Article(id=1156986627395510894, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156963927277003616, articleNumber=null, orderNo=null, doi=10.12404/j.issn.1671-1815.2402154, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1711382400000, receivedDateStr=2024-03-26, revisedDate=1734278400000, revisedDateStr=2024-12-16, acceptedDate=null, acceptedDateStr=null, onlineDate=1753776707743, onlineDateStr=2025-07-29, pubDate=1742227200000, pubDateStr=2025-03-18, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1753776707743, onlineIssueDateStr=2025-07-29, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1753776707743, creator=13701087609, updateTime=1753776707743, updator=13701087609, issue=Issue{id=1156963927277003616, tenantId=1146029695717560320, journalId=1146123166801305609, year='2025', volume='25', issue='8', pageStart='3079', pageEnd='3528', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1753771295613, creator=13701087609, updateTime=1753777038876, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1156988016305726153, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156963927277003616, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1156988016305726154, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1156963927277003616, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=3325, endPage=3332, ext={EN=ArticleExt(id=1156986628708328052, articleId=1156986627395510894, tenantId=1146029695717560320, journalId=1146123166801305609, language=EN, title=A Coke Intelligent Loading Overflow Detection Method Based on Dark Channel Prior Knowledge and ResNet Network, columnId=1156963928489157478, journalTitle=Science Technology and Engineering, columnName=Automation and Computational Technology, runingTitle=null, highlight=null, articleAbstract=

The accurate detection of coke overflow in high-dust environments is a pivotal challenge in achieving intelligent coke loading. A method was proposed to address this issue for the intelligent detection of coke loading overflow, which was based on dark channel prior knowledge and the ResNet network. Firstly, a video collector was used to obtain video information of the coke loading scene, and the original time-series video image frames were processed to obtain the region of interest between the discharge port and the loader. Secondly, the prior knowledge method of dark channels was employed to process the regions of interest. Enhancing the contrast between the target areas and irrelevant areas within the regions of interest, thereby mitigating the effects of dust on subsequent detection models. Moreover, the problem of overflow detection was transformed into a binary classification task by labeling the regions of interest based on the actual loading of coke. Finally, the ResNet network was utilized for modeling, enabling the completion of model training and experimentation during the loading process of newly acquired coke. The experimental results demonstrate that the proposed method exhibits promising performance on new data, achieving an overall accuracy of 86.81%. Specifically, the accuracy, recall, and F1 score for the overflow class are 84. 12%, 90.74%, and 0.8730, respectively. Furthermore, the application of the dark channel prior algorithm in data processing results in a notable increase in the recall rate of the overflow class by 3.31%.

, correspAuthors=Hui-fang HOU, 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=Kang-zhan XIE, Hui-fang HOU, Zi-hao ZHANG, Wen-tao SUN), CN=ArticleExt(id=1156986776343634554, articleId=1156986627395510894, tenantId=1146029695717560320, journalId=1146123166801305609, language=CN, title=基于暗通道先验知识和ResNet网络的焦炭智能装载溢出检测方法, columnId=1154013913436902243, journalTitle=科学技术与工程, columnName=自动化技术、计算机技术, runingTitle=null, highlight=null, articleAbstract=

高粉尘环境下进行精准的焦炭溢出检测是实现焦炭智能装载的重要挑战。针对此问题,提出一种基于暗通道先验知识和ResNet 网络的焦炭智能装载溢出检测方法。首先,利用视频采集器获取焦炭装载场景视频信息,并对原始时间序列视频图像帧进行处理以获得下料口及装载器之间感兴趣区域;其次,提出利用暗通道先验知识方法对感兴趣区域进行处理,提升感兴趣区域中目标区域与无关区域之间对比度,以降低粉尘对后续检测模型的影响。再者,根据焦炭实际装载情况对感兴趣区域进行标注将溢出检测问题转化成二分类。最终,提出利用ResNet 网络建模完成对模型的训练获得训练模型并在新采集焦炭装载过程中进行实验。实验证明所提方法在新的数据上测试结果表现优异,整体准确率达到86. 81%,其中溢出类的精确度、召回率和F1 分数分别为84. 12%、90. 74% 和0. 873 0。并且在使用了暗通道先验算法处理数据后,溢出类的召回率上升了3. 31%。

, correspAuthors=侯惠芳, authorNote=null, correspAuthorsNote=
*侯惠芳(1972-),女,汉族,河南安阳人,博士,教授。研究方向:计算机视觉。E-mail: houhuifang@haut.edu.cn。
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=rto4mWjWwZ3B1odMgavOJg==, magXml=lVNjtYpLYMoTC9MsOdaVwA==, pdfUrl=null, pdf=94IJqMTw+6rkilEX5A1GYg==, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=29t2eiNSOZeyDHhq7p5PPA==, mapNumber=null, authorCompany=null, fund=null, authors=

解康战(1997—),男,汉族,河南周口人,硕士研究生。研究方向:计算机视觉。E-mail: xkz19971010@163.com。

, authorsList=解康战, 侯惠芳, 张自豪, 孙文涛)}, authors=[Author(id=1156986779145429687, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=xkz19971010@163.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1156986779195761338, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779145429687, language=EN, stringName=Kang-zhan XIE, firstName=Kang-zhan, middleName=null, lastName=XIE, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1156986779271258812, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779145429687, language=CN, stringName=解康战, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 河南工业大学 信息科学与工程学院 郑州 450001, bio={"content":"

解康战(1997—),男,汉族,河南周口人,硕士研究生。研究方向:计算机视觉。E-mail: xkz19971010@163.com。

"}, bioImg=null, bioContent=

解康战(1997—),男,汉族,河南周口人,硕士研究生。研究方向:计算机视觉。E-mail: xkz19971010@163.com。

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1156986778998629037, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=1, ext=[AuthorCompanyExt(id=1156986779007017646, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779015406255, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 河南工业大学 信息科学与工程学院 郑州 450001)])]), Author(id=1156986779338367679, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=houhuifang@haut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1156986779439030978, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779338367679, language=EN, stringName=Hui-fang HOU, firstName=Hui-fang, middleName=null, lastName=HOU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, *, address=2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1156986779514528452, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779338367679, language=CN, stringName=侯惠芳, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, *, address=2 河南工业大学 人工智能与大数据学院 郑州 450001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1156986779086709426, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=2, ext=[AuthorCompanyExt(id=1156986779090903731, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779099292340, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 河南工业大学 人工智能与大数据学院 郑州 450001)])]), Author(id=1156986779573248711, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, 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=1156986779648746186, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779573248711, language=EN, stringName=Zi-hao ZHANG, firstName=Zi-hao, middleName=null, lastName=ZHANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1156986779711660748, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779573248711, language=CN, stringName=张自豪, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2 河南工业大学 人工智能与大数据学院 郑州 450001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1156986779086709426, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=2, ext=[AuthorCompanyExt(id=1156986779090903731, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779099292340, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 河南工业大学 人工智能与大数据学院 郑州 450001)])]), Author(id=1156986779770381006, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1156986779841684176, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779770381006, language=EN, stringName=Wen-tao SUN, firstName=Wen-tao, middleName=null, lastName=SUN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1156986779959124689, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, authorId=1156986779770381006, language=CN, stringName=孙文涛, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 河南工业大学 信息科学与工程学院 郑州 450001, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1156986778998629037, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=1, ext=[AuthorCompanyExt(id=1156986779007017646, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779015406255, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 河南工业大学 信息科学与工程学院 郑州 450001)])])], keywords=[Keyword(id=1156986781229998840, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, orderNo=1, keyword=coke intelligent loading overflow detection), Keyword(id=1156986781305496315, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, orderNo=2, keyword=prior knowledge of dark channels), Keyword(id=1156986781385188094, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, orderNo=3, keyword=ResNet network), Keyword(id=1156986781439714047, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, orderNo=1, keyword=焦炭智能装载溢出检测), Keyword(id=1156986781506822914, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, orderNo=2, keyword=暗通道先验知识), Keyword(id=1156986781569737475, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, orderNo=3, keyword=ResNet 网络)], refs=[Reference(id=1156986794391724052, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=王熹, journalName=null, refType=null, unstructuredReference=王熹. 基于YOLOv5 的垃圾溢出检测研究[D]. 武汉: 江汉大学, 2022., articleTitle=基于YOLOv5 的垃圾溢出检测研究, refAbstract=null), Reference(id=1156986794446250005, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=Wang Xi, journalName=null, refType=null, unstructuredReference=Wang Xi. Research on overflow detection of garbage based YOLOv5[D]. Wuhan: Jianghan University, 2022., articleTitle=Research on overflow detection of garbage based YOLOv5, refAbstract=null), Reference(id=1156986794496581654, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2010, volume=33, issue=12, pageStart=2341, pageEnd=2353, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=He K, Sun J, Tang X, journalName=IEEE Transactions on Pattern Analysis and Machine Intelligence, refType=null, unstructuredReference=He K, Sun J, Tang X. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010. 33(12): 2341-2353., articleTitle=Single image haze removal using dark channel prior, refAbstract=null), Reference(id=1156986794542718999, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2022, volume=10, issue=null, pageStart=92726, pageEnd=92739, url=null, language=null, rfNumber=[3], rfOrder=3, authorNames=Guo Q, Wang H, Yang J, journalName=IEEE Access, refType=null, unstructuredReference=Guo Q, Wang H, Yang J. Adaptive dark channel prior enhancement algorithm for different source night visionhalation images[J]. IEEE Access, 2022. 10. 92726-92739., articleTitle=Adaptive dark channel prior enhancement algorithm for different source night visionhalation images, refAbstract=null), Reference(id=1156986794584662040, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=1, pageEnd=6, url=null, language=null, rfNumber=[4], rfOrder=4, authorNames=Ajith A P, Vidyamol K, Devassy BR, journalName=null, refType=null, unstructuredReference=Ajith A P, Vidyamol K, Devassy BR, et al. Dark channel prior based single image dehazing of daylight captures[C]// 2023 Advanced Computing and Communication Technologies for High Performance Applications (ACCTHPA). New York: IEEE, 2023. 1-6., articleTitle=Dark channel prior based single image dehazing of daylight captures, refAbstract=null), Reference(id=1156986794660159513, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=501, pageEnd=506, url=null, language=null, rfNumber=[5], rfOrder=5, authorNames=Samiullah PA, journalName=null, refType=null, unstructuredReference=Samiullah PA. Single image haze removal using dark channel saturation priori model and non-linear diffusion patch method[C]// 2021 International Symposium on Electrical, Electronics and Information Engineering. New York: IEEE, 2021. 501-506., articleTitle=Single image haze removal using dark channel saturation priori model and non-linear diffusion patch method, refAbstract=null), Reference(id=1156986794706296858, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=19, issue=7, pageStart=416, pageEnd=424, url=null, language=null, rfNumber=[6], rfOrder=6, authorNames=Zhang J, Liu H, Ying X, journalName=Optoelectronics Letters, refType=null, unstructuredReference=Zhang J, Liu H, Ying X, et al. Coordinated underwater dark channel prior for artifact removal of challenging image enhancement[J]. Optoelectronics Letters, 2023. 19(7): 416-424., articleTitle=Coordinated underwater dark channel prior for artifact removal of challenging image enhancement, refAbstract=null), Reference(id=1156986794760822811, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=41, issue=3, pageStart=162, pageEnd=172, url=null, language=null, rfNumber=[7], rfOrder=7, authorNames=李玉鑫, 梁天全, 于会山, journalName=计算机仿真, refType=null, unstructuredReference=李玉鑫, 梁天全, 于会山, 等. 融合暗通道与Retinex 算法的水下图像复原研究[J]. 计算机仿真, 2024. 41(3): 162-172., articleTitle=融合暗通道与Retinex 算法的水下图像复原研究, refAbstract=null), Reference(id=1156986794823737373, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=41, issue=3, pageStart=162, pageEnd=172, url=null, language=null, rfNumber=[7], rfOrder=8, authorNames=Li Yuxin, Liang Tianquan, Yu Huishan, journalName=Computer Simulation, refType=null, unstructuredReference=Li Yuxin, Liang Tianquan, Yu Huishan, et al. Research on underwater image restoration combining dark channeland Retinex algorithm[J]. Computer Simulation, 2024. 41(3): 162-172., articleTitle=Research on underwater image restoration combining dark channeland Retinex algorithm, refAbstract=null), Reference(id=1156986794869874719, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2021, volume=15, issue=1, pageStart=016519, pageEnd=null, url=null, language=null, rfNumber=[8], rfOrder=9, authorNames=Zhu H, Xie J, Tang X, journalName=Journal of Applied Remote Sensing, refType=null, unstructuredReference=Zhu H, Xie J, Tang X, et al. Stellar map centroid positioning based on dark channel denoising and feasibility of jitter detection on ZiYuan3 satellite platform[J]. Journal of Applied Remote Sensing, 2021. 15(1): 016519., articleTitle=Stellar map centroid positioning based on dark channel denoising and feasibility of jitter detection on ZiYuan3 satellite platform, refAbstract=null), Reference(id=1156986794957955106, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=38, issue=6, pageStart=36, pageEnd=41, url=null, language=null, rfNumber=[9], rfOrder=10, authorNames=许懿娜, 王义, 黄华平, journalName=遥感信息, refType=null, unstructuredReference=许懿娜, 王义, 黄华平, 等. 一种改进暗通道先验的航空影像快速去雾方法[J]. 遥感信息, 2023. 38(6): 36-41., articleTitle=一种改进暗通道先验的航空影像快速去雾方法, refAbstract=null), Reference(id=1156986795004092452, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=38, issue=6, pageStart=36, pageEnd=41, url=null, language=null, rfNumber=[9], rfOrder=11, authorNames=Xu Yina, Wang Yi, Huang Huaping, journalName=Remote Sensing Information, refType=null, unstructuredReference=Xu Yina, Wang Yi, Huang Huaping, et al. A fast dehaze method for aerial images with improved dark channel prior[J]. Remote Sensing Information, 2023. 38(6): 36-41., articleTitle=A fast dehaze method for aerial images with improved dark channel prior, refAbstract=null), Reference(id=1156986795050229798, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=8, pageStart=115, pageEnd=119, url=null, language=null, rfNumber=[10], rfOrder=12, authorNames=杨卫, 张俊丽, 吴睿, journalName=激光杂志, refType=null, unstructuredReference=杨卫, 张俊丽, 吴睿. 基于暗通道和 Retinex 理论的光斑图像序列增强方法[J]. 激光杂志, 2023. 44(8): 115-119., articleTitle=基于暗通道和 Retinex 理论的光斑图像序列增强方法, refAbstract=null), Reference(id=1156986795092172842, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=8, pageStart=115, pageEnd=119, url=null, language=null, rfNumber=[10], rfOrder=13, authorNames=Yang Wei, Zhang Junli, Wu Rui, journalName=Laser Journal, refType=null, unstructuredReference=Yang Wei, Zhang Junli, Wu Rui. Spot image sequence enhancement method based on dark channel and Retinex theory[J]. Laser Journal, 2023. 44(8): 115-119., articleTitle=Spot image sequence enhancement method based on dark channel and Retinex theory, refAbstract=null), Reference(id=1156986795146698796, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=61, issue=6, pageStart=1407, pageEnd=1418, url=null, language=null, rfNumber=[11], rfOrder=14, authorNames=牛宏侠, 王春智, 梁乐观, journalName=吉林大学学报(理学版), refType=null, unstructuredReference=牛宏侠, 王春智, 梁乐观, 等. 基于改进暗通道先验的沙尘图像清晰化算法[J]. 吉林大学学报(理学版), 2023. 61(6): 1407-1418., articleTitle=基于改进暗通道先验的沙尘图像清晰化算法, refAbstract=null), Reference(id=1156986795213807662, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=61, issue=6, pageStart=1407, pageEnd=1418, url=null, language=null, rfNumber=[11], rfOrder=15, authorNames=Niu Hongxia, Wang Chunzhi, Liang Leguan, journalName=Journal of Jilin University (Science Edition), refType=null, unstructuredReference=Niu Hongxia, Wang Chunzhi, Liang Leguan, et al. Sand dust image clarity algorithm based on improved dark channel priors[J]. Journal of Jilin University (Science Edition), 2023. 61(6): 1407-1418., articleTitle=Sand dust image clarity algorithm based on improved dark channel priors, refAbstract=null), Reference(id=1156986795259945009, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=3, pageStart=12, pageEnd=17, url=null, language=null, rfNumber=[12], rfOrder=16, authorNames=魏清, 陈锦文, journalName=自动化仪表, refType=null, unstructuredReference=魏清, 陈锦文. 基于高分遥感图像处理技术的公路勘查方法研究[J]. 自动化仪表, 2024. 45(3): 12-17., articleTitle=基于高分遥感图像处理技术的公路勘查方法研究, refAbstract=null), Reference(id=1156986795314470962, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=45, issue=3, pageStart=12, pageEnd=17, url=null, language=null, rfNumber=[12], rfOrder=17, authorNames=Wei Qing, Chen Jinwen, journalName=Automation Instrumentation, refType=null, unstructuredReference=Wei Qing, Chen Jinwen. Research on highwaysurvey methods based on high-resolution remote sensing image processing technology[J]. Automation Instrumentation, 2024. 45(3): 12-17., articleTitle=Research on highwaysurvey methods based on high-resolution remote sensing image processing technology, refAbstract=null), Reference(id=1156986795368996915, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=174, issue=null, pageStart=108462, pageEnd=null, url=null, language=null, rfNumber=[13], rfOrder=18, authorNames=Siuly S, Khare SK, Kabir E, journalName=Computers in Biology and Medicine, refType=null, unstructuredReference=Siuly S, Khare SK, Kabir E, et al. An efficient Parkinson's disease detection framework: leveraging time-frequency representation and AlexNet convolutional neural network[J]. Computers in Biology and Medicine, 2024. 174. 108462., articleTitle=An efficient Parkinson's disease detection framework: leveraging time-frequency representation and AlexNet convolutional neural network, refAbstract=null), Reference(id=1156986795444494388, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=12, issue=null, pageStart=null, pageEnd=null, url=10.1109/ACCESS. 2024. 3380839, language=null, rfNumber=[14], rfOrder=19, authorNames=Li Z, Li B, Jahng SG, journalName=IEEE Access, refType=null, unstructuredReference=Li Z, Li B, Jahng SG, et al. Improved VGG algorithm for visual prosthesis image recognition[J]. IEEE Access, 2024. 12., articleTitle=Improved VGG algorithm for visual prosthesis image recognition, refAbstract=null), Reference(id=1156986795499020341, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2016, volume=null, issue=null, pageStart=770, pageEnd=778, url=null, language=null, rfNumber=[15], rfOrder=20, authorNames=He K, Zhang X, Ren S, journalName=null, refType=null, unstructuredReference=He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]// Proceedings of the IEEE Conference on Computer Vision And Pattern Recognition. New York: IEEE, 2016. 770-778., articleTitle=Deep residual learning for image recognition, refAbstract=null), Reference(id=1156986795566129206, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2021, volume=11, issue=6, pageStart=1071, pageEnd=null, url=null, language=null, rfNumber=[16], rfOrder=21, authorNames=Odusami M, Maskeliūnas R, Damaševičius R, journalName=Diagnostics, refType=null, unstructuredReference=Odusami M, Maskeliūnas R, Damaševičius R, et al. Analysis of-features of Alzheimer's disease: detection of early stage from functional brain changes in magnetic resonance images using a finetuned ResNet18 network[J]. Diagnostics, 2021. 11(6): 1071., articleTitle=Analysis of-features of Alzheimer's disease: detection of early stage from functional brain changes in magnetic resonance images using a finetuned ResNet18 network, refAbstract=null), Reference(id=1156986795620655159, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=29, pageStart=12628, pageEnd=12637, url=null, language=null, rfNumber=[17], rfOrder=22, authorNames=刘今子, 杜文颖, 董驰, journalName=科学技术与工程, refType=null, unstructuredReference=刘今子, 杜文颖, 董驰, 等. 基于色彩空间和深度残差网络ResNet-50的复杂岩性油气藏岩石样本智能分类及识别[J]. 科学技术与工程, 2023. 23(29): 12628-12637., articleTitle=基于色彩空间和深度残差网络ResNet-50的复杂岩性油气藏岩石样本智能分类及识别, refAbstract=null), Reference(id=1156986795691958329, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=29, pageStart=12628, pageEnd=12637, url=null, language=null, rfNumber=[17], rfOrder=23, authorNames=Liu Jinzi, Du Wenying, Dong Chi, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Liu Jinzi, Du Wenying, Dong Chi, et al. Intelligent classification and recognition of rock samples based on color space and deep residual network ResNet-50[J]. Science Technology and Engineering, 2023. 23(29): 12628-12637., articleTitle=Intelligent classification and recognition of rock samples based on color space and deep residual network ResNet-50, refAbstract=null), Reference(id=1156986795750678586, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2022, volume=4, issue=null, pageStart=1, pageEnd=13, url=null, language=null, rfNumber=[18], rfOrder=24, authorNames=Zhang Q, journalName=SN Applied Sciences, refType=null, unstructuredReference=Zhang Q. A novel ResNet101 model based on dense dilated convolution for image classification[J]. SN Applied Sciences, 2022. 4. 1-13., articleTitle=A novel ResNet101 model based on dense dilated convolution for image classification, refAbstract=null), Reference(id=1156986795813593147, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=28, pageStart=12101, pageEnd=12107, url=null, language=null, rfNumber=[19], rfOrder=25, authorNames=陈甲华, 王平平, journalName=科学技术与工程, refType=null, unstructuredReference=陈甲华, 王平平. 基于贝叶斯优化与CBAM-ResNet的乏燃料剪切机故障诊断方法[J]. 科学技术与工程, 2023. 23(28): 12101-12107., articleTitle=基于贝叶斯优化与CBAM-ResNet的乏燃料剪切机故障诊断方法, refAbstract=null), Reference(id=1156986795884896316, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=28, pageStart=12101, pageEnd=12107, url=null, language=null, rfNumber=[19], rfOrder=26, authorNames=Chen Jiahua, Wang Pingping, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Chen Jiahua, Wang Pingping. Condition monitoring ofspent fuel shearing tools using bayesian optimized CBAM-ResNet[J]. Science Technology and Engineering, 2023. 23(28): 12101-12107., articleTitle=Condition monitoring ofspent fuel shearing tools using bayesian optimized CBAM-ResNet, refAbstract=null), Reference(id=1156986795943616573, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=568, issue=null, pageStart=127063, pageEnd=null, url=null, language=null, rfNumber=[20], rfOrder=27, authorNames=Su J, Ahmed M, Lu Y, journalName=Neurocomputing, refType=null, unstructuredReference=Su J, Ahmed M, Lu Y, et al. Roformer: enhanced Transformer with rotary position embedding[J]. Neurocomputing, 2024. 568. 127063., articleTitle=Roformer: enhanced Transformer with rotary position embedding, refAbstract=null), Reference(id=1156986796010725438, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2024, volume=28, issue=null, pageStart=null, pageEnd=null, url=10.1007/s11263-024-02034-6, language=null, rfNumber=[21], rfOrder=28, authorNames=Zhang J, Li X, Wang Y, journalName=International Journal of Computer Vision, refType=null, unstructuredReference=Zhang J, Li X, Wang Y, et al. Eatformer: improving vision Transformer inspired by evolutionary algorithm[J]. International Journal of Computer Vision, 2024. 28., articleTitle=Eatformer: improving vision Transformer inspired by evolutionary algorithm, refAbstract=null), Reference(id=1156986796061057088, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=10012, pageEnd=10022, url=null, language=null, rfNumber=[22], rfOrder=29, authorNames=Liu Z, Lin Y, Cao Y, journalName=null, refType=null, unstructuredReference=Liu Z, Lin Y, Cao Y, et al. Swin transformer: Hierarchicalvision transformer using shifted windows[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. New York: IEEE, 2021. 10012-10022., articleTitle=Swin transformer: Hierarchicalvision transformer using shifted windows, refAbstract=null), Reference(id=1156986796119777346, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=11976, pageEnd=11986, url=null, language=null, rfNumber=[23], rfOrder=30, authorNames=Liu Z, Mao H, Wu CY, journalName=null, refType=null, unstructuredReference=Liu Z, Mao H, Wu CY, et al. A convnet for the 2020s[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2022. 11976-11986., articleTitle=A convnet for the 2020s, refAbstract=null)], funds=[Fund(id=1156986794316226579, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, awardId=222103810043, language=CN, fundingSource=河南省级科技研发计划联合基金(应用攻关类)(222103810043), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1156986778998629037, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=1, ext=[AuthorCompanyExt(id=1156986779007017646, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779015406255, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986778998629037, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 河南工业大学 信息科学与工程学院 郑州 450001)]), AuthorCompany(id=1156986779086709426, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, xref=2, ext=[AuthorCompanyExt(id=1156986779090903731, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China), AuthorCompanyExt(id=1156986779099292340, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, companyId=1156986779086709426, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 河南工业大学 人工智能与大数据学院 郑州 450001)])], figs=[ArticleFig(id=1156986792852415488, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 1, caption=Coke overflow detection flowchart, figureFileSmall=nNpVL5+s/Zyez/MjElyb9A==, figureFileBig=BjgopbGvfpHgjxgPGgSQLw==, tableContent=null), ArticleFig(id=1156986792919523328, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图1, caption=焦炭溢出检测流程图, figureFileSmall=nNpVL5+s/Zyez/MjElyb9A==, figureFileBig=BjgopbGvfpHgjxgPGgSQLw==, tableContent=null), ArticleFig(id=1156986792986632193, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 2, caption=Data collection location and original images, figureFileSmall=x9Qq5grnmaMbBcAbuINAXQ==, figureFileBig=F4KyK+Tqptv9i49l2ctaGQ==, tableContent=null), ArticleFig(id=1156986793074712580, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图2, caption=数据采集场地与原始图片, figureFileSmall=x9Qq5grnmaMbBcAbuINAXQ==, figureFileBig=F4KyK+Tqptv9i49l2ctaGQ==, tableContent=null), ArticleFig(id=1156986793150210053, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 3, caption=Visual comparison of extraction of regions of interest and dust removal effects, figureFileSmall=alQS7UxJQNzkeajN2ZFmSw==, figureFileBig=8ULeofOddtfmeLFFveLdVw==, tableContent=null), ArticleFig(id=1156986793217318918, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图3, caption=感兴趣区域的提取与去尘效果的可视化对比, figureFileSmall=alQS7UxJQNzkeajN2ZFmSw==, figureFileBig=8ULeofOddtfmeLFFveLdVw==, tableContent=null), ArticleFig(id=1156986793284427783, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 4, caption=Dark channel prior algorithm process, figureFileSmall=gH7pL9B55vJcF/3vJEvwHw==, figureFileBig=x/gCI1r9jtGkq8144aWqCA==, tableContent=null), ArticleFig(id=1156986793385091080, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图4, caption=暗通道先验算法流程, figureFileSmall=gH7pL9B55vJcF/3vJEvwHw==, figureFileBig=x/gCI1r9jtGkq8144aWqCA==, tableContent=null), ArticleFig(id=1156986793443811337, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 5, caption=ResNet-18 network model diagram, figureFileSmall=21As6KxxXy9/1i1yhUj3VQ==, figureFileBig=L1GVXnn9Fk+cqx/hOifroA==, tableContent=null), ArticleFig(id=1156986793494142986, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图5, caption=ResNet-18 网络模型图, figureFileSmall=21As6KxxXy9/1i1yhUj3VQ==, figureFileBig=L1GVXnn9Fk+cqx/hOifroA==, tableContent=null), ArticleFig(id=1156986793531891723, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 6, caption=Training loss curves of various models, figureFileSmall=H7HWZcrdXt5yKfFIbzx58w==, figureFileBig=X11ewCdlwiH3C1KfwAdtzA==, tableContent=null), ArticleFig(id=1156986793624166412, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图6, caption=不同模型的训练损失下降曲线, figureFileSmall=H7HWZcrdXt5yKfFIbzx58w==, figureFileBig=X11ewCdlwiH3C1KfwAdtzA==, tableContent=null), ArticleFig(id=1156986793687080973, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 7, caption=Loss reduction process of ResNet-18 under different dust removal algorithms, figureFileSmall=d/X9b57v5EdHvDUtFjmANQ==, figureFileBig=icfAziM1ZbsJrM4YMZHWtA==, tableContent=null), ArticleFig(id=1156986793762578446, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图7, caption=ResNet-18 在不同去尘算法处理下的损失下降过程, figureFileSmall=d/X9b57v5EdHvDUtFjmANQ==, figureFileBig=icfAziM1ZbsJrM4YMZHWtA==, tableContent=null), ArticleFig(id=1156986793871630351, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Fig. 8, caption=Speed comparison of dust removal algorithms, figureFileSmall=jeWsdyNXlVxsZYz1MwGNbA==, figureFileBig=JRC+3seQTFSTXRvq75oP6g==, tableContent=null), ArticleFig(id=1156986793938739216, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=图8, caption=不同去尘算法的速度对比, figureFileSmall=jeWsdyNXlVxsZYz1MwGNbA==, figureFileBig=JRC+3seQTFSTXRvq75oP6g==, tableContent=null), ArticleFig(id=1156986793989070865, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=EN, label=Table 1, caption=Test results of different models, figureFileSmall=null, figureFileBig=null, tableContent=
模型 ResNet-18ResNet-50 ResNet-101 Swin- Transformer- Tiny ConvNeXt_ Tiny
准确率 $/\%$ 86.30 83.72 85.89 75.98 81.23
正类精 确度/% 85.59 83.66 84.15 75.87 89.04
负类精 确度 $/\%$ 87.05 83.77 87.81 76.10 76.03
正类召 回率/% 87.43 83.80 88.42 76.19 71.57
负类召 回率/% 85.19 83.63 83.35 75.77 91.23
正类 F1 分数 0.865 0 0.8373 0.8623 0.760 3 0.7935
负类 F1 分数 0.861 1 0.8370 0.8552 0.7593 0.8294
单张图片预测 时间 $/\mathrm{s}$ 0.0163 0.031 0 0.407 9 0.093 2 0.031 0
), ArticleFig(id=1156986794039402514, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1156986627395510894, language=CN, label=表1, caption=不同模型的测试结果, figureFileSmall=null, figureFileBig=null, tableContent=
模型 ResNet-18ResNet-50 ResNet-101 Swin- Transformer- Tiny ConvNeXt_ Tiny
准确率 $/\%$ 86.30 83.72 85.89 75.98 81.23
正类精 确度/% 85.59 83.66 84.15 75.87 89.04
负类精 确度 $/\%$ 87.05 83.77 87.81 76.10 76.03
正类召 回率/% 87.43 83.80 88.42 76.19 71.57
负类召 回率/% 85.19 83.63 83.35 75.77 91.23
正类 F1 分数 0.865 0 0.8373 0.8623 0.760 3 0.7935
负类 F1 分数 0.861 1 0.8370 0.8552 0.7593 0.8294
单张图片预测 时间 $/\mathrm{s}$ 0.0163 0.031 0 0.407 9 0.093 2 0.031 0
)], attaches=null, journal=Journal(id=1146119176004939786, delFlag=0, nameCn=科学技术与工程, nameEn=Science Technology and Engineering, nameHistory1=null, nameHistory2=null, issn=1671-1815, eissn=, cn=11-4688/T, coden=null, periodic=4, language=CN, oaType=是, 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=UKU/O7GSka5polgCTkbIIw==, journalPrice=null, startedYear=null, abbrevIsoEn=Sci Technol Eng, journalRemark=null, publicationField=null, createdTime=null, updatedTime=1754445529766, createdBy=null, updatedBy=13701087609, firstLetterCn=S, firstLetterEn=S, subjectCode=Natural Sciences, subjectName=自然科学, subjectCodeEn=Natural Sciences, subjectNameEn=null, picCn=UKU/O7GSka5polgCTkbIIw==, picEn=5hwlULoNwcbj3xUmVi9MAQ==, jcr=null, cjcr=null, exts=[JournalExt(id=1159791870395564357, language=CN, name=科学技术与工程, 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.stae.com.cn/jsygc/home, createdTime=1754445529793, updatedTime=1754445529793, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=http://www.stae.com.cn/jsygc/site/menus/20090429150146001, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1159791870441701702, language=EN, name=Science Technology and Engineering, 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.stae.com.cn/jsygc/home, createdTime=1754445529804, updatedTime=1754445529804, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://www.stae.com.cn/jsygc/author/login, submissionEditorUrl=http://www.stae.com.cn/jsygc/editor/login, submissionReviewUrl=http://www.stae.com.cn/jsygc/reviewer/login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1146123166801305609, websiteList=[Website(id=1148243202391400884, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, 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/kxjsygc/CN, language=CN, createTime=1751692112777, createBy=18614031015, updateTime=1753520965431, updateBy=18614031015, name=科学技术与工程-中文站点, tplId=1146099689490845704, title=科学技术与工程, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1148622798802673703, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=articleTextType, value=kx, createTime=1751782615614, updateTime=1751782615614, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798781702180, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=banner, value=null, createTime=1751782615609, updateTime=1751782615609, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798769119267, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1751782615606, updateTime=1751782615606, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798794285094, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1751782615612, updateTime=1751782615612, creator=18614031015, updator=18614031015), WebsiteProps(id=1148622798790090789, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1148243202391400884, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1751782615611, updateTime=1751782615611, creator=18614031015, updator=18614031015)]), Website(id=1155914124811976731, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1146123166801305609, 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/kxjsygc/EN, language=EN, createTime=1753521003206, createBy=18614031015, updateTime=1753521003206, updateBy=18614031015, name=科学技术与工程-英文站点, tplId=1146101810881728533, title=Science Technology and Engineering, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1155914371227308235, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=articleTextType, value=kx, createTime=1753521061952, updateTime=1753521061952, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371210531016, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=banner, value=null, createTime=1753521061947, updateTime=1753521061947, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371202142407, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=logo, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic?fileId=j86gbwi+p0Idkyl5SzIlmQ==, createTime=1753521061945, updateTime=1753521061945, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371223113930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/kjdb/CN/file/pic, createTime=1753521061950, updateTime=1753521061950, creator=18614031015, updator=18614031015), WebsiteProps(id=1155914371218919625, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1155914124811976731, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1753521061949, updateTime=1753521061949, creator=18614031015, updator=18614031015)])], journalTitle=科学技术与工程, weixinUrl=null, journalUrl=null, iacademicId=null, status=0, seqNo=null, journalTitleEn=Science Technology and Engineering, journalPhotoCn=UKU/O7GSka5polgCTkbIIw==, journalPhotoEn=5hwlULoNwcbj3xUmVi9MAQ==, journalFirstLetter=S, 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=null, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2402154, detailUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/10.12404/j.issn.1671-1815.2402154, pdfUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/PDF/10.12404/j.issn.1671-1815.2402154, pdfUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/PDF/10.12404/j.issn.1671-1815.2402154, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于暗通道先验知识和ResNet网络的焦炭智能装载溢出检测方法
收藏切换
PDF下载
解康战 1 , 侯惠芳 2, * , 张自豪 2 , 孙文涛 1
科学技术与工程 | 自动化技术、计算机技术 2025,25(8): 3325-3332
收起
收藏切换
科学技术与工程 | 自动化技术、计算机技术 2025, 25(8): 3325-3332
基于暗通道先验知识和ResNet网络的焦炭智能装载溢出检测方法
全屏
解康战1 , 侯惠芳2, * , 张自豪2, 孙文涛1
作者信息
  • 1 河南工业大学 信息科学与工程学院 郑州 450001
  • 2 河南工业大学 人工智能与大数据学院 郑州 450001
  • 解康战(1997—),男,汉族,河南周口人,硕士研究生。研究方向:计算机视觉。E-mail: xkz19971010@163.com。

通讯作者:

*侯惠芳(1972-),女,汉族,河南安阳人,博士,教授。研究方向:计算机视觉。E-mail: houhuifang@haut.edu.cn。
A Coke Intelligent Loading Overflow Detection Method Based on Dark Channel Prior Knowledge and ResNet Network
Kang-zhan XIE1 , Hui-fang HOU2, * , Zi-hao ZHANG2, Wen-tao SUN1
Affiliations
  • 1 School of Information Science and Engineering Henan University of Technology Zhengzhou 450001 China
  • 2 School of Artificial Intelligence and Big Data Henan University of Technology Zhengzhou 450001 China
出版时间: 2025-03-18 doi: 10.12404/j.issn.1671-1815.2402154
文章导航
收藏切换

高粉尘环境下进行精准的焦炭溢出检测是实现焦炭智能装载的重要挑战。针对此问题,提出一种基于暗通道先验知识和ResNet 网络的焦炭智能装载溢出检测方法。首先,利用视频采集器获取焦炭装载场景视频信息,并对原始时间序列视频图像帧进行处理以获得下料口及装载器之间感兴趣区域;其次,提出利用暗通道先验知识方法对感兴趣区域进行处理,提升感兴趣区域中目标区域与无关区域之间对比度,以降低粉尘对后续检测模型的影响。再者,根据焦炭实际装载情况对感兴趣区域进行标注将溢出检测问题转化成二分类。最终,提出利用ResNet 网络建模完成对模型的训练获得训练模型并在新采集焦炭装载过程中进行实验。实验证明所提方法在新的数据上测试结果表现优异,整体准确率达到86. 81%,其中溢出类的精确度、召回率和F1 分数分别为84. 12%、90. 74% 和0. 873 0。并且在使用了暗通道先验算法处理数据后,溢出类的召回率上升了3. 31%。

焦炭智能装载溢出检测  /  暗通道先验知识  /  ResNet 网络

The accurate detection of coke overflow in high-dust environments is a pivotal challenge in achieving intelligent coke loading. A method was proposed to address this issue for the intelligent detection of coke loading overflow, which was based on dark channel prior knowledge and the ResNet network. Firstly, a video collector was used to obtain video information of the coke loading scene, and the original time-series video image frames were processed to obtain the region of interest between the discharge port and the loader. Secondly, the prior knowledge method of dark channels was employed to process the regions of interest. Enhancing the contrast between the target areas and irrelevant areas within the regions of interest, thereby mitigating the effects of dust on subsequent detection models. Moreover, the problem of overflow detection was transformed into a binary classification task by labeling the regions of interest based on the actual loading of coke. Finally, the ResNet network was utilized for modeling, enabling the completion of model training and experimentation during the loading process of newly acquired coke. The experimental results demonstrate that the proposed method exhibits promising performance on new data, achieving an overall accuracy of 86.81%. Specifically, the accuracy, recall, and F1 score for the overflow class are 84. 12%, 90.74%, and 0.8730, respectively. Furthermore, the application of the dark channel prior algorithm in data processing results in a notable increase in the recall rate of the overflow class by 3.31%.

coke intelligent loading overflow detection  /  prior knowledge of dark channels  /  ResNet network
解康战, 侯惠芳, 张自豪, 孙文涛. 基于暗通道先验知识和ResNet网络的焦炭智能装载溢出检测方法. 科学技术与工程, 2025 , 25 (8) : 3325 -3332 . DOI: 10.12404/j.issn.1671-1815.2402154
Kang-zhan XIE, Hui-fang HOU, Zi-hao ZHANG, Wen-tao SUN. A Coke Intelligent Loading Overflow Detection Method Based on Dark Channel Prior Knowledge and ResNet Network[J]. Science Technology and Engineering, 2025 , 25 (8) : 3325 -3332 . DOI: 10.12404/j.issn.1671-1815.2402154
在工业领域, 焦炭是钢铁行业必不可获取的原材料。实现焦炭的智能装载是提高生产效率、降低生产成本的重要途径。实际焦炭智能装载的难点之一是焦炭装载过程中会出现大量粉尘。粉尘会加大数据采集的难度, 降低原始数据中信息价值的占比,同时也严重限制了检测方法中对技术的选择。目前, 焦炭装载过程的溢出检测主要依赖于人工实时关注装载动态来降低焦炭溢出情况的发生, 这种方式存在检测效率低、检测准确率低的问题。然而, 通过查阅相关文献, 目前对于焦炭智能装载方向的研究还存在空白。在相近的研究背景下存在研究范例,王熹等[1] 将智能检测算法用于垃圾箱的溢出检测实现了较好的效果。然而, 受限于焦炭装载场景的高粉尘环境及连续装载情况, 该方法无法满足焦炭装载过程实时溢出检测的需求。因此亟需探索新的焦炭智能装载溢出检测方法。
焦炭装载过程中在下料口与装载器之间将形成较为浓密的焦炭颗粒进而形成浓密的粉尘, 这将严重影响视频采集器对装载情况的清晰采集。因此, 有效解决高粉尘环境对视频信息的影响是实现焦炭高精度溢出检测的关键。He 等[2] 提出暗通道先验知识方法以应对白天和黑夜不同光线下的去雾和优化图像并得到了广泛的应用和推广。在暗通道先验知识方法基础上, Guo 等[3] 提出了一种自适应暗通道先验增强算法, 实验验证了该算法在处理不同夜视光晕场景图像时的有效性和通用性。Ajith 等[4] 使用引导方差加权平均滤波器来临时制作传输图,并生成清晰的图像。该方法的主要关注点是实时和合成日光图片。Samiullah[5] 通过一种新的非线性扩散片方法与暗通道饱和先验模型结合并提出一种简单有效的基于恢复的处雾方法, 可以用于处理照明。在水下环境中也可以使用去雾方法来增强图像质量。Zhang 等[6] 提出了一种协调水下暗通道先方法获得了更高的平均水下图像质量测量分数 (约 2.2748)。李玉鑫等[7] 针对水下图像的细节模糊和色彩失真等问题, 提出一种融合暗通道先验和 Retinex (retinal-cortex theory, 视网膜大脑皮层理论)方法, 得到更清晰的水下图像。在航空和卫星方面也有应用。Zhu 等[8] 基于暗通道去噪验证了星图质心定位及紫原三号卫星平台抖动检测的可行性。许懿娜等[9] 提出一种改进暗通道先验的快速去雾方法,比原始算法效率提升 34~74 倍。实现了对航空影像的快速去雾。在一些恶劣环境和其他一些复杂环境中同样可以使用此算法${}^{\left\lbrack {10}\text{-}{12}\right\rbrack }$。受去雾算法在降低各种环境下噪声影响方面的显著效果启发, 旨在将去雾算法的原理和技术应用于除尘领域,以实现提高原始图片的信息价值占比。
基于传统人工观察的焦炭溢出检测标准, 为智能检测方法提供了丰富的先验知识。如何利用当前网络模型完成检测标准与上述处理后的视频图像特征之间的建模是实现焦炭溢出智能检测的重要步骤。近年来, 大量网络模型得到了广泛的关注及应用。首先是 AlexNet 网络, 该模型使用修正线性单元 (rectified linear unit, Relu) 激活函数、防止过拟合等方法并以此获得 ImageNet 分类大赛冠军。Siuly 等[13] 对其进行改进,使用基于时频表示的 AlexNet 卷积神经网络准确识别与帕金森病相关的复杂脑电模式。然后 VGGNet (visual geometry group network) 在此基础上改进, 它采用重复元素、小卷积核增加了网络深度。Li 等[14] 以该模型作为技术核心,引入果蝇优化算法,为视觉图像分析和视觉假体提供了丰富的理论基础。但随着网络深度的不断增加,硬件难以支持大量运算。He 等[15] 提出使用残差网络(residual neural network, ResNet)来降低网络的复杂度, 同时不损失模型的精度。鉴于该结构能灵活地进行网络层数的增加或减少, 以及与其他技术结合的高效便捷性, 该技术已被广泛应用于多个领域, 展现出卓越的适应性和可扩展性, 成为推动相关领域技术发展的重要力量。Odusami 等[16] 采用 ResNet-18 模型,对大脑磁共振图像中的轻度认知障碍的不同阶段进行了精准分类。刘今子等[17] 提出了一种基于色彩空间和 ResNet-50 的方法, 该方法能够实现对不同颜色类别的岩石样本的精确分类。Zhang[18] 提出了一种基于密集扩张卷积的 ResNet-101 模型, 该模型在医学肝脏肿瘤的分类任务中展现出了优异的性能。陈甲华等[19] 将 CBAM-ResNet (convolutional block attention module-residual network, 卷积块注意力模块残差网络) 模型应用于乏燃料剪切机的故障诊断中,实现了高达 93.67% 的诊断准确率,为乏燃料剪切机的故障预测和维护提供了有效的技术支持。近年来, Transform-er[20] 在自然语言处理领域大放异彩。在计算机视觉领域, 也提出了类似 Transformer 的 VIT (vision Transformer)[21] 结构,它将图片分割成很多规则的块输入网络并在网络中被其他一些类似 Transformer 的模块处理,取得了出色的效果。Swin- Transformer (shift window Transformer)[22] 又在 VIT 基础上进行改进, 提出了移动窗口的方式解决了 VIT 序列过长的问题并减少了计算复杂度。而 ConvNeXt(convolutional neural network next generation)[23] 借鉴 Transformer 结构, 也取得了优秀的效果。
针对高粉尘浓度下焦炭溢出检测困难的问题。提出一种基于暗通道先验知识和 ResNet 网络的焦炭智能装载溢出检测方法。旨在利用暗通道先验知识方法降低粉尘对检测结果的影响, 利用 ResNet 网络对暗通道先验知识处理后的视频信息进行建模并完成训练,最终实现焦炭装载过程中高精度溢出检测。实验结果表明,研究成果为焦炭智能装载系统的发展提供了理论支撑, 具有实际应用价值。
为解决高粉尘浓度环境下焦炭溢出检测困难的问题, 提出一种基于暗通道先验知识和 ResNet 网络的焦炭溢出检测方法。该方法的流程如图1 所示。首先, 利用视频采集器从装载场景中采集带装载场景信息的原始图片;然后,对原始图片进行预处理。预处理过程包括裁剪获得感兴趣区域、去尘处理。将预处理后的图片根据一定的标记规则制作成数据集, 利用该数据集训练 ResNet-18 分类模型网络;最后,将训练好的分类模型用于检测新的带有感兴趣区域信息的图片,得到溢出检测结果。下面部分将对上述步骤进行详细阐述。
采用工业摄像头在焦炭装载场景中获取带有感兴趣区域信息的原始图片。数据采集场地和原始图片和感兴趣区域如图2 所示。
图2(c) 可以看出,原始图片中无关区域相对于感兴趣区域所占面积太大, 这会影响分类模型对特征的提取。为了降低无关区域对最终结果的影响。需要首先对原始图片进行裁剪。以图片左上角为原点,向右为$x$ 轴,向下为$y$ 轴,刻度为像素。在(650,350)位置沿$x$$y$ 轴截取${550}\times {400}$ 的感兴趣区域。裁剪前后如图3 所示。
图2(c) 所示, 在焦炭装载流程中, 粉尘颗粒物的释放将形成显著的视觉干扰。这些颗粒物会导致光线的遮挡和散射, 从而显著降低图像中目标区域与背景区域之间的对比度, 进而在模型训练阶段降低图像信息的质量和价值。为了保障模型训练的有效性和精准度, 必须采用专业的图像处理技术或环境控制措施, 以显著减少粉尘对图像质量的负面影响。粉尘一般是粒径范围为$1 \sim {200\mu }\mathrm{m}$ 的固体小颗粒,而雾是空气中${0.1}\sim {100\mu }\mathrm{m}$ 的小水滴或冰晶组成的水汽凝结物。二者在物理结构上具有相似性。根据这一信息, 利用去雾算法来对裁剪后的图片进行去尘处理。
目前,去雾算法主要有图像增强和图像还原两种方式。在本次研究中, 面临的挑战在于所采集的图像中装载物体产生的粉尘量较大,且由于粉尘与图像中的无关区域在物质形态上相近, 导致两者在色彩上的区分度不够显著。然而, 得益于光线条件的特定影响, 焦炭产生的粉尘与其本体在对比度上表现出一定的差异性。考虑到工业生产对时间效率的高度要求, 经过严格的实验验证, 发现暗通道先验算法在处理此类图像时, 不仅在处理速度上展现出显著的优势, 同时在精度方面也达到了相当高的水平。因此, 为了有效去除图像中的粉尘干扰, 提高图像的清晰度和对比度, 采用暗通道先验算法作为图像处理的主要方法。这一选择旨在确保在维持高处理效率的同时, 实现对粉尘干扰的有效抑制。暗通道先验算法是基于暗通道先验知识, 通过对大量有雾和无雾图像进行统计分析后得出结论: 无雾图像的$\mathrm{R}\text{、}\mathrm{G}\text{、}\mathrm{\;B}$ 三通道一般会有一个值特别低, 而受到雾的影响后, 暗通道会增大, 接近于 127 灰白色。基于此先验理论对有雾图像进行重构,得到清晰图像。暗通道先验算法的计算公式为
$ J\left(k\right)= \mathop{\min }\limits_{{c \in \left\lbrack {\mathrm{R},\mathrm{G},\mathrm{B}}\right\rbrack }}\left\lbrack {\mathop{\min }\limits_{{x \in {w}_{k}}}{J}^{c}\left(x\right)}\right\rbrack \rightarrow 0 $
式(1)中:$J$ 为无雾图像;$k$ 为图像中某个点;$c$$\mathrm{R}$$\mathrm{G}$$\mathrm{B}$ 三通道其中之一;${w}_{k}$ 为某点的邻域窗口;$x$$c$ 的值。
在得到暗通道后, 通过雾气退化模型[式 (2)] 从原始图像中重构去雾后的图像。计算过程如式(3)、式(4),步骤如下。
$ I ={tJ}+ \left({1 - t}\right) A $
$\mathop{\min }\limits_{{c \in \{\mathrm{R},\mathrm{G},\mathrm{B}\}}}\left({\mathop{\min }\limits_{{x \in {w}_{k}}}\frac{{I}^{c}}{{A}^{c}}}\right)= \left\lbrack {\mathop{\min }\limits_{{c \in \left({\mathrm{R},\mathrm{G},\mathrm{B}}\right)}}\left({\mathop{\min }\limits_{{x \in {w}_{k}}}\frac{{J}^{c}}{{A}^{c}}}\right)}\right\rbrack t + 1 - t $
$ t = 1 -\mathop{\min }\limits_{{c \in \{\mathrm{R},\mathrm{G},\mathrm{B}\}}}\left({\mathop{\min }\limits_{{x \in {w}_{k}}}\frac{{I}^{c}}{{A}^{c}}}\right)$
式中:$I$ 为原始拍摄图像;$J$ 为恢复后的图像;$t$ 为图像透射率;$A$ 为全局大气光。
步骤 1 求$I$。想要获取$I$ 需要估计出$A$$t$。全局大气光是场景中没有被任何物体吸收或反射的光线。实际应用中可以通过计算暗通道灰度值最大的像素点对应原图像的像素位置的每个通道的灰度值的平均值来估算。
步骤 2 求$t$。对式 (2) 两边同时除以$A$,然后计算两边暗通道得到式 (3), 将式 (1) 代入式 (3) 得到式 (4)。计算出$t$ 后对其进行细化处理,用$I$ 作引导图进行引导图滤波,使$t$ 更加平滑同时保存边界信息。
步骤 3 计算$J$。结合式 (2) 即可计算$J$,流程如图4 所示。
在得到预处理后的数据, 需要构建用于分类模型训练的数据集。用于训练的第一段装载视频长$7\mathrm{\;{min}}$,视频采集器帧率 30 帧/s。对采集的第一段装载视频时间序列帧剔除装载准备阶段, 然后标记出溢出时间节点, 考虑到连续装载和装载工具的移动,该时间节点的前后$2\mathrm{\;s}$ 标记为溢出,其余为非溢出。模型测试采用另一段长$4\mathrm{\;{min}}{30}\mathrm{\;s}$ 的视频,依据上述方法进行标记,用作测试集。
目前分类网络有很多, 一般在工业应用中对速度的要求比较严格, 因此应该尽可能选择更加小的分类网络。ResNet 因为其著名的残差连接, 可以在相同的训练时间内达到更高的性能水平。相对于传统的卷积神经网络, 它的参数更少并且灵活易扩展, 是一个优秀的轻量化模型, 因此选用 ResNet-18 作为分类网络, 其网络模型如图5 所示。
为了验证本文方法, 采用海康威视工业摄像头对焦炭装载过程进行视频拍摄, 产生的图像分辨率是${1.920}\times {1.080}$。共采集两段视频。用于制作训练集的视频共有 11391 帧,其中溢出的有 2679 帧,非溢出的 8 712 帧。在这些数据中提取出 1 462 帧用作验证集。为了确保验证集中的数据有一定的多样性和代表性, 同时保证训练集和验证集在数据分布上保持一致。采取每隔$n$ 帧提取一帧的方法,溢出类中$n$ 为 4、非溢出类中$n$ 为 11。为了降低类别不平衡对模型训练造成的影响, 使用复制的方法对训练集的溢出类进行过采样。最终训练集共 16000 帧,两类之比为$1 : 1$。测试集视频总共7 400 帧,其中溢出类共 605 帧。对于测试集采用欠采样的方式进行处理, 对非溢出类进行随机删减。最终共得到 1 210 帧,两类之比为$1 : 1$
为了验证不同算法在不同程度的粉尘中的表现。进行 3 组实验,分别取轻微程度、中等程度、重度的粉尘浓度图片,分别对比了基于图像增强的带色彩恢复的多尺度 Retinex 算法、自动色彩均衡算法和基于图像还原的暗通道先验算法。其中在自动色彩均衡算法中对$\mathrm{R}$$\mathrm{G}$$\mathrm{\;B}$ 通道均进行增强,对比度增强因子取 5, 卷积模板半径取 7。3 种方法在不同程度的粉尘浓度下的表现效果如图5 所示。可以看出, 带色彩恢复的多尺度 Retinex 算法处理后清除了表面粉尘,原物体更加清晰。暗通道先验算法、自动色彩均衡算法二者均分别增强了目标区域 (图中亮光区域) 和装载物体的对比度, 但自动色彩均衡算法最大程度上恢复了原始的装载物体颜色、暗通道先验算法调整了目标区域和装载物体的对比度。从装载物体轮廓上来看,暗通道先验算法则不如另外两种。
为了验证不同的分类模型的精度和速度, 使用未除尘的训练集对 ResNet-18、ResNet-50、ResNet-101、Swin-Transformer-Tiny、ConvNeXt-Tiny 模型进行训练。使用 NVIDIA GeForce RTX3060 显卡、CPU 为 12th i7-12700、${16}\mathrm{G}$ 内存。训练参数为轮数 100 、 batch-size 为 8 、采用 Adam 优化器 (Adaptive Moment Estimation,自适应矩估计)、学习率采用 cos 下降方式,初始学习率为$1 \times {10}^{-2}$,最小学习率为初始学习率乘以 0.01 。
不同模型在训练集上的损失下降曲线如图6 所示。在测试集上的结果如表1 所示。根据图和表可知, ResNet-18 的 loss 最终下降到约 0.4, 在 5 个模型中最低。Swin-Transformer-Tiny 最终 loss 下降到约 0.6 , 在 5 个模型中最高、其次是 ConvNeXt-Tiny。这表明残差结构更利于此数据集的目标区域信息的提取。随着 ResNet 层数的增加, 训练效果的趋势出现波动。精度、召回率反正在不同程度的波动。这是因为目标区域对比度信息的变化具有整体性和连续性。在随着感受野的下降和特征被压缩的过小, 反而不利于模型对特征的提取。同时在单张图片预测时间上也是 ResNet-18 更优。本文方法在工业使用过程中,一般比较看重溢出类(正类)召回率, 因为漏报要比误判造成的影响更加严重。综上所述, 选择 ResNet-18 模型作为分类网络。
以 ResNet-18 网络模型为基准。对比不同去尘算法的效果, 对 3 种去尘算法分别进行训练, 参数和硬件保持上述不变。图7 为 ResNet-18 去尘前和 3 种方式去尘后的训练过程的损失变化。可以看出, 去尘前的验证损失收敛后达到 0.2, 而 3 种方式去尘处理后,损失均下降到 0.2 以下。暗通道先验算法和带色彩恢复的多尺度 Retinex 算法损失下降要比自动色彩均衡算法更低。这说明粉尘和装载物体属于相同物体的不同形态, 自动色彩均衡算法恢复了装载物体的原始颜色,这在一定程度上会对训练产生影响。而另外两种方式则增大了二者的对比度。
表2 为 3 种方式在测试集上的结果。可以看出, 经过 3 种算法处理后, 测试集的正类召回率分别上升了${3.31}\%,{1.16}\%,{8.60}\%$。这证明了去雾算法可以适用于高粉尘浓度下的焦炭溢出检测的场景。为了最后选用合适的去尘方法, 随机选取 100 张裁剪后的图片分别使用 3 种去尘算法进行处理, 重复 3 次并计算单张图片处理时间。
图8 为 3 次实验的结果, 可以看出自动色彩均衡算法平均速度约为$3\mathrm{\;s}$、带色彩恢复的多尺度 Ret-inex 算法平均速度约为${1.2}\mathrm{\;s}$、暗通道先验算法平均速度约为${0.1}\mathrm{\;s}$。在速度上暗通道先验算法比其他两种方式快 10~30 倍,而在召回率方面只比带色彩恢复的多尺度 Retinex 算法低约 6%。本文方法一般在工业中使用, 因此采用暗通道先验算法来进行去尘。
为解决焦炭智能装载溢出检测问题, 提出一种基于暗通道先验知识和 ResNet 网络的焦炭智能装载溢出检测方法。得出如下结论。
(1)首先通过裁剪和去尘对原始数据进行预处理, 得到分类数据集。然后利用该数据集对 ResNet-18 分类网络进行训练并在新的数据集上进行测试, 以证明所提方法的有效性。在测试上的整体准确率为 86.81%,其中溢出类的精确度、召回率和 F1 分数分别为${84.12}\%\text{、}{90.74}\%$ 和 0.8730。去尘处理后溢出类的召回率分别上升了 3.31%、1.16%、8.60%。这表明此方法可以适用于焦炭智能装载溢出检测。
(2)由于高粉尘环境的影响光线的强弱、摄像机的角度, 这些都会严重影响最终的结果。实验结果表明, 不同的模型间的差异明显, 在整体准确度上最大相差 5% 以上。从改进分类模型出发,使该模型对于光的强弱改变和位置改变更加灵敏且效果更加稳定。暗通道先验算法是一种图像还原的方式, 速度与相机采集图像的速度相比, 依然差距很大。在实际工业生产中, 只能检测每秒部分帧来增加整体系统运行速度。可以对所提算法进行改进, 使单位时间内处理更多帧,进而增加对噪声的鲁棒性和最终精度。
  • 河南省级科技研发计划联合基金(应用攻关类)(222103810043)
参考文献 引证文献
排序方式:
[1]
王熹. 基于YOLOv5 的垃圾溢出检测研究[D]. 武汉: 江汉大学, 2022.
Wang Xi. Research on overflow detection of garbage based YOLOv5[D]. Wuhan: Jianghan University, 2022.
[2]
He K, Sun J, Tang X. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010. 33(12): 2341-2353.
[3]
Guo Q, Wang H, Yang J. Adaptive dark channel prior enhancement algorithm for different source night visionhalation images[J]. IEEE Access, 2022. 10. 92726-92739.
[4]
Ajith A P, Vidyamol K, Devassy BR, et al. Dark channel prior based single image dehazing of daylight captures[C]// 2023 Advanced Computing and Communication Technologies for High Performance Applications (ACCTHPA). New York: IEEE, 2023. 1-6.
[5]
Samiullah PA. Single image haze removal using dark channel saturation priori model and non-linear diffusion patch method[C]// 2021 International Symposium on Electrical, Electronics and Information Engineering. New York: IEEE, 2021. 501-506.
[6]
Zhang J, Liu H, Ying X, et al. Coordinated underwater dark channel prior for artifact removal of challenging image enhancement[J]. Optoelectronics Letters, 2023. 19(7): 416-424.
[7]
李玉鑫, 梁天全, 于会山, 等. 融合暗通道与Retinex 算法的水下图像复原研究[J]. 计算机仿真, 2024. 41(3): 162-172.
Li Yuxin, Liang Tianquan, Yu Huishan, et al. Research on underwater image restoration combining dark channeland Retinex algorithm[J]. Computer Simulation, 2024. 41(3): 162-172.
[8]
Zhu H, Xie J, Tang X, et al. Stellar map centroid positioning based on dark channel denoising and feasibility of jitter detection on ZiYuan3 satellite platform[J]. Journal of Applied Remote Sensing, 2021. 15(1): 016519.
[9]
许懿娜, 王义, 黄华平, 等. 一种改进暗通道先验的航空影像快速去雾方法[J]. 遥感信息, 2023. 38(6): 36-41.
Xu Yina, Wang Yi, Huang Huaping, et al. A fast dehaze method for aerial images with improved dark channel prior[J]. Remote Sensing Information, 2023. 38(6): 36-41.
[10]
杨卫, 张俊丽, 吴睿. 基于暗通道和 Retinex 理论的光斑图像序列增强方法[J]. 激光杂志, 2023. 44(8): 115-119.
Yang Wei, Zhang Junli, Wu Rui. Spot image sequence enhancement method based on dark channel and Retinex theory[J]. Laser Journal, 2023. 44(8): 115-119.
[11]
牛宏侠, 王春智, 梁乐观, 等. 基于改进暗通道先验的沙尘图像清晰化算法[J]. 吉林大学学报(理学版), 2023. 61(6): 1407-1418.
Niu Hongxia, Wang Chunzhi, Liang Leguan, et al. Sand dust image clarity algorithm based on improved dark channel priors[J]. Journal of Jilin University (Science Edition), 2023. 61(6): 1407-1418.
[12]
魏清, 陈锦文. 基于高分遥感图像处理技术的公路勘查方法研究[J]. 自动化仪表, 2024. 45(3): 12-17.
Wei Qing, Chen Jinwen. Research on highwaysurvey methods based on high-resolution remote sensing image processing technology[J]. Automation Instrumentation, 2024. 45(3): 12-17.
[13]
Siuly S, Khare SK, Kabir E, et al. An efficient Parkinson's disease detection framework: leveraging time-frequency representation and AlexNet convolutional neural network[J]. Computers in Biology and Medicine, 2024. 174. 108462.
[14]
Li Z, Li B, Jahng SG, et al. Improved VGG algorithm for visual prosthesis image recognition[J]. IEEE Access, 2024. 12. 10.1109/ACCESS. 2024. 3380839
[15]
He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]// Proceedings of the IEEE Conference on Computer Vision And Pattern Recognition. New York: IEEE, 2016. 770-778.
[16]
Odusami M, Maskeliūnas R, Damaševičius R, et al. Analysis of-features of Alzheimer's disease: detection of early stage from functional brain changes in magnetic resonance images using a finetuned ResNet18 network[J]. Diagnostics, 2021. 11(6): 1071.
[17]
刘今子, 杜文颖, 董驰, 等. 基于色彩空间和深度残差网络ResNet-50的复杂岩性油气藏岩石样本智能分类及识别[J]. 科学技术与工程, 2023. 23(29): 12628-12637.
Liu Jinzi, Du Wenying, Dong Chi, et al. Intelligent classification and recognition of rock samples based on color space and deep residual network ResNet-50[J]. Science Technology and Engineering, 2023. 23(29): 12628-12637.
[18]
Zhang Q. A novel ResNet101 model based on dense dilated convolution for image classification[J]. SN Applied Sciences, 2022. 4. 1-13.
[19]
陈甲华, 王平平. 基于贝叶斯优化与CBAM-ResNet的乏燃料剪切机故障诊断方法[J]. 科学技术与工程, 2023. 23(28): 12101-12107.
Chen Jiahua, Wang Pingping. Condition monitoring ofspent fuel shearing tools using bayesian optimized CBAM-ResNet[J]. Science Technology and Engineering, 2023. 23(28): 12101-12107.
[20]
Su J, Ahmed M, Lu Y, et al. Roformer: enhanced Transformer with rotary position embedding[J]. Neurocomputing, 2024. 568. 127063.
[21]
Zhang J, Li X, Wang Y, et al. Eatformer: improving vision Transformer inspired by evolutionary algorithm[J]. International Journal of Computer Vision, 2024. 28. 10.1007/s11263-024-02034-6
[22]
Liu Z, Lin Y, Cao Y, et al. Swin transformer: Hierarchicalvision transformer using shifted windows[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. New York: IEEE, 2021. 10012-10022.
[23]
Liu Z, Mao H, Wu CY, et al. A convnet for the 2020s[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. New York: IEEE, 2022. 11976-11986.
2025年第25卷第8期
PDF下载
262
103
引用本文
BibTeX
文章信息
doi: 10.12404/j.issn.1671-1815.2402154
  • 接收时间:2024-03-26
  • 首发时间:2025-07-29
  • 出版时间:2025-03-18
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-03-26
  • 修回日期:2024-12-16
基金
河南省级科技研发计划联合基金(应用攻关类)(222103810043)
作者信息
    1 河南工业大学 信息科学与工程学院 郑州 450001
    2 河南工业大学 人工智能与大数据学院 郑州 450001

通讯作者:

*侯惠芳(1972-),女,汉族,河南安阳人,博士,教授。研究方向:计算机视觉。E-mail: houhuifang@haut.edu.cn。
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2402154
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