Article(id=1251856526474953466, tenantId=1146029695717560320, journalId=1251234268282663017, issueId=1251856520619700745, articleNumber=null, orderNo=null, doi=10.3969/j.issn.1003-3106.2025.11.005, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1752681600000, receivedDateStr=2025-07-17, revisedDate=null, revisedDateStr=null, acceptedDate=null, acceptedDateStr=null, onlineDate=1776395454281, onlineDateStr=2026-04-17, pubDate=1762272000000, pubDateStr=2025-11-05, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1776395454281, onlineIssueDateStr=2026-04-17, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1776395454281, creator=13701087609, updateTime=1776395454281, updator=13701087609, issue=Issue{id=1251856520619700745, tenantId=1146029695717560320, journalId=1251234268282663017, year='2025', volume='55', issue='11', pageStart='2131', pageEnd='2324', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=1, specialIssue=null, createTime=1776395452885, creator=13701087609, updateTime=1776395571911, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1251857019939013255, tenantId=1146029695717560320, journalId=1251234268282663017, issueId=1251856520619700745, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1251857019939013256, tenantId=1146029695717560320, journalId=1251234268282663017, issueId=1251856520619700745, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=2174, endPage=2183, ext={EN=ArticleExt(id=1251856527364145926, articleId=1251856526474953466, tenantId=1146029695717560320, journalId=1251234268282663017, language=EN, title=Target Detection Algorithm for Ship Infrared Images Based on Improved YOLO11n, columnId=1251856523492798993, journalTitle=Radio Engineering, columnName=Signal and Information Processing, runingTitle=null, highlight=null, articleAbstract=

A ship infrared image target detection algorithm based on YOLO11n, named AGT-YOLO, is proposed to address the issues of low model accuracy and recall rate, difficulties in identifying small targets, and multi-scale recognition challenges under complex sea conditions. By introducing an improved GhostHGNetv2 network, the background discrimination capability is enhanced; the designed ASFP2 optimized neck network improves detection capabilities for low-resolution images and very small targets; the proposed Tack Adaptive Alignment Detection Head ( TAADH ) replaces the original detection head, enhancing localization and classification performance;meanwhile, the AFGCAttention mechanism is integrated to improve global information processing capability and the model's generalization ability. Experimental results show that compared to the baseline model YOLO11n, AGT-YOLO achieves a 4.4% increase in recall rate and a 3.1% increase in mean average precision at IoU=0.5 ( mAP@ 50), demonstrating strong multi-scale recognition capability and robustness in complex environments.

, correspAuthors=Cifa CHEN, 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=Peng PENG, Cifa CHEN, Shang ZHANG), CN=ArticleExt(id=1251856537061376109, articleId=1251856526474953466, tenantId=1146029695717560320, journalId=1251234268282663017, language=CN, title=基于改进YOLO11n的船舶红外图像目标检测算法, columnId=1251856523803177493, journalTitle=无线电工程, columnName=信号与信息处理, runingTitle=null, highlight=null, articleAbstract=

提出一种基于YOLO11n的船舶红外图像目标检测算法AGT-YOLO,旨在解决模型精度和召回率偏低、小目标识别困难以及复杂海况下的多尺度识别问题。通过引入改进后的GhostHGNetv2网络,增强背景区分能力;设计ASF-P2优化颈部网络,以提升对低分辨率图像和极小目标的检测能力;研究任务自适应对齐检测头(Task Adaptive Alignment Detection Head,TAADH)替换原有检测头,提升定位和分类性能;同时,融入AFGCAttention注意力机制,提高全局信息处理能力和模型的泛化能力。实验结果表明,与基准模型YOLO11n相比,AGT-YOLO的召回率提高了4.4%,平均精度均值@50(mean Average Precision at IoU=0.5,mAP@50)提高了3.1%,并在复杂环境下展现出较强的多尺度识别能力和鲁棒性。

, correspAuthors=陈慈发, authorNote=null, correspAuthorsNote=
陈慈发 男,(1967—),教授。主要研究方向:嵌入式系统、物联网、计算机测控系统。
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=HkvfXd9CPBx6YBwE/iuGJg==, magXml=iol68sa5645/bFEu6xuYSA==, pdfUrl=null, pdf=KEDH+hOnrpzFrHLQ7DD7dw==, pdfFileSize=9488057, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=HDtVhyUiCcbxe5ztKAmruA==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=p45QUF86Lk+0YXFfGA1G4g==, mapNumber=null, authorCompany=null, fund=null, authors=

彭鹏 男,(2001—),硕士研究生。主要研究方向:图像处理、计算机视觉。

张上 男,(1979—),博士,副教授。主要研究方向:物联网、计算机应用、图像处理。

, authorsList=彭鹏, 陈慈发, 张上)}, authors=[Author(id=1251856537921208495, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, orderNo=0, 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=1251856538047037624, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856537921208495, language=EN, stringName=Peng PENG, firstName=Peng, middleName=null, lastName=PENG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China
2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China
3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251856538210615488, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856537921208495, language=CN, stringName=彭鹏, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002
2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002
3.三峡大学 计算机与信息学院,湖北 宜昌 443002, bio={"content":"

彭鹏 男,(2001—),硕士研究生。主要研究方向:图像处理、计算机视觉。

"}, bioImg=null, bioContent=

彭鹏 男,(2001—),硕士研究生。主要研究方向:图像处理、计算机视觉。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251856537422086278, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=1., ext=[AuthorCompanyExt(id=1251856537443057802, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537455640717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002)]), AuthorCompany(id=1251856537531138193, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=2., ext=[AuthorCompanyExt(id=1251856537539526803, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537547915413, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002)]), AuthorCompany(id=1251856537640190107, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=3., ext=[AuthorCompanyExt(id=1251856537648578717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537656967326, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.三峡大学 计算机与信息学院,湖北 宜昌 443002)])]), Author(id=1251856538298695880, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, 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=1, authorType=1, ext={EN=AuthorExt(id=1251856538411942095, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856538298695880, language=EN, stringName=Cifa CHEN, firstName=Cifa, middleName=null, lastName=CHEN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, 4, *, address=2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China
3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China
4.Big Data Research Center, Jingchu University of Technology, Jingmen 448001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251856538525188310, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856538298695880, language=CN, stringName=陈慈发, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, 4, *, address=2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002
3.三峡大学 计算机与信息学院,湖北 宜昌 443002
4.荆楚理工学院 大数据研究中心,湖北 荆门 448001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251856537531138193, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=2., ext=[AuthorCompanyExt(id=1251856537539526803, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537547915413, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002)]), AuthorCompany(id=1251856537640190107, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=3., ext=[AuthorCompanyExt(id=1251856537648578717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537656967326, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.三峡大学 计算机与信息学院,湖北 宜昌 443002)]), AuthorCompany(id=1251856537774407844, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=4., ext=[AuthorCompanyExt(id=1251856537786990757, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537774407844, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4.Big Data Research Center, Jingchu University of Technology, Jingmen 448001, China), AuthorCompanyExt(id=1251856537795379366, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537774407844, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4.荆楚理工学院 大数据研究中心,湖北 荆门 448001)])]), Author(id=1251856538630045913, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, 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=1251856538881704165, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856538630045913, language=EN, stringName=Shang ZHANG, firstName=Shang, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China
2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China
3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1251856540488122606, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, authorId=1251856538630045913, language=CN, stringName=张上, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002
2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002
3.三峡大学 计算机与信息学院,湖北 宜昌 443002, bio={"content":"

张上 男,(1979—),博士,副教授。主要研究方向:物联网、计算机应用、图像处理。

"}, bioImg=null, bioContent=

张上 男,(1979—),博士,副教授。主要研究方向:物联网、计算机应用、图像处理。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1251856537422086278, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=1., ext=[AuthorCompanyExt(id=1251856537443057802, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537455640717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002)]), AuthorCompany(id=1251856537531138193, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=2., ext=[AuthorCompanyExt(id=1251856537539526803, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537547915413, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002)]), AuthorCompany(id=1251856537640190107, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=3., ext=[AuthorCompanyExt(id=1251856537648578717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537656967326, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.三峡大学 计算机与信息学院,湖北 宜昌 443002)])])], keywords=[Keyword(id=1251856540714615031, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, orderNo=1, keyword=object detection), Keyword(id=1251856540819472637, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, orderNo=2, keyword=YOLO11n), Keyword(id=1251856540949496071, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, orderNo=3, keyword=infrared ship inspection), Keyword(id=1251856541079519501, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, orderNo=4, keyword=multi-scale model), Keyword(id=1251856541192765717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, orderNo=1, keyword=目标检测), Keyword(id=1251856541297623324, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, orderNo=2, keyword=YOLO11n), Keyword(id=1251856541377315105, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, orderNo=3, keyword=红外船舶检测), Keyword(id=1251856541473784104, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, orderNo=4, keyword=多尺度模型)], refs=[Reference(id=1251856545437401541, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=王斌, journalName=海面目标红外检测方法研究, refType=null, unstructuredReference=王斌.海面目标红外检测方法研究[D].大连:大连海事大学,2018., articleTitle=null, refAbstract=null), Reference(id=1251856545504510409, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=古静, journalName=基于深度学习的红外船舶目标检测方法研究, refType=null, unstructuredReference=古静.基于深度学习的红外船舶目标检测方法研究[D].镇江:江苏科技大学,2023., articleTitle=null, refAbstract=null), Reference(id=1251856545626145233, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2023, volume=49, issue=4, pageStart=103, pageEnd=115, url=null, language=null, rfNumber=[3], rfOrder=2, authorNames=王勇, 王柏容, 慕东东, journalName=大连海事大学学报, refType=null, unstructuredReference=王勇,王柏容,慕东东.红外船舶检测的研究现状及展望[J].大连海事大学学报,2023,49(4):103-115., articleTitle=红外船舶检测的研究现状及展望, refAbstract=null), Reference(id=1251856545726808531, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2021, volume=9, issue=null, pageStart=16692, pageEnd=16706, url=null, language=null, rfNumber=[4], rfOrder=3, authorNames=LI H, DENG L B, YANG C, journalName=IEEE Access, refType=null, unstructuredReference=LI H,DENG L B,YANG C,et al. Enhanced YOLO v3 Tiny Network for Real-time Ship Detection from Visual Image[J]. IEEE Access,2021,9:16692-16706., articleTitle=Enhanced YOLO v3 Tiny Network for Real-time Ship Detection from Visual Image, refAbstract=null), Reference(id=1251856545819083222, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2022, volume=22, issue=10, pageStart=3782, pageEnd=null, url=null, language=null, rfNumber=[5], rfOrder=4, authorNames=YE J, YUAN Z Y, QIAN C, journalName=Sensors, refType=null, unstructuredReference=YE J,YUAN Z Y,QIAN C,et al. CAA-YOLO:Combined-attention-augmented YOLO for Infrared Ocean Ships Detection[J]. Sensors,2022,22(10):3782., articleTitle=CAA-YOLO:Combined-attention-augmented YOLO for Infrared Ocean Ships Detection, refAbstract=null), Reference(id=1251856545911357916, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2025, volume=32, issue=1, pageStart=61, pageEnd=67, url=null, language=null, rfNumber=[6], rfOrder=5, authorNames=王海群, 魏培旭, 解浩龙, journalName=电光与控制, refType=null, unstructuredReference=王海群,魏培旭,解浩龙,.基于改进YOLOv8的红外船舶检测[J].电光与控制,2025,32(1):61-67., articleTitle=基于改进YOLOv8的红外船舶检测, refAbstract=null), Reference(id=1251856545986855390, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2024, volume=24, issue=17, pageStart=5789, pageEnd=null, url=null, language=null, rfNumber=[7], rfOrder=6, authorNames=DING Q X, LI W C, XU C C, journalName=Sensors, refType=null, unstructuredReference=DING Q X,LI W C,XU C C,et al. GMS-YOLO:An Algorithm for Multi-scale Object Detection in Complex Environments in Confined Compartments[J]. Sensors,2024, 24(17):5789., articleTitle=GMS-YOLO:An Algorithm for Multi-scale Object Detection in Complex Environments in Confined Compartments, refAbstract=null), Reference(id=1251856546074935777, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2024-11-26, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[8], rfOrder=7, authorNames=HAN D, YE T, HAN Y, journalName=null, refType=null, unstructuredReference=HAN D,YE T,HAN Y,et al. Agent Attention:On the Integration of Softmax and Linear Attention[EB/OL]. (2023-12-14)[2024-11-26]. https://arxiv.org/abs/2312.08874., articleTitle=Agent Attention:On the Integration of Softmax and Linear Attention, refAbstract=null), Reference(id=1251856546154627556, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2024-11-26, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[9], rfOrder=8, authorNames=KANG M, TING C M, TING F F, journalName=null, refType=null, unstructuredReference=KANG M,TING C M,TING F F,et al. ASF-YOLO:A Novel YOLO Model with Attentional Scale Sequence Fusion for Cell Instance Segmentation[EB/OL]. (2023-12-11)[2024-11-26]. https://arxiv.org/abs/2312.06458., articleTitle=ASF-YOLO:A Novel YOLO Model with Attentional Scale Sequence Fusion for Cell Instance Segmentation, refAbstract=null), Reference(id=1251856546259485163, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2021, volume=40, issue=null, pageStart=3437, pageEnd=3455, url=null, language=null, rfNumber=[10], rfOrder=9, authorNames=YANGX, ZHANG Y F, LI T, journalName=Circuits,Systems,and Signal Processing, refType=null, unstructuredReference=YANGX,ZHANG Y F,LI T,et al. Image Super-resolution Based on the Down-sampling Iterative Module and Deep CNN[J]. Circuits,Systems,and Signal Processing,2021, 40:3437-3455., articleTitle=Image Super-resolution Based on the Down-sampling Iterative Module and Deep CNN, refAbstract=null), Reference(id=1251856546339176941, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=9, pageStart=4432, pageEnd=null, url=null, language=null, rfNumber=[11], rfOrder=10, authorNames=PARKH J, KANG J W, KIM B G, journalName=Sensors, refType=null, unstructuredReference=PARKH J,KANG J W,KIM B G. ssFPN:Scale Sequence (S2) Feature-based Feature Pyramid Network for Object Detection[J]. Sensors,2023,23(9):4432., articleTitle=ssFPN:Scale Sequence (S2) Feature-based Feature Pyramid Network for Object Detection, refAbstract=null), Reference(id=1251856546465006065, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=4489, pageEnd=4497, url=null, language=null, rfNumber=[12], rfOrder=11, authorNames=TRAN D, BOURDEV L, FERGUS R, journalName=null, refType=null, unstructuredReference=TRAN D,BOURDEV L,FERGUS R,et al. Learning Spatiotemporal Features with 3D Convolutional Networks[C]//2015 IEEE International Conference on Computer Vision (ICCV). Santiago:IEEE,2015:4489-4497., articleTitle=Learning Spatiotemporal Features with 3D Convolutional Networks, refAbstract=null), Reference(id=1251856546561475059, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2018, volume=107, issue=null, pageStart=3, pageEnd=11, url=null, language=null, rfNumber=[13], rfOrder=12, authorNames=ELFWING S, UCHIBE E, DOYA K, journalName=Neural Networks, refType=null, unstructuredReference=ELFWING S, UCHIBE E, DOYA K. Sigmoid-weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning[J]. Neural Networks,2018, 107:3-11., articleTitle=Sigmoid-weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning, refAbstract=null), Reference(id=1251856546636972534, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2022, volume=121, issue=null, pageStart=108159, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=13, authorNames=JIN X, XIE Y P, WEI X S, journalName=Pattern Recognition, refType=null, unstructuredReference=JIN X,XIE Y P,WEI X S,et al. Delving Deep into Spatial Pooling for Squeeze-and-Excitation Networks[J]. Pattern Recognition,2022,121:108159., articleTitle=Delving Deep into Spatial Pooling for Squeeze-and-Excitation Networks, refAbstract=null), Reference(id=1251856546699887097, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=3490, pageEnd=3499, url=null, language=null, rfNumber=[15], rfOrder=14, authorNames=FENGC J, ZHONG Y J, GAO Y, journalName=null, refType=null, unstructuredReference=FENGC J,ZHONG Y J,GAO Y,et al. Tood:Task-aligned One-stage Object Detection[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Montreal:IEEE,2021:3490-3499., articleTitle=Tood:Task-aligned One-stage Object Detection, refAbstract=null), Reference(id=1251856546754413052, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2024-11-26, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[16], rfOrder=15, authorNames=IOFFES, SZEGEDY C, journalName=null, refType=null, unstructuredReference=IOFFES,SZEGEDY C. Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift[EB/OL]. (2015-02-11)[2024-11-26].https://arxiv.org/abs/1502.03167., articleTitle=Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift, refAbstract=null), Reference(id=1251856546821521921, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=3, pageEnd=19, url=null, language=null, rfNumber=[17], rfOrder=16, authorNames=WU Y X, HE K M, journalName=null, refType=null, unstructuredReference=WU Y X,HE K M.Group Normalization[C]//Proceedings of the European Conference on Computer Vision (ECCV). Munich:Springer,2018:3-19., articleTitle=Group Normalization, refAbstract=null), Reference(id=1251856546892825092, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2020, volume=44, issue=4, pageStart=1922, pageEnd=1933, url=null, language=null, rfNumber=[18], rfOrder=17, authorNames=TIAN Z, SHEN C H, CHEN H, journalName=IEEE Transactions on Pattern Analysis and Machine Intelligence, refType=null, unstructuredReference=TIAN Z,SHEN C H,CHEN H,et al. FCOS:A Simple and Strong Anchor-free Object Detector[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2020, 44(4):1922-1933., articleTitle=FCOS:A Simple and Strong Anchor-free Object Detector, refAbstract=null), Reference(id=1251856546980905477, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2024-11-26, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[19], rfOrder=18, authorNames=SHAW P, USZKOREIT J, VASWANI A, journalName=null, refType=null, unstructuredReference=SHAW P,USZKOREIT J,VASWANI A. Self-attention with Relative Position Representations[EB/OL]. (2018-03-06)[2024-11-26]. https://arxiv.org/abs/1803.02155., articleTitle=Self-attention with Relative Position Representations, refAbstract=null), Reference(id=1251856547068985864, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2025, volume=55, issue=2, pageStart=271, pageEnd=280, url=null, language=null, rfNumber=[20], rfOrder=19, authorNames=杨慧聪, 陈慈发, 张上, journalName=无线电工程, refType=null, unstructuredReference=杨慧聪,陈慈发,张上.基于通道剪枝的YOLOv8n印刷电路板缺陷检测[J].无线电工程,2025,55(2):271-280., articleTitle=基于通道剪枝的YOLOv8n印刷电路板缺陷检测, refAbstract=null), Reference(id=1251856547182232078, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=839, pageEnd=847, url=null, language=null, rfNumber=[21], rfOrder=20, authorNames=CHATTOPADHAY A, SARKAR A, HOWLADER P, journalName=null, refType=null, unstructuredReference=CHATTOPADHAY A, SARKAR A, HOWLADER P,et al.Grad-CAM++:Generalized Gradient-based Visual Explanations for Deep Convolutional Networks[C]//2018 IEEE Winter Conference on Applications of Computer Vision (WACV). Lake Tahoe:IEEE,2018:839-847., articleTitle=Grad-CAM++:Generalized Gradient-based Visual Explanations for Deep Convolutional Networks, refAbstract=null), Reference(id=1251856547291283987, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, doi=null, pmid=null, pmcid=null, year=2025, volume=55, issue=8, pageStart=1560, pageEnd=1570, url=null, language=null, rfNumber=[22], rfOrder=21, authorNames=贾星宇, 李大鹏, journalName=无线电工程, refType=null, unstructuredReference=贾星宇,李大鹏.基于YOLO11的无人机航拍图像小目标检测算法[J].无线电工程,2025,55(8):1560-1570., articleTitle=基于YOLO11的无人机航拍图像小目标检测算法, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1251856537422086278, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=1., ext=[AuthorCompanyExt(id=1251856537443057802, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537455640717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537422086278, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002)]), AuthorCompany(id=1251856537531138193, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=2., ext=[AuthorCompanyExt(id=1251856537539526803, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537547915413, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537531138193, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002)]), AuthorCompany(id=1251856537640190107, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=3., ext=[AuthorCompanyExt(id=1251856537648578717, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China), AuthorCompanyExt(id=1251856537656967326, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537640190107, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.三峡大学 计算机与信息学院,湖北 宜昌 443002)]), AuthorCompany(id=1251856537774407844, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, xref=4., ext=[AuthorCompanyExt(id=1251856537786990757, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537774407844, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4.Big Data Research Center, Jingchu University of Technology, Jingmen 448001, China), AuthorCompanyExt(id=1251856537795379366, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, companyId=1251856537774407844, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4.荆楚理工学院 大数据研究中心,湖北 荆门 448001)])], figs=[ArticleFig(id=1251856541624779059, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.1, caption=Structure of AGT-YOLO network, figureFileSmall=HZE3VQD7FhM1wyDcXaJbmg==, figureFileBig=HDtVhyUiCcbxe5ztKAmruA==, tableContent=null), ArticleFig(id=1251856541704470841, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图1, caption=AGT-YOLO网络结构, figureFileSmall=HZE3VQD7FhM1wyDcXaJbmg==, figureFileBig=HDtVhyUiCcbxe5ztKAmruA==, tableContent=null), ArticleFig(id=1251856541943546180, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.2, caption=Network structure of improved GhostHGNetv2, figureFileSmall=+1nJgdb433my+ExwTqSRYA==, figureFileBig=9O+iBd+oFHP9nTWULxXjsQ==, tableContent=null), ArticleFig(id=1251856542052598090, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图2, caption=改进后的GhostHGNetv2网络结构, figureFileSmall=+1nJgdb433my+ExwTqSRYA==, figureFileBig=9O+iBd+oFHP9nTWULxXjsQ==, tableContent=null), ArticleFig(id=1251856542178427218, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.3, caption=Structure of Ghost_Block module, figureFileSmall=RtjsIAM0s6iozppIFPZEHQ==, figureFileBig=wW+UwRwP+qxvdMZCSE/cgA==, tableContent=null), ArticleFig(id=1251856542279090522, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图3, caption=Ghost_Block模块结构, figureFileSmall=RtjsIAM0s6iozppIFPZEHQ==, figureFileBig=wW+UwRwP+qxvdMZCSE/cgA==, tableContent=null), ArticleFig(id=1251856542354587999, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.4, caption=Structure of Scalseq module, figureFileSmall=VU0ePPMe/tHz4gQ7aAF3fw==, figureFileBig=g1Ej3h2oUIVst2+mj5k6uw==, tableContent=null), ArticleFig(id=1251856542459445605, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图4, caption=Scalseq模块结构, figureFileSmall=VU0ePPMe/tHz4gQ7aAF3fw==, figureFileBig=g1Ej3h2oUIVst2+mj5k6uw==, tableContent=null), ArticleFig(id=1251856542547525995, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.5, caption=Structure of asf_attention module, figureFileSmall=5IJGwwRsAnOq9RNiDKjcwA==, figureFileBig=Mhwwm3AHpnzyudVIU9pe3w==, tableContent=null), ArticleFig(id=1251856542639800688, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图5, caption=asf_attention模块结构, figureFileSmall=5IJGwwRsAnOq9RNiDKjcwA==, figureFileBig=Mhwwm3AHpnzyudVIU9pe3w==, tableContent=null), ArticleFig(id=1251856542702715250, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.6, caption=Structure of TAADH module, figureFileSmall=fcRaLqZHsJH0W9+9eYQlWQ==, figureFileBig=HM/7NENqpHJjbQKLYg5fIw==, tableContent=null), ArticleFig(id=1251856542790795642, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图6, caption=TAADH模块结构, figureFileSmall=fcRaLqZHsJH0W9+9eYQlWQ==, figureFileBig=HM/7NENqpHJjbQKLYg5fIw==, tableContent=null), ArticleFig(id=1251856542878876030, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.7, caption=Structure of task decomposition module, figureFileSmall=80LPZ7gHs3dj9e3HkFKCuw==, figureFileBig=mY2/bjycX6+FNHCH5Xgeaw==, tableContent=null), ArticleFig(id=1251856542958567812, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图7, caption=任务分解模块结构, figureFileSmall=80LPZ7gHs3dj9e3HkFKCuw==, figureFileBig=mY2/bjycX6+FNHCH5Xgeaw==, tableContent=null), ArticleFig(id=1251856543059231112, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.8, caption=Structure of AFGC module, figureFileSmall=J0QXhSV7NRA3qII0oLVIVg==, figureFileBig=oMuDtXU12NkvnYNNaSd1lw==, tableContent=null), ArticleFig(id=1251856543117951374, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图8, caption=AFGC模块结构, figureFileSmall=J0QXhSV7NRA3qII0oLVIVg==, figureFileBig=oMuDtXU12NkvnYNNaSd1lw==, tableContent=null), ArticleFig(id=1251856543206031765, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Fig.9, caption=Heatmap of ship inspection, figureFileSmall=rbyQPDel5EXAJ2LKoVL2iQ==, figureFileBig=tNwgbI8GnavA6lxi0zFfAQ==, tableContent=null), ArticleFig(id=1251856543331860892, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=图9, caption=船舶检测热力图, figureFileSmall=rbyQPDel5EXAJ2LKoVL2iQ==, figureFileBig=tNwgbI8GnavA6lxi0zFfAQ==, tableContent=null), ArticleFig(id=1251856543407358371, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Tab.1, caption=

Results of ablation experiments

, figureFileSmall=null, figureFileBig=null, tableContent=
AliasImprovement strategyPRmAP@0.5mAP@0.5:0.95
GhostHGNetv2TAADHASF-P2AFGCAttention
A××××90.884.990.363.4
B×××92.385.291.064.0
C××91.786.591.664.6
D×91.089.093.264.6
E91.989.393.464.7
), ArticleFig(id=1251856544992805287, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=表1, caption=

消融实验结果

, figureFileSmall=null, figureFileBig=null, tableContent=
AliasImprovement strategyPRmAP@0.5mAP@0.5:0.95
GhostHGNetv2TAADHASF-P2AFGCAttention
A××××90.884.990.363.4
B×××92.385.291.064.0
C××91.786.591.664.6
D×91.089.093.264.6
E91.989.393.464.7
), ArticleFig(id=1251856545076691371, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Tab.2, caption=

Experimental comparison between AGT-YOLO and existing models

, figureFileSmall=null, figureFileBig=null, tableContent=
模型P/%R/%Para/106mAP@0.5/%mAP@0.5:0.95/%FPS
SSD82.080.586.842.833
CenterNet88.191.291.758.839
YOLOv5n90.684.72.5090.362.8135
YOLOv689.383.14.2388.662.3141
YOLOv7-tiny89.287.06.0391.060.4110
YOLOv8n90.485.43.0190.863.7135
YOLOv9t90.885.51.9790.163.771
YOLOv10n88.883.62.7089.862.6112
YOLO11n90.884.92.5890.363.4136
AGT-YOLO91.989.32.5193.464.7121
), ArticleFig(id=1251856545164771762, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=表2, caption=

AGT-YOLO与已有模型实验对比

, figureFileSmall=null, figureFileBig=null, tableContent=
模型P/%R/%Para/106mAP@0.5/%mAP@0.5:0.95/%FPS
SSD82.080.586.842.833
CenterNet88.191.291.758.839
YOLOv5n90.684.72.5090.362.8135
YOLOv689.383.14.2388.662.3141
YOLOv7-tiny89.287.06.0391.060.4110
YOLOv8n90.485.43.0190.863.7135
YOLOv9t90.885.51.9790.163.771
YOLOv10n88.883.62.7089.862.6112
YOLO11n90.884.92.5890.363.4136
AGT-YOLO91.989.32.5193.464.7121
), ArticleFig(id=1251856545252852153, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=EN, label=Tab.3, caption=

Generalization experiment

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@0.5mAP@0.5∶0.95PR
YOLO11n34.719.945.635.6
AGT-YOLO40.523.851.039.7
), ArticleFig(id=1251856545332543933, tenantId=1146029695717560320, journalId=1251234268282663017, articleId=1251856526474953466, language=CN, label=表3, caption=

泛化实验

, figureFileSmall=null, figureFileBig=null, tableContent=
模型mAP@0.5mAP@0.5∶0.95PR
YOLO11n34.719.945.635.6
AGT-YOLO40.523.851.039.7
)], attaches=null, journal=Journal(id=1251231494547484770, delFlag=0, nameCn=无线电工程, nameEn=Radio Engineering, nameHistory1=null, nameHistory2=null, issn=1003-3106, eissn=, cn=13-1097/TN, coden=null, periodic=月刊, language=CN, oaType=1, 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=, officePhone=, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=g0Lcfmfo7oUYDLtQTsujLQ==, journalPrice=null, startedYear=null, abbrevIsoEn=Radio Engineering, journalRemark=null, publicationField=null, createdTime=1776246435060, updatedTime=1776252047215, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=R, firstLetterEn=R, subjectCode=Engineering, subjectName=工程, subjectCodeEn=Engineering, subjectNameEn=null, picCn=g0Lcfmfo7oUYDLtQTsujLQ==, picEn=fe/NRE6hx/Z5ZHweFTw2gA==, jcr=null, cjcr=null, exts=[JournalExt(id=1251255033736676338, language=CN, name=无线电工程, 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=1776252047237, updatedTime=1776252047237, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=1, submissionEditorUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=3, submissionReviewUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=2, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1251255033787007987, language=EN, name=Radio Engineering, 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=1776252047249, updatedTime=1776252047249, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=1, submissionEditorUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=3, submissionReviewUrl=https://wxdg.cbpt.cnki.net/index.aspx?t=2, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1251234268282663017, websiteList=[Website(id=1251257283494232110, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1251234268282663017, 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/wxdgc/CN, language=CN, createTime=1776252583625, createBy=18614031015, updateTime=1776253601640, updateBy=18614031015, name=无线电工程-中文, tplId=1146099689490845704, title=无线电工程, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1251261682236997912, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=articleTextType, value=kx, createTime=1776253632363, updateTime=1776253632363, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682199249173, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=banner, value=null, createTime=1776253632354, updateTime=1776253632354, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682262163739, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=grayFlag, value=0, createTime=1776253632369, updateTime=1776253632369, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682190860564, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=logo, value=https://castjournals.cast.org.cn/joweb/wxdgc/CN/file/pic?fileId=AnvFcVcgnutwkVR+TFSYqg==, createTime=1776253632352, updateTime=1776253632352, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682283135261, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=minRunFlag, value=0, createTime=1776253632374, updateTime=1776253632374, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682224414999, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/wxdgc/CN/file/pic, createTime=1776253632360, updateTime=1776253632360, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682274746652, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=silenceFlag, value=0, createTime=1776253632372, updateTime=1776253632372, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682207637782, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1776253632356, updateTime=1776253632356, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682245386521, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=themeColor, value=null, createTime=1776253632365, updateTime=1776253632365, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261682249580826, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283494232110, code=themeStyle, value=null, createTime=1776253632366, updateTime=1776253632366, creator=18614031015, updator=18614031015)]), Website(id=1251257283603284042, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1251234268282663017, 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/wxdgc/EN, language=EN, createTime=1776252583647, createBy=18614031015, updateTime=1776253597767, updateBy=18614031015, name=无线电工程-英文, tplId=1146101810881728533, title=Radio Engineering, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1251261640977625682, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=articleTextType, value=kx, createTime=1776253622526, updateTime=1776253622526, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640956654159, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=banner, value=null, createTime=1776253622521, updateTime=1776253622521, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261641002791509, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=grayFlag, value=0, createTime=1776253622532, updateTime=1776253622532, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640948265550, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=logo, value=https://castjournals.cast.org.cn/joweb/wxdgc/EN/file/pic?fileId=AnvFcVcgnutwkVR+TFSYqg==, createTime=1776253622519, updateTime=1776253622519, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261641015374423, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=minRunFlag, value=0, createTime=1776253622535, updateTime=1776253622535, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640969237073, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/wxdgc/EN/file/pic, createTime=1776253622524, updateTime=1776253622524, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261641011180118, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=silenceFlag, value=0, createTime=1776253622534, updateTime=1776253622534, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640965042768, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1776253622523, updateTime=1776253622523, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640986014291, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=themeColor, value=null, createTime=1776253622528, updateTime=1776253622528, creator=18614031015, updator=18614031015), WebsiteProps(id=1251261640994402900, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1251257283603284042, code=themeStyle, value=null, createTime=1776253622530, updateTime=1776253622530, creator=18614031015, updator=18614031015)])], journalTitle=无线电工程, weixinUrl=null, journalUrl=https://wxdg.cbpt.cnki.net/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Radio Engineering, journalPhotoCn=g0Lcfmfo7oUYDLtQTsujLQ==, journalPhotoEn=fe/NRE6hx/Z5ZHweFTw2gA==, journalFirstLetter=R, 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/wxdgc/CN/10.3969/j.issn.1003-3106.2025.11.005, detailUrlEn=https://castjournals.cast.org.cn/joweb/wxdgc/EN/10.3969/j.issn.1003-3106.2025.11.005, pdfUrlCn=https://castjournals.cast.org.cn/joweb/wxdgc/CN/PDF/10.3969/j.issn.1003-3106.2025.11.005, pdfUrlEn=https://castjournals.cast.org.cn/joweb/wxdgc/EN/PDF/10.3969/j.issn.1003-3106.2025.11.005, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于改进YOLO11n的船舶红外图像目标检测算法
收藏切换
PDF下载
彭鹏 1, 2, 3 , 陈慈发 2, 3, 4, * , 张上 1, 2, 3
无线电工程 | 信号与信息处理 2025,55(11): 2174-2183
收起
收藏切换
无线电工程 | 信号与信息处理 2025, 55(11): 2174-2183
基于改进YOLO11n的船舶红外图像目标检测算法
全屏
彭鹏1, 2, 3, 陈慈发2, 3, 4, *, 张上1, 2, 3
作者信息
  • 1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002
  • 2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002
  • 3.三峡大学 计算机与信息学院,湖北 宜昌 443002
  • 4.荆楚理工学院 大数据研究中心,湖北 荆门 448001
  • 彭鹏 男,(2001—),硕士研究生。主要研究方向:图像处理、计算机视觉。

    张上 男,(1979—),博士,副教授。主要研究方向:物联网、计算机应用、图像处理。

通讯作者:

陈慈发 男,(1967—),教授。主要研究方向:嵌入式系统、物联网、计算机测控系统。
Target Detection Algorithm for Ship Infrared Images Based on Improved YOLO11n
Peng PENG1, 2, 3, Cifa CHEN2, 3, 4, *, Shang ZHANG1, 2, 3
Affiliations
  • 1.Hubei Key Laboratory of Intelligent Vision Based Monitoring for Hydroelectric Engineering, China Three Gorges University, Yichang 443002, China
  • 2.Hubei Province Engineering Technology Research Center for Construction Quality Testing Equipment, China Three Gorges University, Yichang 443002, China
  • 3.College of Computer and Information Technology, China Three Gorges University, Yichang 443002, China
  • 4.Big Data Research Center, Jingchu University of Technology, Jingmen 448001, China
出版时间: 2025-11-05 doi: 10.3969/j.issn.1003-3106.2025.11.005
文章导航
收藏切换

提出一种基于YOLO11n的船舶红外图像目标检测算法AGT-YOLO,旨在解决模型精度和召回率偏低、小目标识别困难以及复杂海况下的多尺度识别问题。通过引入改进后的GhostHGNetv2网络,增强背景区分能力;设计ASF-P2优化颈部网络,以提升对低分辨率图像和极小目标的检测能力;研究任务自适应对齐检测头(Task Adaptive Alignment Detection Head,TAADH)替换原有检测头,提升定位和分类性能;同时,融入AFGCAttention注意力机制,提高全局信息处理能力和模型的泛化能力。实验结果表明,与基准模型YOLO11n相比,AGT-YOLO的召回率提高了4.4%,平均精度均值@50(mean Average Precision at IoU=0.5,mAP@50)提高了3.1%,并在复杂环境下展现出较强的多尺度识别能力和鲁棒性。

目标检测  /  YOLO11n  /  红外船舶检测  /  多尺度模型

A ship infrared image target detection algorithm based on YOLO11n, named AGT-YOLO, is proposed to address the issues of low model accuracy and recall rate, difficulties in identifying small targets, and multi-scale recognition challenges under complex sea conditions. By introducing an improved GhostHGNetv2 network, the background discrimination capability is enhanced; the designed ASFP2 optimized neck network improves detection capabilities for low-resolution images and very small targets; the proposed Tack Adaptive Alignment Detection Head ( TAADH ) replaces the original detection head, enhancing localization and classification performance;meanwhile, the AFGCAttention mechanism is integrated to improve global information processing capability and the model's generalization ability. Experimental results show that compared to the baseline model YOLO11n, AGT-YOLO achieves a 4.4% increase in recall rate and a 3.1% increase in mean average precision at IoU=0.5 ( mAP@ 50), demonstrating strong multi-scale recognition capability and robustness in complex environments.

object detection  /  YOLO11n  /  infrared ship inspection  /  multi-scale model
彭鹏, 陈慈发, 张上. 基于改进YOLO11n的船舶红外图像目标检测算法. 无线电工程, 2025 , 55 (11) : 2174 -2183 . DOI: 10.3969/j.issn.1003-3106.2025.11.005
Peng PENG, Cifa CHEN, Shang ZHANG. Target Detection Algorithm for Ship Infrared Images Based on Improved YOLO11n[J]. Radio Engineering, 2025 , 55 (11) : 2174 -2183 . DOI: 10.3969/j.issn.1003-3106.2025.11.005
随着我国经济高速增长与国际形势演变,海洋产业快速发展,海上作业规模持续扩大,推动舰船检测技术成为航海监管、渔业管理与国防安全的核心支撑技术[1]。该技术不仅影响海上救援效率,更对民生服务与军事战略实施具有深远意义[2]
红外探测技术凭借远作用距离与强抗干扰性,在船舶检测中占据重要地位,但小目标识别仍是技术瓶颈。远距离船舶因弱特征特性易受海杂波及背景噪声干扰,尤其在复杂海天线背景下,目标边缘模糊导致漏检误判风险加剧。此外,红外图像固有的低分辨率、高背景复杂度及多源噪声问题进一步制约检测精度提升[3]
针对上述问题,近年来众多学者开始探索深度学习在船舶红外图像目标检测中的应用。例如Li等[4]以YOLOv3为基础,通过数据增强、锚框优化及浅层特征融合,提高微弱信号检出率;Ye等[5]则在YOLOv5中新增高分辨率特征层,并嵌入TA(Target Attention)模块抑制复杂背景噪声,实现对小目标更精准定位。然而,Ye等[5]方法的注意力机制过度聚焦于小尺度区域,抑制大目标特征表达,导致大目标精度下降;且网络对场景光照、云层等环境变化敏感,小目标抗干扰能力不足。王海群等[6]在YOLOv8算法中引进双向信息流和可学习权重加强网络以提高特征表达能力,通过双线性插值统一不同尺度的池化结果,并进行拼接和归一化的方法,增强了网络的多尺度特征提取能力,但仍然面临着抗噪声干扰能力不足的问题。
YOLO11n是由Ultralytics团队于2024年9月30日发布的新一代目标检测和图像分割模型,相较于前代YOLOv8模型引入C3k2模块和C2PSA注意力机制,动态调整特征提取,增强小目标检测能力。颈部部分采用了改进的路径聚合网络-特征金字塔网络++(Path Aggregation Network-Feature Pyramid Network++, PAN-FPN++)结构,增强了特征的表达能力和层次感。采用自适应损失函数(VFLLoss+)和动态锚框机制分别提升了对目标方向和位置的识别精度。
本文基于YOLO11n基础模型进行改进,针对Ye等[5]研究中出现多尺度目标识别的问题,提出TAADH,通过任务对齐机制联合优化定位与分类分支,利用动态卷积核增强多尺度特征适配性。针对王海群等[6]研究中出现的复杂海况特征提取不足的问题,采用改进的GhostHGNetv2[7]网络,在降低参数量的同时增强背景区分能力,同时融合AFGCAttention[8]模块提高全局特征聚合效率。对于船舶红外图像检测小目标问题,Li等[4]也做了针对性的研究,但对于小目标的检测精度仍有提升空间,本文设计ASF-P2结构[9],结合自适应通道筛选策略,使小目标检测精度得以提升。因此,本文提出多尺度船舶红外检测算法AGT-YOLO(ASF-P2,GhostHGNet,TAADH)。
红外船舶检测因图像分辨率低、海面杂波干扰及船舶尺度跨度大,原始YOLO11特征提取不足、误检高,难以直接应用,为此提出AGT-YOLO。AGT-YOLO的网络结构如图1所示。
首先,用改进的GhostHGNetv2网络代替原有特征提取网络,解决复杂背景场景下红外目标的检测,提高模型检测的性能,并降低其参数量。其次,引入ASF-YOLO的思想,添加P2特征层,用于解决船舶红外图像小目标因背景干扰以及分辨率限制导致识别困难的问题。然后,运用本文提出的检测头TAADH,动态融合多尺度特征,强化对不同尺寸目标的感知。最后,采用AFGCAttention捕获全局上下文,提升整体模型一致性,使模型在新数据上保持鲁棒性与高泛化能力。
针对船舶红外图像中复杂背景干扰及目标特征弱化问题,本研究提出改进型GhostHGNetv2网络。该网络融合HGNetv2架构与Ghost卷积技术,通过Ghost_HGBlocks模块优化特征提取过程,增强复杂背景下目标与背景的区分能力。
图2所示,GhostHGNetv2基于HGNetv2的四阶段层级结构构建,其核心单元为HGBlock。网络输入端采用HGStem模块,通过深度可分离卷积(DWConv)实现可学习下采样层(LDS Layer)高效特征预处理。其中,LDS层摒弃传统固定池化操作,采用动态权重机制自适应调整下采样策略,以保留更多有效特征信息[10]。Stage 2~Stage 4均集成LDS层,结合Ghost HGBlock的双路径特征复用机制——主路径执行标准卷积,副路径生成轻量化特征图——在降低计算复杂度的同时提升特征表达能力。
针对红外图像背景噪声敏感及目标尺度多样性问题,网络在Stage 3重点优化P4/16特征层处理流程。通过堆叠Ghost_HGBlock模块进行多层次特征融合,结合通道注意力机制动态筛选关键特征,有效抑制海面杂波干扰并增强小目标边缘信息。此外,全网络采用DWConv技术替代标准卷积,进一步压缩模型参数量,兼顾检测精度与实时性需求。该设计通过层级化特征提取与动态自适应机制,显著提升了复杂海况下多尺度船舶目标的检测鲁棒性。
在GhostHGNetv2网络中,HGStem模块的结构见图2。该过程由以下公式表示:
式中:X表示输入的三通道图像数据,Y表示模块输出的特征图,分别表示卷积核大小为3、2、1的一般卷积操作,卷积步长分别为2、1、1;表示窗口大小为2×2,步幅为1的最大池化操作。
此外,另一个网络GhostNet在特征提取上表现优异,主要利用主、副通道。主通道捕捉主要图像信息,而副通道针对细节,卷积层较少,作为正则化机制防过拟合。副通道采用深度可分离卷积降低模型复杂度。最终,两通道输出结合形成特征表示。
Ghost卷积过程可以表示如下:
式中:表示一个深度卷积操作,其内核大小为5×5,步幅为1,组数等于通道数。
Ghost_HGBlock融合GhostNet和Ghost卷积的思想,将Ghost卷积模块整合到HGNetv2骨干网的HGBlock中。采用Ghost卷积模块替换HGBlock Light卷积模块,以实现更好的特征提取,同时也使模型更轻、参数更少、计算需求更少。GhostHG-Block模块的原理及其流程如图3所示。计算过程如下:
式中:G表示前文提到的Ghost卷积操作。
ASF-YOLO是Kang等[9]提出的YOLO框架,融合了尺度和空间特征,适用于快速精确的分割任务。其颈部设计包括尺度序列特征融合(Scale Sequence Feature Fusion,SSFF)模块、时间融合编码器(Temporal Fusion Encoder,TFE)模块和通道与位置注意机制(Channel and Position Attention Mechanism, CPAM)模块。该设计使模型能自适应关注不同尺度小目标的相关通道和空间位置。本文运用卷积和Zoom_cat实现TFE功能,设计Scalseq融合不同尺度特征,避免冗余计算,并引入asf_attention提取代表性特征信息。
船舶目标尺度跨度大,传统FPN虽融合多尺度特征,但仅靠简单拼接/求和,忽视层间深层关联,导致小目标漏检、大目标误检频发,检测精度与鲁棒性不足,亟需强化特征交互。通过设计Scalseq模块,将P2~P5同宽高比特征按尺度顺序排列,串联深层语义与浅层细节;随后用标准差递增的高斯核依次卷积,递进融合跨层信息[11-12],如下所示:
Fσ利用递增标准差σ的二维高斯卷积平滑特征图f生成多尺度表示,将其水平堆叠后,,通过三维卷积(3D-CNN)从尺度序列中提取特征(如图4所示)。基于能提供全面准确信息的P3级特征图设计Scalseq模块:首先调整最高级特征图的通道数和空间维度与P3匹配,然后扩展张量形状并沿深度维度连接,最后用3D卷积、批处理归一化和ReLU[13]完成特征提取。拼接的特征图构成尺度序列。为增强小目标识别能力,额外添加P2级Scalseq,提升模型多尺度特征提取能力。
asf_attention机制专为小目标检测设计,融合了通道注意力和位置注意力网络。通道注意力网络接收PANet后的特征图作为输入1,这些特征图包含卷积提取的详细信息。不同于SENet[14]中的全局平均池化和降维全连接层,提出的新机制采用一维卷积核,不降维地捕捉每个通道及其k个邻近通道的局部交互,提高通道注意力的预测效果。位置注意力网络接收通道注意力网络与Scalseq输出的叠加作为输入2,首先按宽度和高度分割输入特征图,在pw和ph轴上进行特征编码,然后合并生成输出关键位置信息,模块的整体结构如图5所示。
现有YOLO系列算法的解耦头采用分类与定位分支独立设计,虽简化了任务学习流程,却导致两大核心问题:其一,分类任务依赖语义特征,而定位任务需精准空间信息,二者特征共享易引发冲突,尤其在复杂海况下加剧误检与漏检;其二,固定尺度感受野难以适配船舶红外图像中的目标尺度剧烈变化特性(如近岸大船与远距离小艇并存场景),且多分支结构增加了模型存储与计算负担。这些局限性在红外船舶检测任务中尤为突出,需通过特征交互机制与归一化策略优化实现突破。
针对上述问题对检测头进行改进优化,结合任务对齐的单阶段目标检测器(Task-aligned Onestage Object Detection,TOOD)算法[15]的思想,设计TAADH模块,并提出一个改进型组归一化(Group Normalization,GN-Conv)卷积模块, GN-Conv以组归一化替代批量归一化层(Batch Normalization, BN)[16]。GN[17]创新性地通过通道分组统计量计算,有效缓解对批量规模约束,在动态计算图框架及极小批量(如抛处理规模(Batch Size)≤4)场景下仍保持稳定性能。实验表明,相较于BN,GN在模型精度、跨任务泛化能力等维度均展现出显著优势。
TADDH架构采用一种特征提取器,旨在集成来自多个卷积层的特征信息,从而学习并融合出表征任务交互关系的联合特征。在定位分支中,该联合特征被输入到DCNv2(可变形卷积网络V2)模块中,直接用于生成其执行特征对齐所需的偏移量和掩码参数。另一方面,分类分支则利用联合特征驱动一个动态特征选择机制,该机制由1×1卷积层和3×3卷积层组合构成(其具体结构如图6所示),用以筛选与分类任务最相关的特征。
鉴于定位与分类任务的目标差异,二者共享同一组特征时易引发特征干扰,各自所需的特征关注点也存在显著区别。为此,本文设计了一种任务解耦模块(结构如图7所示),基于跨层级任务交互特征动态推导特征权重系数,以促进定位与分类任务的特征解耦。其中,权重系数w由层级间交互特征计算生成,用于建模层级之间的关联特性。
文献[18]研究证实,GN可显著提升检测头在目标定位与分类任务中的性能表现。为进一步优化模型效率,本研究引入卷积权重共享机制,通过复用跨层卷积核减少冗余参数。针对多尺度目标检测的固有挑战,设计特征尺度自适应调节层(Scale层),通过可学习的缩放因子对各层级特征进行动态校准,以适配不同检测头对目标尺度的差异化需求。
借鉴TOOD思想,通过在检测头上实现定制任务对齐结构,避免传统独立分类和定位分支间的交互缺失。TAADH提取任务交互联合特征,DCNv2和交互特征生成offset/mask定位,动态选择分类,动态优化提升检测。
AFGCAttetion模块结构如图8所示,将传统的softmax注意力和线性注意力相结合,使其在计算复杂度上具有线性特性,降低模型的计算成本。通过代理向量AKV中聚合信息,然后将信息广播回Q,这种方式使得模型能够有效地处理全局信息,从而提升模型的整体性能。
该模块继承了softmax和线性注意力的优点。为了更好地利用位置信息,以最大限度地发挥其潜力,模块中有一个经过精心设计的代理偏差,以提高代理的注意力。具体地说,在注意力计算中引入代理偏差,是受相对位置偏置(Relative Position Bias, RPB)[19]的启发,即:
式中:B1∈Rn×N,B2∈RN×n表示原始代理偏差。在代理模型设计中,为提高参数利用效率,每个代理偏差均由3个正交的偏差分量(如系统偏差、随机偏差和模型结构偏差)复合生成。该方法避免直接将原始偏差项设为可训练变量,转而通过分量解耦与重组优化参数学习过程。由于线性注意力受到特征多样性不足的影响,该模块遵循并采用深度可分离卷积(Depthwise Convolution,DWC)模块来保持特征多样性。
因此,该模块由三部分组成,即纯代理注意、代理偏差和DWC模块。本文的模块可以公式化为:
式中:QKV∈RN×C,A∈RN×C,B1∈Rn×N,B2∈RN×n。在默认设置中,代理A通过池化获得,即A=合并(Q)。整体模块复杂度表示为:
式中:Nn为输入特征和代理的数量,k=3为DWC的内核大小,4NC2为预处理阶段,NC为获取代理阶段,2nNC+2NnC为运用代理注意力阶段,k2NC为DWC特征提取阶段。值得注意的是,该模型仅表现出N的线性复杂度。
该模型能够在保持较低计算复杂度的同时,提供较大感受野,从而提升模型的性能。结合代理注意力机制的这些优点,将其应用于船舶红外图像目标检测模型中,进一步提高模型对小目标和低分辨率图像的检测精度。由于代理注意力能够提供更大的感受野,可以帮助模型更好地理解目标与其上下文环境的关系,从而在复杂的海天背景下,更准确地检测和识别船舶目标,提升模型的鲁棒性。
实验在Windows 11操作系统,GPU为NVIDIA GeForce RTX 4070Ti Super显存16 GB,PyTorch1.2.0+CU-DA11.3深度学习框架以及Python3.10.5中实现,实验采用单卡GPU训练架构,批处理规模(batch size)设定为16,输入数据尺寸规范化为640 pixel×640 pixel[20]。模型共执行300次完整训练周期(epoch),参数优化基于随机梯度下降(Stochastic Gradient Descent,SGD)算法实现,初始学习率、动量(momentum)、权重衰减等参数均使用YOLO11n模型中的原始参数。
本研究构建于infiRay发布的海事船舶红外实测数据集,该数据集覆盖海上航道、近岸港口及沿海区域的多场景、多时段、多分辨率船舶红外图像。数据集分为7种船只类别,分别是邮轮、散货船、军舰、帆船、皮划艇、集装箱船、渔船,共计8326张。将数据集图片按照7:3的比例随机划分成训练集和验证集进行实验。
为验证AGT-YOLO算法在跨域场景下的泛化能力,本研究引入天津大学机器学习与数据挖掘团队构建的VisDrone2019基准数据集。该数据集涵盖多尺度目标分布特性(含8599张航拍图像),重点聚焦小目标检测难题。实验采用分层随机抽样策略(Stratified Random Sampling),将原始数据集依据样本类别分布划分为互斥子集,其中训练集占比70%,验证集占比30%,以消除数据分布偏差并验证模型在未知场景下的迁移学习效能。
实验选取目标识别任务中常见的评价指标:准确率(P)、召回率(R)、平均精度均值(mAP@0.5, mAP@ 0.5:0.95)、参数量(Para)和每秒帧数(Frames Per Second,FPS)。各项指标介绍如下。
(1)P表示检测目标准确率,量化模型排除负样本干扰的能力,计算公式为:
式中:TP为正确识别的正样本数量,FP为将负样本误判为正样本的错误率。
R表示检测目标召回率,衡量模型捕捉正样本的灵敏度,计算依赖漏检统计量,计算公式为:
式中:FN表示本应被检出但被模型忽略的正样本。
(2)mAP表示所有类别检出正确率的均值, mAP@0.5表示交并比(Intersection over Union,IoU)为0.5时各类别平均精度(Average Precision,AP)值的算术平均,侧重宽松定位场景的评估;mAP@0.5:0.95通过多阈值积分(步长0.05),考核模型在高精度定位要求(IoU≥0.5)下的综合性能,避免单一阈值评价的偏差,计算公式为:
式中:APi表示第i个类别的AP,c表示数据集中的标签类别数。
(3)Para表示衡量模型的规模和空间复杂度的指标。
(4)FPS反映模型训练速度的刷新率,计算公式为:
式中:preprocess表示预处理时间,inference表示推理时间,postprocess表示后处理时间。
为验证算法AGT-YOLO各个模块的有效性,在infiRay的船舶红外图像数据集上进行消融实验,其中“√”表示使用该方法。以原始YOLO11n算法为基础模型,依次引入GhostHGNetv2、TAADH、ASFP2、AFGCAttention模块。实验结果如表1所示。
表1可以看出,使用改进后的GhostHG-Netv2模块替换YOLO11n的主干特征提取网络,准确率上升了1.5个百分点,提高了模型总体识别目标的准确性,面对多种类别目标的样本,检测精度都有所提高,验证了引入该模块能够提高区分背景和待检测目标差异的能力。在加入GhostHGNetv2方法的基础上对检测头进行定制化任务对齐设计后,mAP@0.5:0.95增加了1.2个百分点,验证了TAADH在锚框选择上精确度更高。通过引入ASF-P2特征融合机制优化YOLO11的颈部网络,实验表明该结构显著提升了多尺度特征融合效率,最终使模型在IoU阈值为0.5时的mAP提升至93.2%。相较于基线模型的对比方法,mAP指标获得1.6个百分点的增益,充分验证了ASF-P2在增强目标检测精度方面的有效性。引入AFGCAttention注意力模块后,准确率、召回率和mAP都有小幅度的提升,验证了模型能够有效处理全局信息,以提升整体性能。
为评估AGT-YOLO算法在目标检测任务上的性能,开展了一系列对比实验,选取9种公认目标检测架构:SSD、CenterNet、YOLOv5n、YOLOv6、YOLOv7-tiny、YOLOv8n、YOLOv9t、YOLOv10n、YOLO11n,这几种经典目标检测算法模型作为比较基准。通过这些对比实验,能够更全面地评估模型的优势和效果。基于多模型对照实验的定量分析结果(如表2所示),系统量化了本研究所提架构在检测精度、推理效率及泛化能力维度的综合性能优。
表2可知,改进算法AGT-YOLO较基线模型YOLO11n, mAP@ 0.5提升3.1%(90.3%→93.4%),mAP@ 0.5:0.95提升1.3%(63.4%→64.7%),印证了多尺度特征融合机制对定位精度的增强效果。在FPS指标上,也能够实现更高的帧率,从而提供更流畅的实时检测体验,满足海上救援等要求快速响应的需求。在参数量上,AGT-YOLO算法也未劣化。总体而言,AGT-YOLO能够解决船舶红外目标检测方法存在的模型精度和召回率较低、小目标因背景干扰以及受分辨率限制导致识别困难、复杂海况下特征提取困难等问题。
对YOLO11n与算法AGT-YOLO在公共数据集VisDrone2019上做泛化实验,结果如表3所示。AGT-YOLO与YOLO11n相比,mAP@0.5从34.7%增加到40.5%,提升了5.8个百分点;mAP@0.5:0.95从19.9%提升为23.8%,提高了3.9个百分点;P从45.6%提升到51.0%,提高了5.4个百分点;R从35.6%提升到39.7%,提高了4.1个百分点。以上数据表明,AGT-YOLO算法不仅能提高复杂场景中不同分辨率下多尺度目标检测率,还能提高日常情况小目标的识别率,说明算法具有较强的通用性。
采用GradCAMPlusPlus[21]技术生成目标检测显著性图。算法以高阶导数重标梯度贡献,将模型聚焦区域映射为色彩梯度:深红对应核心判别特征,橙黄次之,冷蓝则指示对结果影响微弱的冗余信息,直观呈现网络关注焦点。图9展示了在infiRay船舶红外图像数据集上,YOLO11n算法和AGT-YOLO算法在多尺度(图9(a))、遮掩(图9(b))、海天线(图9(c))、小目标(图9(d))等场景下的检测效果热力图,该结果加了边界框和置信度,便于分析[22]
在多尺度目标检测中,基线算法并未完全检测出小尺寸目标,而AGT-YOLO算法能够比较准确地检测出各个尺度的目标。在出现遮掩场景的情况下,基线算法并未检测出靠近被遮掩的目标,而AGT-YOLO算法能准确地识别出目标的重叠部分。在海天线干扰的背景下,基线算法对小目标的检测出现了漏检的现象,而AGT-YOLO算法能够准确地识别出这些小目标。在针对小目标检测下,基线算法出现了大量漏检还有极个别错检现象,而AGT-YOLO算法出现漏检的概率极低,没有错检现象的出现。对比热力图分析可知,AGT-YOLO算法检测效果更好,能够适应复杂场景下多尺度、多目标的船舶红外图像检测。
针对船舶红外图像目标检测精度低、召回率低、小目标识别难及多尺度识别问题,提出的基于YOLO11n的AGT-YOLO改进算法,采用增强GhostHGNetv2网络作为主干特征提取网络,提升背景区分能力。同时,设计ASF-P2优化颈部网络,自研TAADH,并融入AFGCAttention机制,以提高特征提取、小目标检测、定位和分类性能,以及多尺度目标识别和模型泛化能力。实验结果显示,与YOLO11n相比,AGT-YOLO的召回率提高了4.4%,平均精度mAP@50提高了3.1%,且在泛化实验中,mAP@0.5、mAP@0.5:0.95分别提升了5.8、3.9个百分点。总之,AGT-YOLO算法在船舶红外目标检测上表现优异,证明了本文算法的可行性。
参考文献 引证文献
排序方式:
[1]
王斌.海面目标红外检测方法研究[D].大连:大连海事大学,2018.
[2]
古静.基于深度学习的红外船舶目标检测方法研究[D].镇江:江苏科技大学,2023.
[3]
王勇,王柏容,慕东东.红外船舶检测的研究现状及展望[J].大连海事大学学报,2023,49(4):103-115.
[4]
LI H,DENG L B,YANG C,et al. Enhanced YOLO v3 Tiny Network for Real-time Ship Detection from Visual Image[J]. IEEE Access,2021,9:16692-16706.
[5]
YE J,YUAN Z Y,QIAN C,et al. CAA-YOLO:Combined-attention-augmented YOLO for Infrared Ocean Ships Detection[J]. Sensors,2022,22(10):3782.
[6]
王海群,魏培旭,解浩龙,.基于改进YOLOv8的红外船舶检测[J].电光与控制,2025,32(1):61-67.
[7]
DING Q X,LI W C,XU C C,et al. GMS-YOLO:An Algorithm for Multi-scale Object Detection in Complex Environments in Confined Compartments[J]. Sensors,2024, 24(17):5789.
[8]
HAN D,YE T,HAN Y,et al. Agent Attention:On the Integration of Softmax and Linear Attention[EB/OL]. (2023-12-14)[2024-11-26]. https://arxiv.org/abs/2312.08874.
[9]
KANG M,TING C M,TING F F,et al. ASF-YOLO:A Novel YOLO Model with Attentional Scale Sequence Fusion for Cell Instance Segmentation[EB/OL]. (2023-12-11)[2024-11-26]. https://arxiv.org/abs/2312.06458.
[10]
YANGX,ZHANG Y F,LI T,et al. Image Super-resolution Based on the Down-sampling Iterative Module and Deep CNN[J]. Circuits,Systems,and Signal Processing,2021, 40:3437-3455.
[11]
PARKH J,KANG J W,KIM B G. ssFPN:Scale Sequence (S2) Feature-based Feature Pyramid Network for Object Detection[J]. Sensors,2023,23(9):4432.
[12]
TRAN D,BOURDEV L,FERGUS R,et al. Learning Spatiotemporal Features with 3D Convolutional Networks[C]//2015 IEEE International Conference on Computer Vision (ICCV). Santiago:IEEE,2015:4489-4497.
[13]
ELFWING S, UCHIBE E, DOYA K. Sigmoid-weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning[J]. Neural Networks,2018, 107:3-11.
[14]
JIN X,XIE Y P,WEI X S,et al. Delving Deep into Spatial Pooling for Squeeze-and-Excitation Networks[J]. Pattern Recognition,2022,121:108159.
[15]
FENGC J,ZHONG Y J,GAO Y,et al. Tood:Task-aligned One-stage Object Detection[C]//2021 IEEE/CVF International Conference on Computer Vision (ICCV). Montreal:IEEE,2021:3490-3499.
[16]
IOFFES,SZEGEDY C. Batch Normalization:Accelerating Deep Network Training by Reducing Internal Covariate Shift[EB/OL]. (2015-02-11)[2024-11-26].https://arxiv.org/abs/1502.03167.
[17]
WU Y X,HE K M.Group Normalization[C]//Proceedings of the European Conference on Computer Vision (ECCV). Munich:Springer,2018:3-19.
[18]
TIAN Z,SHEN C H,CHEN H,et al. FCOS:A Simple and Strong Anchor-free Object Detector[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2020, 44(4):1922-1933.
[19]
SHAW P,USZKOREIT J,VASWANI A. Self-attention with Relative Position Representations[EB/OL]. (2018-03-06)[2024-11-26]. https://arxiv.org/abs/1803.02155.
[20]
杨慧聪,陈慈发,张上.基于通道剪枝的YOLOv8n印刷电路板缺陷检测[J].无线电工程,2025,55(2):271-280.
[21]
CHATTOPADHAY A, SARKAR A, HOWLADER P,et al.Grad-CAM++:Generalized Gradient-based Visual Explanations for Deep Convolutional Networks[C]//2018 IEEE Winter Conference on Applications of Computer Vision (WACV). Lake Tahoe:IEEE,2018:839-847.
[22]
贾星宇,李大鹏.基于YOLO11的无人机航拍图像小目标检测算法[J].无线电工程,2025,55(8):1560-1570.
2025年第55卷第11期
PDF下载
96
44
引用本文
BibTeX
文章信息
doi: 10.3969/j.issn.1003-3106.2025.11.005
  • 接收时间:2025-07-17
  • 首发时间:2026-04-17
  • 出版时间:2025-11-05
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2025-07-17
基金
作者信息
    1.三峡大学 水电工程智能视觉监测湖北省重点实验室,湖北 宜昌 443002
    2.三峡大学 湖北省建筑质量检测装备工程技术研究中心,湖北 宜昌 443002
    3.三峡大学 计算机与信息学院,湖北 宜昌 443002
    4.荆楚理工学院 大数据研究中心,湖北 荆门 448001

通讯作者:

陈慈发 男,(1967—),教授。主要研究方向:嵌入式系统、物联网、计算机测控系统。
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/wxdgc/CN/10.3969/j.issn.1003-3106.2025.11.005
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