Article(id=1148106707819029331, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106698197295351, articleNumber=1003-3033(2025)02-0066-07, orderNo=null, doi=10.16265/j.cnki.issn1003-3033.2025.02.0880, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1727193600000, receivedDateStr=2024-09-25, revisedDate=1732464000000, revisedDateStr=2024-11-25, acceptedDate=null, acceptedDateStr=null, onlineDate=1751659569934, onlineDateStr=2025-07-05, pubDate=1740672000000, pubDateStr=2025-02-28, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1751659569934, onlineIssueDateStr=2025-07-05, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1751659569934, creator=13701087609, updateTime=1751659569934, updator=13701087609, issue=Issue{id=1148106698197295351, tenantId=1146029695717560320, journalId=1146031787341344770, year='2025', volume='35', issue='2', pageStart='1', pageEnd='252', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1751659567641, creator=13701087609, updateTime=1757401525528, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1172190215188894212, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106698197295351, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1172190215188894213, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106698197295351, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=66, endPage=72, ext={EN=ArticleExt(id=1149767845506106202, articleId=1148106707819029331, tenantId=1146029695717560320, journalId=1146031787341344770, language=EN, title=Aircraft rivet and detachment anomaly detection algorithm based on improved YOLOv8n, columnId=1149733269173878863, journalTitle=China Safety Science Journal, columnName=Safety engineering technology, runingTitle=null, highlight=null, articleAbstract=

In order to address the issue of missed detections during small target detection of aircraft rivets,an improved YOLOv8n algorithm for the detection of aircraft rivets and their anomalies was proposed. First,by adding a small object detection head,the shallow detail information in the backbone network was better fused,enhancing the model's feature fusion capability and its ability to recognize and locate small rivet targets. Second,the first two convolutions in the backbone network were replaced with SPD-Conv,which reduces information loss during down sampling through the combination of feature map reorganization and non-stride convolutions. Finally,large separable kernel attention (LSKA) was integrated into the spatial pyramid pooling fast (SPPF) module,capturing the dependencies between spatial and channel dimensions by calculating spatial and channel weights on each feature map and adjusting the feature maps to enhance the algorithm's ability to extract and recognize rivet feature information. Ablation experiments and comparative experiments were conducted based on a self-built aircraft rivet dataset. The results show that the proposed algorithm can achieve real-time identification of aircraft rivets and their anomalies,with precision,recall,and mean average precision (mAP) values improved by 6.5%,16%,and 15%,respectively,compared to the YOLOv8n algorithm. The detection performance is also significantly better than other mainstream algorithms.

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Zhenghong XIA, Hu HE, Lei YANG, Jianjun WU, Lu LIU), CN=ArticleExt(id=1148106713300984049, articleId=1148106707819029331, tenantId=1146029695717560320, journalId=1146031787341344770, language=CN, title=基于改进YOLOv8n的飞机铆钉及脱落异常检测算法, columnId=1149733269727526997, journalTitle=中国安全科学学报, columnName=安全工程技术, runingTitle=null, highlight=null, articleAbstract=

为解决飞机铆钉小目标检测时易出现漏检的问题,提出一种基于改进YOLOv8n的飞机铆钉及脱落异常检测算法。首先,通过添加小目标检测层,更好地融合骨干网络中的浅层细节信息,提高算法的特征融合能力以及对铆钉小目标的识别与定位性能;其次,将骨干网络中的前2次卷积替换为空间深度转换卷积(SPD-Conv),通过特征图的重组与非跨步卷积的组合,减少算法在下采样过程中的信息丢失;然后,将大型可分离核注意力(LSKA)融入快速空间金字塔池化(SPPF)模块中,通过计算每个特征图上的空间权重和通道权重,捕捉空间与通道之间的依赖关系,并调整特征图,增强算法对铆钉特征信息的提取和识别能力;最后,基于自建的飞机铆钉数据集进行消融试验和对比试验。结果表明:所提算法能实时检测飞机铆钉及脱落异常,较YOLOv8n算法检测结果在精确率、召回率、平均精度均值(mAP)分别提升6.5%、16%、15%,较其他主流算法的检测性能均有较大提升。

, correspAuthors=夏正洪, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=7SgZxXxRUy3Ty/CTFk6pXg==, magXml=Ae8zI6qcHknyFwpHUwTg8w==, pdfUrl=null, pdf=+Cj+ipmafCefuRYgPVJd3w==, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=3cSXEaq4eW3I9GWdObhpGw==, mapNumber=null, authorCompany=null, fund=null, authors=

夏正洪 (1985—),男,四川乐山人,硕士,教授,主要从事基于深度学习的航空运行安全风险评价研究。E-mail:

刘璐 高级工程师

, authorsList=夏正洪, 何琥, 杨磊, 吴建军, 刘璐)}, authors=[Author(id=1165681795061260425, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=66949827@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165681795140952204, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795061260425, language=EN, stringName=Zhenghong XIA, firstName=Zhenghong, middleName=null, lastName=XIA, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165681795191283853, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795061260425, language=CN, stringName=夏正洪, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307, bio={"content":"

夏正洪 (1985—),男,四川乐山人,硕士,教授,主要从事基于深度学习的航空运行安全风险评价研究。E-mail:

"}, bioImg=null, bioContent=

夏正洪 (1985—),男,四川乐山人,硕士,教授,主要从事基于深度学习的航空运行安全风险评价研究。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165681794851545214, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=1, ext=[AuthorCompanyExt(id=1165681794855739519, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China), AuthorCompanyExt(id=1165681794868322432, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307)])]), Author(id=1165681795291947151, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165681795342278801, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795291947151, language=EN, stringName=Hu HE, firstName=Hu, middleName=null, lastName=HE, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165681795384221842, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795291947151, language=CN, stringName=何琥, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165681794851545214, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=1, ext=[AuthorCompanyExt(id=1165681794855739519, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China), AuthorCompanyExt(id=1165681794868322432, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307)])]), Author(id=1165681795447136404, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165681795514245270, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795447136404, language=EN, stringName=Lei YANG, firstName=Lei, middleName=null, lastName=YANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165681795568771223, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795447136404, language=CN, stringName=杨磊, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165681794851545214, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=1, ext=[AuthorCompanyExt(id=1165681794855739519, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China), AuthorCompanyExt(id=1165681794868322432, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307)])]), Author(id=1165681795652657305, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165681795770097819, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795652657305, language=EN, stringName=Jianjun WU, firstName=Jianjun, middleName=null, lastName=WU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 North Regional Administration,Civil Aviation Administration of China,Beijing 100621,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165681795828818076, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795652657305, language=CN, stringName=吴建军, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 中国民用航空华北地区管理局,北京 100621, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165681794918654081, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=2, ext=[AuthorCompanyExt(id=1165681794927042690, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794918654081, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 North Regional Administration,Civil Aviation Administration of China,Beijing 100621,China), AuthorCompanyExt(id=1165681794935431299, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794918654081, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国民用航空华北地区管理局,北京 100621)])]), Author(id=1165681795883344030, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165681795988201632, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795883344030, language=EN, stringName=Lu LIU, firstName=Lu, middleName=null, lastName=LIU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3 Chengdu Aircraft Design and Research Institute,Aviation Industry Corporation of China,Ltd.,Chengdu Sichuan 610091,China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165681796042727585, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, authorId=1165681795883344030, language=CN, stringName=刘璐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3 中国航空工业集团有限公司 成都飞机设计研究所,四川 成都 610091, bio={"content":"

刘璐 高级工程师

"}, bioImg=null, bioContent=

刘璐 高级工程师

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165681795006734469, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=3, ext=[AuthorCompanyExt(id=1165681795010928774, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681795006734469, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Chengdu Aircraft Design and Research Institute,Aviation Industry Corporation of China,Ltd.,Chengdu Sichuan 610091,China), AuthorCompanyExt(id=1165681795019317383, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681795006734469, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 中国航空工业集团有限公司 成都飞机设计研究所,四川 成都 610091)])])], keywords=[Keyword(id=1165681796176945314, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=1, keyword=improved YOLOv8n), Keyword(id=1165681796231471267, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=2, keyword=aircraft rivets), Keyword(id=1165681796290191524, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=3, keyword=detachment), Keyword(id=1165681796336328869, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=4, keyword=anomaly detection), Keyword(id=1165681796390854822, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=5, keyword=space to depth convolution(SPD-Conv)), Keyword(id=1165681796436992167, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, orderNo=6, keyword=ablation experiment), Keyword(id=1165681796487323816, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=1, keyword=改进YOLOv8n), Keyword(id=1165681796533461161, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=2, keyword=飞机铆钉), Keyword(id=1165681796587987114, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=3, keyword=脱落), Keyword(id=1165681796638318763, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=4, keyword=异常检测), Keyword(id=1165681796709621932, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=5, keyword=空间深度转换卷积(SPD-Conv)), Keyword(id=1165681796759953581, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, orderNo=6, keyword=消融试验)], refs=[Reference(id=1165681798337011908, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=67, issue=6, pageStart=55, pageEnd=64, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=吴军, 单腾飞, 黄硕, journalName=航空制造技术, refType=null, unstructuredReference=吴军, 单腾飞, 黄硕, 等. 基于YOLOv7通道冗余改进的飞机蒙皮损伤检测[J]. 航空制造技术, 2024, 67(6):55-64., articleTitle=基于YOLOv7通道冗余改进的飞机蒙皮损伤检测, refAbstract=null), Reference(id=1165681798391537861, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=67, issue=6, pageStart=55, pageEnd=64, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=WU Jun, SHAN Tengfei, HUANG Shuo, journalName=Aeronautical Manufacturing Technology, refType=null, unstructuredReference=WU Jun, SHAN Tengfei, HUANG Shuo, et al. Detection for aircraft skin damage based on improvement of channel redundancy for YOLOv7[J]. Aeronautical Manufacturing Technology, 2024, 67(6):55-64., articleTitle=Detection for aircraft skin damage based on improvement of channel redundancy for YOLOv7, refAbstract=null), Reference(id=1165681798441869510, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2022, volume=909, issue=null, pageStart=69, pageEnd=74, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=KATUNIN A, DRAGAN K, journalName=Applied Mechanics and Materials, refType=null, unstructuredReference=KATUNIN A, DRAGAN K. Qualitative to quantitative non-destructive evaluation: a concept for D-sight inspections of aircraft structures[J]. Applied Mechanics and Materials, 2022, 909: 69-74., articleTitle=Qualitative to quantitative non-destructive evaluation: a concept for D-sight inspections of aircraft structures, refAbstract=null), Reference(id=1165681798517366983, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2022, volume=35, issue=10, pageStart=254, pageEnd=264, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=DING Meng, WU Boer, XU Juan, journalName=Chinese Journal of Aeronautics, refType=null, unstructuredReference=DING Meng, WU Boer, XU Juan, et al. Visual inspection of aircraft skin: automated pixel-level defect detection by instance segmentation[J]. Chinese Journal of Aeronautics, 2022, 35(10): 254-264., articleTitle=Visual inspection of aircraft skin: automated pixel-level defect detection by instance segmentation, refAbstract=null), Reference(id=1165681798567698632, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=66, issue=22, pageStart=69, pageEnd=77, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=周雨轩, 周文彬, 李小丽, journalName=航空制造技术, refType=null, unstructuredReference=周雨轩, 周文彬, 李小丽, 等. 飞机蒙皮缺陷自动化检测研究进展[J]. 航空制造技术, 2023, 66(22):69-77., articleTitle=飞机蒙皮缺陷自动化检测研究进展, refAbstract=null), Reference(id=1165681798626418889, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=66, issue=22, pageStart=69, pageEnd=77, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=ZHOU Yuxuan, ZHOU Wenbin, LI Xiaoli, journalName=Aeronautical Manufacturing Technology, refType=null, unstructuredReference=ZHOU Yuxuan, ZHOU Wenbin, LI Xiaoli, et al. Research progress in automatic detection of aircraft skin defects[J]. Aeronautical Manufacturing Technology, 2023, 66(22):69-77., articleTitle=Research progress in automatic detection of aircraft skin defects, refAbstract=null), Reference(id=1165681798706110666, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2017, volume=34, issue=10, pageStart=1138, pageEnd=1143, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=周志新, journalName=机电工程, refType=null, unstructuredReference=周志新. 机械裂纹无损检测方法综述[J]. 机电工程, 2017, 34(10): 1138-1143., articleTitle=机械裂纹无损检测方法综述, refAbstract=null), Reference(id=1165681798769025227, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2017, volume=34, issue=10, pageStart=1138, pageEnd=1143, url=null, language=null, rfNumber=[5], rfOrder=7, authorNames=ZHOU Zhixin, journalName=Journal of Mechanical & Electrical Engineering, refType=null, unstructuredReference=ZHOU Zhixin. Overview of NDT methods for mechanical cracks[J]. Journal of Mechanical & Electrical Engineering, 2017, 34(10): 1138-1143., articleTitle=Overview of NDT methods for mechanical cracks, refAbstract=null), Reference(id=1165681798819356876, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2017, volume=39, issue=6, pageStart=1137, pageEnd=1149, url=null, language=null, rfNumber=[6], rfOrder=8, authorNames=REN Shaoqing, HE Kaiming, GIRSHICK R, journalName=IEEE Transaction on Pattern Analysis and Machine Intelligence, refType=null, unstructuredReference=REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149., articleTitle=Faster R-CNN: towards real-time object detection with region proposal networks, refAbstract=null), Reference(id=1165681798865494221, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=11, pageStart=71, pageEnd=76, url=null, language=null, rfNumber=[7], rfOrder=9, authorNames=彭林聪, 王克瑞, 周浩, journalName=激光杂志, refType=null, unstructuredReference=彭林聪, 王克瑞, 周浩, 等. 基于改进SSD的目标检测算法[J]. 激光杂志, 2024, 45(11):71-76., articleTitle=基于改进SSD的目标检测算法, refAbstract=null), Reference(id=1165681798911631566, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=11, pageStart=71, pageEnd=76, url=null, language=null, rfNumber=[7], rfOrder=10, authorNames=PENG Lincong, WANG Kerui, ZHOU Hao, journalName=Laser Journal, refType=null, unstructuredReference=PENG Lincong, WANG Kerui, ZHOU Hao, et al. Object detection algorithm based on improved SSD[J]. Laser Journal, 2024, 45(11):71-76., articleTitle=Object detection algorithm based on improved SSD, refAbstract=null), Reference(id=1165681798957768911, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=34, issue=3, pageStart=237, pageEnd=246, url=null, language=null, rfNumber=[8], rfOrder=11, authorNames=左明成, 焦文华, journalName=中国安全科学学报, refType=null, unstructuredReference=左明成, 焦文华. 面向煤矿井下作业场景的安全帽佩戴识别算法[J]. 中国安全科学学报, 2024, 34(3):237-246., articleTitle=面向煤矿井下作业场景的安全帽佩戴识别算法, refAbstract=null), Reference(id=1165681799008100560, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=34, issue=3, pageStart=237, pageEnd=246, url=null, language=null, rfNumber=[8], rfOrder=12, authorNames=ZUO Mingcheng, JIAO Wenhua, journalName=China Safety Science Journal, refType=null, unstructuredReference=ZUO Mingcheng, JIAO Wenhua. Helmet-wearing recognition algorithm for coal mine underground operation scenarios[J]. China Safety Science Journal, 2024, 34(3):237-246., articleTitle=Helmet-wearing recognition algorithm for coal mine underground operation scenarios, refAbstract=null), Reference(id=1165681799142318289, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=10, pageStart=1487, pageEnd=1493, url=null, language=null, rfNumber=[9], rfOrder=13, authorNames=张立国, 袁煜淋, 金梅, journalName=计量学报, refType=null, unstructuredReference=张立国, 袁煜淋, 金梅, 等. 基于改进YOLOv8n的无人机目标检测算法研究[J]. 计量学报, 2024, 45(10):1487-1493., articleTitle=基于改进YOLOv8n的无人机目标检测算法研究, refAbstract=null), Reference(id=1165681799196844242, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=10, pageStart=1487, pageEnd=1493, url=null, language=null, rfNumber=[9], rfOrder=14, authorNames=ZHANG Liguo, YUAN Yulin, JIN Mei, journalName=Acta Metrologica Sinica, refType=null, unstructuredReference=ZHANG Liguo, YUAN Yulin, JIN Mei, et al. Research on UAV target detection algorithm based on improved YOLOv8n[J]. Acta Metrologica Sinica, 2024, 45(10):1487-1493., articleTitle=Research on UAV target detection algorithm based on improved YOLOv8n, refAbstract=null), Reference(id=1165681799255564499, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=60, issue=21, pageStart=205, pageEnd=214, url=null, language=null, rfNumber=[10], rfOrder=15, authorNames=姚景丽, 程光, 万飞, journalName=计算机工程与应用, refType=null, unstructuredReference=姚景丽, 程光, 万飞, 等. 改进YOLOv8的轻量化轴承缺陷检测算法[J]. 计算机工程与应用, 2024, 60(21):205-214., articleTitle=改进YOLOv8的轻量化轴承缺陷检测算法, refAbstract=null), Reference(id=1165681799314284756, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=60, issue=21, pageStart=205, pageEnd=214, url=null, language=null, rfNumber=[10], rfOrder=16, authorNames=YAO Jingli, CHENG Guang, WAN Fei, journalName=Computer Engineering and Applications, refType=null, unstructuredReference=YAO Jingli, CHENG Guang, WAN Fei, et al. Improved lightweight bearing defect detection algorithm of YOLOv8[J]. Computer Engineering and Applications, 2024, 60(21):205-214., articleTitle=Improved lightweight bearing defect detection algorithm of YOLOv8, refAbstract=null), Reference(id=1165681799368810709, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=55, issue=10, pageStart=318, pageEnd=326, url=null, language=null, rfNumber=[11], rfOrder=17, authorNames=肖伸平, 赵倩颖, 曾甲元, journalName=农业机械学报, refType=null, unstructuredReference=肖伸平, 赵倩颖, 曾甲元, 等. 基于YOLO-DCL的复杂环境油茶果遮挡检测与计数研究[J]. 农业机械学报, 2024, 55(10):318-326,480., articleTitle=基于YOLO-DCL的复杂环境油茶果遮挡检测与计数研究, refAbstract=null), Reference(id=1165681799419142358, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2024, volume=55, issue=10, pageStart=318, pageEnd=326, url=null, language=null, rfNumber=[11], rfOrder=18, authorNames=XIAO Shenping, ZHAO Qianying, ZENG Jiayuan, journalName=Transactions of the Chinese Society for Agricultural Machinery, refType=null, unstructuredReference=XIAO Shenping, ZHAO Qianying, ZENG Jiayuan, et al. Camellia oleifera fruits occlusion detection and counting in complex environments based on improved YOLO-DCL[J]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55(10):318-326,480., articleTitle=Camellia oleifera fruits occlusion detection and counting in complex environments based on improved YOLO-DCL, refAbstract=null), Reference(id=1165681799477862615, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=2, pageStart=82, pageEnd=88, url=null, language=null, rfNumber=[12], rfOrder=19, authorNames=夏正洪, 魏汝祥, 李彦冬, journalName=中国安全科学学报, refType=null, unstructuredReference=夏正洪, 魏汝祥, 李彦冬. 改进YOLOv3算法下通航机场场面运动目标检测[J]. 中国安全科学学报, 2023, 33(2):82-88., articleTitle=改进YOLOv3算法下通航机场场面运动目标检测, refAbstract=null), Reference(id=1165681799528194264, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=2, pageStart=82, pageEnd=88, url=null, language=null, rfNumber=[12], rfOrder=20, authorNames=XIA Zhenghong, WEI Ruxiang, LI Yandong, journalName=China Safety Science Journal, refType=null, unstructuredReference=XIA Zhenghong, WEI Ruxiang, LI Yandong. Moving target detection of general aviation airport based on improved YOLOv3 algorithm[J]. China Safety Science Journal, 2023, 33(2):82-88., articleTitle=Moving target detection of general aviation airport based on improved YOLOv3 algorithm, refAbstract=null), Reference(id=1165681799582720217, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2022, volume=37, issue=2, pageStart=179, pageEnd=184, url=null, language=null, rfNumber=[13], rfOrder=21, authorNames=农昌瑞, 张静, 杨智勇, journalName=海军航空大学学报, refType=null, unstructuredReference=农昌瑞, 张静, 杨智勇. 改进的YOLOv4在飞机蒙皮损伤检测中的应用[J]. 海军航空大学学报, 2022, 37(2):179-184,230., articleTitle=改进的YOLOv4在飞机蒙皮损伤检测中的应用, refAbstract=null), Reference(id=1165681799637246170, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2022, volume=37, issue=2, pageStart=179, pageEnd=184, url=null, language=null, rfNumber=[13], rfOrder=22, authorNames=NONG Changrui, ZHANG Jing, YANG Zhiyong, journalName=Journal of Naval Aviation University, refType=null, unstructuredReference=NONG Changrui, ZHANG Jing, YANG Zhiyong. Application of improved YOLOv4 in aircraft skin damage detection[J]. Journal of Naval Aviation University, 2022, 37(2):179-184, 230., articleTitle=Application of improved YOLOv4 in aircraft skin damage detection, refAbstract=null), Reference(id=1165681799687577820, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=60, issue=10, pageStart=260, pageEnd=268, url=null, language=null, rfNumber=[14], rfOrder=23, authorNames=王童, 王宏伟, 王红, journalName=激光与光电子学进展, refType=null, unstructuredReference=王童, 王宏伟, 王红. 基于旋转目标检测的飞机蒙皮损伤检测[J]. 激光与光电子学进展, 2023, 60(10):260-268., articleTitle=基于旋转目标检测的飞机蒙皮损伤检测, refAbstract=null), Reference(id=1165681799737909471, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=60, issue=10, pageStart=260, pageEnd=268, url=null, language=null, rfNumber=[14], rfOrder=24, authorNames=WANG Tong, WANG Hongwei, WANG Hong, journalName=Laser & Optoelectronics Progress, refType=null, unstructuredReference=WANG Tong, WANG Hongwei, WANG Hong. Aircraft skin damage detection based on rotating object detection[J]. Laser & Optoelectronics Progress, 2023, 60(10):260-268., articleTitle=Aircraft skin damage detection based on rotating object detection, refAbstract=null), Reference(id=1165681799792435424, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=19, issue=6, pageStart=199, pageEnd=205, url=null, language=null, rfNumber=[15], rfOrder=25, authorNames=夏正洪, 何琥, 吴建军, journalName=中国安全生产科学技术, refType=null, unstructuredReference=夏正洪, 何琥, 吴建军, 等. 基于深度学习的航空铆钉分类及异常情况检测[J]. 中国安全生产科学技术, 2023, 19(6):199-205., articleTitle=基于深度学习的航空铆钉分类及异常情况检测, refAbstract=null), Reference(id=1165681799855349986, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, doi=null, pmid=null, pmcid=null, year=2023, volume=19, issue=6, pageStart=199, pageEnd=205, url=null, language=null, rfNumber=[15], rfOrder=26, authorNames=XIA Zhenghong, HE Hu, WU Jianjun, journalName=Journal of Safety Science and Technology, refType=null, unstructuredReference=XIA Zhenghong, HE Hu, WU Jianjun, et al. Aviation rivet classification and abnormal situation detection based on deep learning[J]. Journal of Safety Science and Technology, 2023, 19(6):199-205., articleTitle=Aviation rivet classification and abnormal situation detection based on deep learning, refAbstract=null)], funds=[Fund(id=1165681798076965056, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, awardId=2024YFC3014400, language=CN, fundingSource=国家重点研发计划(2024YFC3014400), fundOrder=null, country=null), Fund(id=1165681798135685313, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, awardId=F2024KF04C, language=CN, fundingSource=民航飞行技术与飞行安全科研基地(F2024KF04C), fundOrder=null, country=null), Fund(id=1165681798186016962, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, awardId=2024YFTX0078, language=CN, fundingSource=四川省重点研发计划项目(2024YFTX0078), fundOrder=null, country=null), Fund(id=1165681798232154307, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, awardId=24CAFUC03047, language=CN, fundingSource=中国民用航空飞行学院基本科研项目(24CAFUC03047), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1165681794851545214, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=1, ext=[AuthorCompanyExt(id=1165681794855739519, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China), AuthorCompanyExt(id=1165681794868322432, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794851545214, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307)]), AuthorCompany(id=1165681794918654081, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=2, ext=[AuthorCompanyExt(id=1165681794927042690, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794918654081, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 North Regional Administration,Civil Aviation Administration of China,Beijing 100621,China), AuthorCompanyExt(id=1165681794935431299, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681794918654081, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 中国民用航空华北地区管理局,北京 100621)]), AuthorCompany(id=1165681795006734469, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, xref=3, ext=[AuthorCompanyExt(id=1165681795010928774, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681795006734469, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Chengdu Aircraft Design and Research Institute,Aviation Industry Corporation of China,Ltd.,Chengdu Sichuan 610091,China), AuthorCompanyExt(id=1165681795019317383, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, companyId=1165681795006734469, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 中国航空工业集团有限公司 成都飞机设计研究所,四川 成都 610091)])], figs=[ArticleFig(id=1165681796906754222, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.1, caption=Network architecture of the improvement of YOLOv8n, figureFileSmall=F/aZdPwRVIKqiClu8XmwBw==, figureFileBig=CP9MLMhDh1xMeFOMbUbh2g==, tableContent=null), ArticleFig(id=1165681796961280175, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图1, caption=改进YOLOv8n算法的网络结构

注:SPD-Conv为空间深度转换卷积(Space to Depth Convolution); C2f为跨阶段部分网络融合模块(Cross Stage PartialDarknet53 to 2-stage Feature Pyramid Network);SPPF-LSKA为融入了LSKA的SPPF。

, figureFileSmall=F/aZdPwRVIKqiClu8XmwBw==, figureFileBig=CP9MLMhDh1xMeFOMbUbh2g==, tableContent=null), ArticleFig(id=1165681797011611824, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.2, caption=Rivet samples and bounding box width-height ratio distribution, figureFileSmall=BrsJXxiE6gifyXpeZS1IcQ==, figureFileBig=SxXw21JSyRpzC8BOZ2zmkw==, tableContent=null), ArticleFig(id=1165681797061943473, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图2, caption=铆钉样本及标签框宽高比例分布, figureFileSmall=BrsJXxiE6gifyXpeZS1IcQ==, figureFileBig=SxXw21JSyRpzC8BOZ2zmkw==, tableContent=null), ArticleFig(id=1165681797141635250, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.3, caption=Structure of SPD-Conv, figureFileSmall=1OoLmiLkDBQPcZIeFzAUOA==, figureFileBig=bMY37WCObjZxVv3asPnIDQ==, tableContent=null), ArticleFig(id=1165681797196161203, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图3, caption=SPD-Conv结构, figureFileSmall=1OoLmiLkDBQPcZIeFzAUOA==, figureFileBig=bMY37WCObjZxVv3asPnIDQ==, tableContent=null), ArticleFig(id=1165681797242298548, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.4, caption=Structure of SPPF and SPPF_LSKA, figureFileSmall=GcaUMfzwOpENkFBh65PRAw==, figureFileBig=OI5zEIItlW7bG91qFMXg8w==, tableContent=null), ArticleFig(id=1165681797338767541, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图4, caption=SPPF和SPPF_LSKA模块, figureFileSmall=GcaUMfzwOpENkFBh65PRAw==, figureFileBig=OI5zEIItlW7bG91qFMXg8w==, tableContent=null), ArticleFig(id=1165681797389099190, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.5, caption=Comparison of algorithm improvement effects, figureFileSmall=inS4+t8ZOuLOeW94NVUmYQ==, figureFileBig=rri8bBNxnoRYVF84xA8cNQ==, tableContent=null), ArticleFig(id=1165681797472985271, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图5, caption=算法改进效果对比, figureFileSmall=inS4+t8ZOuLOeW94NVUmYQ==, figureFileBig=rri8bBNxnoRYVF84xA8cNQ==, tableContent=null), ArticleFig(id=1165681797523316920, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Fig.6, caption=Comparison of algorithm detection performance before and after improvement, figureFileSmall=xFdZxf3YPR5Hyo9tpuD5EA==, figureFileBig=9Ac0O75PYkpnr7Hqa4EG/Q==, tableContent=null), ArticleFig(id=1165681797573648569, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=图6, caption=改进前后算法检测效果对比, figureFileSmall=xFdZxf3YPR5Hyo9tpuD5EA==, figureFileBig=9Ac0O75PYkpnr7Hqa4EG/Q==, tableContent=null), ArticleFig(id=1165681797623980218, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Table 1, caption=

Training parameter settings

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 参数
训练总轮数 300 动量 0.937
批量 8 权值衰减 0.000 5
耐心值 50 热身训练轮次 3
输入图像尺寸 640×640 热身训练动量 0.8
优化器 SGD 提前关闭数据
增强的轮数
10
线程数 1 初始学习率 0.01
), ArticleFig(id=1165681797699477691, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=表1, caption=

训练参数设置

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 参数
训练总轮数 300 动量 0.937
批量 8 权值衰减 0.000 5
耐心值 50 热身训练轮次 3
输入图像尺寸 640×640 热身训练动量 0.8
优化器 SGD 提前关闭数据
增强的轮数
10
线程数 1 初始学习率 0.01
), ArticleFig(id=1165681797770780860, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Table 2, caption=

Results of ablation experiment

, figureFileSmall=null, figureFileBig=null, tableContent=
算法 P/
%
R/
%
mAP@
0.5/%
参数
量/106
FPS/
(帧·s-1)
YOLOv8n 80.2 63.2 72.1 3.01 114.7
+P2 86.7 73.7 82.2 2.93 105.3
+P2+SPD-Conv 85.3 80.4 86.7 2.92 104.5
+P2+SPD-Conv
+SPPF-LSKA
86.4 79.2 87.1 3.20 102.0
), ArticleFig(id=1165681797846278333, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=表2, caption=

消融试验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
算法 P/
%
R/
%
mAP@
0.5/%
参数
量/106
FPS/
(帧·s-1)
YOLOv8n 80.2 63.2 72.1 3.01 114.7
+P2 86.7 73.7 82.2 2.93 105.3
+P2+SPD-Conv 85.3 80.4 86.7 2.92 104.5
+P2+SPD-Conv
+SPPF-LSKA
86.4 79.2 87.1 3.20 102.0
), ArticleFig(id=1165681797900804286, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=EN, label=Table 3, caption=

Results of Comparative experiments

, figureFileSmall=null, figureFileBig=null, tableContent=
算法 P/% R/% mAP@0.5/% FPS/
(帧·s-1)
YOLOv8n 80.2 63.2 72.1 114.7
YOLOv8s 84.9 68.7 78.7 112.6
YOLOv5n 83.7 64.6 74.1 118.5
YOLOv5s 88.2 57.4 69.1 117.7
Ours 86.4 79.2 87.1 102.0
), ArticleFig(id=1165681797959524543, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106707819029331, language=CN, label=表3, caption=

对比试验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
算法 P/% R/% mAP@0.5/% FPS/
(帧·s-1)
YOLOv8n 80.2 63.2 72.1 114.7
YOLOv8s 84.9 68.7 78.7 112.6
YOLOv5n 83.7 64.6 74.1 118.5
YOLOv5s 88.2 57.4 69.1 117.7
Ours 86.4 79.2 87.1 102.0
)], attaches=null, journal=Journal(id=1123942128916217864, delFlag=0, nameCn=中国安全科学学报, nameEn=China Safety Science Journal, nameHistory1=null, nameHistory2=null, issn=1003-3033, eissn=, cn=11-2865/X, coden=null, periodic=0, language=CN, oaType=0, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=fkqsFM6VKlHC4gCtS5XqTw==, journalPrice=null, startedYear=null, abbrevIsoEn=Chin Saf Sci J, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1754269350027, createdBy=null, updatedBy=13701087609, firstLetterCn=C, firstLetterEn=C, subjectCode=Engineering, subjectName=工程, subjectCodeEn=Engineering, subjectNameEn=null, picCn=fkqsFM6VKlHC4gCtS5XqTw==, picEn=SHn9HgqSxtJrOcAxqD++4Q==, jcr=null, cjcr=null, exts=[JournalExt(id=1159052918994595848, language=CN, name=中国安全科学学报, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.cssjj.com.cn/, createdTime=1754269350050, updatedTime=1754269350050, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://www.cssjj.com.cn/CN/column/item15.shtml, submissionAuthorUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionReviewUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionCeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionAeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, option={"copyright":""}), JournalExt(id=1159052919040733193, language=EN, name=China Safety Science Journal, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.cssjj.com.cn/EN/1003-3033/home.shtml, createdTime=1754269350061, updatedTime=1754269350061, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=https://synbioj.cip.com.cn/EN/column/column3.shtml, submissionAuthorUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionReviewUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionCeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionAeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, option={"copyright":""})], databaseList=null, tenantJournalId=1146031787341344770, websiteList=[Website(id=1148243202345263519, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146031787341344770, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN, language=CN, createTime=1751692112766, createBy=18614031015, updateTime=1753502583634, updateBy=18614031015, name=《中国安全科学学报》中文站点, tplId=1146099689490845704, title=中国安全科学学报, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148618794941046792, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=articleTextType, value=kx, createTime=1751781661020, updateTime=1751781661020, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794911686661, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=banner, value=null, createTime=1751781661012, updateTime=1751781661012, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794894909444, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=tui0IVO9FMwB61HHtX5scg==, createTime=1751781661008, updateTime=1751781661008, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794932658183, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751781661017, updateTime=1751781661017, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794924269574, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751781661015, updateTime=1751781661015, creator=18614031015, updator=18614031015)]), Website(id=1155836763751993353, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146031787341344770, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN, language=EN, createTime=1753502558893, createBy=18614031015, updateTime=1753524450387, updateBy=18614031015, name=《中国安全科学学报》英文站点, tplId=1146101810881728533, title=China Safety Science Journal, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155895925743669425, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=articleTextType, value=kx, createTime=1753516664205, updateTime=1753516664205, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925722697902, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=banner, value=null, createTime=1753516664200, updateTime=1753516664200, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925714309293, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=tui0IVO9FMwB61HHtX5scg==, createTime=1753516664198, updateTime=1753516664198, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925735280816, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753516664203, updateTime=1753516664203, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925731086511, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753516664202, updateTime=1753516664202, creator=18614031015, updator=18614031015)])], journalTitle=中国安全科学学报, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=China Safety Science Journal, journalPhotoCn=fkqsFM6VKlHC4gCtS5XqTw==, journalPhotoEn=SHn9HgqSxtJrOcAxqD++4Q==, journalFirstLetter=C, journalRecommend=null, journalNew=null, journalCollection=1, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/10.16265/j.cnki.issn1003-3033.2025.02.0880, detailUrlEn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN/10.16265/j.cnki.issn1003-3033.2025.02.0880, pdfUrlCn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/PDF/10.16265/j.cnki.issn1003-3033.2025.02.0880, pdfUrlEn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN/PDF/10.16265/j.cnki.issn1003-3033.2025.02.0880, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于改进YOLOv8n的飞机铆钉及脱落异常检测算法
收藏切换
PDF下载
夏正洪 1 , 何琥 1 , 杨磊 1 , 吴建军 2 , 刘璐 3
中国安全科学学报 | 安全工程技术 2025,35(2): 66-72
收起
收藏切换
中国安全科学学报 | 安全工程技术 2025, 35(2): 66-72
基于改进YOLOv8n的飞机铆钉及脱落异常检测算法
全屏
夏正洪1 , 何琥1, 杨磊1, 吴建军2, 刘璐3
作者信息
  • 1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307
  • 2 中国民用航空华北地区管理局,北京 100621
  • 3 中国航空工业集团有限公司 成都飞机设计研究所,四川 成都 610091
  • 夏正洪 (1985—),男,四川乐山人,硕士,教授,主要从事基于深度学习的航空运行安全风险评价研究。E-mail:

    刘璐 高级工程师

Aircraft rivet and detachment anomaly detection algorithm based on improved YOLOv8n
Zhenghong XIA1 , Hu HE1, Lei YANG1, Jianjun WU2, Lu LIU3
Affiliations
  • 1 School of Air Traffic Management,Civil Aviation Flight University of China,Guanghan Sichuan 618307,China
  • 2 North Regional Administration,Civil Aviation Administration of China,Beijing 100621,China
  • 3 Chengdu Aircraft Design and Research Institute,Aviation Industry Corporation of China,Ltd.,Chengdu Sichuan 610091,China
出版时间: 2025-02-28 doi: 10.16265/j.cnki.issn1003-3033.2025.02.0880
文章导航
收藏切换

为解决飞机铆钉小目标检测时易出现漏检的问题,提出一种基于改进YOLOv8n的飞机铆钉及脱落异常检测算法。首先,通过添加小目标检测层,更好地融合骨干网络中的浅层细节信息,提高算法的特征融合能力以及对铆钉小目标的识别与定位性能;其次,将骨干网络中的前2次卷积替换为空间深度转换卷积(SPD-Conv),通过特征图的重组与非跨步卷积的组合,减少算法在下采样过程中的信息丢失;然后,将大型可分离核注意力(LSKA)融入快速空间金字塔池化(SPPF)模块中,通过计算每个特征图上的空间权重和通道权重,捕捉空间与通道之间的依赖关系,并调整特征图,增强算法对铆钉特征信息的提取和识别能力;最后,基于自建的飞机铆钉数据集进行消融试验和对比试验。结果表明:所提算法能实时检测飞机铆钉及脱落异常,较YOLOv8n算法检测结果在精确率、召回率、平均精度均值(mAP)分别提升6.5%、16%、15%,较其他主流算法的检测性能均有较大提升。

改进YOLOv8n  /  飞机铆钉  /  脱落  /  异常检测  /  空间深度转换卷积(SPD-Conv)  /  消融试验

In order to address the issue of missed detections during small target detection of aircraft rivets,an improved YOLOv8n algorithm for the detection of aircraft rivets and their anomalies was proposed. First,by adding a small object detection head,the shallow detail information in the backbone network was better fused,enhancing the model's feature fusion capability and its ability to recognize and locate small rivet targets. Second,the first two convolutions in the backbone network were replaced with SPD-Conv,which reduces information loss during down sampling through the combination of feature map reorganization and non-stride convolutions. Finally,large separable kernel attention (LSKA) was integrated into the spatial pyramid pooling fast (SPPF) module,capturing the dependencies between spatial and channel dimensions by calculating spatial and channel weights on each feature map and adjusting the feature maps to enhance the algorithm's ability to extract and recognize rivet feature information. Ablation experiments and comparative experiments were conducted based on a self-built aircraft rivet dataset. The results show that the proposed algorithm can achieve real-time identification of aircraft rivets and their anomalies,with precision,recall,and mean average precision (mAP) values improved by 6.5%,16%,and 15%,respectively,compared to the YOLOv8n algorithm. The detection performance is also significantly better than other mainstream algorithms.

improved YOLOv8n  /  aircraft rivets  /  detachment  /  anomaly detection  /  space to depth convolution(SPD-Conv)  /  ablation experiment
夏正洪, 何琥, 杨磊, 吴建军, 刘璐. 基于改进YOLOv8n的飞机铆钉及脱落异常检测算法. 中国安全科学学报, 2025 , 35 (2) : 66 -72 . DOI: 10.16265/j.cnki.issn1003-3033.2025.02.0880
Zhenghong XIA, Hu HE, Lei YANG, Jianjun WU, Lu LIU. Aircraft rivet and detachment anomaly detection algorithm based on improved YOLOv8n[J]. China Safety Science Journal, 2025 , 35 (2) : 66 -72 . DOI: 10.16265/j.cnki.issn1003-3033.2025.02.0880
铆钉在飞机制造领域扮演着至关重要的角色,它们不仅是飞机结构连接的重要组成部分,还直接影响着飞机的安全性和飞行性能。据统计,一架商用飞机的铆钉数量高达上百万颗。然而,高速飞行及气流的影响会产生剧烈震动,可能导致铆钉松动、脱落和失效,不仅降低飞机蒙皮结构的完整性和飞行性能,也使得运营维护成本及安全风险隐患急剧增加。因此,研究先进的小目标检测技术,快速准确检测飞机铆钉松动及脱落异常情况,对保障飞机的安全飞行具有重要意义。
传统的飞机蒙皮及铆钉检测手段主要通过机务工程师目视[1],缺点是效率低、主观性强、受环境因素影响大[2-4],存在错检和漏检问题[5]。随着深度神经网络的发展和硬件设备的进步,卷积神经网络(Convolutional Neural Networks,CNN)越来越广泛地被应用在目标检测、图像分类等任务中。目前,主流算法如基于区域的更快卷积神经网络[6](Faster Region-based Convolutional Neural Network,Faster R-CNN)、单次多框检测器[7-8](Single Shot MultiBox Detector,SSD)、你只看一次[9-12](You Only Look Once,YOLO)系列等逐渐被运用到航空航天等工业领域,以实现对飞机机身损伤等的检测。农昌瑞等[13]通过改进YOLOv4算法,在增加少量计算成本的情况下提高了YOLOv4对机身多个类型损伤目标的检测能力;王童等[14]采用特征金字塔网络(Feature Pyramid Network,FPN)、引入坐标注意力(Coordinate Attention,CA)模块、采用边界框边缘感知向量描述损伤位置,准确定位任意方向的损伤,提高了对铆钉损伤的检测能力;夏正洪等[15]提出了一种基于深度学习的航空铆钉分类及脱落异常情况检测方法,基于YOLOX识别出航空铆钉分类及脱落异常,但随着输入目标尺寸变化时易出现漏检问题。虽现有研究在铆钉识别与定位上取得一定成果,但算法对铆钉特征信息的提取能力还有待提升,其检测精度与召回率还有较大提升空间。
因此,笔者拟提出一种基于改进YOLOv8n的飞机铆钉及脱落异常检测算法,从加入小目标检测层、替换骨干网络的卷积模块、集成大型可分离核注意力(Large Separable Kernel Attention,LSKA)到快速空间金字塔池化(Spatial Pyramid Pooling Fast,SPPF)模块3方面改进YOLOv8n算法,以期实现对飞机铆钉小目标的高效、准确定位,及时发现铆钉脱落异常情况,提升飞机的运行安全保障能力。
YOLOv8算法由骨干特征提取网络、颈部、检测头3个部分共同构成。骨干网络负责从输入图像中提取特征,将图像转化为包含丰富语义信息的特征表示;颈部网络负责融合来自骨干网络的特征,提高算法性能;检测头负责识别与定位目标。按照其算法大小从小到大可分为v8n、v8s、v8m、v8l、v8x共5个版本,算法越大则宽度与深度越大,检测效果越好,但算法复杂度也随之增加。为减少计算量、加快算法推理速度,选择在YOLOv8n算法的基础上改进。
飞机铆钉分布密集且尺寸很小,随着YOLOv8骨干网络卷积的层层递进,特征图大小逐层递减,特征图中铆钉的细节特征不断丢失,骨干网络中提取到的有效特征少,会进一步影响到颈部网络特征融合,导致算法对飞机铆钉的检测效果较差。为解决算法对飞机铆钉检测效果不佳的问题,分别从骨干网络、颈部网络与检测头3方面改进YOLOv8n算法(图1)。
1) 颈部网络中,在原来最后一个上采样基础上继续增加上采样模块,与骨干网络第1次C2f处理后的特征图作特征通道拼接,通过进一步充分融合骨干网络提取到的细节特征,加强算法对小尺寸目标的感知能力。同时,在检测头部分增加一个160×160大小的检测头,识别与定位小目标。
2) 骨干网络中,将前2次卷积模块替换为专门针对小目标的SPD-Conv,通过特征图重组与非跨步卷积的组合,避免小目标特征信息随着特征图的缩小不断丢失,增强对飞机铆钉的感知能力。
3) 将LSKA融合到SPPF模块中,通过计算每个特征图上的空间权重和通道权重,捕捉空间与通道之间的依赖关系,并调整特征图,增强算法对复杂多尺度特征的捕捉能力。
飞机铆钉尺寸较小,分布密集,部分铆钉分布情况如图2a所示,且在数据集中绝大部分铆钉边界框宽高均小于原图宽高的0.1倍,铆钉标签框宽高比例分布情况如图2b所示。小目标相比于其他目标所包含信息更少,算法难以学习到其具有鉴别力的特征,特别是在原始YOLOv8n算法的骨干特征提取网络中下采样倍率较大,随着骨干特征提取网络的层层深入,小目标特征信息逐渐丢失,算法难以提取到充分的铆钉特征信息,并进行精确识别与定位。
为此,提出增加一个小目标检测层,即在算法原有基础上增加特征图尺寸为160×160的检测层。首先,在颈部网络中第2个C2f后继续增加上采样层,将特征图尺寸进一步放大到160×160,通过特征拼接层将特征图与骨干网络第1个C2f层处理后大小为160×160特征图进行通道维度的拼接,让颈部网络融合更多浅层更细节和包含更丰富小目标的特征信息。然后,经过C2f层处理后,传入160×160的检测头中,提高算法对小目标的感知能力与识别和定位性能。同时,在颈部网络中堆叠的小目标特征信息将会进一步沿着下采样路径传递到其他3个尺度的检测头中,增强算法的特征融合能力与对不同尺寸目标的检测性能。
普通卷积由于使用了跨步卷积的方式,即步长≥2,往往会导致细粒度信息的丢失和特征的学习效果较差,而面对飞机铆钉这种小目标,其分辨率较低且可供算法学习的上下文信息有限,经过骨干网络的几次卷积操作之后其特征信息不断丢失,导致算法对铆钉的感知能力较弱,检测性能差。因此,引入SPD-Conv替换骨干网络的第1次与第2次的卷积操作。SPD-Conv由空间到深度层和一个非跨步卷积层共同组成(图3),其中,S为特征图高和宽的值,C为特征图通道数。
在空间到深度(Space to Depth,SPD)层中,定义特征图X大小为S×S×C1,将特征图X划分为一系列子特征图:
f 0,0 = X [ 0 : S : k 0 : S : k ]
f 1,0 = X 1 : S : k 0 : S : k
f 0,1 = X 0 : S : k 1 : S : k
f k - 1,0 = X k - 1 : S : k 0 : S : k
f 0 k - 1 = X [ 0 : S : k k - 1 : S : k ]
f k - 1 k - 1 = X [ k - 1 : S : k k - 1 : S : k ]
式中:fij(i=0,1,k-1,j=0,1,k-1)为X划分出的子特征图,当k=2,此时特征图X就被划分为f0,0f1,0f0,1f1,1 4个子特征图,每个子特征图高和宽的值S均为原特征图X的一半,即S/2,通道数C1不发生变化。接着将f0,0f1,0f0,1f1,14个子特征图沿着特征通道维度拼接,得到新的特征图X1,此时X1大小为(S/2)×(S/2)×4×C1。在SPD层后,紧接着对X1进行步长为1通道数为C2的非跨步卷积操作,最终得到特征图X2,其形状为(S/2)×(S/2)×C2。整个SPD-Conv过程中,由于进行的是特征图的重组与非跨步卷积的操作,既实现了传统卷积下采样的效果,又保留了特征图中所有的细节信息,这对铆钉小目标识别任务非常有利。
SPPF模块通过融合骨干网络中提取到的不同尺度的特征图,可增强算法对待检测目标尺度变化的感知能力,提高算法对多尺度特征的提取效果,是YOLOv8中的重要组成部分。将LSKA融入SPPF中,通过计算每个特征图上的空间权重和通道权重,捕捉空间与通道之间的依赖关系,并调整特征图,使算法能够更加关注与飞机铆钉相关的特征,提高算法检测精度,如图4所示。
所采用的数据集是由机务工程师在实际机身检查任务中拍摄的铆钉图像数据,使用图像标注软件Labelimg制作而成。该数据集共包含2类目标,rivet为正常铆钉,abn为机身异常铆钉。其中,铆钉实例近9 000个,异常铆钉实例近500个,绝大部分实例均为小目标。数据集总共包含509张机身铆钉图像数据,按照7∶2∶1比例划分为训练集、测试集与验证集。参数设置见表1
整个试验过程均采用以上参数训练算法,输入图像尺寸均为640×640,训练总轮数为300,耐心值设置为50,即在算法训练过程中算法性能如果超过50轮次均未提升则触发早停而停止训练。批量值设置为8,每次将8张图像合并传入算法训练,优化器选择SGD,动量设置为0.937,权值衰减为0.000 5。同时,采用热身训练策略,设置前3次训练方式为热身训练,帮助算法收敛,热身训练动量设置为0.8。训练过程中开启马赛克数据增强策略,将多张图像经过随机裁剪、旋转等操作后合并为一张图像传入算法,以增加数据多样性,提高算法鲁棒性和对小目标检测的性能。
为评估算法性能,引入精确率P、召回率R、平均精度(Average Precision,AP)、平均精度均值(Mean Average Precision,mAP)及每秒帧数(Frames Per Second,FPS)作为评价指标,公式如下:
P = T P T P + F P
R = T P T P + F N
$\mathrm{AP}=\int_0^1P(R)\mathrm{d}R$
m A P = i = 1 M A P i M
F P S = N ( p ) T ( p )
式中:TP为正类被预测为正类的样本数量;FP为负类被预测为正类的样本数量;FN为正类被预测为负类的样本量;P(R)为精确率-召回率曲线;M为数据集中类别数量;N(p)为处理图像的总数;T(p)为处理图像的时间。
为验证各个改进模块对算法铆钉检测性能的提升效果,设计4组消融试验,分别在YOLOv8n基础上添加小目标检测层(P2)、SPD-Conv,在SPPF中融入LSKA模块,将各个改进模块依次加入YOLOv8n算法中,最终得到试验结果,见表2
可见:在加入P2小目标检测层后,算法更好地融合了骨干网络中的浅层细节信息,其颈部网络特征融合能力得到提高;由于增加了对应的检测头,铆钉小目标的识别与定位性能也得到增强,算法对小目标的感知能力得到显著提升,具体表现为PR、mAP@0.5分别提升6.5%、10.5%、10.1%,PR达到86.7%、73.7%。在加入P2检测层后继续替换SPD-Conv,此时算法对小目标的检测性能得到进一步提高,R、mAP@0.5分别提升6.7%、4.5%,R高达80.4%,算法参数量降低至2.92×106。最后,将LSKA融入SPPF中,P和mAP@0.5较前一步试验增长1.1%、0.4%,mAP@0.5高达87.1%。虽然最终算法的PR未达到试验过程中的最佳值,但是算法性能却达到更好的平衡,综合性能达到更优。总的来看,相对于基础YOLOv8n算法,改进后的算法P提升6.2%,R提升16%,mAP@0.5提升15%,而由于性能提升所带来的算法参数增长只有0.19×106,且FPS在整个试验过程中都高于100,说明在算法的改进过程中仅增加少许的复杂度,能够较大提升算法性能。
为更直观地观察算法改进后的效果,将原算法与改进后的算法在相同的试验环境中测试,从PR、mAP@0.5与mAP@0.5∶0.95这 4个方面改进前后算法,对比结果如图5所示。
图5可知:算法改进后收敛速度更快,且改进后算法的PR、mAP值比原算法更高,算法性能更好,改进后算法在训练的第258轮时触发早停机制,近50轮训练算法性能没有进一步提升,节省了计算资源。
为验证文中算法的有效性,对所提算法与YOLOv8n、YOLOv8s以及目前主流目标检测算法YOLOv5n、YOLOv5s,对比其PR、mAP值与FPS,结果见表3
表3可知:改进后的算法在机身铆钉及异常情况检测的任务中,R与mAP值在5种算法中达到了最高,分别为79.2%与87.1%。YOLOv5s算法P较高是损失R实现的,即采取较为保守的预测方式,虽然P在5种算法中达到了最高,但是R仅为57.4%,处于最低水平。改进后的算法较尺寸更大的YOLOv8s与YOLOv5s算法mAP值分别提升8.4%、18%,且FPS为102帧/s,满足实时检测的要求。证明改进后的算法在机身铆钉及异常检测任务中性能更具优越性。
为更直观地观察算法改进后在铆钉识别与异常检测任务中的性能提升,将原始YOLOv8n算法与改进后的算法对尾翼(图6a图6d)、机身顶部(图6b图6e)、机翼位置(图6c图6f)铆钉进行检测,检测图像拍摄距离均大于50cm,为更加清晰地展示检测效果对比,均作了放大处理。检测结果如图6所示。
可见:在对尾翼、机身顶部、机翼的检测中,改进后的算法对细小铆钉的感知能力得到显著增强,无论在机翼、机身顶部、尾翼部分检测结果PR均大幅度高于YOLOv8n原算法,特别是尾翼(图6d)、机身顶部(图6e)区域铆钉分布密集的区域对比效果更加明显,有效减少误检与漏检情况。同时,改进后的算法对于机身异常铆钉的检测能力也得到了增强,在铆钉分布密集、尺寸小的情况下也能准确识别与定位机身存在异常情况的铆钉(图6e)。
1) 提出一种基于改进YOLOv8n的飞机铆钉及脱落异常检测算法,从加入小目标检测层、替换骨干网络的卷积模块、集成LSKA到SPPF模块3方面改进YOLOv8n,该算法可实现对飞机铆钉小目标及其异常情况的特征提取、识别与定位。
2) 通过消融试验验证所提算法,该算法较YOLOv8n检测结果在PR、mAP值分别提升6.5%、16%、15%;对比试验结果表明:所提算法较现有主流目标检测算法的检测性能均有优势;下一步将研究不利天气(如雨天/雾天)影响下的铆钉及脱落异常识别方法。
  • 国家重点研发计划(2024YFC3014400)
  • 民航飞行技术与飞行安全科研基地(F2024KF04C)
  • 四川省重点研发计划项目(2024YFTX0078)
  • 中国民用航空飞行学院基本科研项目(24CAFUC03047)
参考文献 引证文献
排序方式:
[1]
吴军, 单腾飞, 黄硕, 等. 基于YOLOv7通道冗余改进的飞机蒙皮损伤检测[J]. 航空制造技术, 2024, 67(6):55-64.
WU Jun, SHAN Tengfei, HUANG Shuo, et al. Detection for aircraft skin damage based on improvement of channel redundancy for YOLOv7[J]. Aeronautical Manufacturing Technology, 2024, 67(6):55-64.
[2]
KATUNIN A, DRAGAN K. Qualitative to quantitative non-destructive evaluation: a concept for D-sight inspections of aircraft structures[J]. Applied Mechanics and Materials, 2022, 909: 69-74.
[3]
DING Meng, WU Boer, XU Juan, et al. Visual inspection of aircraft skin: automated pixel-level defect detection by instance segmentation[J]. Chinese Journal of Aeronautics, 2022, 35(10): 254-264.
[4]
周雨轩, 周文彬, 李小丽, 等. 飞机蒙皮缺陷自动化检测研究进展[J]. 航空制造技术, 2023, 66(22):69-77.
ZHOU Yuxuan, ZHOU Wenbin, LI Xiaoli, et al. Research progress in automatic detection of aircraft skin defects[J]. Aeronautical Manufacturing Technology, 2023, 66(22):69-77.
[5]
周志新. 机械裂纹无损检测方法综述[J]. 机电工程, 2017, 34(10): 1138-1143.
ZHOU Zhixin. Overview of NDT methods for mechanical cracks[J]. Journal of Mechanical & Electrical Engineering, 2017, 34(10): 1138-1143.
[6]
REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transaction on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[7]
彭林聪, 王克瑞, 周浩, 等. 基于改进SSD的目标检测算法[J]. 激光杂志, 2024, 45(11):71-76.
PENG Lincong, WANG Kerui, ZHOU Hao, et al. Object detection algorithm based on improved SSD[J]. Laser Journal, 2024, 45(11):71-76.
[8]
左明成, 焦文华. 面向煤矿井下作业场景的安全帽佩戴识别算法[J]. 中国安全科学学报, 2024, 34(3):237-246.
ZUO Mingcheng, JIAO Wenhua. Helmet-wearing recognition algorithm for coal mine underground operation scenarios[J]. China Safety Science Journal, 2024, 34(3):237-246.
[9]
张立国, 袁煜淋, 金梅, 等. 基于改进YOLOv8n的无人机目标检测算法研究[J]. 计量学报, 2024, 45(10):1487-1493.
ZHANG Liguo, YUAN Yulin, JIN Mei, et al. Research on UAV target detection algorithm based on improved YOLOv8n[J]. Acta Metrologica Sinica, 2024, 45(10):1487-1493.
[10]
姚景丽, 程光, 万飞, 等. 改进YOLOv8的轻量化轴承缺陷检测算法[J]. 计算机工程与应用, 2024, 60(21):205-214.
YAO Jingli, CHENG Guang, WAN Fei, et al. Improved lightweight bearing defect detection algorithm of YOLOv8[J]. Computer Engineering and Applications, 2024, 60(21):205-214.
[11]
肖伸平, 赵倩颖, 曾甲元, 等. 基于YOLO-DCL的复杂环境油茶果遮挡检测与计数研究[J]. 农业机械学报, 2024, 55(10):318-326,480.
XIAO Shenping, ZHAO Qianying, ZENG Jiayuan, et al. Camellia oleifera fruits occlusion detection and counting in complex environments based on improved YOLO-DCL[J]. Transactions of the Chinese Society for Agricultural Machinery, 2024, 55(10):318-326,480.
[12]
夏正洪, 魏汝祥, 李彦冬. 改进YOLOv3算法下通航机场场面运动目标检测[J]. 中国安全科学学报, 2023, 33(2):82-88.
XIA Zhenghong, WEI Ruxiang, LI Yandong. Moving target detection of general aviation airport based on improved YOLOv3 algorithm[J]. China Safety Science Journal, 2023, 33(2):82-88.
[13]
农昌瑞, 张静, 杨智勇. 改进的YOLOv4在飞机蒙皮损伤检测中的应用[J]. 海军航空大学学报, 2022, 37(2):179-184,230.
NONG Changrui, ZHANG Jing, YANG Zhiyong. Application of improved YOLOv4 in aircraft skin damage detection[J]. Journal of Naval Aviation University, 2022, 37(2):179-184, 230.
[14]
王童, 王宏伟, 王红. 基于旋转目标检测的飞机蒙皮损伤检测[J]. 激光与光电子学进展, 2023, 60(10):260-268.
WANG Tong, WANG Hongwei, WANG Hong. Aircraft skin damage detection based on rotating object detection[J]. Laser & Optoelectronics Progress, 2023, 60(10):260-268.
[15]
夏正洪, 何琥, 吴建军, 等. 基于深度学习的航空铆钉分类及异常情况检测[J]. 中国安全生产科学技术, 2023, 19(6):199-205.
XIA Zhenghong, HE Hu, WU Jianjun, et al. Aviation rivet classification and abnormal situation detection based on deep learning[J]. Journal of Safety Science and Technology, 2023, 19(6):199-205.
2025年第35卷第2期
PDF下载
418
175
引用本文
BibTeX
文章信息
doi: 10.16265/j.cnki.issn1003-3033.2025.02.0880
  • 接收时间:2024-09-25
  • 首发时间:2025-07-05
  • 出版时间:2025-02-28
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-09-25
  • 修回日期:2024-11-25
基金
国家重点研发计划(2024YFC3014400)
民航飞行技术与飞行安全科研基地(F2024KF04C)
四川省重点研发计划项目(2024YFTX0078)
中国民用航空飞行学院基本科研项目(24CAFUC03047)
作者信息
    1 中国民用航空飞行学院 空中交通管理学院,四川 广汉 618307
    2 中国民用航空华北地区管理局,北京 100621
    3 中国航空工业集团有限公司 成都飞机设计研究所,四川 成都 610091
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/10.16265/j.cnki.issn1003-3033.2025.02.0880
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数

Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)

Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏