基于改进YOLOv5s网络的斜拉桥拉索表面缺陷检测

TP391; 针对人工检测斜拉桥拉索表面缺陷效率低、安全性差,而现有目标检测方法速度慢、精度低,受拉索表面污垢干扰容易导致错检、漏检等问题,本文改进YOLOv5s网络以实现拉索表面缺陷快速准确检测.在主干网络增加TRANS模块,获取单幅图像更多特征,提高缺陷检测精度;为减少参数量、提高计算速度,将颈部网络的CSP模块替换为GhostBottleneck模块,同时利用深度可分离卷积代替普通卷积;利用SIOU损失函数减少边界框震荡,提高预测框和真实框重叠度计算结果准确性,增加模型稳定性.实验结果表明:改进YOLOv5s网络的mAP和FPS分别达到 94.26%和 68 f/s,优于Faster-R...

Full description

Saved in:
Bibliographic Details
Published in光电工程 Vol. 51; no. 5; pp. 中插1 - 20
Main Authors 王鹏峰, 李运堂, 黄永勇, 朱文凯, 林婕, 王斌锐
Format Journal Article
LanguageChinese
Published 中国计量大学现代科技学院,浙江 金华 322002%中国计量大学机电工程学院,浙江 杭州 310018 2024
Subjects
TRANS模块
损失函数
loss function
cable-stayed bridge cable
defects detection
斜拉桥拉索
TRANS module
YOLOv5s network
YOLOv5s网络
缺陷检测
Online AccessGet full text
ISSN1003-501X
DOI10.12086/oee.2024.240028

Cover

Abstract TP391; 针对人工检测斜拉桥拉索表面缺陷效率低、安全性差,而现有目标检测方法速度慢、精度低,受拉索表面污垢干扰容易导致错检、漏检等问题,本文改进YOLOv5s网络以实现拉索表面缺陷快速准确检测.在主干网络增加TRANS模块,获取单幅图像更多特征,提高缺陷检测精度;为减少参数量、提高计算速度,将颈部网络的CSP模块替换为GhostBottleneck模块,同时利用深度可分离卷积代替普通卷积;利用SIOU损失函数减少边界框震荡,提高预测框和真实框重叠度计算结果准确性,增加模型稳定性.实验结果表明:改进YOLOv5s网络的mAP和FPS分别达到 94.26%和 68 f/s,优于Faster-RCNN、YOLOv4和常规YOLOv5等网络,满足斜拉桥拉索表面缺陷检测精度和实时性要求.
AbstractList TP391; 针对人工检测斜拉桥拉索表面缺陷效率低、安全性差,而现有目标检测方法速度慢、精度低,受拉索表面污垢干扰容易导致错检、漏检等问题,本文改进YOLOv5s网络以实现拉索表面缺陷快速准确检测.在主干网络增加TRANS模块,获取单幅图像更多特征,提高缺陷检测精度;为减少参数量、提高计算速度,将颈部网络的CSP模块替换为GhostBottleneck模块,同时利用深度可分离卷积代替普通卷积;利用SIOU损失函数减少边界框震荡,提高预测框和真实框重叠度计算结果准确性,增加模型稳定性.实验结果表明:改进YOLOv5s网络的mAP和FPS分别达到 94.26%和 68 f/s,优于Faster-RCNN、YOLOv4和常规YOLOv5等网络,满足斜拉桥拉索表面缺陷检测精度和实时性要求.
Abstract_FL An improved YOLOv5s network for defects detection for the cable surface of cable-stayed bridge fast and accurately is proposed.This overcomes the problems of low efficiency and poor safety of manual inspection,slow and inaccuracy of existing target detection methods because of the interference of dirt leading to wrong and missed detections.The TRANS module is added to the backbone network of conventional YOLOv5s to obtain more features of a single image and improve defect detection accuracy.Moreover,the CSP module of the neck network is replaced by the GhostBottleneck module and ordinary convolution is replaced by depth-separable convolution to reduce parameters and improve the computational speed of the network.Furthermore,the SIOU loss function is used for suppressing the oscillation of the bounding box and improving the calculation accuracy of repeatability between the prediction and the real box,which can increase the model stability.The experiments show that mAP and FPS of improved YOLOv5s network are 94.26%and 68 frames per second,respectively.The performance is better than that of Faster-RCNN,YOLOv4,and conventional YOLOv5,and it can find the surface defect for the cable of the cable-stayed bridge accurately and timely.
Author 王斌锐
王鹏峰
黄永勇
林婕
李运堂
朱文凯
AuthorAffiliation 中国计量大学现代科技学院,浙江 金华 322002%中国计量大学机电工程学院,浙江 杭州 310018
AuthorAffiliation_xml – name: 中国计量大学现代科技学院,浙江 金华 322002%中国计量大学机电工程学院,浙江 杭州 310018
Author_FL Wang Pengfeng
Lin Jie
Wang Binrui
Zhu Wenkai
Huang Yongyong
Li Yuntang
Author_FL_xml – sequence: 1
  fullname: Wang Pengfeng
– sequence: 2
  fullname: Li Yuntang
– sequence: 3
  fullname: Huang Yongyong
– sequence: 4
  fullname: Zhu Wenkai
– sequence: 5
  fullname: Lin Jie
– sequence: 6
  fullname: Wang Binrui
Author_xml – sequence: 1
  fullname: 王鹏峰
– sequence: 2
  fullname: 李运堂
– sequence: 3
  fullname: 黄永勇
– sequence: 4
  fullname: 朱文凯
– sequence: 5
  fullname: 林婕
– sequence: 6
  fullname: 王斌锐
BookMark eNrjYmDJy89LZWCQMDTQMzQysDDTz09N1TMyMDLRMzIxMDCyYGHgNDQwMNY1NTCM4GDgLS7OTDIwMDGyNLcwM-dksH06f9eTXX3Ppux8sX92pL-Pf5lp8fO9E5_vnvN8VsuzaXOedXc-W7gUSD7fsujFwhUv5y56vmfXy5nbny1ueLa1m4eBNS0xpziVF0pzM4S6uYY4e-j6-Lt7Ojv66BYDrTbSNU1LTDJLS00xM0hMNTYxMDc3NjI2NU01tjQzNTK2MExJTE5MNDI3NUlOTbJMtEw2SjQ2S05OSU5MNk4yMDezTDHmZlCFmFuemJeWmJcen5VfWpQHtDE-PSU9GeRZA1OgRcYAIy1eww
ClassificationCodes TP391
ContentType Journal Article
Copyright Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
Copyright_xml – notice: Copyright © Wanfang Data Co. Ltd. All Rights Reserved.
DBID 2B.
4A8
92I
93N
PSX
TCJ
DOI 10.12086/oee.2024.240028
DatabaseName Wanfang Data Journals - Hong Kong
WANFANG Data Centre
Wanfang Data Journals
万方数据期刊 - 香港版
China Online Journals (COJ)
China Online Journals (COJ)
DatabaseTitleList
DeliveryMethod fulltext_linktorsrc
DocumentTitle_FL Defects detection for cable surface of cable-stayed bridge based on improved YOLOv5s network
EndPage 20
ExternalDocumentID gdgc202405002
GroupedDBID 2B.
4A8
8FE
8FG
92I
93N
ABJCF
ABUWG
AFKRA
ALMA_UNASSIGNED_HOLDINGS
BENPR
BGLVJ
BPHCQ
BVBZV
CCPQU
HCIFZ
L6V
M7S
PHGZM
PHGZT
PIMPY
PMFND
PQQKQ
PROAC
PSX
PTHSS
TCJ
ID FETCH-LOGICAL-s1002-5fab6fed60ae3407732355e39652381dacaa2754ceb9a9c2a36ccdcac3b0769d3
ISSN 1003-501X
IngestDate Thu May 29 03:55:49 EDT 2025
IsPeerReviewed true
IsScholarly true
Issue 5
Keywords TRANS模块
损失函数
loss function
cable-stayed bridge cable
defects detection
斜拉桥拉索
TRANS module
YOLOv5s network
YOLOv5s网络
缺陷检测
Language Chinese
LinkModel OpenURL
MergedId FETCHMERGED-LOGICAL-s1002-5fab6fed60ae3407732355e39652381dacaa2754ceb9a9c2a36ccdcac3b0769d3
ParticipantIDs wanfang_journals_gdgc202405002
PublicationCentury 2000
PublicationDate 2024
PublicationDateYYYYMMDD 2024-01-01
PublicationDate_xml – year: 2024
  text: 2024
PublicationDecade 2020
PublicationTitle 光电工程
PublicationTitle_FL Opto-Electronic Engineering
PublicationYear 2024
Publisher 中国计量大学现代科技学院,浙江 金华 322002%中国计量大学机电工程学院,浙江 杭州 310018
Publisher_xml – name: 中国计量大学现代科技学院,浙江 金华 322002%中国计量大学机电工程学院,浙江 杭州 310018
SSID ssib004297867
ssib023646518
ssib036437391
ssib023167165
ssj0002964646
ssib002258422
ssib001102639
ssib051369860
ssib000459782
Score 2.3985171
Snippet TP391;...
SourceID wanfang
SourceType Aggregation Database
StartPage 中插1
Title 基于改进YOLOv5s网络的斜拉桥拉索表面缺陷检测
URI https://d.wanfangdata.com.cn/periodical/gdgc202405002
Volume 51
hasFullText 1
inHoldings 1
isFullTextHit
isPrint
link http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwnR3LahRBcAjx4kUUFZ8hBxsPYeNMTz8PHqY3MwZRc0kgnsI8Yy4bMImHnDx4kYAgiKBGJBHBD1DEqPgzupv8hVU97c5ERWMuQ1NbXV2Ppbuqu7ra8y6VmfZ1ziBSxSmQVRCwwiqoOyzMqoJnLA8q3Ie8eUtMz7Hr83x-ZPRyK2tpbTWbzNf_eK_kMFYFGNgVb8n-h2WHRAEAbbAvfMHC8D2QjUnMiU6IiUjM8KtiEguioa1JrIhJiDa3Z27M3OMrJJbETBEd2IYhuosNDX2Y7SMsRBBliNLYiAIS8RYEejESUaSLPykSa6KnLAR-6lomAKKJkbZ7iDkU0DAcKLRdYGRacUcTeeUIgV44nETKangMhAAQClgA2iCUSixuSIzfoAjkA0WvJfYRJfKJog2KRpFrSY1PjLI8AETuowJCBE4ZSloUYCdp74vQZkfUqhx0MWWNYFC5qJoYtQPjQdea2Qi0ZokBZiScQMAEdjeoJpRZWsPAsJHV2hDZKhRUTLtOldraBtjUyYQbBnsCq6iCCZg3ba1bfljuai1EfzHNwbmzdqk5ABIauMNKXO1lEDMWuW8fMxquk64w8FI7E8Euei2RBHKv6zvvzp-ydx1_X6mpb08Ol0usVUvZJOYyuzIB--ufLxaLOWL43BaNPUKlDDB3VyXX2nGJbj_LAD4tFa1yt7BoKdaqUwlOmFSN302xHETQHPfjowqCN05biKfbYVNUigeh0MrtK6DLh7kLor7J-FNzLiUCxbzyi5D2fmCvSnuLLVd29rh3zMWg41E9oZzwRtbvnPSufn-1823nUf_Jx92vL9ykMfjyePBpc_D8Qf_pZn_jYX_rDXwH77Z3t97uvdwefN7Ze_ah__p-__3GKW8uiWe70x33tkpnBYsud3iVZqIqC-GnZch8KUMKkUcZasHRiS_SPE2p5CyHyTzVOU1DkedFnuZh5kuhi_C0N9pb7pVnvHFelIyVEMYFWB8x5ylEkVUVUk05K1UpznpjTtwFN3euLOwz6rl_IZz3jmK73vm84I2u3l0rL0IssJqN2f_BD3_nuOM
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=%E5%9F%BA%E4%BA%8E%E6%94%B9%E8%BF%9BYOLOv5s%E7%BD%91%E7%BB%9C%E7%9A%84%E6%96%9C%E6%8B%89%E6%A1%A5%E6%8B%89%E7%B4%A2%E8%A1%A8%E9%9D%A2%E7%BC%BA%E9%99%B7%E6%A3%80%E6%B5%8B&rft.jtitle=%E5%85%89%E7%94%B5%E5%B7%A5%E7%A8%8B&rft.au=%E7%8E%8B%E9%B9%8F%E5%B3%B0&rft.au=%E6%9D%8E%E8%BF%90%E5%A0%82&rft.au=%E9%BB%84%E6%B0%B8%E5%8B%87&rft.au=%E6%9C%B1%E6%96%87%E5%87%AF&rft.date=2024&rft.pub=%E4%B8%AD%E5%9B%BD%E8%AE%A1%E9%87%8F%E5%A4%A7%E5%AD%A6%E7%8E%B0%E4%BB%A3%E7%A7%91%E6%8A%80%E5%AD%A6%E9%99%A2%2C%E6%B5%99%E6%B1%9F+%E9%87%91%E5%8D%8E+322002%25%E4%B8%AD%E5%9B%BD%E8%AE%A1%E9%87%8F%E5%A4%A7%E5%AD%A6%E6%9C%BA%E7%94%B5%E5%B7%A5%E7%A8%8B%E5%AD%A6%E9%99%A2%2C%E6%B5%99%E6%B1%9F+%E6%9D%AD%E5%B7%9E+310018&rft.issn=1003-501X&rft.volume=51&rft.issue=5&rft.spage=%E4%B8%AD%E6%8F%921&rft.epage=20&rft_id=info:doi/10.12086%2Foee.2024.240028&rft.externalDocID=gdgc202405002
thumbnail_s http://utb.summon.serialssolutions.com/2.0.0/image/custom?url=http%3A%2F%2Fwww.wanfangdata.com.cn%2Fimages%2FPeriodicalImages%2Fgdgc%2Fgdgc.jpg