[1]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,(06):655-660.[doi:10.16357/j.cnki.issn1000-5862.2022.06.14]
　HE Jinbao,HU Qiubao,FU Zhichao,et al.The Bridge Apparent Defects Detection Based on DCGAN and Improved YOLOv5s[J].Journal of Jiangxi Normal University:Natural Science Edition,2022,(06):655-660.[doi:10.16357/j.cnki.issn1000-5862.2022.06.14]
点击复制

基于DCGAN和改进YOLOv5s的桥梁表面缺陷检测识别()
分享到：

《江西师范大学学报》（自然科学版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2022年06期
页码:
655-660
栏目:
信息科学与技术
出版日期:
2022-11-25

文章信息/Info

Title:
The Bridge Apparent Defects Detection Based on DCGAN and Improved YOLOv5s
文章编号:
1000-5862(2022)06-0655-06
作者:
何金宝1; 胡秋宝1; 付志超2; 赖 毅2*; 刘知远1
1.江西省港口集团有限公司,江西 南昌 330008; 2.江西省路港检测中心有限公司,江西 南昌 330200)
Author(s):
HE Jinbao1; HU Qiubao1; FU Zhichao2; LAI Yi2*; LIU Zhiyuan1
1.Jiangxi Port Group Company Limited,Nanchang Jiangxi 330008,China; 2.Jiangxi Port and Shipping Quality Inspection Center,Nanchang Jiangxi 330200,China)
关键词:
表面缺陷检测;  深度卷积生成式对抗网络;  注意力机制;  YOLOv5s
Keywords:
surface defect detection;  deep convolutional generative confrontation network;  attention mechanism;  YOLOv5s
分类号:
TH 133.3
DOI:
10.16357/j.cnki.issn1000-5862.2022.06.14
文献标志码:
A
摘要:
针对人工检测桥梁表面缺陷存在精度低、速度慢和漏检率高等问题,该文提出了基于深度卷积生成式对抗网络(deep convolutional generative adversarial networks,DCGAN)和改进YOLOv5s的桥梁表面缺陷检测识别模型.首先,通过DCGAN网络对自主采集的桥梁表面缺陷图像进行数据增强,建立每类缺陷样本数量较为均衡的数据集; 其次,在YOLOv5s模型基础上嵌入CBAM注意力机制模块,使模型将注意力集中于缺陷区域,从而提升图像分类的准确率; 最后,为验证所提方法的适用性,将包含4类桥梁表面缺陷的数据集进行训练与测试.实验结果表明:该文提出的模型在桥梁表面缺陷检测上的准确率为92%,相比其他深度学习模型具有更高的检测精度和鲁棒性.
Abstract:
In order to solve the problems of low accuracy,slow speed and high missed detection rate in current bridge surface defect detection methods,the bridge surface defect detection method based on deep convolutional generative adversarial networks(DCGAN)and improved YOLOv5s is proposed.recognition methods.Firstly,through the deep convolutional generative confrontation network,the self-collected bridge surface defect images are enhanced to establish a more balanced data set with rich sample features.Secondly,the CBAM attention mechanism module is embedded on the basis of the YOLOv5s model to learn the importance of each channel and spatial feature independently,thereby improving the performance of image classification.Finally,in order to verify the applicability of the proposed method,four data sets of bridge surface defects will be included for training and testing.Experimental results show that the accuracy of the proposed model in the detection of bridge surface defects is 92%,which has higher detection accuracy and robustness than other deep learning models.

参考文献/References:

[1] 吉伯海,傅中秋.近年国内桥梁倒塌事故原因分析 [J].土木工程学报,2010,43(S1):495-498.
[2] 胡兴意,陈波,刘国强,等.融合BIM与影像建模技术的桥梁检测方法及其应用研究 [J].中外公路,2020,40(1):107-112.
[3] YAO Jinbao,CHEN Jinsong,WANG Wensi,et al.Damage detection of bridge structure based on SVM [J].Mathematical Problems in Engineering,2013,2013(16):490372.
[4] 陈飞飞,张宇峰,韩晓健.基于图像特征值的混凝土桥梁表面病害图像分类 [J].结构工程师,2018,34(1):59-63.
[5] 魏武,王俊杰,蔡钊雄.基于小波和Radon变换的桥梁裂缝检测 [J].计算机工程与设计,2013,34(9):3151-3157.
[6] PENG Xiong,ZHONG Xingu,ZHAO Chao,et al.A UAV-based machine vision method for bridge crack recognition and width quantification through hybrid feature learning [J].Construction and Building Materials,2021,299(5):123896.
[7] ARCHANA M,VINAYAK E,SHAILAJA P.Advances in deep learning for medical image analysis [M].Florida:CRC Press,2022.
[8] TURAN M,ALMALIOGLU Y,ARAUJO H,et al.Deep EndoVO:a recurrent convolutional neural network(RCNN)based visual odometry approach for endoscopic capsule robots [J].Neurocomputing,2018,275:1861-1870.
[9] CHEN Kuanbing,XUAN Ying,LIN Aijun,et al.Esophageal cancer detection based on classification of gastrointestinal CT images using improved Faster RCNN [J].Computer Methods and Programs in Biomedicine,2021,207:106172.
[10] REGHUKUMAR A,ANBARASI L,PRASSANNA J,et al.Vision based segmentation and classification of cracks using deep neural networks [J].International Journal of Uncertainty,Fuzziness and Knowledge-Based Systems,2021,29(S1):141-156.
[11] DENG Jianghua,LU Ye,LEE V C.Concrete crack detection with handwriting script interferences using faster region-based convolutional neural network [J].Computer-Aided Civil and Infrastructure Engineering,2020,35(4):373-388.
[12] GAO Hui,WANG Wenhao,YANG Chengjin,et al.Traffic signal image detection technology based on YOLO [J].Journal of Physics:Conference Series,2021,1961(1):012012.
[13] SUN Zhenyu,GUO Xiaoming,ZHANG Xiaoyang,et al.Research on robot target recognition based on deep learning [J].Journal of Physics:Conference Series,2021,1948(1):012056.
[14] 周清松,董绍江,罗家元,等.改进YOLOv3的桥梁表观病害检测识别 [J].重庆大学学报,2022,45(6):121-130.
[15] 孙朝云,马志丹,李伟,等.基于深度卷积神经网络融合模型的路面裂缝识别方法 [J].长安大学学报(自然科学版),2020,40(4):1-13.
[16] ALYOUBI W L,ABULKHAIR M F,SHALASH W M,et al.Image classification and lesions localization system using deep learning [J].Sensors,2021,21(11):367-375.
[17] SHRIVASTAVA N,HANIF M A,MITTAL S,et al.A survey of hardware architectures for generative adversarial networks [J].Journal of Systems Architecture,2021,118:102227.
[18] Jia Wei,Xu Shiquan,Liang Zhen,et al.Real-time automatic helmet detection of motorcyclists in urban traffic using improved YOLOv5 detector [J].IET Image Processing,2021,15(14):3623-3637.
[19] ZHENG Wenbo,YAN Lan,GOU Chao,et al.Pay attention to doctor-patient dialogues:multi-modal knowledge graph attention image-text embedding for COVID-19 diagnosis [J].Information Fusion,2021,75:168-185.
[20] ZHANG Kaihua,SHEN Haikuo.Solder joint defect detection in the connectors using improved Faster-RCNN algorithm [J].Applied Sciences,2021,11(2):576.
[21] 薛勇,王立扬,张瑜.基于GoogLeNet深度迁移学习的苹果缺陷检测方法 [J].农业机械学报,2020,51(7):30-35.
[22] TENG Bowen,ZHAO Hongjian,JIA Pei,et al.Research on ceramic sanitary ware defect detection method based on improved VGG network [J].Journal of Physics:Conference Series,2020,1650(2):022084.

备注/Memo

备注/Memo:
收稿日期:2022-04-16
基金项目:江西省交通运输厅科技课题(2022S0033,2022S0036)资助项目.
作者简介:何金宝(1969—),男,江西德兴人,高级工程师,主要从事港口与航道项目管理研究.E-mail:2369341717@qq.com
通信作者:赖 毅(1981—),男,江西会昌人,高级工程师,主要从事桥梁工程、港口工程健康监测与管理研究.E-mail:11592227@qq.com
更新日期/Last Update: 2022-11-25