基于数据增广与改进YOLOv8的桥梁缺陷检测

doi:10.16088/j.issn.1001-6600.2024071003

摘要/Abstract

摘要： 为解决干扰背景下桥梁表面缺陷检测精度低、漏检率及误检率高等问题,本文提出一种数据增广与改进YOLOv8的桥梁缺陷检测方法。通过StyleGAN3和深度图像融合方法对少样本数据进行增广。在YOLOv8主干中加入SPD-Conv模块,提升对低分辨率缺陷的特征提取能力;颈部在AFPN结构的基础上,设计出AFPN_UCG结构,使网络能更好地处理多尺度信息;在C2f中引入RFCBAMConv和DLKA模块,构建C2f_RD模块,使其精准传递梯度信息,同时能够让网络更有效地捕捉小目标信息;通过DCNv3模块与Dynamic Head相结合设计出新的检测头,其将尺度、空间和任务3种注意力机制结合并使用DCNv3动态调整,进一步提升模型对不规则缺陷的预测性能。经实验,数据增广后mAP@0.5提升了2.4个百分点,改进后的YOLOv8准确率为93.2%,mAP@0.5为91.3%,较原模型分别提高了4.2和4.3个百分点,能够更加精准检测桥梁缺陷。

关键词: 桥梁缺陷检测, StyleGAN3, YOLOv8, 特征融合, 注意力卷积, 信息交互

Abstract: In order to solve the problems of low detection accuracy, high missed detection rate and high false detection rate of bridge surface defects under the background of interference, a bridge defect detection method based on data enhancement and improved YOLOv8 is proposed. The small sample data is augmented by StyleGAN3 and depth image fusion. The SPD-Conv module is added to the YOLOv8 backbone to improve the feature extraction capability of low-resolution defects. Based on AFPN structure, AFPN_UCG structure is designed to make the network handle multi-scale information better. In C2f, RFCBAMConv and DLKA modules are introduced to construct C2f_RD module, which can transmit gradient information accurately and capture small target information more effectively. A new detection Head is designed by combining DCNv3 module with Dynamic Head, which combines three attention mechanisms of scale, space and task and uses DCNv3 dynamic adjustment to further improve the prediction performance of the model for irregular defects. Through experiments, mAP@0.5 increases by 2.4 percentage points after the data is expanded, and the accuracy rate of the improved YOLOv8 is 93.2% and mAP@0.5 is 91.3%, respectively, which are 4.2 and 4.3 percentage points higher than that of the original model, which can detect bridge defects more accurately.

Key words: bridge defect detection, StyleGAN3, YOLOv8, feature fusion, convolution of attention;information interactive

中图分类号: TP391.41

梁胤杰, 南新元, 蔡鑫, 李云鹏, 勾海光. 基于数据增广与改进YOLOv8的桥梁缺陷检测[J]. 广西师范大学学报（自然科学版）, 2025, 43(3): 84-97.

LIANG Yinjie, NAN Xinyuan, CAI Xin, LI Yunpeng, GOU Haiguang. Bridge Defect Detection Based on Data Augmentation and Improved YOLOv8[J]. Journal of Guangxi Normal University(Natural Science Edition), 2025, 43(3): 84-97.

参考文献

[1] 《中国公路学报》编辑部.中国桥梁工程学术研究综述·2021[J].中国公路学报,2021,34(2):1-97. DOI: 10.3969/j.issn.1001-7372.2021.02.002.
[2]陈榕峰,徐群丽,秦凯强.桥梁裂缝智能检测系统的研究新进展[J].公路,2019,64(5):101-105.
[3]李学军,权林霏,刘冬梅,等.基于Faster-RCNN改进的交通标志检测算法[J/OL].吉林大学学报(工学版):1-10[2024-06-30].https://doi.org/10.13229/j.cnki.jdxbgxb.20230553. DOI: 10.13229/j.cnki.jdxbgxb.20230553.
[4]茅智慧,朱佳利,吴鑫,等.基于YOLO的自动驾驶目标检测研究综述[J].计算机工程与应用,2022,58(15):68-77. DOI: 10.3778/j.issn.1002-8331.2203-0310.
[5]郭翔羽,石天怡,陈燕楠,等.基于YOLO-CDBW模型的列车接触网异物检测研究[J].广西师范大学学报(自然科学版),2025,43(2): 56-69. DOI: 10.16088/j.issn.1001-6600.2024040102.
[6]ALI R, KANG D, SUH G, et al. Real-time multiple damage mapping using autonomous UAV and deep faster region-based neural networks for GPS-denied structures[J]. Automation in Construction, 2021, 130: 103831. DOI: 10.1016/j.autcon.2021.103831.
[7]LI R X, YU J Y, LI F, et al. Automatic bridge crack detection using Unmanned aerial vehicle and Faster R-CNN[J]. Construction and Building Materials, 2023, 362: 129659. DOI: 10.1016/j.conbuildmat.2022.129659.
[8]毛莺池,唐江红,王静,等.基于Faster R-CNN的多任务增强裂缝图像检测方法[J].智能系统学报,2021,16(2):286-293. DOI: 10.11992/tis.201910004.
[9]ZHANG C B, CHANG C C, JAMSHIDI M. Concrete bridge surface damage detection using a single-stage detector[J]. Computer-Aided Civil and Infrastructure Engineering, 2020, 35(4): 389-409. DOI: 10.1111/mice.12500.
[10]YU L C, HE S H, LIU X S, et al. Intelligent crack detection and quantification in the concrete bridge: A deep learning-assisted image processing approach[J]. Advances in Civil Engineering, 2022,2022(1): 1813821. DOI: 10.1155/2022/1813821.
[11]何金宝,胡秋宝,付志超,等.基于DCGAN和改进YOLOv5s的桥梁表面缺陷检测识别[J].江西师范大学学报(自然科学版),2022,46(6):655-660. DOI: 10.16357/j.cnki.issn1000-5862.2022.06.14.
[12]彭雨诺,刘敏,万智,等.基于改进YOLO的双网络桥梁表观病害快速检测算法[J].自动化学报,2022,48(4):1018-1032. DOI: 10.16383/j.aas.c210807.
[13]李松,史涛,井方科.改进YOLOv8的道路损伤检测算法[J].计算机工程与应用,2023,59(23):165-174. DOI: 10.3778/j.issn.1002-8331.2306-0205.
[14]张彬,周粤川,张敏,等.生成对抗网络改进角度与应用研究综述[J].计算机应用研究,2023,40(3):649-658. DOI: 10.19734/j.issn.1001-3695.2022.08.0410.
[15]KARRAS T, AITTALA M, LAINE S, et al. Alias-free generative adversarial networks[C]// Proceedings of the 35th International Conference on Neural Information Processing Systems. Red Hook: Curran Associates Inc., 2021, 34: 852-863.
[16]ZHANG L Z, WEN T, SHI J B. Deep image blending[C]// 2020 IEEE Winter Conference on Applications of Computer Vision (WACV). Los Alamitos, CA: IEEE Computer Society, 2020: 231-240. DOI: 10.1109/WACV45572.2020.9093632.
[17]SUNKARA R, LUO T. No more strided convolutions or pooling: a new CNN building block for Low-Resolution images and small objects[C]// Joint European conference on machine learning and knowledge discovery in databases. Cham: Springer Nature Switzerland, 2023: 443-459. DOI: 10.1007/978-3-031-26409-2_27.
[18]YANG G Y, LEI J, ZHU Z K, et al. AFPN: asymptotic feature pyramid network for object detection[C]// 2023 IEEE International Conference on Systems, Man, and Cybernetics (SMC). Los Alamitos, CA: IEEE Computer Society, 2023: 2184-2189. DOI: 10.1109/SMC53992.2023.10394415.
[19]WU T Y, TANG S, ZHANG R, et al. CGNet: a light-weight context guided network for semantic segmentation[J]. IEEE Transactions on Image Processing, 2021, 30: 1169-1179. DOI: 10.1109/TIP.2020.3042065.
[20]ZHANG X, LIU C, YANG D G, et al. RFAConv: innovating spatial attention and standard convolutional operation[EB/OL]. (2024-03-28)[2024-07-04]. https://arxiv.org/abs/2304.03198. DOI: 10.48550/arXiv.2304.03198.
[21]AZAD R, NIGGEMEIER L, HÜTTEMANN M, et al. Beyond self-attention: deformable large kernel attention for medical image segmentation[C]// 2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV). Los Alamitos, CA: IEEE Computer Society, 2024: 1276-1286. DOI: 10.1109/WACV57701.2024.00132.
[22]DAI X Y, CHEN Y P, XIAO B, et al. Dynamic head: unifying object detection heads with attentions[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2021: 7369-7378. DOI: 10.1109/CVPR46437.2021.00729.
[23]WANG W H, DAI J F, CHEN Z, et al. Internimage: exploring large-scale vision foundation models with deformable convolutions[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2023: 14408-14419. DOI: 10.1109/CVPR52729.2023.01385.
[24]吴利刚,陈乐,刘泽鹏,等.基于YOLOv8-ABW的黄花成熟度检测方法[J].农业工程学报,2024,40(13):262-272. DOI: 10.11975/j.issn.1002-6819.202401074.
[25]张林鍹,巴音塔娜,曾庆松.基于StyleGAN2-ADA和改进YOLO v7的葡萄叶片早期病害检测方法[J].农业机械学报,2024,55(1):241-252.
[26]WANG A, CHEN H, LIU L H, et al. YOLOv10: real-time end-to-end object detection[EB/OL]. (2024-05-23)[2024-07-04]. https://arxiv.org/abs/2405.14458. DOI: 10.48550/arXiv.2405.14458.
[27]WANG C Y, YEH I H, LIAO H Y M. YOLOv9: learning what you want to learn using programmable gradient information[EB/OL]. (2024-02-29)[2024-07-04]. https://arxiv.org/abs/2402.13616. DOI: 10.48550/arXiv.2402.13616.
[28]ZHAO Y A, LV W Y, XU S L, et al. DETRs beat YOLOs on real-time object detection[C]// 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2024: 16965-16974. DOI: 10.1109/CVPR52733.2024.01605.
[29]XU X Z, JIANG Y Q, CHEN W H, et al. DAMO-YOLO: a report on real-time object detection design[EB/OL]. (2023-04-24)[2024-07-04]. https://arxiv.org/abs/2211.15444. DOI: 10.48550/arXiv.2211.15444.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed