YOLOv11n-CDL: accurate and lightweight pavement defect detection via enhanced multi-scale attention and feature fusion
DOI: https://doi.org/10.3846/jcem.2026.26166Abstract
Pavement defect detection requires both high accuracy and real-time performance in complex road environments, yet existing lightweight models often struggle with blurred textures, background interference, and small cracks. To address these limitations, this study proposes YOLOv11n-CDL, an enhanced lightweight detector integrating three targeted improvements. First, the ConvSmart module expands the receptive field and strengthens multi-scale feature extraction, improving the representation of defects of varying sizes. Second, a Double-Stage Attention (DSA) mechanism, embedded at the deepest backbone stage, iteratively highlights discriminative crack patterns while suppressing shadows, markings, and texture noise. Third, a P2-level small-object detection path provides high-resolution features that significantly improve sensitivity to fine cracks and micro-potholes. Experiments on IRRDD show that YOLOv11n-CDL achieves 75.3% mAP@0.5 and 44.6% mAP@0.5:0.95, outperforming the baseline by 3.0 and 1.1 percentage points, and exceeding YOLOv8n and YOLOv7-tiny in both precision and recall. Additional results on RDD2022 and low-power devices confirm strong generalization and real-time deployability. These improvements demonstrate that YOLOv11n-CDL offers an effective balance between accuracy, robustness, and efficiency for practical pavement inspection applications.
Keywords:
pavement maintenance, repair, YOLOv11n, attention mechanism, pavement defect detection, multi-scale perception, lightweight convolution module, small object detectionHow to Cite
Share
License
Copyright (c) 2026 The Author(s). Published by Vilnius Gediminas Technical University.
This work is licensed under a Creative Commons Attribution 4.0 International License.
References
Bai, F., Ma, Q. L., & Zhao, M. (2025). AC-YOLO for aerial pavement crack detection. Computer Engineering and Applications, 61(1), 153–164.
Cai, Y., Liang, D., Luo, D., He, X., Yang, X., & Bai, X. (2023). A discrepancy aware framework for robust anomaly detection. IEEE Transactions on Industrial Informatics, 20(3), 3986–3995. https://doi.org/10.1109/TII.2023.3318302
Chen, X.-D., Jiang, N., Dong, L. tian, W., Wu, X., Zeng, P., & Yang, F. (2019). Automatic identification of asphalt pavement diseases in plateau mountainous areas based on YOLO deep learning model. Highway Transportation Technology (Applied Technology Edition), 15(11), 75–78.
Chen, X.-D., Ai, D.-H., Zhang, J.-C., Cai, H.-Y., & Cui, K.-R. (2020). Pavement crack detection method based on Gabor filtering fused with convolutional neural networks. Chinese Optics, 13(6), 1293–1301. https://doi.org/10.37188/CO.2020-0041
Ding, K., Ding, Z., Zhang, Z., Yuan, M., Ma, G., & Lv, G. (2024). SCD-YOLO: A novel object detection method for efficient road crack detection. Multimedia Systems, 30(6), Article 351. https://doi.org/10.1007/s00530-024-01538-y
Dong, H. Z., Lin, S. X., & She, Y. N. (2025a). Research progress on YOLO detection technology for traffic targets. Journal of Zhejiang University (Engineering Science), 59(2), 249–260.
Du, Y., Pan, N., Xu, Z., Zhou, Y., & Chen, Y. (2021). Pavement distress detection and classification based on YOLO network. International Journal of Pavement Engineering, 22(13), 1659–1672. https://doi.org/10.1080/10298436.2020.1714047
Duan, Z. X., He, Y., Zhang, X., & Gao, J. (2024). Pavement crack detection method based on effective feature extraction and cascade optimization. Journal of Computer-Aided Design and Graphics, 36(12), 2020–2028. https://doi.org/10.3724/SP.J.1089.2024.20129
Girshick, R. (2015). Fast R-CNN. In Proceedings of 2015 IEEE International Conference on Computer Vision (pp. 1440–1448), Santiago, Chile. IEEE. https://doi.org/10.1109/ICCV.2015.169
He, T. J., & Li, H. E. (2024). Pavement disease detection model based on improved YOLOv5. China Civil Engineering Journal, 57(2), 96–106.
Hu, J., Shen, L., & Sun, G. (2020). Squeeze-and-excitation networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 42(8), 2011–2023. https://doi.org/10.1109/TPAMI.2019.2913372
Hu, X. W., Yan, Y. X., Wang, D. W., & Zhang, Y. H. (2024). Lightweight pavement disease detection method based on YOLOM algorithm. China Journal of Highway and Transport, 37(12), 381–391.
Lang, H., Wen, T., Lu, J., Ding, S., & Chen, S. (2021). 3D pavement crack defect detection method based on deep learning. Journal of Southeast University (Natural Science Edition), 51(1), 53–60.
Miao, Y., Zhang, Z., Wang, H., Dai, W., Zhao, Z., Wang, X., Yang, C., & Shi, Y. (2023). Pavement leaf detection method based on AC-YOLO. Control and Decision, 38(7), 1878–1886.
Mnih, V., Heess, N., Graves, A., & Kavukcuoglu, K. (2014). Recurrent models of visual attention (arXiv:1406.6247). https://doi.org/10.48550/arXiv.1406.6247
Qin, L., Tan, Z. F., Lei, G. P., & Chen, Q. (2025). EMF-YOLO: Lightweight multi-scale feature extraction pavement defect detection algorithm. Computer Engineering and Applications. http://kns.cnki.net/kcms/detail/11.2127.tp.20250321.1618.010.html
Qiu, Q., & Lau, D. (2023). Real-time detection of cracks in tiled sidewalks using YOLO-based method applied to UAV images. Automation in Construction, 147, Article 104745. https://doi.org/10.1016/j.autcon.2023.104745
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016). You only look once: Unified, real-time object detection. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 779–788), Las Vegas, NV, USA. IEEE. https://doi.org/10.1109/CVPR.2016.91
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.2016.2577031
Shen, S. Y., Hua, B., & Huang, R. W. (2023). Research on improved YOLOv5 pavement crack detection model. Electronic Measurement Technology, 46(21), 132–142.
Sun, Z. Y., Pei, L. L., Li, W., et al. (2020). Pavement sealing crack detection method based on improved Faster R-CNN. Journal of South China University of Technology (Natural Science Edition), 48(2), 84–93.
Ultralytics. (2025). YOLOv11: Ultralytics YOLO models documentation. https://docs.ultralytics.com/models/yolo11/
Wang, B. X., Bai, S. X., & Zhao, W. G. (2022). Pavement crack defect visual detection method based on feature enhancement learning. Journal of Railway Science and Engineering, 19(7), 1927–1935.
Wang, G.-Q., Zhang, C.-Z., Chen, M.-S., Lin, Y.-C., Tan, X.-H., Liang, P., Kang, Y.-X., Zeng, W.-D., & Wang, Q. (2023). YOLO-MSAPF: Multiscale alignment fusion with parallel feature filtering model for high accuracy weld defect detection. IEEE Transactions on Instrumentation and Measurement, 72, Article 5022914. https://doi.org/10.1109/TIM.2023.3302372
Wang, A., Chen, H., Liu, L., Li, M., & Zhao, X. (2024a). YOLOv10: Real-time end-to-end object detection (arXiv:2405.14458). https://doi.org/10.48550/arXiv.2405.14458
Wang, G., Chen, M., Lin, Y. C., Tan, X., Zhang, C., Yao, W., Gao, B., Li, K., Li, Z., & Zeng, W. (2024b). Efficient multi-branch dynamic fusion network for super-resolution of industrial component image. Displays, 82, Article 102633. https://doi.org/10.1016/j.displa.2023.102633
Wang, G.-Q., Zhang, C.-Z., Chen, M.-S., Lin, Y. C., Tan, X.-H., Kang, Y.-X., Wang, Q., Zeng, W.-D., & Zhao, W.-W. (2024c). A high-accuracy and lightweight detector based on a graph convolution network for strip surface defect detection. Advanced Engineering Informatics, 59, Article 102280. https://doi.org/10.1016/j.aei.2023.102280
Wang, X. Q., Gao, H. B., & Jia, Z. M. (2024d). Pavement defect detection algorithm based on improved YOLOv8. Computer Engineering and Applications, 60(17), 179–190.
Xing, Y., Han, X., Pan, X., Wang, H., & Li, C. (2024). EMG-YOLO: Road crack detection algorithm for edge computing devices. Frontiers in Neurorobotics, 18, Article 1423738. https://doi.org/10.3389/fnbot.2024.1423738
Xue, M., Chen, M., Peng, D., & Chen, S. (2021). One spatio-temporal sharpening attention mechanism for light-weight YOLO models based on sharpening spatial attention. Sensors, 21(23), Article 7949. https://doi.org/10.3390/s21237949
Yao, J. L., Cheng, G., & Wan, F. (2024). Lightweight bearing defect detection algorithm based on improved YOLOv8. Computer Engineering and Applications, 60(21), 205–214.
Zhang, F., Zhu, H., Cen, Y., Kan, S., Zhang, L., Vadakkepat, P., & Lee, T. H. (2024a). Low-shot unsupervised visual anomaly detection via sparse feature representation. IEEE Transactions on Neural Networks and Learning Systems, 36(85), 7903–7917. https://doi.org/10.1109/TNNLS.2024.3420818
Zhang, M. X., Xu, J., Liu, X. P., Zhang, Y., Zhang, C., & Ning, X. (2024b). Pavement crack detection method based on improved U-Net. Computer Engineering and Applications, 60(24), 306–313.
Zhou, J. X., Zhang, Y., Jia, Z. H., & He, Y. (2024). Pavement defect detection algorithm based on improved YOLOv8. Electronic Measurement Technology, 47(19), 146–154. https://doi.org/10.19651/j.cnki.emt.2416638
Zhu, Y., Lin, Y. C., Tan, X. H., & Li, S. X. (2025). A lightweight defect detection network for titanium strip via efficient convolution and feature-driven mechanism. Engineering Applications of Artificial Intelligence, 158, Article 111503. https://doi.org/10.1016/j.engappai.2025.111503
Zuo, H., & Niu, X. W. (2023). Pavement defect detection algorithm based on improved YOLOv5. Information Technology and Informatization, (1), 50–53. Notations Abbreviations DSA – Double Stage Attention Model; FPS – Frames Per Second; mAP – Mean Average Precision; IoU – Intersection over Union.
View article in other formats
CrossMark check
Published
Issue
Section
Copyright
Copyright (c) 2026 The Author(s). Published by Vilnius Gediminas Technical University.
License
This work is licensed under a Creative Commons Attribution 4.0 International License.