Article(id=1251458154177970272, tenantId=1146029695717560320, journalId=1251194880429441115, issueId=1251458153020342360, articleNumber=null, orderNo=null, doi=10.3979/j.issn.1673-825X.202412060286, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1733414400000, receivedDateStr=2024-12-06, revisedDate=1757260800000, revisedDateStr=2025-09-08, acceptedDate=null, acceptedDateStr=null, onlineDate=1776300474923, onlineDateStr=2026-04-16, pubDate=null, pubDateStr=null, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1776300474923, onlineIssueDateStr=2026-04-16, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1776300474923, creator=13041195026, updateTime=1776300474923, updator=13041195026, issue=Issue{id=1251458153020342360, tenantId=1146029695717560320, journalId=1251194880429441115, year='2025', volume='37', issue='5', pageStart='627', pageEnd='780', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=1, specialIssue=null, createTime=1776300474648, creator=13041195026, updateTime=1776311939434, updator=13041195026, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1251506239914586238, tenantId=1146029695717560320, journalId=1251194880429441115, issueId=1251458153020342360, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1251506239914586239, tenantId=1146029695717560320, journalId=1251194880429441115, issueId=1251458153020342360, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=729, endPage=740, ext={EN=ArticleExt(id=1251458154425434210, articleId=1251458154177970272, tenantId=1146029695717560320, journalId=1251194880429441115, language=EN, title=Improved underwater target detection algorithm for YOLOv8, columnId=1251458154354131041, journalTitle=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), columnName=Artificial Intelligenceand Big Data, runingTitle=null, highlight=null, articleAbstract=

To address the poor performance of underwater object detection caused by light attenuation and scattering, this paper proposes an enhanced underwater object detection framework based on YOLOv8, named ERMS-YOLOv8, aiming to improve detection accuracy. The backbone is replaced with an efficient vision transformer(EfficientViT)to strengthen feature extraction of underwater organisms and reduce information loss. The neck adopts a reparameterized generalized-directional feature pyramid network(RepGFPN)to enhance the fusion of high-level semantic and low-level spatial features, enabling richer feature representation. A mixed local channel attention for object detection(MLCA)is introduced to integrate channel, spatial, local, and global channel information, thereby boosting the model's representational capacity. Additionally, a scalable intersection over union loss(SIoU)is employed to improve boundary prediction accuracy. Experimental re sults demonstrate that the proposed method achieves mAP values of 83.9% on the UPRC2021 dataset and 84.4% on the DUO dataset, outperforming the original YOLOv8 and exhibiting superior performance in underwater object detection.

, 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=Quan YUAN, Qingquan YANG, Yalong YUAN, Fengjuan LIU), CN=ArticleExt(id=1251458167150952898, articleId=1251458154177970272, tenantId=1146029695717560320, journalId=1251194880429441115, language=CN, title=改进YOLOv8的水下目标检测算法, columnId=1251458154492543075, journalTitle=重庆邮电大学学报(自然科学版), columnName=人工智能与大数据, runingTitle=null, highlight=null, articleAbstract=

针对水下光线衰减、散射等影响导致水下目标检测效果不佳的问题,提出一种基于YOLOv8的水下目标检测框架ERMS-YOLOv8,提升水下目标检测性能。主干网络采用高效视觉transformer网络(efficient vision transformer,EfficientViT),增强模型对水下生物的特征提取能力,减少特征信息丢失;Neck部分采用高效重参数化广义特征金字塔网络(reparameterized generalized-directional feature pyramid network,RepGFPN),增强模型对水下生物高层语义和低级空间特征的提取和融合能力,使得模型获取更加丰富的特征信息;引入混合局部通道注意力机制(mixed local channel attention for object detection,MLCA),使得模型同时融合通道信息、空间信息、局部通道信息和全局通道信息,增强了模型的表征能力;引入可扩展交并比损失函数(scalable intersection over union loss,SIoU),提升模型对目标边界信息的提取能力,从而进一步提高检测精度。实验结果表明,改进后的算法在UPRC2021和DUO数据集上mAP值分别达到83.9%和84.4%,与基准YOLOv8算法相比都有提高,在水下目标检测中具有优越的性能。

, correspAuthors=null, authorNote=null, correspAuthorsNote=
袁泉
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=ynC2IfLErXRsYEhIeIHu+Q==, magXml=u0ZdvdQ0klIUPkvWpbY7JA==, pdfUrl=null, pdf=YDzmcdxADYugL1G1pZfwPA==, pdfFileSize=15988388, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=M1NJINuZX18868yTYSmzXA==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=lk2S+kiaEHawo15OORSLvQ==, mapNumber=null, authorCompany=null, fund=null, authors=

袁泉,正高级工程师,硕士,主要研究方向为大数据分析、自然语言处理。E-mail:

杨清泉,硕士研究生,主要研究方向为图像处理、目标检测。E-mail:

袁亚隆,硕士研究生,主要研究方向为目标检测。E-mail:

刘凤娟,硕士研究生,主要研究方向为目标检测。E-mail:

, authorsList=袁泉, 杨清泉, 袁亚隆, 刘凤娟)}, authors=[Author(id=1251458167671046626, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=quanyuan@cqupt.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1251458167759127016, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458167671046626, language=EN, stringName=Quan YUAN, firstName=Quan, middleName=null, lastName=YUAN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251458167843013102, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458167671046626, language=CN, stringName=袁泉, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1重庆邮电大学 通信与信息工程学院,重庆 400065
2重庆邮电大学 通信新技术应用研究中心,重庆 400065, bio={"content":"

袁泉,正高级工程师,硕士,主要研究方向为大数据分析、自然语言处理。E-mail:

"}, bioImg=null, bioContent=

袁泉,正高级工程师,硕士,主要研究方向为大数据分析、自然语言处理。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251458167444554194, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=1, ext=[AuthorCompanyExt(id=1251458167452942802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167461331411, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1重庆邮电大学 通信与信息工程学院,重庆 400065)]), AuthorCompany(id=1251458167570383323, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=2, ext=[AuthorCompanyExt(id=1251458167578771930, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167587160539, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2重庆邮电大学 通信新技术应用研究中心,重庆 400065)])]), Author(id=1251458167964647922, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=2673507123@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1251458168073699830, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458167964647922, language=EN, stringName=Qingquan YANG, firstName=Qingquan, middleName=null, lastName=YANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251458168161780218, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458167964647922, language=CN, stringName=杨清泉, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1重庆邮电大学 通信与信息工程学院,重庆 400065
2重庆邮电大学 通信新技术应用研究中心,重庆 400065, bio={"content":"

杨清泉,硕士研究生,主要研究方向为图像处理、目标检测。E-mail:

"}, bioImg=null, bioContent=

杨清泉,硕士研究生,主要研究方向为图像处理、目标检测。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251458167444554194, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=1, ext=[AuthorCompanyExt(id=1251458167452942802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167461331411, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1重庆邮电大学 通信与信息工程学院,重庆 400065)]), AuthorCompany(id=1251458167570383323, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=2, ext=[AuthorCompanyExt(id=1251458167578771930, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167587160539, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2重庆邮电大学 通信新技术应用研究中心,重庆 400065)])]), Author(id=1251458169755615743, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=1426355739@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1251458169860473351, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458169755615743, language=EN, stringName=Yalong YUAN, firstName=Yalong, middleName=null, lastName=YUAN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251458169973719563, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458169755615743, language=CN, stringName=袁亚隆, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1重庆邮电大学 通信与信息工程学院,重庆 400065
2重庆邮电大学 通信新技术应用研究中心,重庆 400065, bio={"content":"

袁亚隆,硕士研究生,主要研究方向为目标检测。E-mail:

"}, bioImg=null, bioContent=

袁亚隆,硕士研究生,主要研究方向为目标检测。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251458167444554194, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=1, ext=[AuthorCompanyExt(id=1251458167452942802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167461331411, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1重庆邮电大学 通信与信息工程学院,重庆 400065)]), AuthorCompany(id=1251458167570383323, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=2, ext=[AuthorCompanyExt(id=1251458167578771930, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167587160539, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2重庆邮电大学 通信新技术应用研究中心,重庆 400065)])]), Author(id=1251458170086965775, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=2164319393@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1251458170204406296, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458170086965775, language=EN, stringName=Fengjuan LIU, firstName=Fengjuan, middleName=null, lastName=LIU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251458170279903774, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, authorId=1251458170086965775, language=CN, stringName=刘凤娟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, address=1重庆邮电大学 通信与信息工程学院,重庆 400065
2重庆邮电大学 通信新技术应用研究中心,重庆 400065, bio={"content":"

刘凤娟,硕士研究生,主要研究方向为目标检测。E-mail:

"}, bioImg=null, bioContent=

刘凤娟,硕士研究生,主要研究方向为目标检测。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251458167444554194, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=1, ext=[AuthorCompanyExt(id=1251458167452942802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167461331411, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1重庆邮电大学 通信与信息工程学院,重庆 400065)]), AuthorCompany(id=1251458167570383323, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=2, ext=[AuthorCompanyExt(id=1251458167578771930, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167587160539, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2重庆邮电大学 通信新技术应用研究中心,重庆 400065)])])], keywords=[Keyword(id=1251458170405732903, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, orderNo=1, keyword=YOLOv8), Keyword(id=1251458170472841772, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, orderNo=2, keyword=EfficientViT), Keyword(id=1251458170539950641, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, orderNo=3, keyword=RepGFPN), Keyword(id=1251458170607059509, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, orderNo=4, keyword=attention mechanism), Keyword(id=1251458170690945594, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, orderNo=5, keyword=loss function), Keyword(id=1251458170791608893, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, orderNo=1, keyword=YOLOv8), Keyword(id=1251458170871300675, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, orderNo=2, keyword=EfficientViT), Keyword(id=1251458170976158281, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, orderNo=3, keyword=重参数化广义特征金字塔网络), Keyword(id=1251458171043267147, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, orderNo=4, keyword=注意力机制), Keyword(id=1251458171106181709, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, orderNo=5, keyword=损失函数)], refs=[Reference(id=1251458176420365040, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=宋明龙, journalName=基于改进SSD的水下目标检测算法研究, refType=null, unstructuredReference=宋明龙.基于改进SSD的水下目标检测算法研究[D].淮南:安徽理工大学, 2023., articleTitle=null, refAbstract=null), Reference(id=1251458176487473908, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=SONG M L, journalName=Research on underwater target detection algorithm based on improved SSD, refType=null, unstructuredReference=SONG M L. Research on underwater target detection algorithm based on improved SSD[D]. Huainan: Anhui University of Science and Technology, 2023., articleTitle=null, refAbstract=null), Reference(id=1251458176554582775, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2015, volume=521, issue=7553, pageStart=436, pageEnd=444, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=LECUN Y, BENGIO Y, HINTON G, journalName=Nature, refType=null, unstructuredReference=LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444., articleTitle=Deep learning, refAbstract=null), Reference(id=1251458176621691643, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2014, volume=null, issue=null, pageStart=580, pageEnd=587, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=GIRSHICK R, DONAHUE J, DARRELL T, journalName=null, refType=null, unstructuredReference=GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA: IEEE, 2014: 580-587., articleTitle=Rich feature hierarchies for accurate object detection and semantic segmentation, refAbstract=null), Reference(id=1251458176718160638, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=1440, pageEnd=1448, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=GIRSHICK R, journalName=null, refType=null, unstructuredReference=GIRSHICK R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile: IEEE, 2015: 1440-1448., articleTitle=Fast R-CNN, refAbstract=null), Reference(id=1251458176823018240, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2017, volume=39, issue=6, pageStart=1137, pageEnd=1149, url=null, language=null, rfNumber=[5], rfOrder=5, authorNames=REN S Q, HE K M, GIRSHICK R, journalName=IEEE Transactions on Pattern Analysis and Machine Intelligence, refType=null, unstructuredReference=REN S Q, HE K M, GIRSHICK R, et al. Faster RCNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149., articleTitle=Faster RCNN: Towards real-time object detection with region proposal networks, refAbstract=null), Reference(id=1251458176885932802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=779, pageEnd=788, url=null, language=null, rfNumber=[6], rfOrder=6, authorNames=REDMON J, DIVVALA S, GIRSHICK R, journalName=null, refType=null, unstructuredReference=REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: Unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016: 779-788., articleTitle=You only look once: Unified, real-time object detection, refAbstract=null), Reference(id=1251458176961430277, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=21, pageEnd=37, url=null, language=null, rfNumber=[7], rfOrder=7, authorNames=LIU W, ANGUELOV D, ERHAN D, journalName=null, refType=null, unstructuredReference=LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision. Amsterdam, Netherlands: Springer, 2016: 21-37., articleTitle=SSD: single shot multibox detector, refAbstract=null), Reference(id=1251458177041122058, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=38, issue=04, pageStart=747, pageEnd=754, url=null, language=null, rfNumber=[8], rfOrder=8, authorNames=强伟, 贺昱曜, 郭玉锦, journalName=西北工业大学学报, refType=null, unstructuredReference=强伟,贺昱曜,郭玉锦,.基于改进SSD的水下目标检测算法研究[J].西北工业大学学报, 2020, 38(04):747-754., articleTitle=基于改进SSD的水下目标检测算法研究, refAbstract=null), Reference(id=1251458178651734795, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=38, issue=04, pageStart=747, pageEnd=754, url=null, language=null, rfNumber=[8], rfOrder=9, authorNames=QIANG W, HE Y Y, GUO Y J, journalName=Journal of Northwestern Polytechnical University, refType=null, unstructuredReference=QIANG W, HE Y Y, GUO Y J, et al. Research on underwater target detection algorithm based on improved SSD[J]. Journal of Northwestern Polytechnical University, 2020, 38(04): 747-754., articleTitle=Research on underwater target detection algorithm based on improved SSD, refAbstract=null), Reference(id=1251458178769175311, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=2, pageStart=499, pageEnd=507, url=null, language=null, rfNumber=[9], rfOrder=10, authorNames=陈宇梁, 董绍江, 孙世政, journalName=北京航空航天大学学报, refType=null, unstructuredReference=陈宇梁,董绍江,孙世政,.改进YOLOv5s的弱光水下生物目标检测算法[J].北京航空航天大学学报, 2024, 50(2): 499-507., articleTitle=改进YOLOv5s的弱光水下生物目标检测算法, refAbstract=null), Reference(id=1251458178911781649, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=2, pageStart=499, pageEnd=507, url=null, language=null, rfNumber=[9], rfOrder=11, authorNames=CHEN Y L, DONG S J, SUN S Z, journalName=Journal of Beijing University of Aeronautics and Astronautics, refType=null, unstructuredReference=CHEN Y L, DONG S J, SUN S Z, et al. Enhanced YOLOv5s for low-light underwater biological-object detection[J]. Journal of Beijing University of Aeronautics and Astronautics, 2024, 50(2): 499-507., articleTitle=Enhanced YOLOv5s for low-light underwater biological-object detection, refAbstract=null), Reference(id=1251458179100525332, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=60, issue=13, pageStart=200, pageEnd=208, url=null, language=null, rfNumber=[10], rfOrder=12, authorNames=陶洋, 朱腾, 钟邦乾, journalName=计算机工程与应用, refType=null, unstructuredReference=陶洋,朱腾,钟邦乾,. RepViTS-YOLOX:水下模糊及遮挡目标检测方法[J].计算机工程与应用, 2024, 60(13): 200-208., articleTitle=RepViTS-YOLOX:水下模糊及遮挡目标检测方法, refAbstract=null), Reference(id=1251458179247325974, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=60, issue=13, pageStart=200, pageEnd=208, url=null, language=null, rfNumber=[10], rfOrder=13, authorNames=TAO Y, ZHU T, ZHONG B Q, journalName=Computer Engineering and Applications, refType=null, unstructuredReference=TAO Y, ZHU T, ZHONG B Q, et al. RepViTS-YOLOX:detection method for blurred and occluded underwater targets[J]. Computer Engineering and Applications, 2024, 60(13): 200-208., articleTitle=RepViTS-YOLOX:detection method for blurred and occluded underwater targets, refAbstract=null), Reference(id=1251458179347989274, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=44, issue=11, pageStart=3610, pageEnd=3616, url=null, language=null, rfNumber=[11], rfOrder=14, authorNames=李大海, 李冰涛, 王振东, journalName=计算机应用, refType=null, unstructuredReference=李大海,李冰涛,王振东.基于改进YOLOv8的水下目标检测算法[J].计算机应用, 2024, 44(11): 3610-3616., articleTitle=基于改进YOLOv8的水下目标检测算法, refAbstract=null), Reference(id=1251458179427681053, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=44, issue=11, pageStart=3610, pageEnd=3616, url=null, language=null, rfNumber=[11], rfOrder=15, authorNames=LIDH, LIBT, WANGZD, journalName=Journal of Computer Applications, refType=null, unstructuredReference=LIDH,LIBT,WANGZD. Underwater object detection algorithm based on improved YOLOv8[J]. Journal of Computer Applications, 2024, 44(11): 3610-3616., articleTitle=Underwater object detection algorithm based on improved YOLOv8, refAbstract=null), Reference(id=1251458179536732962, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=1, pageEnd=7, url=null, language=null, rfNumber=[12], rfOrder=16, authorNames=SHI K, W ANG J, ZHU H, journalName=null, refType=null, unstructuredReference=SHI K, W ANG J, ZHU H, et al. Intelligent target detection of underwater optical images based on image enhancement[C]//Proceedings of Fourteenth International Conference on Information Optics and Photonics (CIOP 2023). Xi'an, China: SPIE, 2023: 1-7., articleTitle=Intelligent target detection of underwater optical images based on image enhancement, refAbstract=null), Reference(id=1251458179641590566, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=null, issue=17, pageStart=1734, pageEnd=1747, url=null, language=null, rfNumber=[13], rfOrder=17, authorNames=YI H, LIU B, ZHAO B, journalName=IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, refType=null, unstructuredReference=YI H, LIU B, ZHAO B, et al. Small object detection algorithm based on improved YOLOv8 for remote sensing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024 (17): 1734-1747., articleTitle=Small object detection algorithm based on improved YOLOv8 for remote sensing, refAbstract=null), Reference(id=1251458179725476649, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=9, pageStart=2165, pageEnd=2174, url=null, language=null, rfNumber=[14], rfOrder=18, authorNames=李姝, 李思远, 刘国庆, journalName=小型微型计算机系统, refType=null, unstructuredReference=李姝,李思远,刘国庆.基于YOLOv8无人机航拍图像的小目标检测算法研究[J].小型微型计算机系统, 2024, 45(9): 2165-2174., articleTitle=基于YOLOv8无人机航拍图像的小目标检测算法研究, refAbstract=null), Reference(id=1251458179855500077, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=9, pageStart=2165, pageEnd=2174, url=null, language=null, rfNumber=[14], rfOrder=19, authorNames=LI S, LI S Y, LIU G Q, journalName=Journal of Chinese Systems, refType=null, unstructuredReference=LI S, LI S Y,LIU G Q. Research on small target detection algorithm based on YOLOv8 UAV aerial images[J]. Journal of Chinese Systems, 2024, 45(9): 2165-2174., articleTitle=Research on small target detection algorithm based on YOLOv8 UAV aerial images, refAbstract=null), Reference(id=1251458179951969072, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=33, pageStart=21002, pageEnd=21012, url=null, language=null, rfNumber=[15], rfOrder=20, authorNames=LIX, W ANG W, WU L, journalName=Advances in Neural Information Processing Systems, refType=null, unstructuredReference=LIX, W ANG W, WU L, et alGeneralized focal loss:Learning qualified and distributed bounding boxes for dense object detection[J]. Advances in Neural Information Processing Systems, 2020(33): 21002-21012., articleTitle=Generalized focal loss:Learning qualified and distributed bounding boxes for dense object detection, refAbstract=null), Reference(id=1251458180027466547, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=516, pageEnd=520, url=null, language=null, rfNumber=[16], rfOrder=21, authorNames=YU J H, JIANG Y N, WANG Z Y, journalName=null, refType=null, unstructuredReference=YU J H, JIANG Y N, WANG Z Y, et al. Unitbox: An advanced object detection network[C]//Proceedings of the 24th ACM international conference on Multimedia. New York, USA: ACM, 2016: 516-520., articleTitle=Unitbox: An advanced object detection network, refAbstract=null), Reference(id=1251458180123935541, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=14420, pageEnd=14430, url=null, language=null, rfNumber=[17], rfOrder=22, authorNames=LIU X, PENG H, ZHENG N, journalName=null, refType=null, unstructuredReference=LIU X, PENG H, ZHENG N, et al. EfficientViT: Memory efficient vision transformer with cascaded group attention[C]//Proceedings of 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver:IEEE, 2023: 14420-14430., articleTitle=EfficientViT: Memory efficient vision transformer with cascaded group attention, refAbstract=null), Reference(id=1251458180237181752, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=8759, pageEnd=8768, url=null, language=null, rfNumber=[18], rfOrder=23, authorNames=LIU S, QI L, QIN H F, journalName=null, refType=null, unstructuredReference=LIU S, QI L, QIN H F, et al. Path aggregation network for instance segmentation[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA: IEEE: 8759-8768., articleTitle=Path aggregation network for instance segmentation, refAbstract=null), Reference(id=1251458180304290618, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=10778, pageEnd=10787, url=null, language=null, rfNumber=[19], rfOrder=24, authorNames=TAN M, PANG R, LE Q V, journalName=null, refType=null, unstructuredReference=TAN M, PANG R, LE Q V. EfficientDet: Scalable and efficient object detection[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recogn ition. Seattle,WA, USA: IEEE, 2020: 10778-10787., articleTitle=EfficientDet: Scalable and efficient object detection, refAbstract=null), Reference(id=1251458180375593788, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024-12-01, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[20], rfOrder=25, authorNames=JIANG Y Q, TAN Z Y, WANG J Y, journalName=null, refType=null, unstructuredReference=JIANG Y Q, TAN Z Y, WANG J Y. GiraffeDet: A heavy-neck paradigm for object detection[EB/OL].[2024-12-01]. https://arxiv.org/abs/2202.04256., articleTitle=GiraffeDet: A heavy-neck paradigm for object detection, refAbstract=null), Reference(id=1251458180467868478, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024-12-01, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[21], rfOrder=26, authorNames=XU X Z, JIANG Y Q, CHEN W H, journalName=null, refType=null, unstructuredReference=XU X Z, JIANG Y Q, CHEN W H, et al. Damo-YOLO:A report on real-time object detection design[EB/OL].[2024-12-01]. https://arxiv.org/abs/2211.15444., articleTitle=Damo-YOLO:A report on real-time object detection design, refAbstract=null), Reference(id=1251458180576920385, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=390, pageEnd=391, url=null, language=null, rfNumber=[22], rfOrder=27, authorNames=WANG C Y, LIAOHYM, WUYH, journalName=null, refType=null, unstructuredReference=WANG C Y, LIAOHYM,WUYH, et al. CSPNet: A new backbone that can enhance learning capability of CNN[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Seattle, WA, USA: IEEE, 2020:390-391., articleTitle=CSPNet: A new backbone that can enhance learning capability of CNN, refAbstract=null), Reference(id=1251458180685972291, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=null, issue=8, pageStart=187, pageEnd=194, url=null, language=null, rfNumber=[23], rfOrder=28, authorNames=龙阳, 肖小玲, journalName=制造技术与机床, refType=null, unstructuredReference=龙阳,肖小玲.改进YOLOv8的金属表面缺陷检测模型[J].制造技术与机床, 2024(8): 187-194., articleTitle=改进YOLOv8的金属表面缺陷检测模型, refAbstract=null), Reference(id=1251458180786635590, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=null, issue=8, pageStart=187, pageEnd=194, url=null, language=null, rfNumber=[23], rfOrder=29, authorNames=LONG Y, XIAO X L, journalName=Manufacturing Technology &Machine Tool, refType=null, unstructuredReference=LONG Y, XIAO X L. Enhanced YOLOv8 for metal-surface defect inspection[J]. Manufacturing Technology &Machine Tool, 2024(8): 187-194., articleTitle=Enhanced YOLOv8 for metal-surface defect inspection, refAbstract=null), Reference(id=1251458180908270410, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=8, pageStart=91, pageEnd=98, url=null, language=null, rfNumber=[24], rfOrder=30, authorNames=陈腾杰, 李永安, 张之好, journalName=工矿自动化, refType=null, unstructuredReference=陈腾杰,李永安,张之好,.基于改进YOLOv8n+DeepSORT的带式输送机异物检测及计数方法[J].工矿自动化, 2024, 50(8): 91-98., articleTitle=基于改进YOLOv8n+DeepSORT的带式输送机异物检测及计数方法, refAbstract=null), Reference(id=1251458181038293838, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2024, volume=50, issue=8, pageStart=91, pageEnd=98, url=null, language=null, rfNumber=[24], rfOrder=31, authorNames=CHEN T J, LI Y A, ZHANG Z H, journalName=Industry and Mine Automation, refType=null, unstructuredReference=CHEN T J, LI Y A, ZHANG Z H, et al. Foreign-object detection and counting on belt conveyors via improved YOLOv8n plus DeepSORT[J]. Industry and Mine Automation, 2024, 50(8): 91-98., articleTitle=Foreign-object detection and counting on belt conveyors via improved YOLOv8n plus DeepSORT, refAbstract=null), Reference(id=1251458181105402705, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=12993, pageEnd=13000, url=null, language=null, rfNumber=[25], rfOrder=32, authorNames=ZHENG Z H, WANG P, LIU W, journalName=null, refType=null, unstructuredReference=ZHENG Z H, WANG P, LIU W, et al. Distance-IoU loss: Faster and better learning for bounding box regression[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence. PaloAlto California, USA: AAAI Press, 2020: 12993-13000., articleTitle=Distance-IoU loss: Faster and better learning for bounding box regression, refAbstract=null), Reference(id=1251458181176705876, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=1, pageEnd=3, url=null, language=null, rfNumber=[26], rfOrder=33, authorNames=DANG Y J, CHEN S L, MIAO H C, journalName=null, refType=null, unstructuredReference=DANG Y J, CHEN S L, MIAO H C, et al. Target detection from drone perspectives: Enhancing YOLOv5_3S with SIoU loss and SPD modules[C]//Cross Strait Radio Science and Wireless Technology Conference (CSRSWTC). Guilin, China: IEEE, 2023: 1-3., articleTitle=Target detection from drone perspectives: Enhancing YOLOv5_3S with SIoU loss and SPD modules, refAbstract=null), Reference(id=1251458181256397656, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2025, volume=65, issue=1, pageStart=54, pageEnd=62, url=null, language=null, rfNumber=[27], rfOrder=34, authorNames=施克权, 李祺, 隋皓, journalName=电讯技术, refType=null, unstructuredReference=施克权,李祺,隋皓,. IEMAyoloViT:基于改进YOLOv8的水下目标检测算法[J].电讯技术, 2025, 65(1): 54-62., articleTitle=IEMAyoloViT:基于改进YOLOv8的水下目标检测算法, refAbstract=null), Reference(id=1251458181340283739, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, doi=null, pmid=null, pmcid=null, year=2025, volume=65, issue=1, pageStart=54, pageEnd=62, url=null, language=null, rfNumber=[27], rfOrder=35, authorNames=SHI K Q, LI Q, SUI H, journalName=Telecommunication Engineering, refType=null, unstructuredReference=SHI K Q, LI Q, SUI H, et al. IEMA-yoloViT: An improved YOLOv8-based underwater object detection algorithm[J]. Telecommunication Engineering, 2025, 65(1): 54-62., articleTitle=IEMA-yoloViT: An improved YOLOv8-based underwater object detection algorithm, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1251458167444554194, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=1, ext=[AuthorCompanyExt(id=1251458167452942802, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167461331411, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167444554194, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1重庆邮电大学 通信与信息工程学院,重庆 400065)]), AuthorCompany(id=1251458167570383323, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, xref=2, ext=[AuthorCompanyExt(id=1251458167578771930, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China), AuthorCompanyExt(id=1251458167587160539, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, companyId=1251458167570383323, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2重庆邮电大学 通信新技术应用研究中心,重庆 400065)])], figs=[ArticleFig(id=1251458171227816532, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.1, caption=Network structure of YOLOv8, figureFileSmall=6rx20f2H9bmmF6xmspaKjw==, figureFileBig=M1NJINuZX18868yTYSmzXA==, tableContent=null), ArticleFig(id=1251458171294925400, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图1, caption=YOLOv8网络结构, figureFileSmall=6rx20f2H9bmmF6xmspaKjw==, figureFileBig=M1NJINuZX18868yTYSmzXA==, tableContent=null), ArticleFig(id=1251458171496252000, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.2, caption=Module structure of C2f, figureFileSmall=OhvLQSYHwOy8dozo+a5WZw==, figureFileBig=u5k8xfM26UWS0UhlnL6ODA==, tableContent=null), ArticleFig(id=1251458171584332389, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图2, caption=C2f模块结构, figureFileSmall=OhvLQSYHwOy8dozo+a5WZw==, figureFileBig=u5k8xfM26UWS0UhlnL6ODA==, tableContent=null), ArticleFig(id=1251458171684995690, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.3, caption=Network Structure of ERMS-YOLOv8, figureFileSmall=1MS3Xr3Zqc2yaC/Zg8MwyA==, figureFileBig=p/weCgG3TE+YoyT9h6K7KA==, tableContent=null), ArticleFig(id=1251458171756298861, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图3, caption=ERMS-YOLOv8的网络结构, figureFileSmall=1MS3Xr3Zqc2yaC/Zg8MwyA==, figureFileBig=p/weCgG3TE+YoyT9h6K7KA==, tableContent=null), ArticleFig(id=1251458171831796337, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.4, caption=Backbone network structure of EfficientVit, figureFileSmall=KeQIwR5NC17T5i5p0Sbvnw==, figureFileBig=y6gk6WUbd4+/UyWbEqBU9g==, tableContent=null), ArticleFig(id=1251458171932459639, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图4, caption=EfficientVit主干网络结构, figureFileSmall=KeQIwR5NC17T5i5p0Sbvnw==, figureFileBig=y6gk6WUbd4+/UyWbEqBU9g==, tableContent=null), ArticleFig(id=1251458172028928635, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.5, caption=Evolutionary design of feature pyramid network, figureFileSmall=mOAVd3fPommGXqDERlltNQ==, figureFileBig=C9S6djouVGhwuZxB4v6LmQ==, tableContent=null), ArticleFig(id=1251458172129591936, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图5, caption=特征金字塔网络演化设计, figureFileSmall=mOAVd3fPommGXqDERlltNQ==, figureFileBig=C9S6djouVGhwuZxB4v6LmQ==, tableContent=null), ArticleFig(id=1251458172213478022, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.6, caption=Structure of RepGFPN, figureFileSmall=1/MVlmGMSpBkXbDqQ77QUQ==, figureFileBig=YgQhSj+TSNRkSZK+mPCiIA==, tableContent=null), ArticleFig(id=1251458172301558409, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图6, caption=RepGFPN结构, figureFileSmall=1/MVlmGMSpBkXbDqQ77QUQ==, figureFileBig=YgQhSj+TSNRkSZK+mPCiIA==, tableContent=null), ArticleFig(id=1251458172377055886, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.7, caption=Structure of CSPStage, figureFileSmall=DAzHLFp0p1VmWpe4PkygtQ==, figureFileBig=AihA++wdy0se0UorbPe0Og==, tableContent=null), ArticleFig(id=1251458172486107793, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图7, caption=CSPStage结构, figureFileSmall=DAzHLFp0p1VmWpe4PkygtQ==, figureFileBig=AihA++wdy0se0UorbPe0Og==, tableContent=null), ArticleFig(id=1251458172607742612, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.8, caption=Structure of Rep, figureFileSmall=uCr87lwFEwyzSi5O8iRPIQ==, figureFileBig=BcSRFkvt9+jJAENKXm7D4w==, tableContent=null), ArticleFig(id=1251458174214161048, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图8, caption=Rep结构, figureFileSmall=uCr87lwFEwyzSi5O8iRPIQ==, figureFileBig=BcSRFkvt9+jJAENKXm7D4w==, tableContent=null), ArticleFig(id=1251458174314824346, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.9, caption=Partial images of dataset, figureFileSmall=BpzVuUD7Pp7ZFKmfPQzZJw==, figureFileBig=w8t5nVPChY+Tz93MZiK5Pw==, tableContent=null), ArticleFig(id=1251458174419681952, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图9, caption=数据集的部分图片, figureFileSmall=BpzVuUD7Pp7ZFKmfPQzZJw==, figureFileBig=w8t5nVPChY+Tz93MZiK5Pw==, tableContent=null), ArticleFig(id=1251458174591648418, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.10, caption=Comparison of detection effects, figureFileSmall=qKClhvRciglNwvZWiiK7og==, figureFileBig=T0EDE/kNeGCYvGWhvOywIg==, tableContent=null), ArticleFig(id=1251458174679728804, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图10, caption=检测效果对比, figureFileSmall=qKClhvRciglNwvZWiiK7og==, figureFileBig=T0EDE/kNeGCYvGWhvOywIg==, tableContent=null), ArticleFig(id=1251458174788780711, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.11, caption=Comparison of images with different intensities of Gaussian noise, figureFileSmall=HUFoeyd0pVvMnJX4zSZEWQ==, figureFileBig=WA3wC99udBtGm9n/XAMwEQ==, tableContent=null), ArticleFig(id=1251458174872666794, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图11, caption=不同强度高斯噪声的图像对比, figureFileSmall=HUFoeyd0pVvMnJX4zSZEWQ==, figureFileBig=WA3wC99udBtGm9n/XAMwEQ==, tableContent=null), ArticleFig(id=1251458174952358574, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Fig.12, caption=Comparison of images under different intensities of lighting, figureFileSmall=MzQXCbRVox+h8frVhwuQlQ==, figureFileBig=Holr7Rej+niYl3hNyY6loQ==, tableContent=null), ArticleFig(id=1251458175015273138, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=图12, caption=不同强度光照的图像对比, figureFileSmall=MzQXCbRVox+h8frVhwuQlQ==, figureFileBig=Holr7Rej+niYl3hNyY6loQ==, tableContent=null), ArticleFig(id=1251458175120130741, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.1, caption=

Experimental environment configuration

, figureFileSmall=null, figureFileBig=null, tableContent=
实验环境环境配置
编译环境Python3.8
GPUNVIDIA RTX 3080
CPUIntel Core i9-10980XE
内存128G
集成开发环境PyCharm
), ArticleFig(id=1251458175199822522, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表1, caption=

实验环境配置

, figureFileSmall=null, figureFileBig=null, tableContent=
实验环境环境配置
编译环境Python3.8
GPUNVIDIA RTX 3080
CPUIntel Core i9-10980XE
内存128G
集成开发环境PyCharm
), ArticleFig(id=1251458175283708606, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.2, caption=

Comparison of experimental results of different algorithms under URPC2021 dataset

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@50/%mAP@50-95/%GFLOPs参数量/108
Faster-RCNN69.438.827.842.0
YOLOv580.246.57.12.5
YOLOv679.645.811.84.2
YOLOv881.047.18.13.0
WCA-YOLOv883.749.013.16.6
IEMAyoloViT83.345.79.54.0
本文算法83.949.19.64.3
), ArticleFig(id=1251458175422120642, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表2, caption=

不同算法在URPC2021数据集的实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@50/%mAP@50-95/%GFLOPs参数量/108
Faster-RCNN69.438.827.842.0
YOLOv580.246.57.12.5
YOLOv679.645.811.84.2
YOLOv881.047.18.13.0
WCA-YOLOv883.749.013.16.6
IEMAyoloViT83.345.79.54.0
本文算法83.949.19.64.3
), ArticleFig(id=1251458175535366855, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.3, caption=

Comparison of experimental results of different algorithms under DUO dataset

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@50/%mAP@50-95/%GFLOPs参数量/108
Faster-RCNN70.154.027.842.0
YOLOv581.359.07.12.5
YOLOv680.959.511.84.2
YOLOv882.362.48.13.0
WCA-YOLOv884.364.513.16.6
IEMAyoloViT83.863.79.54.0
本文算法84.465.19.64.3
), ArticleFig(id=1251458175615058633, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表3, caption=

不同算法在DUO数据集下实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@50/%mAP@50-95/%GFLOPs参数量/108
Faster-RCNN70.154.027.842.0
YOLOv581.359.07.12.5
YOLOv680.959.511.84.2
YOLOv882.362.48.13.0
WCA-YOLOv884.364.513.16.6
IEMAyoloViT83.863.79.54.0
本文算法84.465.19.64.3
), ArticleFig(id=1251458175703139021, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.4, caption=

Comparison of experimental results of YOLOv8 introducing different attention mechanisms under URPC2021 dataset

, figureFileSmall=null, figureFileBig=null, tableContent=
注意力机制mAP@50/%mAP@50-95/%GFLOPs参数量
CPCA81.347.18.4127040
SimAm81.147.08.20
DA81.447.28.4266624
MLCA81.447.38.210
), ArticleFig(id=1251458175782830801, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表4, caption=

YOLOv8引入不同注意力机制在URPC2021数据集下实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
注意力机制mAP@50/%mAP@50-95/%GFLOPs参数量
CPCA81.347.18.4127040
SimAm81.147.08.20
DA81.447.28.4266624
MLCA81.447.38.210
), ArticleFig(id=1251458175883494101, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.5, caption=

Results of ablation experiment

, figureFileSmall=null, figureFileBig=null, tableContent=
模型EfficientViTRepGFPNMLCASIoUmAP@50/%mAP@50-95/%GFLOPs参数量/108
YOLOv881.047.18.13.0
模型181.647.59.54.0
模型282.047.78.53.3
模型381.447.38.23.0
模型481.347.28.13.0
模型583.048.39.44.3
模型683.648.89.54.3
模型783.949.19.64.3
), ArticleFig(id=1251458175967380186, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表5, caption=

消融实验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
模型EfficientViTRepGFPNMLCASIoUmAP@50/%mAP@50-95/%GFLOPs参数量/108
YOLOv881.047.18.13.0
模型181.647.59.54.0
模型282.047.78.53.3
模型381.447.38.23.0
模型481.347.28.13.0
模型583.048.39.44.3
模型683.648.89.54.3
模型783.949.19.64.3
), ArticleFig(id=1251458176072237792, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.6, caption=

Comparison of experimental results of different Gaussian noises

, figureFileSmall=null, figureFileBig=null, tableContent=
高斯噪声强度mAP@50/%召回率/%
083.977.3
0.0183.777.0
0.0281.274.5
0.0575.468.5
), ArticleFig(id=1251458176147735267, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表6, caption=

不同高斯噪声实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
高斯噪声强度mAP@50/%召回率/%
083.977.3
0.0183.777.0
0.0281.274.5
0.0575.468.5
), ArticleFig(id=1251458176231621350, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=EN, label=Tab.7, caption=

Comparison of experimental results under different light intensities

, figureFileSmall=null, figureFileBig=null, tableContent=
光照强度mAP@50/%召回率/%
-0.283.777.0
-0.183.977.2
083.977.3
0.184.177.6
0.284.077.3
0.578.272.1
), ArticleFig(id=1251458176302924524, tenantId=1146029695717560320, journalId=1251194880429441115, articleId=1251458154177970272, language=CN, label=表7, caption=

不同光照强度实验结果对比

, figureFileSmall=null, figureFileBig=null, tableContent=
光照强度mAP@50/%召回率/%
-0.283.777.0
-0.183.977.2
083.977.3
0.184.177.6
0.284.077.3
0.578.272.1
)], attaches=null, journal=Journal(id=1251194162217791579, delFlag=0, nameCn=重庆邮电大学学报(自然科学版), nameEn=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), nameHistory1=null, nameHistory2=null, issn=1673-825X, eissn=null, cn=50-1181/N, coden=null, periodic=1, language=CN, oaType=null, 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=eIjswk9Qxcq7+V27dEZ90g==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), journalRemark=null, publicationField=null, createdTime=1776237534337, updatedTime=1776238167705, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=J, firstLetterEn=J, subjectCode=Natural Sciences, subjectName=null, subjectCodeEn=Natural Sciences, subjectNameEn=null, picCn=eIjswk9Qxcq7+V27dEZ90g==, picEn=PN1NU8XEwLhVlBzwQEoroA==, jcr=null, cjcr=null, exts=[JournalExt(id=1251196818869137642, language=CN, name=重庆邮电大学学报(自然科学版), nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1776238167732, updatedTime=1776238167732, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/author/login, submissionEditorUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/editor/login, submissionReviewUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1251196818936246507, language=EN, name=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1776238167747, updatedTime=1776238167747, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/author/login, submissionEditorUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/editor/login, submissionReviewUrl=http://journal2.cqupt.edu.cn/jcuptnsecn/jcuptnse/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1251194880429441115, websiteList=[Website(id=1251197148419797361, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1251194880429441115, 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/cqyddxxb/CN, language=CN, createTime=1776238246302, createBy=18614031015, updateTime=1776238652182, updateBy=18614031015, name=重庆邮电大学学报(自然科学版)-中文, tplId=1146099689490845704, title=重庆邮电大学学报(自然科学版), delFlag=0, indexPage=/home, props=[WebsiteProps(id=1251198981565530670, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=articleTextType, value=kx, createTime=1776238683358, updateTime=1776238683358, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981540364843, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=banner, value=null, createTime=1776238683352, updateTime=1776238683352, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981590696497, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=grayFlag, value=0, createTime=1776238683364, updateTime=1776238683364, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981531976234, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=logo, value=https://castjournals.cast.org.cn/joweb/cqyddxxb/CN/file/pic?fileId=X24gYotabwCg03WC1YllbA==, createTime=1776238683350, updateTime=1776238683350, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981607473715, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=minRunFlag, value=0, createTime=1776238683368, updateTime=1776238683368, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981557142061, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/cqyddxxb/CN/file/pic, createTime=1776238683356, updateTime=1776238683356, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981599085106, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=silenceFlag, value=0, createTime=1776238683366, updateTime=1776238683366, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981548753452, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1776238683354, updateTime=1776238683354, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981578113583, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=themeColor, value=null, createTime=1776238683361, updateTime=1776238683361, creator=18614031015, updator=18614031015), WebsiteProps(id=1251198981582307888, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148419797361, code=themeStyle, value=null, createTime=1776238683362, updateTime=1776238683362, creator=18614031015, updator=18614031015)]), Website(id=1251197148516266372, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1251194880429441115, 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/cqyddxxb/EN, language=EN, createTime=1776238246325, createBy=18614031015, updateTime=1776238647495, updateBy=18614031015, name=重庆邮电大学学报(自然科学版)-英文, tplId=1146101810881728533, title=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), delFlag=0, indexPage=/home, props=[WebsiteProps(id=1251199007897371192, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=articleTextType, value=kx, createTime=1776238689636, updateTime=1776238689636, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007872205365, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=banner, value=null, createTime=1776238689630, updateTime=1776238689630, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007914148411, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=grayFlag, value=0, createTime=1776238689640, updateTime=1776238689640, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007863816756, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=logo, value=https://castjournals.cast.org.cn/joweb/cqyddxxb/EN/file/pic?fileId=X24gYotabwCg03WC1YllbA==, createTime=1776238689628, updateTime=1776238689628, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007930925629, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=minRunFlag, value=0, createTime=1776238689644, updateTime=1776238689644, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007888982583, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/cqyddxxb/EN/file/pic, createTime=1776238689634, updateTime=1776238689634, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007922537020, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=silenceFlag, value=0, createTime=1776238689642, updateTime=1776238689642, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007880593974, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1776238689632, updateTime=1776238689632, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007901565497, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=themeColor, value=null, createTime=1776238689637, updateTime=1776238689637, creator=18614031015, updator=18614031015), WebsiteProps(id=1251199007909954106, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251197148516266372, code=themeStyle, value=null, createTime=1776238689639, updateTime=1776238689639, creator=18614031015, updator=18614031015)])], journalTitle=重庆邮电大学学报(自然科学版), weixinUrl=null, journalUrl=http://journal2.cqupt.edu.cn/jcuptnsecn, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), journalPhotoCn=eIjswk9Qxcq7+V27dEZ90g==, journalPhotoEn=PN1NU8XEwLhVlBzwQEoroA==, journalFirstLetter=J, journalRecommend=null, journalNew=null, journalCollection=null, 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=, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/cqyddxxb/CN/10.3979/j.issn.1673-825X.202412060286, detailUrlEn=https://castjournals.cast.org.cn/joweb/cqyddxxb/EN/10.3979/j.issn.1673-825X.202412060286, pdfUrlCn=https://castjournals.cast.org.cn/joweb/cqyddxxb/CN/PDF/10.3979/j.issn.1673-825X.202412060286, pdfUrlEn=https://castjournals.cast.org.cn/joweb/cqyddxxb/EN/PDF/10.3979/j.issn.1673-825X.202412060286, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
改进YOLOv8的水下目标检测算法
收藏切换
PDF下载
袁泉 1, 2 , 杨清泉 1, 2 , 袁亚隆 1, 2 , 刘凤娟 1, 2
重庆邮电大学学报(自然科学版) | 人工智能与大数据 2025,37(5): 729-740
收起
收藏切换
重庆邮电大学学报(自然科学版) | 人工智能与大数据 2025, 37(5): 729-740
改进YOLOv8的水下目标检测算法
全屏
袁泉1, 2 , 杨清泉1, 2 , 袁亚隆1, 2 , 刘凤娟1, 2
作者信息
  • 1重庆邮电大学 通信与信息工程学院,重庆 400065
  • 2重庆邮电大学 通信新技术应用研究中心,重庆 400065
  • 袁泉,正高级工程师,硕士,主要研究方向为大数据分析、自然语言处理。E-mail:

    杨清泉,硕士研究生,主要研究方向为图像处理、目标检测。E-mail:

    袁亚隆,硕士研究生,主要研究方向为目标检测。E-mail:

    刘凤娟,硕士研究生,主要研究方向为目标检测。E-mail:

通讯作者:

Improved underwater target detection algorithm for YOLOv8
Quan YUAN1, 2 , Qingquan YANG1, 2 , Yalong YUAN1, 2 , Fengjuan LIU1, 2
Affiliations
  • 1School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
  • 2Research Center for New Communication Technology Applications, Chongqing University of Posts and Telecommunications, Chongqing 400065, P. R. China
doi: 10.3979/j.issn.1673-825X.202412060286
文章导航
收藏切换

针对水下光线衰减、散射等影响导致水下目标检测效果不佳的问题,提出一种基于YOLOv8的水下目标检测框架ERMS-YOLOv8,提升水下目标检测性能。主干网络采用高效视觉transformer网络(efficient vision transformer,EfficientViT),增强模型对水下生物的特征提取能力,减少特征信息丢失;Neck部分采用高效重参数化广义特征金字塔网络(reparameterized generalized-directional feature pyramid network,RepGFPN),增强模型对水下生物高层语义和低级空间特征的提取和融合能力,使得模型获取更加丰富的特征信息;引入混合局部通道注意力机制(mixed local channel attention for object detection,MLCA),使得模型同时融合通道信息、空间信息、局部通道信息和全局通道信息,增强了模型的表征能力;引入可扩展交并比损失函数(scalable intersection over union loss,SIoU),提升模型对目标边界信息的提取能力,从而进一步提高检测精度。实验结果表明,改进后的算法在UPRC2021和DUO数据集上mAP值分别达到83.9%和84.4%,与基准YOLOv8算法相比都有提高,在水下目标检测中具有优越的性能。

YOLOv8  /  EfficientViT  /  重参数化广义特征金字塔网络  /  注意力机制  /  损失函数

To address the poor performance of underwater object detection caused by light attenuation and scattering, this paper proposes an enhanced underwater object detection framework based on YOLOv8, named ERMS-YOLOv8, aiming to improve detection accuracy. The backbone is replaced with an efficient vision transformer(EfficientViT)to strengthen feature extraction of underwater organisms and reduce information loss. The neck adopts a reparameterized generalized-directional feature pyramid network(RepGFPN)to enhance the fusion of high-level semantic and low-level spatial features, enabling richer feature representation. A mixed local channel attention for object detection(MLCA)is introduced to integrate channel, spatial, local, and global channel information, thereby boosting the model's representational capacity. Additionally, a scalable intersection over union loss(SIoU)is employed to improve boundary prediction accuracy. Experimental re sults demonstrate that the proposed method achieves mAP values of 83.9% on the UPRC2021 dataset and 84.4% on the DUO dataset, outperforming the original YOLOv8 and exhibiting superior performance in underwater object detection.

YOLOv8  /  EfficientViT  /  RepGFPN  /  attention mechanism  /  loss function
袁泉, 杨清泉, 袁亚隆, 刘凤娟. 改进YOLOv8的水下目标检测算法. 重庆邮电大学学报(自然科学版), 2025 , 37 (5) : 729 -740 . DOI: 10.3979/j.issn.1673-825X.202412060286
Quan YUAN, Qingquan YANG, Yalong YUAN, Fengjuan LIU. Improved underwater target detection algorithm for YOLOv8[J]. Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition), 2025 , 37 (5) : 729 -740 . DOI: 10.3979/j.issn.1673-825X.202412060286
随着经济与科技的发展,人们越来越重视海洋资源开发,水下目标检测一直是其中一个重要方向[1]。与陆地环境不同的是,水下环境更加复杂。
基于深度学习的目标检测算法是计算机视觉领域的重要研究方向之一,主要分为两阶段检测方法和单阶段检测方法[2]。两阶段目标检测主要包括区域卷积神经网络(regions with convolutional neural networks,R-CNN)[3]、快速区域神经网络(fast regions with convolutional neural networks,Fast R-CNN)[4]、更快速区域卷积神经网络(faster regions with convolutional neural networks,Faster R-CNN)[5]等方法;单阶段目标检测的代表是YOLO(you only look once)算法[6]和SSD(single shot multibox detector)算法[7]
目前,水下目标检测通常使用单阶段目标检测方法。文献[8]提出了一种基于改进SSD的水下目标检测算法,将SSD中卷积神经网络替换为ResNet,利用深度可变形卷积模块进行特征提取。文献[9]针对弱光条件下的水下生物,提出了YOLOv5sunderwater方法,引入了限制对比度自适应直方图均衡以及快速空间金字塔池化模块,极大缓解了弱光水下光线衰弱的问题。文献[10]针对水下目标的模糊和遮挡问题,提出了一种基于RepViTSYOLOX的水下目标检测算法,引入空间和通道重构模块与上下文解耦头,对水下遮挡目标的定位和分类更加精准。文献[11]提出了WCA-YOLOv8方法,设计了特征融合模块和下采样增强模块,不仅获得了更高的检测准确性,还可达到更快的检测速度。
近年来深度学习算法在水下目标检测中发挥着重要的作用,但是面对水下复杂的环境仍然存在实时性较差、检测精度较低等问题。本文以YOLOv8为基础,提出一种具有更高精度并适用于实时检测的ERMS-YOLOv8新算法,该算法主要改进如下:
1)将YOLOv8的主干网络替换为高效视觉transformer网络(efficient vision transformer,EfficientViT),融入高效的多尺度注意力机制,增强模型对水下目标全局特征的提取能力,在一定程度上减少计算量,满足水下检测的实时性要求;
2)引入一种重参数化广义特征金字塔网络(reparameterized generalized-directional feature pyramid network,RepGFPN)来替换特征融合网络(Neck)网络中原有的路径聚合网络(path aggregation network,PANet),增强特征图之间的信息交流和整合能力,提升模型对目标在不同尺度和分辨率下的感知能力;
3)引入混合局部通道注意力(mixed local channel attention,MLCA)机制,以提高网络的表达效果,从而满足水下目标检测的检测性能;
4)将损失函数用SIoU替代CIoU(Complete-IoU),利用不同尺度的辅助边界,加速实现边框的回归。
YOLOv8算法建立在YOLO系列算法成功的基础上,结合Darknet-53网络结构思想并改进YOLOv5中C3模块为C2f模块,抛弃了前几代模型的Anchor-Base部分,提供了一个全新的SOTA模型[12]。YOLOv8可以完成检测、分类、分割任务[13]。YOLOv8被分为n、s、m、l、x这五种模型大小,它们主要区别在于模型大小和检测精度上的平衡,而n、s、m是在目标检测中常用的三种模型。
YOLOv8的结构分为四个部分:输入端、主干网络(Backbone)、Neck和检测头(Head)输出,其网络结构如图1所示。YOLOv8在Backbone和Neck部分参考了YOLOv7的ELAN设计思想,最重要的改进是将YOLOv5的跨阶段部分连接模块(cross stage partial,CSP)换成了梯度流更丰富的跨阶段部分融合模块(cross stage partial network fusion,C2f)。C2f模块由2个卷积-批归一化激活模块(CONv-BN-SiLU,CBS)以及多个BottleNeck构成,而CBS由一个卷积层、一个批量归一化层以及激活函数构成[14]。C2f的结构如图2所示。
YOLOv8的损失函数主要用于训练神经网络,使其能够准确预测图像中物体的位置和类别。损失函数通常包括坐标回归损失、分类损失以及边界框损失等部分。
分类损失使用二元交叉熵(BCE)判断“是否为此类”,并输出置信度,表示为
式(1)中:yipi分别代表样本标签的积分值和样本预测为正类的概率;N代表样本总数。
由于去掉了对象损失,在输出中也去掉了“对象置信度”,因此对直接输出的各个类别的“置信度分数”求最大值,将其作为anchor框的置信度。
在回归损失中,使用分布焦点损失(distribution focal loss,DFL)将预测位置迅速聚焦到目标位置附近[15],表示为
式(2)中:SiSi+1分别为网络输出的预测值、临近预测值;yyiyi+1分别为标签的实际值、标签积分值和临近标签积分值。
当预测位置聚集在目标附近时,通过CIoU损失函数fCIoU进一步精确位置PCIoU
式(3)—(4)中:为预测框b和真实框bgt中心点欧氏距离的归一化值;α为正权衡参数;ν为长宽比一致性参数;IIoU为交并比(intersection over union,IoU)值[16]。交并比是一种用于衡量目标检测算法精度的常见指标,在评估模型预测边界框与真实边界框之间重叠程度时非常有用,具体计算方式为交集面积与并集面积之比。其中,交集面积定义为模型预测的边界框与真实边界框重叠部分的面积;并集面积定义为模型预测的边界框和真实边界框总覆盖的面积,即两个边界框各自的面积之和减去交集面积。IoU取值范围在0到1之间,如果接近于1,表示模型预测的边界框与真实边界框高度重叠,预测准确性较高;如果接近于0,表示模型预测的边界框与真实边界框几乎没有重叠,预测准确性较低。αν的具体计算为
式(5)—(6)中:wh分别为预测框的宽和高;wgthgt分别为真实值的宽和高。
本文以YOLOv8为基础模型,提出了ERMS YOLOv8目标检测方法,专注于改善由于水下目标模糊导致检测精度低的问题。ERMS-YOLOv8的框架如图3所示。首先,替换主干网络为EfficientViT,增强模型对水下生物的特征提取能力,减少特征信息丢失;其次,在Neck部分采用RepGFPN网络模块,增强模型对水下生物高层语义和低级空间特征的提取和融合能力,从而使得模型获取更加丰富的特征信息;随后,引入MLCA注意力机制,使得模型同时融合通道信息,空间信息、局部通道信息和全局通道信息,增强模型的表征能力;最后,引入SIoU损失函数,提升模型对目标边界信息的提取能力,从而进一步提高检测精度。ERMS-YOLOv8的输入为水下图片,首先在EfficientViT进行特征提取,得到4张不同尺度大小的特征图;然后,将这4张特征图送入RepGFPN的Neck中,通过特征融合提取特征信息的表征能力;最后,通过解耦检测头对目标进行分类和定位,最终得到水下目标检测结果。
YOLOv8的传统主干网络结构由卷积层和残差模块相叠加构成。由于自身模型的计算冗余会限制训练和推理速度,因此YOLOv8不适合在水下这种复杂场景下部署。同时,传统的主干网络在处理跨尺度信息方面存在一定的局限,难以准确捕捉到水下目标的全局特征。为了解决上述问题,文献[17]提出一种高效视觉Transformer网络:EfficientViT,其主干网络结构如图4所示。
三明治布局块在前馈网络层(feed forward network,FFN)之间应用单个内存绑定多头注意力机制层(multi-head self-attention,MHSA),减少了MHSA中内存受限带来的时间成本,应用更多的FFN层来允许不同信道之间进行通信,从而提高了内存效率。具体来说,三明治布局应用了单一自注意力层进行空间混合,该层夹在FFN层的之间,计算公式为
式(7)中,Xi表示第i块的完整输入特征,布局将Xi变化为Xi+1,在单一自注意力层前后分别有N个FFN层的。该设计减少了模型中自注意力层造成的内存时间损耗,并采用了更多的FFN层来实现不同特征通道之间的高效通信。
为了解决MHSA中注意力头冗余导致的计算效率低下的问题,文献[17]参考Efficient CNNS的设计,提出了级联注意力模块(CGA)为每个头部提供完整特征的不同分割,从而显式地分解头部之间的注意力计算。这种注意力可以表示为
式(8)—(9)中:Xij表示输入特征Xi的第j次分裂;为输出特征;h为注意力头部的总数;分别是映射将输入特征分割不同子空间的投影层;是将连接输出特征投影与输入维度一致的线性层;Attn和Concat分别为注意力计算函数和合并数组函数。
特征金字塔网络大多数采用卷积神经网络(convolutional neural networks,CNN)作为主干网络来提取多尺度特征。研究发现,相比于CNN,特征金字塔网络(feature pyramid network,FPN)更强调高层次语义和低层次空间信息交换。因此,在目标检测模型中一般使用FPN。
图5展示了三种特征金字塔网络的设计。图5中,虚线框表示设计的每个FPN层。传统FPN引入了自上而下的路径来融合多尺度特征。考虑到单向信息流的限制,文献[18]提出了PANet网络结构,增加了一个额外的自下而上的路径聚合网络。为了实现高效的跨尺度连接以及加权特征图融合,文献[19]提出了BiFPN网络结构,去除只有一条输入边的节点,并在同一层次上从原始输入添加额外的边。为了增加跨层和跨尺度的连接,文献[20]提出了GFPN结构。
GFPN虽然能够充分交换不同层次的信息,但是带来了大量额外的上采样和下采样计算量。为了进一步提高目标检测的能力,本文引入了RepGFPN[21],结构如图6所示。
RepGFPN网络结构在GFPN的基础上减少了额外的上采样操作,通过5个CSPStage[22]模块接受更多节点输入,同时融合来自不同相邻层的不同尺度图像特征,在保证一定计算量的同时,增强特征复用及表达能力,提高目标检测的精准度。CSPStage模块通过引入递归操作以及重参数化机制,不断优化特征的融合过程减少模型推理的时间。CSPStage模块如图7所示。
RepGFPN网络中Rep的结构如图8所示。该结构推理阶段的主体仅由3×3卷积和ReLU堆栈组成,在训练阶段利用多分支形式的过参数化网络,在推理过程中将多分支融合为一个分支[23]。多分支架构的优点是利于训练,缺点是不利于推理,通过结构重参数化技术实现训练时间和推理时间结构的解耦。
注意力机制是计算机视觉中应用最广泛的组成部分之一,它可以帮助神经网络突出重要元素,抑制无关元素。现有的注意力模块主要集中在通道领域,这两种注意力机制与人类大脑中基于特征和基于空间的注意力完全对应。然而,绝大多数通道注意力机制只包含通道特征信息,忽略了空间特征信息,而空间特征信息对于构建空间注意力图又至关重要。对空间特征信息的忽略导致模型表示效果或目标检测性能较差。空间注意力模块往往复杂且昂贵,一些注意力过程虽然包含空间方面的信息,但计算量和参数要求太大,无法直接将这些信息纳入轻量级网络模型。文献[24]提出了一种轻量级的混合局部通道注意力(MLCA)模块来提高目标检测网络的性能。该模块可以同时包含信道信息和空间信息,以及局部信息和全局信息,采用一维卷积加速方法来减少计算量和参数的数量,以提高网络的表达效果。
MLCA的工作流程:第1步,将输入特征图经过局部平均池化和全局平均池化处理,局部池化提取局部区域的特征信息,全局池化捕捉整个特征图的统计信息;第2步,双分支结构处理,对全局信息分支使用全局平均池化将特征图压缩从而保留全局信息,对局部空间信息分支通过局部平均池化将特征图划分为多个局部块从而提取局部空间信息;第3步,对两个分支的特征分别经过一维卷积处理,压缩通道维度,同时保持空间维度不变;第4步,混合注意力融合,将局部和全局信息融合,生成最终的注意力特征图。
YOLOv8采用CIoU[25]作为边界框回归损失函数。尽管CIoU在提高精度方面有优势,但是也存在一些缺点,尤其是在处理不同大小的目标框时可能表现不佳,本文引入SIoU[26]作为新的边界框损失函数。
SIoU一般由4部分构成,包括角度损失、距离损失、形状损失和IoU损失。SIoU的定义为
式(10)中,角度损失为
式(11)—(12)中:σ为预测框中点与真实框中点之间的距离;Cx为两框中点之间的高度差。距离损失的计算为
形状损失为
式(13)—(16)中:分别为真实框中心坐标以及预测框中心坐标;wgthgtwh分别为真实框和预测框的宽和高;CwCh分别为真实框和预测框中心点的宽度差和高度差。
本文实验使用URPC2021和DUO公开水下数据集。其中,URPC2021数据集包含6575张图片以及对应的标注文件,检测目标包含海胆、海星、海参、扇贝四种水下生物,涵盖了真实的水下检测场景,本文将该数据集按照7∶3的比例进行划分,其中4585张作为训练集,1990张作为验证集。DUO数据集包含7782张图片以及对应的标注文件,同样涵盖海胆、海星、海参、扇贝这四种典型的水下生物,将该数据集同样按照7∶3的比例进行划分,5400张作为训练集,2382张作为验证集。URPC2021和DUO数据集的部分图片如图9所示。
本实验的实验环境如表1所示。本文训练模型采用随机梯度下降(SGD)作为优化器,输入图像尺寸为640×640,批量大小为16,训练周期为300轮。
为了测试本文基于YOLOv8改进模型的效果,本文利用平均精度(mean average precision,mAP)指标,其定义为
式(17)—(18)中:AP为精准度值中的召回率R以及精准度P相关,RP的定义分别为
式(19)—(20)中:PTP为当检测结果与真值之间的IoU值超过预定义的阈值时,认为该检测为真阳性(true positive,TP),表明目标识别成功;相反,PFP表示如果IoU低于该阈值,则检测被分类为假阳性(FP),表明识别错误;NFN为错误预测的样本,被称为真阴性(FN)。
为了验证改进的算法是否满足水下目标检测实时性的要求,本文将从模型计算量(giga floating point operations,GFLOPs)和每秒帧率(frame per second,FPS)来衡量确定。FPS越大,表示检测速度越快。
为了验证本文目标检测算法的有效性,本文将对比不同算法在URPC2021数据集上的表现。具体实验结果如表2所示。
对比Faster-RCNN以及不同版本YOLO系列算法在URPC2021数据集上的mAP@ 50、GFLOPS、FPS、参数量评价指标可见,YOLOv8仅以3.0×108的参数量和8.1GFLPOs的计算量,在mAP@50上达到了81.0%,成为最佳基础模型。
随后,通过对比结果分析可知,本文所提出的算法相较于基础模型YOLOv8,虽然浮点计算量和网络参数量有所增加,但在mAP@50上提高了2.9百分点,在mAP@50-95上提高了3.3百分点。同时,经过实验可以得到本文算法FPS达到了230虽然在一定程度上牺牲了处理图像特征的时间成本,但是仍然满足实时性需求。
最后,为了验证本文算法的优越性,与一些同类算法在相同实验设备和数据集的情况下进行对比实验,包括WCA-YOLOv8[11]算法,IEMAyoloViT[27]算法。实验结果表明,ERMS-YOLOv8相较于这两个算法具有一定的优越性。
为了验证ERMS-YOLOv8的泛化能力,本文对比不同算法在DUO数据集上的表现,具体结果如表3所示,通过对比分析可知,ERMS-YOLOv8相较于基础模型YOLOv8在mAP@50上提高了2.1百分点,在mAP@50-95上提高了2.7百分点。显然,本文算法在不同的数据集上都具有较好的检测效果。
为了验证MLCA的有效性,本文使用MLCA和一些常用的注意力机制对YOLOv8进行对比实验。实验结果如表4所示。实验结果表明,当加入ML CA时,YOLOv8在URPC2021数据集上表现出优异的性能,在几乎不增加模型参数规模的前提下,提升目标检测的精度。
下面通过消融实验验证本文算法的优越性,消融实验结果如表5所示。表5中模型1将YOLOv8主干网络替换为EfficientViT,相较于YOLOv8,mAP50%提升了0.6百分点,mAP@50-95%提升了0.4百分点,增强了模型的整体特征提取能力。模型2将YOLOv8的颈部网络替换为RepGFPN,相对于YOLOv8,mAP@50%提升了1.0百分点,mAP@50-95%提升了0.6百分点,增强了多尺度特征融合,提高了模型的多尺度检测能力,在不带来额外计算负担的条件下,实现更高的精度。模型3在主干网络中添加MLCA注意力机制,mAP@50%提升了0.4百分点,mAP@50-95%提升了0.2百分点,提高了目标信息的提取能力。模型4将损失函数CIoU替换为SIoU,mAP@50%和mAP@50-95%得到了小幅度提升。模型5、模型6、模型7展示了将这些改进结合在一起会使得目标检测的效果得到更大幅度的提升。
为了更加清晰地展示改进算法的检测效果,本文将选取部分照片进行对比验证,检测效果如图10所示。图10中第一列为原图、第二列为YOLOv8的检测图、第三列为改进算法的检测图。由图10可知,YOLOv8算法在进行水下目标检测时存在漏检和误检的问题,检测效果不佳。本文算法有效改善了这些问题,在保持水下检测实时性的同时,提高模型对水下目标的检测精度。
为了验证模型的抗干扰能力,本文选用两种场景进行鲁棒性实验。
第一个鲁棒性实验通过改变高斯噪声的标准差来对图像添加不同强度的高斯噪声,不同强度高斯噪声的图像对比如图11所示,实验结果如表6所示。在低噪声条件下(高斯噪声强度为0.01),模型的检测精准度与召回率几乎没有受到影响,说明模型具有一定的抗干扰能力;在高斯噪声强度为0.05的条件下,模型的检测精准度与召回率大幅下降。这是因为高斯噪声会使图像变得模糊,目标的边缘和细节变得不清晰,导致模型难以识别目标。
第二个鲁棒性实验测试不同光照强度对检测精准度与召回率的影响。不同强度光照的图像对比如图12所示。图12中,光照强度-0.2表示图像变暗原图的20%,0.2表示变亮原图的20%,实验结果如表7所示。当略微降低光照强度时,模型的检测精准度与召回率几乎没有受到影响,说明模型具有一定的抗干扰能力,并且当适量提升光照强度时,检测精准度与召回率略有提升。
针对水下复杂环境导致出现水下目标模糊等情况,本文提出了一种改进YOLOv8的目标检测算法来提高水下目标的检测精度。本文引入了EfficientViT主干网络、RepGFPN颈部网络和MLCA注意力机制来增强模型对水下目标生物的特征提取能力和跨尺度信息的融合能力,有效提升了对水下模糊目标的检测精度;替换模型损失函数为SIoU,在不损失计算量的前提下,提升模型对边界信息的提取能力,有效提升了目标检测的精度。通过一系列的对比实验和消融实验,验证了所提出算法的有效性。本文方法仍有进一步改进空间,未来将通过相关技术进一步提高水下目标的检测精度,将模型部署在水下的实际场景中并得到应用。
参考文献 引证文献
排序方式:
[1]
宋明龙.基于改进SSD的水下目标检测算法研究[D].淮南:安徽理工大学, 2023.
SONG M L. Research on underwater target detection algorithm based on improved SSD[D]. Huainan: Anhui University of Science and Technology, 2023.
[2]
LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553): 436-444.
[3]
GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA: IEEE, 2014: 580-587.
[4]
GIRSHICK R. Fast R-CNN[C]//2015 IEEE International Conference on Computer Vision (ICCV). Santiago, Chile: IEEE, 2015: 1440-1448.
[5]
REN S Q, HE K M, GIRSHICK R, et al. Faster RCNN: Towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[6]
REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: Unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas, NV, USA: IEEE, 2016: 779-788.
[7]
LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Proceedings of the 14th European Conference on Computer Vision. Amsterdam, Netherlands: Springer, 2016: 21-37.
[8]
强伟,贺昱曜,郭玉锦,.基于改进SSD的水下目标检测算法研究[J].西北工业大学学报, 2020, 38(04):747-754.
QIANG W, HE Y Y, GUO Y J, et al. Research on underwater target detection algorithm based on improved SSD[J]. Journal of Northwestern Polytechnical University, 2020, 38(04): 747-754.
[9]
陈宇梁,董绍江,孙世政,.改进YOLOv5s的弱光水下生物目标检测算法[J].北京航空航天大学学报, 2024, 50(2): 499-507.
CHEN Y L, DONG S J, SUN S Z, et al. Enhanced YOLOv5s for low-light underwater biological-object detection[J]. Journal of Beijing University of Aeronautics and Astronautics, 2024, 50(2): 499-507.
[10]
陶洋,朱腾,钟邦乾,. RepViTS-YOLOX:水下模糊及遮挡目标检测方法[J].计算机工程与应用, 2024, 60(13): 200-208.
TAO Y, ZHU T, ZHONG B Q, et al. RepViTS-YOLOX:detection method for blurred and occluded underwater targets[J]. Computer Engineering and Applications, 2024, 60(13): 200-208.
[11]
李大海,李冰涛,王振东.基于改进YOLOv8的水下目标检测算法[J].计算机应用, 2024, 44(11): 3610-3616.
LIDH,LIBT,WANGZD. Underwater object detection algorithm based on improved YOLOv8[J]. Journal of Computer Applications, 2024, 44(11): 3610-3616.
[12]
SHI K, W ANG J, ZHU H, et al. Intelligent target detection of underwater optical images based on image enhancement[C]//Proceedings of Fourteenth International Conference on Information Optics and Photonics (CIOP 2023). Xi'an, China: SPIE, 2023: 1-7.
[13]
YI H, LIU B, ZHAO B, et al. Small object detection algorithm based on improved YOLOv8 for remote sensing[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2024 (17): 1734-1747.
[14]
李姝,李思远,刘国庆.基于YOLOv8无人机航拍图像的小目标检测算法研究[J].小型微型计算机系统, 2024, 45(9): 2165-2174.
LI S, LI S Y,LIU G Q. Research on small target detection algorithm based on YOLOv8 UAV aerial images[J]. Journal of Chinese Systems, 2024, 45(9): 2165-2174.
[15]
LIX, W ANG W, WU L, et alGeneralized focal loss:Learning qualified and distributed bounding boxes for dense object detection[J]. Advances in Neural Information Processing Systems, 2020(33): 21002-21012.
[16]
YU J H, JIANG Y N, WANG Z Y, et al. Unitbox: An advanced object detection network[C]//Proceedings of the 24th ACM international conference on Multimedia. New York, USA: ACM, 2016: 516-520.
[17]
LIU X, PENG H, ZHENG N, et al. EfficientViT: Memory efficient vision transformer with cascaded group attention[C]//Proceedings of 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver:IEEE, 2023: 14420-14430.
[18]
LIU S, QI L, QIN H F, et al. Path aggregation network for instance segmentation[C]//Proceedings of 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, USA: IEEE: 8759-8768.
[19]
TAN M, PANG R, LE Q V. EfficientDet: Scalable and efficient object detection[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recogn ition. Seattle,WA, USA: IEEE, 2020: 10778-10787.
[20]
JIANG Y Q, TAN Z Y, WANG J Y. GiraffeDet: A heavy-neck paradigm for object detection[EB/OL].[2024-12-01]. https://arxiv.org/abs/2202.04256.
[21]
XU X Z, JIANG Y Q, CHEN W H, et al. Damo-YOLO:A report on real-time object detection design[EB/OL].[2024-12-01]. https://arxiv.org/abs/2211.15444.
[22]
WANG C Y, LIAOHYM,WUYH, et al. CSPNet: A new backbone that can enhance learning capability of CNN[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. Seattle, WA, USA: IEEE, 2020:390-391.
[23]
龙阳,肖小玲.改进YOLOv8的金属表面缺陷检测模型[J].制造技术与机床, 2024(8): 187-194.
LONG Y, XIAO X L. Enhanced YOLOv8 for metal-surface defect inspection[J]. Manufacturing Technology &Machine Tool, 2024(8): 187-194.
[24]
陈腾杰,李永安,张之好,.基于改进YOLOv8n+DeepSORT的带式输送机异物检测及计数方法[J].工矿自动化, 2024, 50(8): 91-98.
CHEN T J, LI Y A, ZHANG Z H, et al. Foreign-object detection and counting on belt conveyors via improved YOLOv8n plus DeepSORT[J]. Industry and Mine Automation, 2024, 50(8): 91-98.
[25]
ZHENG Z H, WANG P, LIU W, et al. Distance-IoU loss: Faster and better learning for bounding box regression[C]//Proceedings of the 34th AAAI Conference on Artificial Intelligence. PaloAlto California, USA: AAAI Press, 2020: 12993-13000.
[26]
DANG Y J, CHEN S L, MIAO H C, et al. Target detection from drone perspectives: Enhancing YOLOv5_3S with SIoU loss and SPD modules[C]//Cross Strait Radio Science and Wireless Technology Conference (CSRSWTC). Guilin, China: IEEE, 2023: 1-3.
[27]
施克权,李祺,隋皓,. IEMAyoloViT:基于改进YOLOv8的水下目标检测算法[J].电讯技术, 2025, 65(1): 54-62.
SHI K Q, LI Q, SUI H, et al. IEMA-yoloViT: An improved YOLOv8-based underwater object detection algorithm[J]. Telecommunication Engineering, 2025, 65(1): 54-62.
2025年第37卷第5期
PDF下载
69
37
引用本文
BibTeX
文章信息
doi: 10.3979/j.issn.1673-825X.202412060286
  • 接收时间:2024-12-06
  • 首发时间:2026-04-16
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-12-06
  • 修回日期:2025-09-08
基金
作者信息
    1重庆邮电大学 通信与信息工程学院,重庆 400065
    2重庆邮电大学 通信新技术应用研究中心,重庆 400065

通讯作者:

参考文献
分享链接
https://castjournals.cast.org.cn/joweb/cqyddxxb/CN/10.3979/j.issn.1673-825X.202412060286
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