection. IEEE Sens. J. 24, 4863–4873. https://doi.org/10.1109/JSEN.2023.3346470 (2024).
Lv, Z. et al. Steel surface defect detection based on MobileViTv2 and YOLOv8. J. Supercomput. 80, 18919–18941. https://doi.org/10.1007/s11227-024-06248-w (2024).
Vasan, V., Sridharan, N. V., Vaithiyanathan, S. & Aghaei, M. Detection and classification of surface defects on hot-rolled steel using vision transformers. Heliyon 10, e38498. https://doi.org/10.1016/j.heliyon.2024.e38498 (2024).
Mao, H. & Gong, Y. Steel surface defect detection based on the lightweight improved RT-DETR algorithm. J. Real-Time Image Process. 22, 28. https://doi.org/10.1007/s11554-024-01611-9 (2025).
Wu, S. et al. SH-DETR: Enhancing steel surface defect detection and classification with an improved transformer architecture. PLoS ONE 20, e0334048. https://doi.org/10.1371/journal.pone.0334048 (2025).
Luo, H. & Xia, Y. Improved RT-DETR approach for steel surface defect identification. Int. J. Sci. Eng. Appl. 13, 11–16 (2024).
Zhang, X., Song, Y., Song, T., Yang, D., Ye, Y., Zhou, J. & Zhang, L. AKConv: Convolutional kernel with arbitrary sampled shapes and arbitrary number of parameters. arXiv 2311.11587 (2023).
Zhao, W., Chen, F., Huang, H., Li, D. & Cheng, W. A new steel defect detection algorithm based on deep learning. Comput. Intell. Neurosci. 2021, 5592878 (2021).
Zhao, Y., Lv, W., Xu, S., Wei, J., Wang, G., Dang, Q. et al. DETRs beat YOLOs on real-time object detection. In Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit., 16965–16974 (2024).
Qian, K., Zou, L., Wang, Z. & Wang, W. Metallic surface defect recognition network based on global feature aggregation and dual context decoupled head. Appl. Soft Comput. 158, 111589 (2024).
Zheng, H., Chen, X., Cheng, H., Du, Y. & Jiang, Z. MD-YOLO: Surface defect detector for industrial complex environments. Opt. Laser Eng. 178, 108170 (2024).
Funding
Partial financial support was received from the Guangdong Provincial Science and Technology Innovation Strategy Special Fund—“Climbing Program” (Grant No. pdjh2023b0680), the 2025 Guangdong Higher Education Key Discipline Construction Project for Research Capacity Enhancement (Grant No. 2025ZDJS096), the University-level Teaching and Education Reform Project (Grant No. 2023XGXK008), the University-level Student Sustainable Science and Technology Innovation Project “Smart Campus Integrated Service Platform for Student Affairs” and “College Student Labor Education Management System Based on Multi-dimensional Data Collection and Fusion”, as well as the Guangzhou Institute of Technology Fund (Grant No. XJ2025010001).
Author information
Authors and Affiliations
Contributions
W.P. designed the model architecture and overall experimental strategy and also contributed to the experiments. R.Z. and Y.Li. performed the experiments and assisted with analysis. J.H. and W.Z. prepared the figures and visualizations. T.L. and Y.Liu. (corresponding authors) organized the research narrative, supervised the study, and revised the manuscript. All authors discussed the results, reviewed, and approved the final manuscript.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Pan, W., Zhong, R., Huang, J. et al. DEENet: an edge-enhanced CNN–Transformer dual-encoder model for steel surface defect detection. Sci Rep (2026). https://doi.org/10.1038/s41598-026-36390-9
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41598-026-36390-9
