Article(id=1148106706472661679, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106697601704181, articleNumber=1003-3033(2025)01-0075-09, orderNo=null, doi=10.16265/j.cnki.issn1003-3033.2025.01.0127, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1724083200000, receivedDateStr=2024-08-20, revisedDate=1729785600000, revisedDateStr=2024-10-25, acceptedDate=null, acceptedDateStr=null, onlineDate=1751659569614, onlineDateStr=2025-07-05, pubDate=1737993600000, pubDateStr=2025-01-28, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1751659569614, onlineIssueDateStr=2025-07-05, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1751659569614, creator=13701087609, updateTime=1751659569614, updator=13701087609, issue=Issue{id=1148106697601704181, tenantId=1146029695717560320, journalId=1146031787341344770, year='2025', volume='35', issue='1', pageStart='1', pageEnd='252', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1751659567499, creator=13701087609, updateTime=1757401533944, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1172190250475573883, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106697601704181, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1172190250475573884, tenantId=1146029695717560320, journalId=1146031787341344770, issueId=1148106697601704181, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=75, endPage=83, ext={EN=ArticleExt(id=1149757693193273976, articleId=1148106706472661679, tenantId=1146029695717560320, journalId=1146031787341344770, language=EN, title=A lightweight forest fire detection algorithm based on YOLOv5s, columnId=1149733269173878863, journalTitle=China Safety Science Journal, columnName=Safety engineering technology, runingTitle=null, highlight=null, articleAbstract=

In order to solve the problems of complex structure, large scale and difficulty in balancing detection accuracy and efficiency of the current forest fire detection algorithm based on deep learning, a lightweight forest fire detection algorithm based on YOLOv5s was proposed. Firstly, an optimized background difference technique was used to eliminate the interference of fire-like objects in the background image, thus reducing the time required for image analysis. Secondly, a group blending strategy was designed to optimize the conventional convolution, and an efficient channel attention (ECA) mechanism and depthwise separable convolution were incorporated into the C3 module of feature extraction, which enhanced the ability of image feature extraction and fusion and at the same time effectively reduces the number of model parameters. Then, a dynamic non-monotonic focusing mechanism was used to optimize the WIOU loss function, reducing the harmful gradients generated by low-quality samples. Finally, sufficient experimental comparisons between the proposed algorithm and other algorithms on the constructed forest fire dataset. The results show that the proposed algorithm shows good generalization in various scenarios, and the detection accuracy of the flame target can reach 86.1%, which is 2.7% higher than that of the standard YOLOv5s, and the detection speed is increased by 11.4%, which effectively reduces the fire false alarm rate and enhances the detection performance of the model.

, correspAuthors=Huazhang WEI, 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=Huilin LIU, Qiong FANG, Yu JIANG, Huazhang WEI, Tao WANG, Shuchuan ZHANG), CN=ArticleExt(id=1148106716354441835, articleId=1148106706472661679, tenantId=1146029695717560320, journalId=1146031787341344770, language=CN, title=基于YOLOv5s的轻量化森林火灾探测算法, columnId=1149733269727526997, journalTitle=中国安全科学学报, columnName=安全工程技术, runingTitle=null, highlight=null, articleAbstract=

为解决当前基于深度学习的森林火灾探测算法存在结构复杂、规模庞大,且难以兼顾检测精度和效率的问题,提出一种基于YOLOv5s的轻量化森林火灾探测算法。首先,采用优化的背景差分技术消除背景图像中类火物体的干扰,减少分析图像所需的时间;其次,设计分组混洗策略优化常规卷积,并在特征提取的C3模块中融入高效通道注意力(ECA)机制和深度可分离卷积,增强图像特征提取与融合能力的同时有效降低模型的参数量;然后,采用动态非单调聚焦机制优化Wise-交并比(WIOU)损失函数,减少低质量样本产生的有害梯度;最后,在构建的森林火灾数据集上将所提算法与其他算法做充分的试验对比。结果表明: 所提算法在各类场景均展现出良好的泛化性,对火焰目标的检测精度达到86.1%,较标准YOLOv5s检测精度提升2.7%,检测速度提升11.4%,有效降低了火灾误报率,增强了模型的检测性能。

, correspAuthors=魏华章, authorNote=null, correspAuthorsNote=
**魏华章(1996—),男,安徽淮南人,硕士研究生,研究方向为计算机视觉。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=gplUb9Dw/tHrG6xQtjMdNA==, magXml=oaW6C51jhiP9jtic7i8hXQ==, pdfUrl=null, pdf=49PUfReZ2YkFzAarzF3E8g==, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=DwzsIs1c+lVWKHLs2pPzKg==, mapNumber=null, authorCompany=null, fund=null, authors=

刘惠临 (1979—),男,安徽合肥人,博士,高级实验师,主要从事计算机视觉、图像处理、多模态数据融合等方面的研究。E-mail:

王涛 副教授

张树川 副教授

, authorsList=刘惠临, 方琼, 江宇, 魏华章, 王涛, 张树川)}, authors=[Author(id=1165721959108194405, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=liuhl@aust.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165721959171108967, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959108194405, language=EN, stringName=Huilin LIU, firstName=Huilin, middleName=null, lastName=LIU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721959221440616, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959108194405, language=CN, stringName=刘惠临, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001, bio={"img":"5JnyzViF/716wSJwPMTh7Q==","content":"

刘惠临 (1979—),男,安徽合肥人,博士,高级实验师,主要从事计算机视觉、图像处理、多模态数据融合等方面的研究。E-mail:

"}, bioImg=5JnyzViF/716wSJwPMTh7Q==, bioContent=

刘惠临 (1979—),男,安徽合肥人,博士,高级实验师,主要从事计算机视觉、图像处理、多模态数据融合等方面的研究。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721958864924760, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=1, ext=[AuthorCompanyExt(id=1165721958869119065, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958877507674, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001)])]), Author(id=1165721959301132394, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165721959410184300, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959301132394, language=EN, stringName=Qiong FANG, firstName=Qiong, middleName=null, lastName=FANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721959456321645, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959301132394, language=CN, stringName=方琼, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721958864924760, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=1, ext=[AuthorCompanyExt(id=1165721958869119065, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958877507674, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001)])]), Author(id=1165721959502458991, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, 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=1165721959556984945, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959502458991, language=EN, stringName=Yu JIANG, firstName=Yu, middleName=null, lastName=JIANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721959615705202, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959502458991, language=CN, stringName=江宇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721958864924760, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=1, ext=[AuthorCompanyExt(id=1165721958869119065, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958877507674, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001)])]), Author(id=1165721959661842548, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=toprosewhz@163.com, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1165721959720562806, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959661842548, language=EN, stringName=Huazhang WEI, firstName=Huazhang, middleName=null, lastName=WEI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, **, address=2 School of Intelligence and Electrical Engineering, Huainan Vocational Technical College, Huainan Anhui 232001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721959770894455, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959661842548, language=CN, stringName=魏华章, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, **, address=2 淮南职业技术学院 智能与电气工程学院,安徽 淮南 232001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721958919450715, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=2, ext=[AuthorCompanyExt(id=1165721958923645020, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958919450715, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Intelligence and Electrical Engineering, Huainan Vocational Technical College, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958932033629, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958919450715, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 淮南职业技术学院 智能与电气工程学院,安徽 淮南 232001)])]), Author(id=1165721959821226105, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1165721959879946363, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959821226105, language=EN, stringName=Tao WANG, firstName=Tao, middleName=null, lastName=WANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3 Key Laboratory of Unmanned Emergency Equipment and Digital Reconstruction of Disaster Processes in Anhui Province, Chuzhou College, Chuzhou Anhui 239099, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721959930278012, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959821226105, language=CN, stringName=王涛, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3 滁州学院 无人应急装备与灾害过程数字化重建安徽省联合共建学科重点实验室,安徽 滁州 239099, bio={"content":"

王涛 副教授

"}, bioImg=null, bioContent=

王涛 副教授

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721958978170974, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=3, ext=[AuthorCompanyExt(id=1165721958982365279, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958978170974, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Key Laboratory of Unmanned Emergency Equipment and Digital Reconstruction of Disaster Processes in Anhui Province, Chuzhou College, Chuzhou Anhui 239099, China), AuthorCompanyExt(id=1165721958990753888, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958978170974, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 滁州学院 无人应急装备与灾害过程数字化重建安徽省联合共建学科重点实验室,安徽 滁州 239099)])]), Author(id=1165721959980609662, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, orderNo=5, 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=1165721960051912832, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959980609662, language=EN, stringName=Shuchuan ZHANG, firstName=Shuchuan, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=4, address=4 School of Safety Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1165721960106438785, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, authorId=1165721959980609662, language=CN, stringName=张树川, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=4, address=4 安徽理工大学 安全科学与工程学院,安徽 淮南 232001, bio={"content":"

张树川 副教授

"}, bioImg=null, bioContent=

张树川 副教授

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1165721959036891233, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=4, ext=[AuthorCompanyExt(id=1165721959045279842, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721959036891233, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4 School of Safety Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721959049474147, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721959036891233, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4 安徽理工大学 安全科学与工程学院,安徽 淮南 232001)])])], keywords=[Keyword(id=1165721960223879298, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=1, keyword=YOLOv5s), Keyword(id=1165721960307765379, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=2, keyword=lightweighting), Keyword(id=1165721960379068548, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=3, keyword=forest fire detection), Keyword(id=1165721960429400197, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=4, keyword=depthwise separable convolution), Keyword(id=1165721960483926150, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=5, keyword=attention), Keyword(id=1165721960538452103, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, orderNo=6, keyword=wise intersection over union(WIOU)), Keyword(id=1165721960597172360, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=1, keyword=YOLOv5s), Keyword(id=1165721960647504009, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=2, keyword=轻量化), Keyword(id=1165721960697835658, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=3, keyword=森林火灾探测), Keyword(id=1165721960743973003, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=4, keyword=深度可分离卷积), Keyword(id=1165721960794304652, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=5, keyword=注意力), Keyword(id=1165721960848830605, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, orderNo=6, keyword=Wise-交并比(WIOU))], refs=[Reference(id=1165721962950176947, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2022, volume=9, issue=2, pageStart=793, pageEnd=804, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=CHOI M, KIM C, OH H, journalName=Journal of Computational Design and Engineering, refType=null, unstructuredReference=CHOI M, KIM C, OH H. A video-based SlowFastMTB model for detection of small amounts of smoke from incipient forest fires[J]. Journal of Computational Design and Engineering, 2022, 9(2):793-804., articleTitle=A video-based SlowFastMTB model for detection of small amounts of smoke from incipient forest fires, refAbstract=null), Reference(id=1165721963013091508, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2020, volume=39, issue=1, pageStart=143, pageEnd=145, url=null, language=null, rfNumber=[2], rfOrder=1, authorNames=熊昊, 李伟, journalName=传感器与微系统, refType=null, unstructuredReference=熊昊, 李伟. 基于SVM的视频火焰检测算法[J]. 传感器与微系统, 2020, 39(1):143-145, 149., articleTitle=基于SVM的视频火焰检测算法, refAbstract=null), Reference(id=1165721963067617462, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2020, volume=39, issue=1, pageStart=143, pageEnd=145, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=XIONG Hao, LI Wei, journalName=Sensors and Microsystems, refType=null, unstructuredReference=XIONG Hao, LI Wei. Video flame detection algorithm based on svm[J]. Sensors and Microsystems, 2020, 39(1):143-145,149., articleTitle=Video flame detection algorithm based on svm, refAbstract=null), Reference(id=1165721963126337720, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2022, volume=24, issue=1, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=GUAN Zhenwei, MIN Feng, HE Wei, journalName=Entropy, refType=null, unstructuredReference=GUAN Zhenwei, MIN Feng, HE Wei, et al. Forest fire detection via feature entropy guided neural network[J]. Entropy, 2022, 24(1):DOI:10.3390/e24010128., articleTitle=Forest fire detection via feature entropy guided neural network, refAbstract=null), Reference(id=1165721963176669369, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=10, pageStart=147, pageEnd=159, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=李海, 孙鹏, journalName=中国安全科学学报, refType=null, unstructuredReference=李海, 孙鹏. 基于知识图谱的火灾图像识别研究[J]. 中国安全科学学报, 2023, 33(10):147-159., articleTitle=基于知识图谱的火灾图像识别研究, refAbstract=null), Reference(id=1165721963247972538, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=10, pageStart=147, pageEnd=159, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=LI Hai, SUN Peng, journalName=China Safety Science Journal, refType=null, unstructuredReference=LI Hai, SUN Peng. Research on fire image recognition based on scientific knowledge graph[J]. China Safety Science Journal, 2023, 33(10):147-159., articleTitle=Research on fire image recognition based on scientific knowledge graph, refAbstract=null), Reference(id=1165721963306692795, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2017, volume=39, issue=6, pageStart=1137, pageEnd=1149, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=REN Shaoqing, HE Kaiming, GIRSHICK R, journalName=IEEE Transactions on Pattern Analysis and Machine Intelligence, refType=null, unstructuredReference=REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6):1137-1149., articleTitle=Faster R-CNN: towards real-time object detection with region proposal networks, refAbstract=null), Reference(id=1165721963369607356, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2020, volume=53, issue=9, pageStart=976, pageEnd=983, url=null, language=null, rfNumber=[6], rfOrder=7, authorNames=张为, 魏晶晶, journalName=天津大学学报, refType=null, unstructuredReference=张为, 魏晶晶. 嵌入DenseNet结构和空洞卷积模块的改进YOLOv3火灾检测算法[J]. 天津大学学报, 2020, 53(9):976-983., articleTitle=嵌入DenseNet结构和空洞卷积模块的改进YOLOv3火灾检测算法, refAbstract=null), Reference(id=1165721963424133309, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2020, volume=53, issue=9, pageStart=976, pageEnd=983, url=null, language=null, rfNumber=[6], rfOrder=8, authorNames=ZHANG Wei, WEI Jingjing, journalName=Journal of Tianjin University, refType=null, unstructuredReference=ZHANG Wei, WEI Jingjing. Improved YOLOv3 fire detection-algorithm with embedded DenseNet structure and cavity convolution module[J]. Journal of Tianjin University, 2020, 53(9):976-983., articleTitle=Improved YOLOv3 fire detection-algorithm with embedded DenseNet structure and cavity convolution module, refAbstract=null), Reference(id=1165721963474464958, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=21, pageEnd=37, url=null, language=null, rfNumber=[7], rfOrder=9, authorNames=LIU Wei, ANGUELOV D, ERHAN D, journalName=European Conference on Computer Vision (ECCV), refType=null, unstructuredReference=LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]. European Conference on Computer Vision (ECCV), 2016:21-37., articleTitle=SSD: single shot multibox detector, refAbstract=null), Reference(id=1165721963537379519, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=183, pageEnd=193, url=null, language=null, rfNumber=[8], rfOrder=10, authorNames=SHARMA J, GRANMO O, GOODWIN M, journalName=Engineering Applications of Neural Networks:18th International Conference, refType=null, unstructuredReference=SHARMA J, GRANMO O, GOODWIN M, et al. Deep convolutional neural networks for fire detection in images[C]. Engineering Applications of Neural Networks:18th International Conference, 2017:183-193., articleTitle=Deep convolutional neural networks for fire detection in images, refAbstract=null), Reference(id=1165721963596099779, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=61, issue=11, pageStart=106, pageEnd=110, url=null, language=null, rfNumber=[9], rfOrder=11, authorNames=邱思远, 王孝兰, 张伟伟, journalName=农业装备与车辆工程, refType=null, unstructuredReference=邱思远, 王孝兰, 张伟伟, 等. 基于YOLOX的森林火灾早期预警研究[J]. 农业装备与车辆工程, 2023, 61(11):106-110., articleTitle=基于YOLOX的森林火灾早期预警研究, refAbstract=null), Reference(id=1165721963659014342, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=61, issue=11, pageStart=106, pageEnd=110, url=null, language=null, rfNumber=[9], rfOrder=12, authorNames=QIU Siyuan, WANG Xiaolan, ZHANG Weiwei, journalName=Agricultural Equipment and Vehicle Engineering, refType=null, unstructuredReference=QIU Siyuan, WANG Xiaolan, ZHANG Weiwei, et al. Research on early of forest fire based on YOLOX[J]. Agricultural Equipment and Vehicle Engineering, 2023, 61(11):106-110., articleTitle=Research on early of forest fire based on YOLOX, refAbstract=null), Reference(id=1165721963721928903, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=6, pageStart=152, pageEnd=158, url=null, language=null, rfNumber=[10], rfOrder=13, authorNames=王喆, 李享, 杨栋梁, journalName=中国安全科学学报, refType=null, unstructuredReference=王喆, 李享, 杨栋梁, 等. 基于YOLOv5s的城镇森林交界域火灾探测模型[J]. 中国安全科学学报, 2023, 33(6):152-158., articleTitle=基于YOLOv5s的城镇森林交界域火灾探测模型, refAbstract=null), Reference(id=1165721963784843467, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=33, issue=6, pageStart=152, pageEnd=158, url=null, language=null, rfNumber=[10], rfOrder=14, authorNames=WANG Zhe, LI Xiang, YANG Dongliang, journalName=China Safety Science Journal, refType=null, unstructuredReference=WANG Zhe, LI Xiang, YANG Dongliang, et al. Fire detection model of wildland-urban interface based on YOLOv5s[J]. China Safety Science Journal, 2023, 33(6):152-158., articleTitle=Fire detection model of wildland-urban interface based on YOLOv5s, refAbstract=null), Reference(id=1165721963851952334, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=61, pageEnd=null, url=null, language=null, rfNumber=[11], rfOrder=15, authorNames=HOSSEINI A, HASHEMZADEH M, FARAJZADEH N, journalName=Journal of Computational Science, refType=null, unstructuredReference=HOSSEINI A, HASHEMZADEH M, FARAJZADEH N. UFS-Net: a unified flame and smoke detection method for early detection of fire in video surveillance applications using cnns[J]. Journal of Computational Science, 2022, 61: DOI:10.1016/j.jocs.2022.101638., articleTitle=UFS-Net: a unified flame and smoke detection method for early detection of fire in video surveillance applications using cnns, refAbstract=null), Reference(id=1165721963906478288, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=8, issue=6, pageStart=145, pageEnd=153, url=null, language=null, rfNumber=[12], rfOrder=16, authorNames=高源, 杨望, 曹磊, journalName=林业工程学报, refType=null, unstructuredReference=高源, 杨望, 曹磊, 等. 基于改进YOLOv5的特征抑制多尺度林火检测算法[J]. 林业工程学报, 2023, 8(6):145-153., articleTitle=基于改进YOLOv5的特征抑制多尺度林火检测算法, refAbstract=null), Reference(id=1165721963956809938, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=8, issue=6, pageStart=145, pageEnd=153, url=null, language=null, rfNumber=[12], rfOrder=17, authorNames=GAO Yuan, YANG Wang, CAO Lei, journalName=Journal of Forestry Engineering, refType=null, unstructuredReference=GAO Yuan, YANG Wang, CAO Lei, et al. Feature suppression multi-scale forest fire detection algorithm based on improved YOLOv5[J]. Journal of Forestry Engineering, 2023, 8(6):145-153., articleTitle=Feature suppression multi-scale forest fire detection algorithm based on improved YOLOv5, refAbstract=null), Reference(id=1165721964040696021, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=2, pageStart=397, pageEnd=405, url=null, language=null, rfNumber=[13], rfOrder=18, authorNames=张全, 张伟, 杨先凤, journalName=安全与环境学报, refType=null, unstructuredReference=张全, 张伟, 杨先凤, 等. 融合YOLOv5-ResNet级联网络的烟火检测方法[J]. 安全与环境学报, 2023, 23(2):397-405., articleTitle=融合YOLOv5-ResNet级联网络的烟火检测方法, refAbstract=null), Reference(id=1165721964091027671, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=2, pageStart=397, pageEnd=405, url=null, language=null, rfNumber=[13], rfOrder=19, authorNames=ZHANG Quan, ZHANG Wei, YANG Xianfeng, journalName=Journal of Safety and Environment, refType=null, unstructuredReference=ZHANG Quan, ZHANG Wei, YANG Xianfeng, et al. Smoke and fire detection method fusing YOLOv5-ResNet cascade network[J]. Journal of Safety and Environment, 2023, 23(2):397-405., articleTitle=Smoke and fire detection method fusing YOLOv5-ResNet cascade network, refAbstract=null), Reference(id=1165721964145553625, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=12, issue=16, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[14], rfOrder=20, authorNames=LIU Huilin, WEI Huazhang, YANG Gaoming, journalName=Electronics, refType=null, unstructuredReference=LIU Huilin, WEI Huazhang, YANG Gaoming, et al. An improved vibe algorithm based on adaptive thresholding and the deep learning-driven frame difference method[J]. Electronics, 2023, 12(16):DOI:10.3390/electronics12163481., articleTitle=An improved vibe algorithm based on adaptive thresholding and the deep learning-driven frame difference method, refAbstract=null), Reference(id=1165721964208468187, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=11 531, pageEnd=11 539, url=null, language=null, rfNumber=[15], rfOrder=21, authorNames=WANG Qilong, WU Banggu, ZHU Pengfei, journalName=Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, refType=null, unstructuredReference=WANG Qilong, WU Banggu, ZHU Pengfei, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020:11 531-11 539., articleTitle=ECA-Net: efficient channel attention for deep convolutional neural networks, refAbstract=null), Reference(id=1165721964262994141, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2024, volume=185, issue=null, pageStart=101, pageEnd=107, url=null, language=null, rfNumber=[16], rfOrder=22, authorNames=CHO Y, journalName=Pattern Recognition Letters, refType=null, unstructuredReference=CHO Y. Weighted intersection over union (WIOU) for evaluating image segmentation[J]. Pattern Recognition Letters, 2024, 185: 101-107., articleTitle=Weighted intersection over union (WIOU) for evaluating image segmentation, refAbstract=null), Reference(id=1165721964317520096, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=287, pageEnd=296, url=null, language=null, rfNumber=[17], rfOrder=23, authorNames=REDDY E, THALE S, journalName=Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics, refType=null, unstructuredReference=REDDY E, THALE S. Pedestrian detection using YOLOv5 and complete-IoU loss for autonomous driving applications[C]. Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics, 2023:287-296., articleTitle=Pedestrian detection using YOLOv5 and complete-IoU loss for autonomous driving applications, refAbstract=null), Reference(id=1165721964384628962, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=53, issue=10, pageStart=2303, pageEnd=2310, url=null, language=null, rfNumber=[18], rfOrder=24, authorNames=章曙光, 唐锐, 邵政瑞, journalName=无线电工程, refType=null, unstructuredReference=章曙光, 唐锐, 邵政瑞, 等. 一种基于小波变换的YOLOv5火灾检测改进算法[J]. 无线电工程, 2023, 53(10):2303-2310., articleTitle=一种基于小波变换的YOLOv5火灾检测改进算法, refAbstract=null), Reference(id=1165721964439154916, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, doi=null, pmid=null, pmcid=null, year=2023, volume=53, issue=10, pageStart=2303, pageEnd=2310, url=null, language=null, rfNumber=[18], rfOrder=25, authorNames=ZHANG Shuguang, TANG Rui, SHAO Zhengrui, journalName=Radio Engineering, refType=null, unstructuredReference=ZHANG Shuguang, TANG Rui, SHAO Zhengrui, et al. A wavelet transform-based improved algorithm for YOLOv5 fire detection[J]. Radio Engineering, 2023, 53(10):2303-2310., articleTitle=A wavelet transform-based improved algorithm for YOLOv5 fire detection, refAbstract=null)], funds=[Fund(id=1165721962769821868, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, awardId=2023g07020007, language=CN, fundingSource=安徽省重点研究与开发计划项目(2023g07020007), fundOrder=null, country=null), Fund(id=1165721962824347823, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, awardId=2024cx2111, language=CN, fundingSource=安徽理工大学研究生创新基金资助(2024cx2111), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1165721958864924760, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=1, ext=[AuthorCompanyExt(id=1165721958869119065, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958877507674, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958864924760, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001)]), AuthorCompany(id=1165721958919450715, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=2, ext=[AuthorCompanyExt(id=1165721958923645020, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958919450715, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Intelligence and Electrical Engineering, Huainan Vocational Technical College, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721958932033629, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958919450715, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 淮南职业技术学院 智能与电气工程学院,安徽 淮南 232001)]), AuthorCompany(id=1165721958978170974, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=3, ext=[AuthorCompanyExt(id=1165721958982365279, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958978170974, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Key Laboratory of Unmanned Emergency Equipment and Digital Reconstruction of Disaster Processes in Anhui Province, Chuzhou College, Chuzhou Anhui 239099, China), AuthorCompanyExt(id=1165721958990753888, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721958978170974, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 滁州学院 无人应急装备与灾害过程数字化重建安徽省联合共建学科重点实验室,安徽 滁州 239099)]), AuthorCompany(id=1165721959036891233, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, xref=4, ext=[AuthorCompanyExt(id=1165721959045279842, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721959036891233, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4 School of Safety Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China), AuthorCompanyExt(id=1165721959049474147, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, companyId=1165721959036891233, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=4 安徽理工大学 安全科学与工程学院,安徽 淮南 232001)])], figs=[ArticleFig(id=1165721960995631246, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.1, caption=Flowchart of the improved ViBe algorithm, figureFileSmall=x+sMs8O1dMqS6d3pro8fNg==, figureFileBig=/7FIXoNRxnFsatI4oTfeng==, tableContent=null), ArticleFig(id=1165721961054351503, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图1, caption=改进ViBe算法的流程, figureFileSmall=x+sMs8O1dMqS6d3pro8fNg==, figureFileBig=/7FIXoNRxnFsatI4oTfeng==, tableContent=null), ArticleFig(id=1165721961129848976, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.2, caption=Network structure of group shuffle convolution, figureFileSmall=S0sCzNwSf1wXRMyWOUpnag==, figureFileBig=iiOwUtBpNuT685rG+I485Q==, tableContent=null), ArticleFig(id=1165721961184374929, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图2, caption=分组混洗卷积的网络结构, figureFileSmall=S0sCzNwSf1wXRMyWOUpnag==, figureFileBig=iiOwUtBpNuT685rG+I485Q==, tableContent=null), ArticleFig(id=1165721961264066706, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.3, caption=C3-ECA network structure, figureFileSmall=K/pNB9GrYaXc/3RUBcWpkA==, figureFileBig=MnEjBvr9TpvG9+L11OZsKw==, tableContent=null), ArticleFig(id=1165721961326981267, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图3, caption=C3-ECA网络结构, figureFileSmall=K/pNB9GrYaXc/3RUBcWpkA==, figureFileBig=MnEjBvr9TpvG9+L11OZsKw==, tableContent=null), ArticleFig(id=1165721961389895828, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.4, caption=Comparison of the results of the prediction frames for the test samples, figureFileSmall=Wq84qxNxicddJgcRRwgapw==, figureFileBig=+fFKzqt2TGJMMHowoui72A==, tableContent=null), ArticleFig(id=1165721961469587605, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图4, caption=检测样本的预测框结果对比, figureFileSmall=Wq84qxNxicddJgcRRwgapw==, figureFileBig=+fFKzqt2TGJMMHowoui72A==, tableContent=null), ArticleFig(id=1165721961524113558, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.5, caption=Flowchart of YOLOv5s in this paper, figureFileSmall=EKGRBOdE8o0IOHOnweHUrQ==, figureFileBig=Na92IySK2LE7jrnO3+SpRg==, tableContent=null), ArticleFig(id=1165721961587028119, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图5, caption=本文YOLOv5s流程, figureFileSmall=EKGRBOdE8o0IOHOnweHUrQ==, figureFileBig=Na92IySK2LE7jrnO3+SpRg==, tableContent=null), ArticleFig(id=1165721961637359768, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.6, caption=Precision-recall graph and F1 value graph, figureFileSmall=GvdcZpmkECzTensH7Sj2Qw==, figureFileBig=g9g1fAb8vHVNROYmfmkcvg==, tableContent=null), ArticleFig(id=1165721961691885721, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图6, caption=精确率-召回率曲线及F1值曲线, figureFileSmall=GvdcZpmkECzTensH7Sj2Qw==, figureFileBig=g9g1fAb8vHVNROYmfmkcvg==, tableContent=null), ArticleFig(id=1165721961738023066, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.7, caption=Forest fire detection effect, figureFileSmall=UHjCAj8kLFBV9HKxP/CjUA==, figureFileBig=6TisFrSMxfywmV69kMBRDQ==, tableContent=null), ArticleFig(id=1165721961817714843, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图7, caption=森林火灾探测效果, figureFileSmall=UHjCAj8kLFBV9HKxP/CjUA==, figureFileBig=6TisFrSMxfywmV69kMBRDQ==, tableContent=null), ArticleFig(id=1165721961884823708, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.8, caption=Effectiveness of forest fire detection in dark scenes, figureFileSmall=6FcrJGcDcSBK6lYdO7yT8A==, figureFileBig=tjrrLvPOmAsJxVB3GQEMfw==, tableContent=null), ArticleFig(id=1165721961939349661, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图8, caption=黑暗场景下森林火灾探测效果, figureFileSmall=6FcrJGcDcSBK6lYdO7yT8A==, figureFileBig=tjrrLvPOmAsJxVB3GQEMfw==, tableContent=null), ArticleFig(id=1165721961993875614, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Fig.9, caption=Effectiveness of early forest fire detection, figureFileSmall=KEm2oURLEHC7cKoo7EsRwA==, figureFileBig=/Iy043O76l8iovjcMmxOFA==, tableContent=null), ArticleFig(id=1165721962044207263, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=图9, caption=早期森林火灾探测效果, figureFileSmall=KEm2oURLEHC7cKoo7EsRwA==, figureFileBig=/Iy043O76l8iovjcMmxOFA==, tableContent=null), ArticleFig(id=1165721962107121824, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Table 1, caption=

Parametric quantities and computational analysis of the two convolutions

, figureFileSmall=null, figureFileBig=null, tableContent=
模型计算 参数量 计算量 占比
常规卷积 P1=294 912 F1=301 989 888 0.559
分组混洗策略优化
下的常规卷积
P2=164 992 F2=186 908 672 0.619
), ArticleFig(id=1165721962161647777, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=表1, caption=

2种卷积的参数量和计算量分析

, figureFileSmall=null, figureFileBig=null, tableContent=
模型计算 参数量 计算量 占比
常规卷积 P1=294 912 F1=301 989 888 0.559
分组混洗策略优化
下的常规卷积
P2=164 992 F2=186 908 672 0.619
), ArticleFig(id=1165721962262311074, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Table 2, caption=

Comparison of bounding box loss performance

, figureFileSmall=null, figureFileBig=null, tableContent=
算法模型 AP75 AP50 AP
WIOUv1 52.24 62.37 45.16
WIOUv2
(r = 0.5)
53.34 63.86 45.56
本文WIOU
(α = 1.7, δ = 4)
53.61 64.04 45.58
), ArticleFig(id=1165721962375557283, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=表2, caption=

边界框损失性能比较

, figureFileSmall=null, figureFileBig=null, tableContent=
算法模型 AP75 AP50 AP
WIOUv1 52.24 62.37 45.16
WIOUv2
(r = 0.5)
53.34 63.86 45.56
本文WIOU
(α = 1.7, δ = 4)
53.61 64.04 45.58
), ArticleFig(id=1165721962446860452, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Table 3, caption=

Detection results of different algorithms

, figureFileSmall=null, figureFileBig=null, tableContent=
算法模型 P R FPR FNR m 模型大小/MB 检测用时/ms
Faster-RCNN 0.794 0.748 0.194 0.252 0.812 209.3 52.5
SSD 0.787 0.743 0.201 0.257 0.804 180.3 60.7
YOLOv4-tiny 0.785 0.739 0.202 0.261 0.811 27.3 8.4
YOLOv5s 0.798 0.745 0.189 0.255 0.834 14.3 9.9
YOLOv5s+小波变换 0.803 0.748 0.184 0.252 0.843 22.5 12.4
YOLOv5s+CIOU 0.814 0.842 0.192 0.158 0.846 20.2 11.7
本文算法 0.826 0.857 0.181 0.143 0.861 9.2 4.6
), ArticleFig(id=1165721962509775013, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=表3, caption=

不同算法的检测结果

, figureFileSmall=null, figureFileBig=null, tableContent=
算法模型 P R FPR FNR m 模型大小/MB 检测用时/ms
Faster-RCNN 0.794 0.748 0.194 0.252 0.812 209.3 52.5
SSD 0.787 0.743 0.201 0.257 0.804 180.3 60.7
YOLOv4-tiny 0.785 0.739 0.202 0.261 0.811 27.3 8.4
YOLOv5s 0.798 0.745 0.189 0.255 0.834 14.3 9.9
YOLOv5s+小波变换 0.803 0.748 0.184 0.252 0.843 22.5 12.4
YOLOv5s+CIOU 0.814 0.842 0.192 0.158 0.846 20.2 11.7
本文算法 0.826 0.857 0.181 0.143 0.861 9.2 4.6
), ArticleFig(id=1165721962593661094, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=EN, label=Table 4, caption=

Ablation experiments

, figureFileSmall=null, figureFileBig=null, tableContent=
模型 m 速度/
(帧·s-1)
YOLOv5s 0.834 35
YOLOv5s-ViBe 0.838 38
YOLOv5s-ViBe-分组混洗 0.839 42
YOLOv5s-ViBe-分组混洗-(ECA-C3) 0.847 39
YOLOv5s-ViBe-分组混洗-
(ECA-C3)-WIOU
0.861 39
), ArticleFig(id=1165721962648187048, tenantId=1146029695717560320, journalId=1146031787341344770, articleId=1148106706472661679, language=CN, label=表4, caption=

消融性试验

, figureFileSmall=null, figureFileBig=null, tableContent=
模型 m 速度/
(帧·s-1)
YOLOv5s 0.834 35
YOLOv5s-ViBe 0.838 38
YOLOv5s-ViBe-分组混洗 0.839 42
YOLOv5s-ViBe-分组混洗-(ECA-C3) 0.847 39
YOLOv5s-ViBe-分组混洗-
(ECA-C3)-WIOU
0.861 39
)], attaches=null, journal=Journal(id=1123942128916217864, delFlag=0, nameCn=中国安全科学学报, nameEn=China Safety Science Journal, nameHistory1=null, nameHistory2=null, issn=1003-3033, eissn=, cn=11-2865/X, coden=null, periodic=0, language=CN, oaType=0, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=fkqsFM6VKlHC4gCtS5XqTw==, journalPrice=null, startedYear=null, abbrevIsoEn=Chin Saf Sci J, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1754269350027, createdBy=null, updatedBy=13701087609, firstLetterCn=C, firstLetterEn=C, subjectCode=Engineering, subjectName=工程, subjectCodeEn=Engineering, subjectNameEn=null, picCn=fkqsFM6VKlHC4gCtS5XqTw==, picEn=SHn9HgqSxtJrOcAxqD++4Q==, jcr=null, cjcr=null, exts=[JournalExt(id=1159052918994595848, language=CN, name=中国安全科学学报, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.cssjj.com.cn/, createdTime=1754269350050, updatedTime=1754269350050, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://www.cssjj.com.cn/CN/column/item15.shtml, submissionAuthorUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionReviewUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionCeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionAeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, option={"copyright":""}), JournalExt(id=1159052919040733193, language=EN, name=China Safety Science Journal, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=null, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=http://www.cssjj.com.cn/EN/1003-3033/home.shtml, createdTime=1754269350061, updatedTime=1754269350061, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=https://synbioj.cip.com.cn/EN/column/column3.shtml, submissionAuthorUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionReviewUrl=https://zgaqkxxbauthor.manuscriptcloud.com/login, submissionCeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, submissionAeEditorUrl=https://zgaqkxxbeditor.manuscriptcloud.com/login, option={"copyright":""})], databaseList=null, tenantJournalId=1146031787341344770, websiteList=[Website(id=1148243202345263519, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146031787341344770, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN, language=CN, createTime=1751692112766, createBy=18614031015, updateTime=1753502583634, updateBy=18614031015, name=《中国安全科学学报》中文站点, tplId=1146099689490845704, title=中国安全科学学报, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148618794941046792, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=articleTextType, value=kx, createTime=1751781661020, updateTime=1751781661020, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794911686661, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=banner, value=null, createTime=1751781661012, updateTime=1751781661012, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794894909444, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=tui0IVO9FMwB61HHtX5scg==, createTime=1751781661008, updateTime=1751781661008, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794932658183, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751781661017, updateTime=1751781661017, creator=18614031015, updator=18614031015), WebsiteProps(id=1148618794924269574, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202345263519, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751781661015, updateTime=1751781661015, creator=18614031015, updator=18614031015)]), Website(id=1155836763751993353, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146031787341344770, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN, language=EN, createTime=1753502558893, createBy=18614031015, updateTime=1753524450387, updateBy=18614031015, name=《中国安全科学学报》英文站点, tplId=1146101810881728533, title=China Safety Science Journal, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155895925743669425, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=articleTextType, value=kx, createTime=1753516664205, updateTime=1753516664205, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925722697902, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=banner, value=null, createTime=1753516664200, updateTime=1753516664200, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925714309293, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=tui0IVO9FMwB61HHtX5scg==, createTime=1753516664198, updateTime=1753516664198, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925735280816, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753516664203, updateTime=1753516664203, creator=18614031015, updator=18614031015), WebsiteProps(id=1155895925731086511, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155836763751993353, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753516664202, updateTime=1753516664202, creator=18614031015, updator=18614031015)])], journalTitle=中国安全科学学报, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=China Safety Science Journal, journalPhotoCn=fkqsFM6VKlHC4gCtS5XqTw==, journalPhotoEn=SHn9HgqSxtJrOcAxqD++4Q==, journalFirstLetter=C, journalRecommend=null, journalNew=null, journalCollection=1, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/10.16265/j.cnki.issn1003-3033.2025.01.0127, detailUrlEn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN/10.16265/j.cnki.issn1003-3033.2025.01.0127, pdfUrlCn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/PDF/10.16265/j.cnki.issn1003-3033.2025.01.0127, pdfUrlEn=https://castjournals.cast.org.cn/joweb/zgaqkxxb/EN/PDF/10.16265/j.cnki.issn1003-3033.2025.01.0127, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于YOLOv5s的轻量化森林火灾探测算法
收藏切换
PDF下载
刘惠临 1 , 方琼 1 , 江宇 1 , 魏华章 2, ** , 王涛 3 , 张树川 4
中国安全科学学报 | 安全工程技术 2025,35(1): 75-83
收起
收藏切换
中国安全科学学报 | 安全工程技术 2025, 35(1): 75-83
基于YOLOv5s的轻量化森林火灾探测算法
全屏
刘惠临1 , 方琼1, 江宇1, 魏华章2, ** , 王涛3, 张树川4
作者信息
  • 1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001
  • 2 淮南职业技术学院 智能与电气工程学院,安徽 淮南 232001
  • 3 滁州学院 无人应急装备与灾害过程数字化重建安徽省联合共建学科重点实验室,安徽 滁州 239099
  • 4 安徽理工大学 安全科学与工程学院,安徽 淮南 232001
  • 刘惠临 (1979—),男,安徽合肥人,博士,高级实验师,主要从事计算机视觉、图像处理、多模态数据融合等方面的研究。E-mail:

    王涛 副教授

    张树川 副教授

通讯作者:

**魏华章(1996—),男,安徽淮南人,硕士研究生,研究方向为计算机视觉。E-mail:
A lightweight forest fire detection algorithm based on YOLOv5s
Huilin LIU1 , Qiong FANG1, Yu JIANG1, Huazhang WEI2, ** , Tao WANG3, Shuchuan ZHANG4
Affiliations
  • 1 School of Computer Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China
  • 2 School of Intelligence and Electrical Engineering, Huainan Vocational Technical College, Huainan Anhui 232001, China
  • 3 Key Laboratory of Unmanned Emergency Equipment and Digital Reconstruction of Disaster Processes in Anhui Province, Chuzhou College, Chuzhou Anhui 239099, China
  • 4 School of Safety Science and Engineering, Anhui University of Science and Technology, Huainan Anhui 232001, China
出版时间: 2025-01-28 doi: 10.16265/j.cnki.issn1003-3033.2025.01.0127
文章导航
收藏切换

为解决当前基于深度学习的森林火灾探测算法存在结构复杂、规模庞大,且难以兼顾检测精度和效率的问题,提出一种基于YOLOv5s的轻量化森林火灾探测算法。首先,采用优化的背景差分技术消除背景图像中类火物体的干扰,减少分析图像所需的时间;其次,设计分组混洗策略优化常规卷积,并在特征提取的C3模块中融入高效通道注意力(ECA)机制和深度可分离卷积,增强图像特征提取与融合能力的同时有效降低模型的参数量;然后,采用动态非单调聚焦机制优化Wise-交并比(WIOU)损失函数,减少低质量样本产生的有害梯度;最后,在构建的森林火灾数据集上将所提算法与其他算法做充分的试验对比。结果表明: 所提算法在各类场景均展现出良好的泛化性,对火焰目标的检测精度达到86.1%,较标准YOLOv5s检测精度提升2.7%,检测速度提升11.4%,有效降低了火灾误报率,增强了模型的检测性能。

YOLOv5s  /  轻量化  /  森林火灾探测  /  深度可分离卷积  /  注意力  /  Wise-交并比(WIOU)

In order to solve the problems of complex structure, large scale and difficulty in balancing detection accuracy and efficiency of the current forest fire detection algorithm based on deep learning, a lightweight forest fire detection algorithm based on YOLOv5s was proposed. Firstly, an optimized background difference technique was used to eliminate the interference of fire-like objects in the background image, thus reducing the time required for image analysis. Secondly, a group blending strategy was designed to optimize the conventional convolution, and an efficient channel attention (ECA) mechanism and depthwise separable convolution were incorporated into the C3 module of feature extraction, which enhanced the ability of image feature extraction and fusion and at the same time effectively reduces the number of model parameters. Then, a dynamic non-monotonic focusing mechanism was used to optimize the WIOU loss function, reducing the harmful gradients generated by low-quality samples. Finally, sufficient experimental comparisons between the proposed algorithm and other algorithms on the constructed forest fire dataset. The results show that the proposed algorithm shows good generalization in various scenarios, and the detection accuracy of the flame target can reach 86.1%, which is 2.7% higher than that of the standard YOLOv5s, and the detection speed is increased by 11.4%, which effectively reduces the fire false alarm rate and enhances the detection performance of the model.

YOLOv5s  /  lightweighting  /  forest fire detection  /  depthwise separable convolution  /  attention  /  wise intersection over union(WIOU)
刘惠临, 方琼, 江宇, 魏华章, 王涛, 张树川. 基于YOLOv5s的轻量化森林火灾探测算法. 中国安全科学学报, 2025 , 35 (1) : 75 -83 . DOI: 10.16265/j.cnki.issn1003-3033.2025.01.0127
Huilin LIU, Qiong FANG, Yu JIANG, Huazhang WEI, Tao WANG, Shuchuan ZHANG. A lightweight forest fire detection algorithm based on YOLOv5s[J]. China Safety Science Journal, 2025 , 35 (1) : 75 -83 . DOI: 10.16265/j.cnki.issn1003-3033.2025.01.0127
森林火灾是全球生态系统面临的重大威胁之一,不仅对自然环境造成巨大破坏,还对人类的生命财产安全构成严重威胁。随着气候变化和人类活动的增加,森林火灾的频率和强度呈上升趋势。因此,及时准确地实现森林火灾探测尤为重要。早期,森林火灾主要依靠地面巡逻和人工瞭望,因其覆盖范围有限、反应速度慢,已无法满足现代火灾预防和应急响应的需求。随着计算机技术的发展,基于感温、感光、感烟等传感器件[1]实现了对森林火灾的探测,将获取的森林火灾发生时空气温度、烟雾浓度等一系列参数的分析结果转化为电信号,进而判断火情。然而,这种火灾探测方式易受环境因素影响,导致检测精度低,存在大量漏检误检问题。
一些学者采用机器学习法探测森林火灾,首先手动提取火焰的静态和动态特征,并将提取的特征导入分类器分类,通过反复迭代训练,使其适用于不同场景。常用的分类器有支持向量机[2]、随机森林[3]等。但这种方法需要手动提取图像的火焰特征,过度依赖人力,效率较低。伴随着人工智能技术的发展,基于深度学习的森林火灾探测技术能够大幅提高火灾探测精度[4]。卷积神经网络是深度学习中最常用的一种技术架构,具有很强的自学习能力和泛化能力,响应速度快且抗干扰能力强,可适用于复杂的视觉任务。基于卷积网络的经典检测算法有快速卷积神经网络(Faster Region-based Convolutional Neural Networks,Faster-RCNN)[5]、YOLO[6]、单次多框检测器(Single Shot multiBox Detector,SSD)[7]等。然而,这些方法都存在一定的局限性,如严重依赖于高性能计算设备、模型容易过拟合等。当前,基于深度学习的森林火灾探测技术的优化策略大致分为2类,分别是提升检测精度和模型轻量化。SHARMA等[8]通过数据集拼接复制现实场景,增强预训练VGG16和ResNet50网络,提升火灾探测精度。邱思远等[9]在YOLOX的Neck网络中添加了注意力模块,增强了森林火灾中火焰的特征提取能力。王喆等[10]将协调注意力机制引入YOLOv5s的主干网络,使模型能够获取不同位置之间的关系,精准定位起火点位置。HOSSEINI[11]将每帧图像分为8类,并对其进行实时分类,以减少网络误报。高源等[12]使用Mosaic对森林火灾图像进行数据增强,在YOLOv5模型中添加完全交并比(Complete Intersection Over Union,CIOU)损失函数,实现了复杂场景下的森林火灾探测。以上基于深度学习的森林火灾探测方法虽然取得不错效果,但很难同时兼顾精度和算力资源要求。此外,光线反射、散射等自然环境干扰因素,也会造成误检。
鉴于当前森林火灾探测中存在精度低、误检率高且难以部署在算力有限监控设备中等问题,笔者拟设计一种基于改进YOLOv5s的森林火灾图像探测算法。首先,通过采用视觉背景提取器(Visual Background extractor,ViBe)提取火焰目标,消除背景中类火物体干扰;然后,设计分组混洗卷积策略减少模型的参数量,并引入高效通道注意力(Efficient Channel Attention,ECA)机制增强高层特征图中火焰的细节特征;最后,通过动态非单调聚焦机制优化Wise-交并比(Wise Intersection Over Union,WIOU)损失函数修正不同质量的锚框,减少产生的有害梯度并增强模型的预测能力,以期在保证精度的同时,实现森林火灾的轻量化探测。
YOLOv5[13]拥有多种不同大小的算法模型结构,如n, s, m, l, x等5种基础类型,其中,ns为轻量化模型,可部署在一些低算力设备上,后3种需要算力支持,本文以YOLOv5s为基线。YOLOv5结构包括3部分:BackBone、Neck和Head。以YOLOv5s结构为例,输入图像经过跨阶段局部网络进行特征图提取,可有效解决网络优化中梯度信息重复的问题,降低模型的参数量和计算量;特征图进入网络后,会得到5个不同尺度的特征图,再经空间金字塔池化快速(Spatial Pyramid Pooling Fast,SPPF)结构提高网络的感受野,加快响应速度。网络的Neck部分由特征金字塔网络和路径聚合网络结构组成,通过多次和特征图进行拼接,确保生成的最终特征图不仅包含丰富的语义信息,还包含物体位置信息。
可视化背景差分技术是一种基于视频图像二值化像素点的目标检测方法,其过程主要包括初始化背景建模、前景分割、背景更新。将该技术用于火灾图像的预处理阶段,主要是提取目标区域火焰的前景,消除背景中类火物体的干扰。虽然标准ViBe的执行效率高,前景探知能力强,但易产生鬼影现象。因此,采用先验操作优化ViBe[14],增强图像质量并抑制鬼影,思路如下:首先采用直方图相似度度量初始几帧图像,并选择邻近差异较大帧进行帧差操作;其次通过形态学滤波填充前景图并在邻域像素内随机采样完成背景建模,最后引入时空采样因子和帧间均速值自适应调整背景变化率与模型更新率的关系,从而消除鬼影,实现火灾图像的动态区域分割,具体流程如图1所示。
在预处理操作中,采用标准ViBe在提取复杂背景下的动态目标时,由于自身方法局限易引起鬼影,导致背景区域像素被误判成前景像素,影响后续检测精度,而改进的ViBe采用形态学滤波抑制背景噪声干扰,使得前景区域更饱满,通过设置前景区域矩形框,框定火灾图像动态区域并映射到原视频帧,相当于分块处理火灾图像,有效降低分析背景图像花费的时间,提高了检测精度,为后续试验提供保障。
网络模型越深非线性表达能力就越强,参数量也越大,对算力资源的需求也就更高。实际上火灾视频监控的边缘端设备算力通常有限,因此,模型的轻量化优化至关重要。目前,主流轻量化探测算法通常采用深度可分离卷积代替常规卷积,如Xception、MobileNet、ShuffleNet等。这种方法虽可大幅减少参数量,但特征图之间的通道信息是分离的,会导致特征提取与融合能力变弱。
文中提出分组混洗卷积策略实现通道间交互。首先进行常规卷积的下采样,将得到的特征图进行深度可分离卷积操作,再将该特征图与原始下采样特征图拼接,通过维度变化和重组等操作生成最终的特征图,具体流程如图2所示。该过程使用混洗卷积将常规卷积生成的信息(密集卷积操作)渗透到深度可分离卷积中。
下式为分组混洗卷积策略与常规卷积的参数量和计算量。
P 1 = K h × K w × C i n × C o u t
F 1 = K h · K w · C i n · C o u t · O h · O w
P 2 = ( K h · K w + C o u t ) · C i n
F 2 = 2 ( K h · K w + C o u t ) · O h · O w · C i n
式中:P1P2分别表示常规卷积和分组混洗卷积策略优化下的参数量;F1F2分别表示其计算量;KhKw为卷积核的高和宽;CinCout分别为输入、输出特征图的通道数;OhOw分别表示输出特征图的高和宽。此处,设定卷积核尺寸为3,Cin为128,Cout为256,OhOw为32。通过式(1)计算的参数量和计算量见表1,采用分组混洗策略优化常规卷积,参数量约为常规卷积的一半,计算量降低至60%,有效改善了火灾图像检测网络模型体量大的问题。
在通道注意力基础上优化ECA机制[15],可实现不降维局部跨信道交互。ECA机制采用一维卷积替代全连接层,只与部分通道交互,极大降低了参数量,更符合轻量化的探测需求。
C3模块主要是学习残差特征行,获取特征图间上下文信息,它含有3个卷积-批归一化-SiLU结构(Convolution-Batch normalization-SiLU, CBS)。改进的结构如图3所示,文中采用深度可分离操作优化C3中的常规卷积,得到深度可分离卷积-批归一化-SiLU结构(Depthwise separable convolution-Batch normalization-SiLU,DBS)。其中,左侧分支含有一个DBS结构,右侧分支由DBS和N个残差模块组成,每个分支可捕捉不同尺度的特征信息,将不同分支获取的特征图拼接,输出最终特征图。为满足火灾探测过程中轻量化需求,采用深度可分离操作优化C3中的常规卷积,减少模型的参数。此外,在融合过程中引入ECA机制增大目标区域的特征权重,更好地捕捉高层特征图中的语义信息,提高火灾图像检测的精度,降低误检率。
边界框损失函数是目标检测算法损失函数中重要的组成部分,目前研究强调边界框对数据集中目标的拟合能力,忽略了当存在低质量样本,过强的拟合能力反而会损害模型效果。在真实的火灾检测场景中,环境因素会产生低质量样本,使模型的泛化能力下降。采用动态非单调聚焦机制优化的WIOU损失函数将注意力分配到低质量的锚框上,从而提升检测的整体性能。
交并比(Intersection Over Union,IOU)是用来度量目标检测任务中预测框与真实框的重叠程度的重要指标。假设预测框=[x y w h],真实框=[xgt ygt wgt hgt],则2框的交集和IOU损失定义分别如下:
S u = w h + w g t h g t - w a h a
L I O U = 1 - I O U = 1 - w a h a S u
式中:wh分别为预测框的宽和高;wgt h g t分别为真实框的宽和高;waha分别为预测框和真实框重叠区域的宽和高。
CHO[16]通过构建距离注意力,得到2层注意力机制的WIOUv1,可有效解决质量差异大的样本间框回归平衡问题。WIOUv1损失定义如下式:
L W I O U v 1 = R W I O U v 1 L I O U R W I O U v 1 = e x p ( x - x g t ) 2 + ( y - y g t ) 2 W 2 g + H 2 g
式中: R   W I O U v 1∈[e]表示中心点连接归一化长度的惩罚项;LIOU∈[0,1],表示IOU损失;(x,y)为预测框的中心坐标;(xgt,ygt)为真实框的中心坐标;WgHg为封闭框的宽和高。
为聚焦更复杂示例,采用交叉熵单调聚焦机制WIOUv2降低简单示例对损失值的影响,获得分类性能提升。采用下式计算WIOUv2损失:
L W I O U v 2 = L I O U * L - I O U r L W I O U v 1
式中: L I O U *为梯度增益; L - I O ULIOU的均值;r为系数;LWIOUv1表示WIOUv1损失。
WIOUv2采用静态聚焦机制,未充分挖掘非单调聚焦机制的潜能,采用动态非单调聚焦机制优化WIOU损失函数,使用“离群度”替代IOU评估锚框质量,使模型能够动态聚焦不同质量的锚框。WIOU损失定义见下式,通过构造一个非单调的聚焦系数β动态聚焦不同质量的锚框。
β = L I O U * L - I O U [ 0 , ]
L ' W I O U = β δ α β - δ L W I O U v 1
式中 α δ均为系数。
将改进损失函数与WIOUv1和WIOUv2这2种边界框损失函数进行质量评估,评估结果如图4所示,可知:所提算法的预测效果比WIOUv1和WIOUv2更好。
性能损失的定量比较结果见表2,采用平均精度(Average Precision,AP)作为评价指标,AP75表示IOU阈值为0.75的平均精度,AP50表示IOU阈值为0.5的平均精度。采用动态非单调聚焦机制优化边界框损失函数使模型能够动态聚焦不同质量的锚框,增强火灾探测算法型预测的准确性。
以YOLOv5s为基准,结合森林火灾探测过程中面临的关键问题优化该算法。首先,使用背景差分技术预处理火灾图像,抑制背景中类火物体干扰;其次,采用分组混洗卷积策略优化常规卷积,在保证精度的同时减少参数量,ECA机制融入深度可分离卷积改进的C3模块,突出火焰细节特征信息;最后,采用非单调聚焦机制优化的WIOU损失函数度量边界框损失,提升模型对低质量样本的学习能力。上述的改进策略,大幅提升了火灾图像的检测精度和速度,增强火灾预警能力。图5展示了YOLOv5s的改进流程。
为保证模型训练的鲁棒性,通过多种来源构建一个包含8 000余张图像的森林火灾数据集,包括研究团队自己采集标注的数据及一些公开数据集,将数据集按7∶2∶1比例划分为训练集、验证集及测试集。
在试验设计方面,将Batchsize值设为16,Epoch值设为300,通过优化和深入学习模型,使其能够全面理解森林火灾图像特征,进而为火灾识别与预警提供可靠的研究基础。
评价目标检测模型使用的指标有精确率P、召回率R、平均精度均值m,火灾探测中还要考虑误检和漏检情况,引入误检率(False Positive Rate,FPR)和漏检率(False Negative Rate,FNR),综合分析火灾探测算法的性能。指标计算如下:
P = T P / ( T P + F P )
R = T P / ( T P + F N )
F P R = F P / ( F P + T N )
F N R = F N / ( T P + F N )
m = 1 c i = 1 c A P i
式中:TP为正确检测到火灾的样本数量;FP为被误检成火灾样本数量;FN为未检测到火灾样本数量;TN为火灾样本误检为负样本数量;c为样本总数;APi为平均检测精度,表示PR曲线与坐标轴所围起来的面积。
采用AP来检测模型的质量,并利用F1得分评价模型的学习性能,得分越高,说明模型的学习性能越好。F1得分的计算公式如下:
F 1 = 2 × P × R / ( P + R )
图6为精确率-召回率曲线及F1值曲线,其中,置信度表示模型对预测框内存在火灾的确信程度。由图6可知:所提算法均展现出了较为稳定且良好的性能表现。
以YOLOv5s模型为基准改进,与主流目标检测算法对比,包括Faster-RCNN、SSD、YOLOv4-tiny、标准YOLOv5s、CIOU损失函数优化的YOLOv5s检测算法及基于小波变换改进的YOLOv5s检测算法,森林火灾正负样本数量设为1∶1,最终试验数据见表3。由表3可知:所提模型m为86.1%,FPR降至18.1%,FNR降至14.3%,检测时间提升至4.6ms,模型大小降至9.2 MB,检测精度和速度明显提升,较对比其他目标检测模型性能更好。采用分组混洗策略优化常规卷积降低模型的计算量和参数量,ECA机制融入深度可分离卷积改进的C3模块,突出火焰细节特征的同时抑制背景噪声干扰;采用非单调聚焦机制优化WIOU损失函数使模型预测框更加逼近真实框,有效改善了低质量样本导致的过拟合问题。在YOLOv5s模型基础上优化特征提取与融合,在降低漏检率的同时降低误检率。与YOLOv5s相比,所提模型的漏检率下降11.2%,误检率也下降了0.8%,较其他算法,漏检率也有一定程度下降。综上所述,所提模型能更好地区分火焰区域特征,提高检测精度的同时降低误检率和漏检率。REDDY[17]和章曙光[18]等对YOLOv5s模型的改进虽然均实现了精度的提升,但检测效果不如本文算法。
为充分验证所提模型在各种实际应用场景下的适用性,针对多个复杂场景下的森林火灾进行试验。
图7为森林火灾场景下的探测效果图,由图可知:所提模型能及时且精准地定位火情,有效降低了火灾扩散的风险,减少森林资源的损失以及对周边生态环境可能造成的破坏。
为进一步评估模型的泛化性,针对黑暗场景下的森林火灾探测效果开展试验。由于光线不足,摄像机拍摄的图像质量会大幅下降,导致火灾特征难以准确识别,同时限制了火焰以外区域的可视性,黑暗场景下的森林火灾探测能够显著提高夜间火灾发现的及时性和准确性,从而有效减少了森林火灾造成的损失。针对上述问题,在背景差分预处理过程中融合了对比增强、图像滤波等功能,抑制了背景对火焰目标的干扰,改善了黑暗场景下的火灾图像数据质量。图8为黑暗场景下的森林火灾探测效果,可以看出,火焰目标均能被准确识别,表明模型的预测性能符合预期,泛化性较为优异。
早期森林火灾的烟雾和火焰较小,形状多变,极易受到背景信息的干扰。为提高模型的抗干扰能力,使用注意力机制增强火焰区域特征,优化边界框损失函数。结果如图9所示,在早期森林火灾场景下,火灾早期较小火焰目标均能够被有效检测到,说明所提模型在该场景中的有效性。
表4为本文算法的消融性试验结果,所提算法较标准YOLOv5s的精度和推理速度均有提升。ViBe技术消除了静态背景中类火物体的干扰,并节省了后续分析图像花费的时间;采用分组混洗策略优化常规卷积,推理速度提升20%;ECA机制融入C3结构使得检测精度提升0.8%,与分组混洗卷积相比,虽然推理速度略有降低,但较标准YOLOv5s的推理速度仍有提升;动态非单调聚焦机制优化的WIOU损失函数,使得精度再次提升约1.4%。综上所述,所提算法在提升检测精度的同时降低模型体量,大幅提升了复杂场景中的森林火灾的探测性能。
1) 所提算法通过改进YOLOv5s结构,在降低参数量的同时提高检测精度,适用于低算力设备。
2) 在自建的森林火灾数据集上针对所提算法开展试验,精确率为82.6%,召回率为85.7%,误检率为18.1%,漏检率为14.3%,平均精度均值为86.1%。
3) 改进后的算法在黑暗场景和早期森林火灾场景中均取得了良好的检测效果。然而,改进ViBe技术主要是针对静态场景下的前景提取,未来工作将进一步探究动态场景下的火灾探测技术。
  • 安徽省重点研究与开发计划项目(2023g07020007)
  • 安徽理工大学研究生创新基金资助(2024cx2111)
参考文献 引证文献
排序方式:
[1]
CHOI M, KIM C, OH H. A video-based SlowFastMTB model for detection of small amounts of smoke from incipient forest fires[J]. Journal of Computational Design and Engineering, 2022, 9(2):793-804.
[2]
熊昊, 李伟. 基于SVM的视频火焰检测算法[J]. 传感器与微系统, 2020, 39(1):143-145, 149.
XIONG Hao, LI Wei. Video flame detection algorithm based on svm[J]. Sensors and Microsystems, 2020, 39(1):143-145,149.
[3]
GUAN Zhenwei, MIN Feng, HE Wei, et al. Forest fire detection via feature entropy guided neural network[J]. Entropy, 2022, 24(1):DOI:10.3390/e24010128.
[4]
李海, 孙鹏. 基于知识图谱的火灾图像识别研究[J]. 中国安全科学学报, 2023, 33(10):147-159.
LI Hai, SUN Peng. Research on fire image recognition based on scientific knowledge graph[J]. China Safety Science Journal, 2023, 33(10):147-159.
[5]
REN Shaoqing, HE Kaiming, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6):1137-1149.
[6]
张为, 魏晶晶. 嵌入DenseNet结构和空洞卷积模块的改进YOLOv3火灾检测算法[J]. 天津大学学报, 2020, 53(9):976-983.
ZHANG Wei, WEI Jingjing. Improved YOLOv3 fire detection-algorithm with embedded DenseNet structure and cavity convolution module[J]. Journal of Tianjin University, 2020, 53(9):976-983.
[7]
LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]. European Conference on Computer Vision (ECCV), 2016:21-37.
[8]
SHARMA J, GRANMO O, GOODWIN M, et al. Deep convolutional neural networks for fire detection in images[C]. Engineering Applications of Neural Networks:18th International Conference, 2017:183-193.
[9]
邱思远, 王孝兰, 张伟伟, 等. 基于YOLOX的森林火灾早期预警研究[J]. 农业装备与车辆工程, 2023, 61(11):106-110.
QIU Siyuan, WANG Xiaolan, ZHANG Weiwei, et al. Research on early of forest fire based on YOLOX[J]. Agricultural Equipment and Vehicle Engineering, 2023, 61(11):106-110.
[10]
王喆, 李享, 杨栋梁, 等. 基于YOLOv5s的城镇森林交界域火灾探测模型[J]. 中国安全科学学报, 2023, 33(6):152-158.
WANG Zhe, LI Xiang, YANG Dongliang, et al. Fire detection model of wildland-urban interface based on YOLOv5s[J]. China Safety Science Journal, 2023, 33(6):152-158.
[11]
HOSSEINI A, HASHEMZADEH M, FARAJZADEH N. UFS-Net: a unified flame and smoke detection method for early detection of fire in video surveillance applications using cnns[J]. Journal of Computational Science, 2022, 61: DOI:10.1016/j.jocs.2022.101638.
[12]
高源, 杨望, 曹磊, 等. 基于改进YOLOv5的特征抑制多尺度林火检测算法[J]. 林业工程学报, 2023, 8(6):145-153.
GAO Yuan, YANG Wang, CAO Lei, et al. Feature suppression multi-scale forest fire detection algorithm based on improved YOLOv5[J]. Journal of Forestry Engineering, 2023, 8(6):145-153.
[13]
张全, 张伟, 杨先凤, 等. 融合YOLOv5-ResNet级联网络的烟火检测方法[J]. 安全与环境学报, 2023, 23(2):397-405.
ZHANG Quan, ZHANG Wei, YANG Xianfeng, et al. Smoke and fire detection method fusing YOLOv5-ResNet cascade network[J]. Journal of Safety and Environment, 2023, 23(2):397-405.
[14]
LIU Huilin, WEI Huazhang, YANG Gaoming, et al. An improved vibe algorithm based on adaptive thresholding and the deep learning-driven frame difference method[J]. Electronics, 2023, 12(16):DOI:10.3390/electronics12163481.
[15]
WANG Qilong, WU Banggu, ZHU Pengfei, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020:11 531-11 539.
[16]
CHO Y. Weighted intersection over union (WIOU) for evaluating image segmentation[J]. Pattern Recognition Letters, 2024, 185: 101-107.
[17]
REDDY E, THALE S. Pedestrian detection using YOLOv5 and complete-IoU loss for autonomous driving applications[C]. Proceedings of International Conference on Paradigms of Communication, Computing and Data Analytics, 2023:287-296.
[18]
章曙光, 唐锐, 邵政瑞, 等. 一种基于小波变换的YOLOv5火灾检测改进算法[J]. 无线电工程, 2023, 53(10):2303-2310.
ZHANG Shuguang, TANG Rui, SHAO Zhengrui, et al. A wavelet transform-based improved algorithm for YOLOv5 fire detection[J]. Radio Engineering, 2023, 53(10):2303-2310.
2025年第35卷第1期
PDF下载
385
167
引用本文
BibTeX
文章信息
doi: 10.16265/j.cnki.issn1003-3033.2025.01.0127
  • 接收时间:2024-08-20
  • 首发时间:2025-07-05
  • 出版时间:2025-01-28
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-08-20
  • 修回日期:2024-10-25
基金
安徽省重点研究与开发计划项目(2023g07020007)
安徽理工大学研究生创新基金资助(2024cx2111)
作者信息
    1 安徽理工大学 计算机科学与工程学院,安徽 淮南 232001
    2 淮南职业技术学院 智能与电气工程学院,安徽 淮南 232001
    3 滁州学院 无人应急装备与灾害过程数字化重建安徽省联合共建学科重点实验室,安徽 滁州 239099
    4 安徽理工大学 安全科学与工程学院,安徽 淮南 232001

通讯作者:

**魏华章(1996—),男,安徽淮南人,硕士研究生,研究方向为计算机视觉。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/zgaqkxxb/CN/10.16265/j.cnki.issn1003-3033.2025.01.0127
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