[1]王井阳,徐勇超,张 波,等.改进 YOLOv10n 的轻量化道路裂缝检测模型[J].郑州大学学报(工学版),2026,47(4):42-49.[doi:10.13705/j.issn.1671-6833.2026.04.007]
　WANG Jingyang,XU Yongchao,ZHANG Bo,et al.Improved YOLOv10n Lightweight Road Crack Detection Model[J].Journal of Zhengzhou University (Engineering Science),2026,47(4):42-49.[doi:10.13705/j.issn.1671-6833.2026.04.007]
点击复制

改进 YOLOv10n 的轻量化道路裂缝检测模型()
分享到：

《郑州大学学报(工学版)》[ISSN:1671-6833/CN:41-1339/T]

卷:
47
期数:
2026年4期
页码:
42-49
栏目:
出版日期:
2026-07-10

文章信息/Info

Title:
Improved YOLOv10n Lightweight Road Crack Detection Model
文章编号:
1671-6833(2026)04-0042-08
作者:
王井阳1, 徐勇超1, 张 波2, 王珏3, 黄敏1
1. 河北科技大学 信息科学与工程学院,河北 石家庄 050018;2. 河北工程技术学院 网络空间安全学院,河北 石家庄 050091;3. 中国电信股份有限公司 石家庄分公司,河北 石家庄 050035
Author(s):
WANG Jingyang1, XU Yongchao1, ZHANG Bo2, WANG Jue3, HUANG Min1
1. School of Information Science and Engineering, Hebei University of Science and Technology, Shijiazhuang 050018, China; 2. School of Cyberspace Security, Hebei University of Engineering Science, Shijiazhuang 050091, China; 3. Shijiazhuang Branch,Chi-naTelecom Corporation Limited, Shijiazhuang 050035, China
关键词:
道路裂缝检测;  YOLOv10n;  注意力机制;  损失函数;  轻量化
Keywords:
road crack detection;  YOLOv10n;  attention mechanism;  loss function;  lightweight
分类号:
TP391. 4U418. 6
DOI:
10.13705/j.issn.1671-6833.2026.04.007
文献标志码:
A
摘要:
针对现有道路裂缝检测模型不能有效平衡检测精度、计算复杂度与检测速度,实际应用效果差的问题,提出了一种基于改进 YOLOv10n 的轻量级道路裂缝检测模型 YOLO-CGVE。 首先,利用坐标注意力( CA) 模块替换部分自注意力( PSA)模块,从而更好地捕捉空间上的局部和全局关系,增强特征提取能力;其次,通过使用轻量级的GSConv 替换主干网络和颈部网络中的部分标准卷积,降低了计算复杂度;再次,在颈部网络采用 VoV-GSCSP 模块替换 C2f 模块,实现对不同阶段的特征图的有效融合,在保证精度的同时进一步降低计算复杂度;最后,使用 ECIoU代替原损失函 数, 提 高 检 测 框 定 位 精 度 和 收 敛 速 度。 在 RDD2022 _ China 数 据 集 上 的 实 验 结 果 表 明: 相 较 于YOLOv10n,YOLO-CGVE 的 mAP@ 0. 5 提高了 2. 4 百分点,达到了 75. 9%,参数量与计算量分别减少了 11. 1% 和9. 8%,同时保持了较高的检测速度。 YOLO-CGVE 可以更好地满足在计算资源有限环境下的应用需求。
Abstract:
Aiming at the problem that the existing road crack detection model cannot effectively balance the detec-tionaccuracy, computational complexity and detection speed with poor practical application effect, a lightweight road crack detection model YOLO-CGVE based on improved YOLOv10n was proposed. Firstly, the coordinate at-tention(CA) module was used to replace the partial self-attention ( PSA) module to better capture the local and global relationships in space and improve the capacity to extract features. Secondly, the computational complexity was reduced by using lightweight GSConv to replace some standard convolutions in the backbone and neck net-works.Thirdly, the original C2f structure in the neck network was replaced by VoV-GSCSP, which allowed for the efficient merging of feature maps from various stages and further minimized computing complexity while maintaining accuracy. Finally, the ECIoU loss function was used to replace the original loss function to improve the detection box positioning accuracy and convergence speed. The experimental results on RDD2022_China dataset showed that compared with YOLOv10n, while keeping a high detection speed, the mAP@ 0. 5 of YOLO-CGVE was improved by 2. 4 percentage points, reaching 75. 9%, and the number of parameters and the amount of computation were de-creasedby 11. 1% and 9. 8%, respectively. YOLO-CGVE could better meet the application needs in environments with limited computing resources.

参考文献/References:

[1] Sun Zhen, Zhu Lingxi, Qin Su, et al. Road surface defect detection algorithm based on YOLOv8[J]. Electronics, 2024, 13(12): 2413.
[2] Li Jun, Zhou Keyu, Zou Jun, et al. Protective equipment wearing detection algorithm in construction scenarios based on YOLOv8n[J]. Journal of Zhengzhou University (Engineering Science), 2025, 46(3): 19–25. [李军，周科宇，邹军，等. 基于改进YOLOv8n的施工场景下防护装备佩戴检测算法[J]. 郑州大学学报(工学版)，2025，46(3)：19–25.]
[3] Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the 2014 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2014: 580–587.
[4] Ren Shaoqing, He Kaiming, Girshick R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137–1149.
[5] Yang Ziyi, Lan Xin, Wang Hui. Comparative analysis of YOLO series algorithms for UAV-based highway distress inspection: performance and application insights [J]. Sensors, 2025, 25(5): 1475.
[6] Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2016: 779–788.
[7] Liu Wei, Anguelov D, Erhan D, et al. SSD: single shot MultiBox detector[C]//Computer Vision – ECCV 2016. Cham: Springer, 2016: 21–37.
[8] Wang Xueqiu, Gao Huanbing, Jia Zemeng. Improved road defect detection algorithm based on YOLOv8[J]. Computer Engineering and Applications, 2024, 60(17): 179–190. [王雪秋，高焕兵，郑泽萌. 改进YOLOv8的道路缺陷检测算法[J]. 计算机工程与应用，2024，60(17)：179–190.]
[9] Hu Fengkuo, Ye Lan, Tan Xianfeng, et al. A refined YOLOv8-based algorithm for lightweight pavement disease detection[J]. Journal of Graphics, 2024, 45(5): 892–900. [胡凤阔，叶兰，谭显峰，等. 一种基于改进YOLOv8的轻量化路面病害检测算法[J]. 图学学报，2024，45(5)：892–900.]
[10] Li Shanbin, Huang Yujun. Damage detection algorithm based on Faster-RCNN[C]//The 5th International Conference on Electronics and Communication, Network and Computer Technology (ECNCT). Piscataway: IEEE, 2023: 177–180.
[11] Wang Jinlei, Meng Ruifeng, Huang Yuanhao, et al. Road defect detection based on improved YOLOv8s model[J]. Scientific Reports, 2024, 14: 16758.
[12] Youwai S, Chaiyaphat A, Chaipetch P. YOLO9tr: a lightweight model for pavement damage detection utilizing a generalized efficient layer aggregation network and attention mechanism[J]. Journal of Real-Time Image Processing, 2024, 21(5): 163.
[13] Carion N, Massa F, Synnaeve G, et al. End-to-end object detection with transformers[C]//Computer Vision – ECCV 2020. Cham: Springer, 2020: 213–229.
[14] Zhao Yan, Lyu Wenyu, Xu Shangliang, et al. DETRs beat YOLOs on real-time object detection [C]//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2024: 16965–16974.
[15] Luo Xianglong, Liu Ruchen, He Xibin, et al. A road damage detection model based on improved RT-DETR for complex environments[J]. IEEE Transactions on Instrumentation and Measurement, 2025, 74: 5037413.
[16] Wang Ao, Chen Hui, Liu Lihao, et al. YOLOv10: real-time end-to-end object detection [PP/OL]. V2. arXiv (2024-10-30)[2026-02-13]. https://arxiv.org/abs/2405.14458.
[17] Hou Qibin, Zhou Daquan, Feng Jiashi. Coordinate attention for efficient mobile network design[C]//Proceedings of the 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway: IEEE, 2021: 13708–13717.
[18] Li Hulin, Li Jun, Wei Hanbing, et al. Slim-neck by GSConv: a lightweight-design for real-time detector architectures[J]. Journal of Real-Time Image Processing, 2024, 21(3): 62.
[19] Zheng Zhaohui, Wang Ping, Ren Dongwei, et al. Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J]. IEEE Transactions on Cybernetics, 2022, 52(8): 8574–8586.
[20] Zhang Yifan, Ren Weiqiang, Zhang Zhang, et al. Focal and efficient IOU loss for accurate bounding box regression[J]. Neurocomputing, 2022, 506: 146–157.
[21] Chen Xinlin, Lian Qingwang, Chen Xuanlai, et al. Surface crack detection method for coal rock based on improved YOLOv5[J]. Applied Sciences, 2022, 12(19): 9695.
[22] Arya D, Maeda H, Ghosh S K, et al. RDD2022: a multi-national image dataset for automatic road damage detection[J]. Geoscience Data Journal, 2024, 11(4): 846–862.
[23] Yang Fan, Zhang Lei, Yu Sijia, et al. Feature pyramid and hierarchical boosting network for pavement crack detection[J]. IEEE Transactions on Intelligent Transportation Systems, 2020, 21(4): 1525–1535.
[24] Zhang Lei, Yang Fan, Daniel Zhang Y, et al. Road crack detection using deep convolutional neural network[C]//Proceedings of the 2016 IEEE International Conference on Image Processing (ICIP). Piscataway: IEEE, 2016: 3708–3712.

更新日期/Last Update: 2026-10-21