Article(id=1228279674149728369, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1228279664221815452, articleNumber=null, orderNo=null, doi=10.12404/j.issn.1671-1815.2404720, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1719158400000, receivedDateStr=2024-06-24, revisedDate=1745683200000, revisedDateStr=2025-04-27, acceptedDate=null, acceptedDateStr=null, onlineDate=1770774294649, onlineDateStr=2026-02-11, pubDate=1754582400000, pubDateStr=2025-08-08, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1770774294649, onlineIssueDateStr=2026-02-11, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1770774294649, creator=13701087609, updateTime=1770774294649, updator=13701087609, issue=Issue{id=1228279664221815452, tenantId=1146029695717560320, journalId=1146123166801305609, year='2025', volume='25', issue='22', pageStart='9211', pageEnd='9648', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=1, specialIssue=null, createTime=1770774292283, creator=13701087609, updateTime=1770777611996, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1228293588207992892, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1228279664221815452, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1228293588207992893, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1228279664221815452, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=9463, endPage=9470, ext={EN=ArticleExt(id=1228279674787262612, articleId=1228279674149728369, tenantId=1146029695717560320, journalId=1146123166801305609, language=EN, title=Three-threshold Image Segmentation of Welding Defects Based on Improved Particle Swarm Algorithm, columnId=1228279665928897192, journalTitle=Science Technology and Engineering, columnName=Papers·Automation and Computational Technology, runingTitle=null, highlight=null, articleAbstract=

To address the issues of distortion and poor segmentation results in weld defect image segmentation, the crack and porosity welding defect images in the rim production process were taken as the research object. An improved particle swarm optimization algorithm based on simulated annealing is proposed for the three-threshold image segmentation of welding defects. First, a three-dimensional Otsu model is constructed using the grayscale value, average grayscale value, and median grayscale value of the image. Next, an adaptive inertia weight and asymmetric learning factor were introduced and integrated into the SA strategy to enhance the algorithm’s solving efficiency and ability to escape local optima. Finally, the SA-IPSO algorithm was used to optimize the three-dimensional Otsu model to obtain the optimal threshold and corresponding defect segmentation image. Various algorithms and models are employed to segment welding defect images. The results show that for crack and porosity defect images, the proposed improved algorithm outperforms the comparison algorithms in terms of peak signal-to-noise ratio and structural similarity evaluation metrics. The proposed method accelerates algorithm convergence while preventing distortion in segmentation results, thereby improving segmentation accuracy.

, correspAuthors=Chao-hua WU, 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=Wei LUO, Chao-hua WU, Jun XIAO, Shu CAI, Xiao-liang SHI), CN=ArticleExt(id=1228279679480689078, articleId=1228279674149728369, tenantId=1146029695717560320, journalId=1146123166801305609, language=CN, title=基于改进粒子群算法的焊接缺陷三阈值图像分割方法, columnId=1228279666075697835, journalTitle=科学技术与工程, columnName=论文·自动化技术、计算机技术, runingTitle=null, highlight=null, articleAbstract=

为解决焊接缺陷图像分割的结果出现失真、分割效果差的问题,以轮辋生产过程中的裂纹和气孔焊接缺陷图像为研究对象,提出了一种基于模拟退火(simulated annealing, SA)策略改进粒子群算法(improved particle swarm optimization, IPSO)的焊接缺陷三阈值图像分割方法。首先通过灰度值、平均灰度值和中值灰度值建立图像的三维最大类间方差(Otsu)模型;其次引入自适应惯性权重和非对称学习因子并融入SA策略增强算法求解效率和跳出局部最优的能力;最后利用SA-IPSO算法优化三维Otsu模型求解得到最佳阈值对应的缺陷分割图像。采用不同算法和模型对焊接缺陷图像进行分割,结果表明:对于裂纹和气孔焊接缺陷图像,本文算法在峰值信噪比(peak signal to noise ratio, PSNR)和结构相似性(structural similarity, SSIM)评价指标上均优于对比算法,在加快算法收敛的同时避免分割结果失真,提高了分割精度。

, correspAuthors=吴超华, authorNote=null, correspAuthorsNote=
* 吴超华(1980—),男,汉族,湖北孝感人,博士,副教授。研究方向:智能装备系统设计与优化,图像分析与识别。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=BZaQ1NbqpRIMzcwgRPAfAg==, magXml=nIXNWmqvw8aMKnsoERe05Q==, pdfUrl=null, pdf=ZOyuaIwN59mfgjiEdSwtjA==, pdfFileSize=5870190, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=RsBVZ7wOrRmQgsahe8/Z9Q==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=fa0buEB6L0o3I5mHEYDSKg==, mapNumber=null, authorCompany=null, fund=null, authors=

罗威(2000—),男,汉族,江西九江人,硕士研究生。研究方向:机器视觉与图像处理技术。E-mail:

, authorsList=罗威, 吴超华, 肖俊, 蔡舒, 史晓亮)}, authors=[Author(id=1228369851501179318, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=luowei332673@whut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1228369851589259709, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369851501179318, language=EN, stringName=Wei LUO, firstName=Wei, middleName=null, lastName=LUO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1228369851723477443, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369851501179318, language=CN, stringName=罗威, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学机电工程学院, 武汉 430070, bio={"content":"

罗威(2000—),男,汉族,江西九江人,硕士研究生。研究方向:机器视觉与图像处理技术。E-mail:

"}, bioImg=null, bioContent=

罗威(2000—),男,汉族,江西九江人,硕士研究生。研究方向:机器视觉与图像处理技术。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])]), Author(id=1228369851920609740, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=wuchaohua@whut.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1228369852054827474, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369851920609740, language=EN, stringName=Chao-hua WU, firstName=Chao-hua, middleName=null, lastName=WU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1228369852142907863, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369851920609740, language=CN, stringName=吴超华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=武汉理工大学机电工程学院, 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])]), Author(id=1228369852222599643, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, 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=1228369852302291426, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369852222599643, language=EN, stringName=Jun XIAO, firstName=Jun, middleName=null, lastName=XIAO, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1228369852394566117, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369852222599643, language=CN, stringName=肖俊, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学机电工程学院, 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])]), Author(id=1228369852503618027, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, 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=1228369852604281327, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369852503618027, language=EN, stringName=Shu CAI, firstName=Shu, middleName=null, lastName=CAI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1228369852704944630, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369852503618027, language=CN, stringName=蔡舒, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学机电工程学院, 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])]), Author(id=1228369854084870655, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, 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=1228369854181339652, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369854084870655, language=EN, stringName=Xiao-liang SHI, firstName=Xiao-liang, middleName=null, lastName=SHI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1228369854290391563, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, authorId=1228369854084870655, language=CN, stringName=史晓亮, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=武汉理工大学机电工程学院, 武汉 430070, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])])], keywords=[Keyword(id=1228369854508495384, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, orderNo=1, keyword=threshold segmentation), Keyword(id=1228369854600770084, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, orderNo=2, keyword=three-dimensional Otsu), Keyword(id=1228369854705627687, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, orderNo=3, keyword=particle swarm optimization algorithm), Keyword(id=1228369854802096686, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, orderNo=4, keyword=simulated annealing strategy), Keyword(id=1228369854881788468, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, orderNo=5, keyword=welding defects), Keyword(id=1228369854961480252, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, orderNo=1, keyword=阈值分割), Keyword(id=1228369855032783428, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, orderNo=2, keyword=三维Otsu), Keyword(id=1228369855120863819, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, orderNo=3, keyword=粒子群优化算法), Keyword(id=1228369855234110031, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, orderNo=4, keyword=模拟退火策略), Keyword(id=1228369855334773334, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, orderNo=5, keyword=焊接缺陷)], refs=[Reference(id=1228369858857988848, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2019, volume=48, issue=7, pageStart=91, pageEnd=93, url=null, language=null, rfNumber=[1], rfOrder=0, authorNames=陈楠, 贺春, 张楠, journalName=焊接技术, refType=null, unstructuredReference=陈楠, 贺春, 张楠, 等. X80钢螺旋焊管主要焊接缺陷分析及预控措施[J]. 焊接技术, 2019, 48(7): 91-93., articleTitle=X80钢螺旋焊管主要焊接缺陷分析及预控措施, refAbstract=null), Reference(id=1228369858983817977, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2019, volume=48, issue=7, pageStart=91, pageEnd=93, url=null, language=null, rfNumber=[1], rfOrder=1, authorNames=Chen Nan, He Chun, Zhang Nan, journalName=Welding Technology, refType=null, unstructuredReference=Chen Nan, He Chun, Zhang Nan, et al. Analysis of main weld defect of X80 spiral welded pipe and pre-control measures[J]. Welding Technology, 2019, 48(7): 91-93., articleTitle=Analysis of main weld defect of X80 spiral welded pipe and pre-control measures, refAbstract=null), Reference(id=1228369859088675584, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=52, issue=6, pageStart=39, pageEnd=47, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=朱子洵, 张洁, 汪俊亮, journalName=华中科技大学学报(自然科学版), refType=null, unstructuredReference=朱子洵, 张洁, 汪俊亮. 基于显著性检测的坯布疵点图像自适应分割方法[J]. 华中科技大学学报(自然科学版), 2024, 52(6): 39-47., articleTitle=基于显著性检测的坯布疵点图像自适应分割方法, refAbstract=null), Reference(id=1228369859189338886, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=52, issue=6, pageStart=39, pageEnd=47, url=null, language=null, rfNumber=[2], rfOrder=3, authorNames=Zhu Zixun, Zhang Jie, Wang Junliang, journalName=Journal of Huazhong University of Science and Technology(Natural Science Edition), refType=null, unstructuredReference=Zhu Zixun, Zhang Jie, Wang Junliang. Adaptive segmentation method of grey fabric defect image based on saliency detection[J]. Journal of Huazhong University of Science and Technology(Natural Science Edition), 2024, 52(6): 39-47., articleTitle=Adaptive segmentation method of grey fabric defect image based on saliency detection, refAbstract=null), Reference(id=1228369859264836362, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=20, pageStart=175, pageEnd=183, url=null, language=null, rfNumber=[3], rfOrder=4, authorNames=陈浩然, 范方辉, 牟天, journalName=食品研究与开发, refType=null, unstructuredReference=陈浩然, 范方辉, 牟天. 基于OTSU图像分割算法的碎米检测[J]. 食品研究与开发, 2023, 44(20): 175-183, 8., articleTitle=基于OTSU图像分割算法的碎米检测, refAbstract=null), Reference(id=1228369859394859792, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=20, pageStart=175, pageEnd=183, url=null, language=null, rfNumber=[3], rfOrder=5, authorNames=Chen Haoran, Fan Fanghui, Mou Tian, journalName=Food Research and Development, refType=null, unstructuredReference=Chen Haoran, Fan Fanghui, Mou Tian. Broken rice detection base on OTSU image segmentation algorithm[J]. Food Research and Development, 2023, 44(20): 175-183, 8., articleTitle=Broken rice detection base on OTSU image segmentation algorithm, refAbstract=null), Reference(id=1228369859487134487, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=1, pageStart=232, pageEnd=239, url=null, language=null, rfNumber=[4], rfOrder=6, authorNames=林宁宁, 高心丹, journalName=计算机工程与设计, refType=null, unstructuredReference=林宁宁, 高心丹. DeepLab V3+改进的树木图像分割[J]. 计算机工程与设计, 2023, 44(1): 232-239., articleTitle=DeepLab V3+改进的树木图像分割, refAbstract=null), Reference(id=1228369859575214878, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=44, issue=1, pageStart=232, pageEnd=239, url=null, language=null, rfNumber=[4], rfOrder=7, authorNames=Lin Ningning, Gao Xindan, journalName=Computer Engineering and Design, refType=null, unstructuredReference=Lin Ningning, Gao Xindan. Improved trees image segmentation based on DeepLab V3+[J]. Computer Engineering and Design, 2023, 44(1): 232-239., articleTitle=Improved trees image segmentation based on DeepLab V3+, refAbstract=null), Reference(id=1228369859709432612, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=12, issue=5, pageStart=75, pageEnd=81, url=null, language=null, rfNumber=[5], rfOrder=8, authorNames=田根源, journalName=蚌埠学院学报, refType=null, unstructuredReference=田根源. 改进黏菌算法优化图像熵的图像分割[J]. 蚌埠学院学报, 2023, 12(5): 75-81., articleTitle=改进黏菌算法优化图像熵的图像分割, refAbstract=null), Reference(id=1228369859801707308, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=12, issue=5, pageStart=75, pageEnd=81, url=null, language=null, rfNumber=[5], rfOrder=9, authorNames=Tian Genyuan, journalName=Journal of Bengbu University, refType=null, unstructuredReference=Tian Genyuan. Image segmentation based on improved slime mould algorithm optimizing image entropy[J]. Journal of Bengbu University, 2023, 12(5): 75-81., articleTitle=Image segmentation based on improved slime mould algorithm optimizing image entropy, refAbstract=null), Reference(id=1228369859910759216, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=16, pageStart=6983, pageEnd=6992, url=null, language=null, rfNumber=[6], rfOrder=10, authorNames=马小晶, 贺航, 王宏伟, journalName=科学技术与工程, refType=null, unstructuredReference=马小晶, 贺航, 王宏伟, 等. 基于改进麻雀搜索算法的最大指数熵分割方法[J]. 科学技术与工程, 2023, 23(16): 6983-6992., articleTitle=基于改进麻雀搜索算法的最大指数熵分割方法, refAbstract=null), Reference(id=1228369860028199737, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=23, issue=16, pageStart=6983, pageEnd=6992, url=null, language=null, rfNumber=[6], rfOrder=11, authorNames=Ma Xiaojing, He Hang, Wang Hongwei, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Ma Xiaojing, He Hang, Wang Hongwei, et al. Maximum exponential entropy segmentation method based on improved sparrow search algorithm[J]. Science Technology and Engineering, 2023, 23(16): 6983-6992., articleTitle=Maximum exponential entropy segmentation method based on improved sparrow search algorithm, refAbstract=null), Reference(id=1228369860124668738, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=24, issue=12, pageStart=5059, pageEnd=5068, url=null, language=null, rfNumber=[7], rfOrder=12, authorNames=贺航, 许连杰, 李高源, journalName=科学技术与工程, refType=null, unstructuredReference=贺航, 许连杰, 李高源, 等. 基于改进蜉蝣算法优化多阈值图像分割[J]. 科学技术与工程, 2024, 24(12): 5059-5068., articleTitle=基于改进蜉蝣算法优化多阈值图像分割, refAbstract=null), Reference(id=1228369860271469386, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=24, issue=12, pageStart=5059, pageEnd=5068, url=null, language=null, rfNumber=[7], rfOrder=13, authorNames=He Hang, Xu Lianjie, Li Gaoyuan, journalName=Science Technology and Engineering, refType=null, unstructuredReference=He Hang, Xu Lianjie, Li Gaoyuan, et al. Optimized multi-thresho-ld image segmentation based on improved mayfly algorithm[J]. Science Technology and Engineering, 2024, 24(12): 5059-5068., articleTitle=Optimized multi-thresho-ld image segmentation based on improved mayfly algorithm, refAbstract=null), Reference(id=1228369860376326989, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=8, issue=2, pageStart=87, pageEnd=111, url=null, language=null, rfNumber=[8], rfOrder=14, authorNames=Fan Q, Ma Y, Wang P, journalName=Fractal and Fractional, refType=null, unstructuredReference=Fan Q, Ma Y, Wang P, et al. Otsu image segmentation based on a fractional order moth-flame optimization algorithm[J]. Fractal and Fractional, 2024, 8(2): 87-111., articleTitle=Otsu image segmentation based on a fractional order moth-flame optimization algorithm, refAbstract=null), Reference(id=1228369860489573206, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2024, volume=12, issue=6, pageStart=854, pageEnd=872, url=null, language=null, rfNumber=[9], rfOrder=15, authorNames=Dong Y, Li M, Zhou M, journalName=Mathematics, refType=null, unstructuredReference=Dong Y, Li M, Zhou M. Multi-threshold image segmentation based on the improved dragonfly algorithm[J]. Mathematics, 2024, 12(6): 854-872., articleTitle=Multi-threshold image segmentation based on the improved dragonfly algorithm, refAbstract=null), Reference(id=1228369860627985245, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2017, volume=47, issue=1, pageStart=255, pageEnd=261, url=null, language=null, rfNumber=[10], rfOrder=16, authorNames=肖明尧, 李雄飞, journalName=吉林大学学报(工学版), refType=null, unstructuredReference=肖明尧, 李雄飞. 基于高斯分解的多尺度3D Otsu阈值分割算法[J]. 吉林大学学报(工学版), 2017, 47(1): 255-261., articleTitle=基于高斯分解的多尺度3D Otsu阈值分割算法, refAbstract=null), Reference(id=1228369860732842853, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2017, volume=47, issue=1, pageStart=255, pageEnd=261, url=null, language=null, rfNumber=[10], rfOrder=17, authorNames=Xiao Mingyao, Li Xiongfei, journalName=Journal of Jinlin University(Engineering and Technology Edition), refType=null, unstructuredReference=Xiao Mingyao, Li Xiongfei. Multi-scale 3D Otsu thresholding algorithm based on gaussion decomposition[J]. Journal of Jinlin University(Engineering and Technology Edition), 2017, 47(1): 255-261., articleTitle=Multi-scale 3D Otsu thresholding algorithm based on gaussion decomposition, refAbstract=null), Reference(id=1228369860820923245, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2021, volume=80, issue=13, pageStart=19399, pageEnd=19420, url=null, language=null, rfNumber=[11], rfOrder=18, authorNames=Neha S, Kumar A B, Vinod I K, journalName=Multimedia Tools and Applications, refType=null, unstructuredReference=Neha S, Kumar A B, Vinod I K. Fusion-based contextually selected 3D Otsu thresholding for image segmentation[J]. Multimedia Tools and Applications, 2021, 80(13): 19399-19420., articleTitle=Fusion-based contextually selected 3D Otsu thresholding for image segmentation, refAbstract=null), Reference(id=1228369860934169459, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2020, volume=30, issue=4, pageStart=94, pageEnd=99, url=null, language=null, rfNumber=[12], rfOrder=19, authorNames=曹爽, 贺玉珍, 安建成, journalName=计算机技术与发展, refType=null, unstructuredReference=曹爽, 贺玉珍, 安建成. 基于改进狼群算法的三维OTSU阈值法[J]. 计算机技术与发展, 2020, 30(4): 94-99., articleTitle=基于改进狼群算法的三维OTSU阈值法, refAbstract=null), Reference(id=1228369861051609975, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2020, volume=30, issue=4, pageStart=94, pageEnd=99, url=null, language=null, rfNumber=[12], rfOrder=20, authorNames=Cao Shuang, He Yuzhen, An Jiancheng, journalName=Computer Technology and Development, refType=null, unstructuredReference=Cao Shuang, He Yuzhen, An Jiancheng. A three-dimension OTSU threshold algorithm based on improved wolf pack algorithm[J]. Computer Technology and Development, 2020, 30(4): 94-99., articleTitle=A three-dimension OTSU threshold algorithm based on improved wolf pack algorithm, refAbstract=null), Reference(id=1228369861173244799, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=40, issue=8, pageStart=1202, pageEnd=1209, url=null, language=null, rfNumber=[13], rfOrder=21, authorNames=刘文长, 吴超华, 陈兴安, journalName=机电工程, refType=null, unstructuredReference=刘文长, 吴超华, 陈兴安. 双立柱堆垛机立柱挠度模型及其参数优化[J]. 机电工程, 2023, 40(8): 1202-1209., articleTitle=双立柱堆垛机立柱挠度模型及其参数优化, refAbstract=null), Reference(id=1228369861286491013, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2023, volume=40, issue=8, pageStart=1202, pageEnd=1209, url=null, language=null, rfNumber=[13], rfOrder=22, authorNames=Liu Wenchang, Wu Chaohua, Chen Xingan, journalName=Journal of Mechanical & Electrical Engineering, refType=null, unstructuredReference=Liu Wenchang, Wu Chaohua, Chen Xingan. Vertical column deflection model and its parameter optimization of double-column stacker[J]. Journal of Mechanical & Electrical Engineering, 2023, 40(8): 1202-1209., articleTitle=Vertical column deflection model and its parameter optimization of double-column stacker, refAbstract=null), Reference(id=1228369861403931531, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2016, volume=21, issue=3, pageStart=49, pageEnd=53, url=null, language=null, rfNumber=[14], rfOrder=23, authorNames=于桂芹, 李刘东, 袁永峰, journalName=哈尔滨理工大学学报, refType=null, unstructuredReference=于桂芹, 李刘东, 袁永峰. 一种结合自适应惯性权重的混合粒子群算法[J]. 哈尔滨理工大学学报, 2016, 21(3): 49-53., articleTitle=一种结合自适应惯性权重的混合粒子群算法, refAbstract=null), Reference(id=1228369861512983440, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2016, volume=21, issue=3, pageStart=49, pageEnd=53, url=null, language=null, rfNumber=[14], rfOrder=24, authorNames=Yu Guiqin, Li Liudong, Yuan Yongfeng, journalName=Journal of Haebin University of Science and Technology, refType=null, unstructuredReference=Yu Guiqin, Li Liudong, Yuan Yongfeng. A hybrid particle swarm optimization algorithm with adaptive inertia weight[J]. Journal of Haebin University of Science and Technology, 2016, 21(3): 49-53., articleTitle=A hybrid particle swarm optimization algorithm with adaptive inertia weight, refAbstract=null), Reference(id=1228369861634618262, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2021, volume=35, issue=1, pageStart=91, pageEnd=96, url=null, language=null, rfNumber=[15], rfOrder=25, authorNames=吴永红, 曾志高, 邓彬, journalName=湖南工业大学学报, refType=null, unstructuredReference=吴永红, 曾志高, 邓彬. 基于惯性权重和学习因子动态调整的粒子群算法[J]. 湖南工业大学学报, 2021, 35(1): 91-96., articleTitle=基于惯性权重和学习因子动态调整的粒子群算法, refAbstract=null), Reference(id=1228369863039710106, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2021, volume=35, issue=1, pageStart=91, pageEnd=96, url=null, language=null, rfNumber=[15], rfOrder=26, authorNames=Wu Yonghong, Zeng Zhigao, Deng Bin, journalName=Journal of Hunan University of Technology, refType=null, unstructuredReference=Wu Yonghong, Zeng Zhigao, Deng Bin. Particle swarm optimization algorithm based on dynamic adjustment of inertial weight and learning factors[J]. Journal of Hunan University of Technology, 2021, 35(1): 91-96., articleTitle=Particle swarm optimization algorithm based on dynamic adjustment of inertial weight and learning factors, refAbstract=null), Reference(id=1228369863178122147, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2021, volume=48, issue=4, pageStart=120, pageEnd=127, url=null, language=null, rfNumber=[16], rfOrder=27, authorNames=闫群民, 马瑞卿, 马永翔, journalName=西安电子科技大学学报, refType=null, unstructuredReference=闫群民, 马瑞卿, 马永翔, 等. 一种自适应模拟退火粒子群优化算法[J]. 西安电子科技大学学报, 2021, 48(4): 120-127., articleTitle=一种自适应模拟退火粒子群优化算法, refAbstract=null), Reference(id=1228369863282979750, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2021, volume=48, issue=4, pageStart=120, pageEnd=127, url=null, language=null, rfNumber=[16], rfOrder=28, authorNames=Yan Qunmin, Ma Ruiqing, Ma Yongxiang, journalName=Journal of Xidian University, refType=null, unstructuredReference=Yan Qunmin, Ma Ruiqing, Ma Yongxiang, et al. Adaptive simulated annealing particle swarm optimization algorithm[J]. Journal of Xidian University, 2021, 48(4): 120-127., articleTitle=Adaptive simulated annealing particle swarm optimization algorithm, refAbstract=null), Reference(id=1228369863421391787, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, doi=null, pmid=null, pmcid=null, year=2004, volume=13, issue=4, pageStart=600, pageEnd=612, url=null, language=null, rfNumber=[17], rfOrder=29, authorNames=Zhou W, Conrad A B, Rahim H S, journalName=IEEE Transactions on Image Processing: A Publication of the IEEE Signal Processing Society, refType=null, unstructuredReference=Zhou W, Conrad A B, Rahim H S, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing: A Publication of the IEEE Signal Processing Society, 2004, 13(4): 600-612., articleTitle=Image quality assessment: from error visibility to structural similarity, refAbstract=null)], funds=[Fund(id=1228369857180267223, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, awardId=52375201, language=CN, fundingSource=国家自然科学基金(52375201), fundOrder=null, country=null), Fund(id=1228369858568581855, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, awardId=20231h0544, language=CN, fundingSource=武汉理工大学产学研科技合作项(20231h0544), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1228369851387933102, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, xref=null, ext=[AuthorCompanyExt(id=1228369851396321712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China), AuthorCompanyExt(id=1228369851408904625, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, companyId=1228369851387933102, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=武汉理工大学机电工程学院, 武汉 430070)])], figs=[ArticleFig(id=1228369855536099938, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.1, caption=Classic Lena image and its histogram, figureFileSmall=8HK9BSJLe6Sm9Wd4lLzMpg==, figureFileBig=ZYen5g0tRXLs/WinpaSjVw==, tableContent=null), ArticleFig(id=1228369855607403108, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图1, caption=经典Lena图及直方图, figureFileSmall=8HK9BSJLe6Sm9Wd4lLzMpg==, figureFileBig=ZYen5g0tRXLs/WinpaSjVw==, tableContent=null), ArticleFig(id=1228369855754203759, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.2, caption=SA-IPSO algorithm optimization flowchart for three-dimensional Otsu threshold segmentation, figureFileSmall=yj2kZgZH5P9lUIK3G8kkQQ==, figureFileBig=iYWf/ipbcsvWaja+WG9H0A==, tableContent=null), ArticleFig(id=1228369855854867064, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图2, caption=SA-IPSO算法优化三维Otsu阈值分割流程图, figureFileSmall=yj2kZgZH5P9lUIK3G8kkQQ==, figureFileBig=iYWf/ipbcsvWaja+WG9H0A==, tableContent=null), ArticleFig(id=1228369855938753153, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.3, caption=Threshold segmentation of crack image, figureFileSmall=SIdv4P0qZhx+ZvMNIWkPxg==, figureFileBig=AeywfqCP8PPGkIJuPqJYjg==, tableContent=null), ArticleFig(id=1228369856039416459, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图3, caption=裂纹图像阈值分割, figureFileSmall=SIdv4P0qZhx+ZvMNIWkPxg==, figureFileBig=AeywfqCP8PPGkIJuPqJYjg==, tableContent=null), ArticleFig(id=1228369856127496850, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.4, caption=Threshold segmentation of pore image, figureFileSmall=qNVOt/9ksg/4xX6MEwtOmg==, figureFileBig=pk8vtOK+ia/HZwQ9XPcCWw==, tableContent=null), ArticleFig(id=1228369856215577243, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图4, caption=气孔图像阈值分割, figureFileSmall=qNVOt/9ksg/4xX6MEwtOmg==, figureFileBig=pk8vtOK+ia/HZwQ9XPcCWw==, tableContent=null), ArticleFig(id=1228369856295269025, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.5, caption=Iteration convergence curve of crack image segmentation, figureFileSmall=02+HCTmP6Yw0NvphB2gFMA==, figureFileBig=DVH7Hh2BITJtflCS1Z8y/A==, tableContent=null), ArticleFig(id=1228369856395932329, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图5, caption=裂纹图像分割迭代收敛曲线, figureFileSmall=02+HCTmP6Yw0NvphB2gFMA==, figureFileBig=DVH7Hh2BITJtflCS1Z8y/A==, tableContent=null), ArticleFig(id=1228369856484012720, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Fig.6, caption=Iteration convergence curve of pore image segmentation, figureFileSmall=DxIuC9T4NG2y+D03cotD7Q==, figureFileBig=Hp+dSRkUXVvlzGiuaUcB5w==, tableContent=null), ArticleFig(id=1228369856576287417, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=图6, caption=气孔图像分割迭代收敛曲线, figureFileSmall=DxIuC9T4NG2y+D03cotD7Q==, figureFileBig=Hp+dSRkUXVvlzGiuaUcB5w==, tableContent=null), ArticleFig(id=1228369856676950717, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Table 1, caption=

Segmentation evaluation metrics for each algorithm

, figureFileSmall=null, figureFileBig=null, tableContent=
图像 算法 PSNR SSIM
PSO+3D 29.158 2 0.345 1
裂纹 SA-IPSO+2D 31.091 4 0.365 6
SA-IPSO+3D 33.698 3 0.374 2
PSO+3D 15.981 6 0.258 3
气孔 SA-IPSO+2D 16.454 2 0.282 6
SA-IPSO+3D 23.872 7 0.337 4
), ArticleFig(id=1228369856773419714, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=表1, caption=

各算法分割评价指标

, figureFileSmall=null, figureFileBig=null, tableContent=
图像 算法 PSNR SSIM
PSO+3D 29.158 2 0.345 1
裂纹 SA-IPSO+2D 31.091 4 0.365 6
SA-IPSO+3D 33.698 3 0.374 2
PSO+3D 15.981 6 0.258 3
气孔 SA-IPSO+2D 16.454 2 0.282 6
SA-IPSO+3D 23.872 7 0.337 4
), ArticleFig(id=1228369856899248839, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=EN, label=Table 2, caption=

Optimization results of each algorithm

, figureFileSmall=null, figureFileBig=null, tableContent=
图像 算法 最佳阈值 最优迭代
次数
适应
度值
PSO+3D 26、91、171 14 166.83
裂纹 SA-IPSO+2D 69、159 8 167.34
SA-IPSO+3D 30、100、163 10 167.92
PSO+3D 20、74、186 17 303.14
气孔 SA-IPSO+2D 25、135 9 304.63
SA-IPSO+3D 25、80、182 12 306.35
), ArticleFig(id=1228369856999912142, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1228279674149728369, language=CN, label=表2, caption=

各算法优化结果

, figureFileSmall=null, figureFileBig=null, tableContent=
图像 算法 最佳阈值 最优迭代
次数
适应
度值
PSO+3D 26、91、171 14 166.83
裂纹 SA-IPSO+2D 69、159 8 167.34
SA-IPSO+3D 30、100、163 10 167.92
PSO+3D 20、74、186 17 303.14
气孔 SA-IPSO+2D 25、135 9 304.63
SA-IPSO+3D 25、80、182 12 306.35
)], 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.1671-1815.2404720, detailUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/10.12404/j.issn.1671-1815.2404720, pdfUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/PDF/10.12404/j.issn.1671-1815.2404720, pdfUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/PDF/10.12404/j.issn.1671-1815.2404720, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于改进粒子群算法的焊接缺陷三阈值图像分割方法
收藏切换
PDF下载
罗威 , 吴超华 * , 肖俊 , 蔡舒 , 史晓亮
科学技术与工程 | 论文·自动化技术、计算机技术 2025,25(22): 9463-9470
收起
收藏切换
科学技术与工程 | 论文·自动化技术、计算机技术 2025, 25(22): 9463-9470
基于改进粒子群算法的焊接缺陷三阈值图像分割方法
全屏
罗威 , 吴超华* , 肖俊, 蔡舒, 史晓亮
作者信息
  • 武汉理工大学机电工程学院, 武汉 430070
  • 罗威(2000—),男,汉族,江西九江人,硕士研究生。研究方向:机器视觉与图像处理技术。E-mail:

通讯作者:

* 吴超华(1980—),男,汉族,湖北孝感人,博士,副教授。研究方向:智能装备系统设计与优化,图像分析与识别。E-mail:
Three-threshold Image Segmentation of Welding Defects Based on Improved Particle Swarm Algorithm
Wei LUO , Chao-hua WU* , Jun XIAO, Shu CAI, Xiao-liang SHI
Affiliations
  • School of Mechanicaland Electronic Engineering, Wuhan University of Technology, Wuhan 430070, China
出版时间: 2025-08-08 doi: 10.12404/j.issn.1671-1815.2404720
文章导航
收藏切换

为解决焊接缺陷图像分割的结果出现失真、分割效果差的问题,以轮辋生产过程中的裂纹和气孔焊接缺陷图像为研究对象,提出了一种基于模拟退火(simulated annealing, SA)策略改进粒子群算法(improved particle swarm optimization, IPSO)的焊接缺陷三阈值图像分割方法。首先通过灰度值、平均灰度值和中值灰度值建立图像的三维最大类间方差(Otsu)模型;其次引入自适应惯性权重和非对称学习因子并融入SA策略增强算法求解效率和跳出局部最优的能力;最后利用SA-IPSO算法优化三维Otsu模型求解得到最佳阈值对应的缺陷分割图像。采用不同算法和模型对焊接缺陷图像进行分割,结果表明:对于裂纹和气孔焊接缺陷图像,本文算法在峰值信噪比(peak signal to noise ratio, PSNR)和结构相似性(structural similarity, SSIM)评价指标上均优于对比算法,在加快算法收敛的同时避免分割结果失真,提高了分割精度。

阈值分割  /  三维Otsu  /  粒子群优化算法  /  模拟退火策略  /  焊接缺陷

To address the issues of distortion and poor segmentation results in weld defect image segmentation, the crack and porosity welding defect images in the rim production process were taken as the research object. An improved particle swarm optimization algorithm based on simulated annealing is proposed for the three-threshold image segmentation of welding defects. First, a three-dimensional Otsu model is constructed using the grayscale value, average grayscale value, and median grayscale value of the image. Next, an adaptive inertia weight and asymmetric learning factor were introduced and integrated into the SA strategy to enhance the algorithm’s solving efficiency and ability to escape local optima. Finally, the SA-IPSO algorithm was used to optimize the three-dimensional Otsu model to obtain the optimal threshold and corresponding defect segmentation image. Various algorithms and models are employed to segment welding defect images. The results show that for crack and porosity defect images, the proposed improved algorithm outperforms the comparison algorithms in terms of peak signal-to-noise ratio and structural similarity evaluation metrics. The proposed method accelerates algorithm convergence while preventing distortion in segmentation results, thereby improving segmentation accuracy.

threshold segmentation  /  three-dimensional Otsu  /  particle swarm optimization algorithm  /  simulated annealing strategy  /  welding defects
罗威, 吴超华, 肖俊, 蔡舒, 史晓亮. 基于改进粒子群算法的焊接缺陷三阈值图像分割方法. 科学技术与工程, 2025 , 25 (22) : 9463 -9470 . DOI: 10.12404/j.issn.1671-1815.2404720
Wei LUO, Chao-hua WU, Jun XIAO, Shu CAI, Xiao-liang SHI. Three-threshold Image Segmentation of Welding Defects Based on Improved Particle Swarm Algorithm[J]. Science Technology and Engineering, 2025 , 25 (22) : 9463 -9470 . DOI: 10.12404/j.issn.1671-1815.2404720
工件在焊接过程中容易受到环境、焊材、焊接参数的影响,产生表面裂纹、表面气孔、夹杂、未熔合等焊接缺陷[1]。焊接缺陷检测在工件加工的质量控制中起着至关重要的作用。传统的检测方法多为人工目测检查,效率较低且检测质量无法得到保证。通过机器视觉技术检测焊接缺陷一直是焊接领域的热门问题。在焊接缺陷检测任务中,阈值分割技术通过分析缺陷图像的灰度特征,分割出缺陷边缘,进而将焊接缺陷区域从图像背景中提取出来,便于分辨焊接缺陷类型及辅助评价缺陷对质量的影响程度[2]
随着图像处理技术的成熟,研究人员对图像分割提出了诸多方法。陈浩然等[3]建立基于Otsu图像分割算法的逻辑回归模型用以检测碎米,该模型较国标法能更好地区分碎米与整米。林宁宁等[4]提出基于改进的DeepLab V3+的街道树木图像分割算法,该算法在自制数据集和Cityscapes公共数据集上相较于其他主流网络精度有所提高。田根源[5]提出改进黏菌算法优化Cross熵的多阈值图像分割方法,并利用伯克利图像库验证算法的有效性。马小晶等[6]将Circle混沌映射和随机游走策略引入麻雀优化算法并用于多阈值图像分割,提高了最佳阈值求解的效率和多阈值图像分割精度。贺航等[7]提出一种改进蜉蝣算法的图像分割方法,通过改进的蜉蝣算法求解指数熵为目标函数的最优分割阈值,能快速有效地解决复杂多目标图像的多阈值分割问题。Fan等[8]提出分数阶飞蛾火焰算法与二维OTSU算法相结合,利用分数阶微分具有记忆性和遗传性的优点,较传统的MFO(moth-flame optimization)算法相比,改进后的算法收敛速度和图像分割精度均得到提高。Dong等[9]提出一种改进的蜻蜓算法(improved dragonfly algorithm, IDA)并应用于OTSU多阈值图像分割,采用精英对抗的学习优化增强种群多样性,引入正弦余弦策略防止算法陷入局部最优,引入自适应t分布策略增强全局和局部搜索之间的平衡提高算法的收敛速度,该算法较其他优化算法在收敛速度和分割精度方面都有较好的表现。
然而现有研究针对图像分割运用于焊接缺陷检测领域研究较少,且阈值分割多停留在一维、二维空间,难以获得图像的深层次特征信息,影响图像分割精度。现提出三维Otsu阈值分割算法对焊接缺陷图像进行分割,并引入模拟退火策略与改进的粒子群优化算法相结合提高算法收敛速度和分割精度。
最大类间方差(Otsu)是经典的图像阈值方法之一。三维Otsu算法考虑了像素的空间信息,具有一定的抗噪性能,分割效果优于一维Otsu和二维Otsu算法[10]。三维Otsu算法以像素灰度值、邻域像素平均灰度值、邻域像素中值灰度值三个维度构建三维直方图实现阈值分割[11]。以经典图像处理Lena图为例,构建的三维直方图如图1所示。
对于一幅大小为M×N、灰度级为L的数字图像,像素点(x,y)处的灰度值用h(x,y)表示,像素点(x,y)的a×a区域的平均灰度值用f(x,y)表示,其公式[12]
$\mathrm{f}(\mathrm{x},\mathrm{y})=\frac{1}{{a}^{2}}\stackrel{a/2}{\sum _{m=-a/2}}\stackrel{a/2}{\sum _{n=-a/2}}h(x+m,y+n)$
式(1)中:m=[-a/2,a/2],n=[-a/2,a/2]分别表示取整后的结果;$(x+m)\in [0,M-1],(y+n)\in [0,N-1]。$
像素点(x,y)的a×a区域的中值灰度值用g(x,y)表示,其公式为
$g(x,y)=\mathrm{m}\mathrm{e}\mathrm{d}\left[h\right(x+m,y+n\left)\right]$
式(2)中:med为计算数组中值函数median;h(x,y)、f(x,y)、g(x,y)的取值范围为[0,L-1]。
h(x,y)、 f(x,y)、g(x,y)组成的向量(i,j,k)构成图像的三维直方图,向量(i,j,k)出现的概率p(i,j,k)计算公式为
$p(i,j,k)=\frac{n(i,j,k)}{MN}$
式(3)中:n(i,j,k)为向量(i,j,k)出现的频数;ijk的取值范围为[0,L-1]。
假设三维直方图中阈值向量(r,s,t)将图像分为目标C0和背景C1,发生的概率分别为PC0PC1,计算公式为
$\left\{\begin{array}{l}{P}_{{C}_{0}}(\mathrm{r},\mathrm{s},\mathrm{t})=\stackrel{r}{\sum _{i=0}}\stackrel{s}{\sum _{j=0}}\stackrel{t}{\sum _{k=0}}{p}_{ijk}\\ {P}_{{C}_{1}}(r,s,t)=\stackrel{L-1}{\sum _{i=r+1}}\stackrel{L-1}{\sum _{j=s+1}}\stackrel{L-1}{\sum _{k=t+1}}{p}_{ijk}\end{array}\right.$
式(4)中:pijk为式(3)中p(i,j,k)的简写;rst的取值范围为[0,L-1];且${P}_{{C}_{0}}(r,s,t)+{P}_{{C}_{1}}(r,s,t)=1。$
目标C0和背景C1的均值矢量分别为μ0μ1,其计算公式为
$\begin{array}{l}{\mu }_{0}=({\mu }_{0i},{\mu }_{0j},{\mu }_{0k}{)}^{\mathrm{T}}\\ =[\frac{\stackrel{r}{\sum _{i=0}}\stackrel{s}{\sum _{j=0}}\stackrel{t}{\sum _{k=0}}i{p}_{ijk}}{{P}_{{C}_{0}}(r,s,t)},\frac{\stackrel{r}{\sum _{i=0}}\stackrel{s}{\sum _{j=0}}\stackrel{t}{\sum _{k=0}}j{p}_{ijk}}{{P}_{{C}_{0}}(r,s,t)},\\  \frac{\stackrel{r}{\sum _{i=0}}\stackrel{s}{\sum _{j=0}}\stackrel{t}{\sum _{k=0}}k{p}_{ijk}}{{P}_{{C}_{0}}(r,s,t)}]{}^{\mathrm{T}}\end{array}$
$\begin{array}{l}{\mu }_{1}=({\mu }_{1i},{\mu }_{1j},{\mu }_{1k}{)}^{\mathrm{T}}\\ =[\frac{\stackrel{L-1}{\sum _{i=r+1}}\stackrel{L-1}{\sum _{j=s+1}}\stackrel{L-1}{\sum _{k=t+1}}i{p}_{ijk}}{{P}_{{C}_{1}}(r,s,t)},\frac{\stackrel{L-1}{\sum _{i=r+1}}\stackrel{L-1}{\sum _{j=s+1}}\stackrel{L-1}{\sum _{k=t+1}}j{p}_{ijk}}{{P}_{{C}_{1}}(r,s,t)},\\  \frac{\stackrel{L-1}{\sum _{i=r+1}}\stackrel{L-1}{\sum _{j=s+1}}\stackrel{L-1}{\sum _{k=t+1}}k{p}_{ijk}}{{P}_{{C}_{1}}(r,s,t)}]{}^{\mathrm{T}}\end{array}$
三维直方图总均值矢量μ计算公式为
$\begin{array}{l}\mu =({\mu }_{i},{\mu }_{j},{\mu }_{k}{)}^{\mathrm{T}}\\ =(\stackrel{L-1}{\sum _{i=0}}\stackrel{L-1}{\sum _{j=0}}\stackrel{L-1}{\sum _{k=0}}i{p}_{ijk},\stackrel{L-1}{\sum _{i=0}}\stackrel{L-1}{\sum _{j=0}}\stackrel{L-1}{\sum _{k=0}}j{p}_{ijk},\\  \stackrel{L-1}{\sum _{i=0}}\stackrel{L-1}{\sum _{j=0}}\stackrel{L-1}{\sum _{k=0}}k{p}_{ijk}){}^{\mathrm{T}}\end{array}$
目标C0和背景C1之间的离散度矩阵计算公式为
$\begin{array}{l}\sigma (r,s,t)={P}_{{C}_{0}}\left[\right({\mu }_{0i}-{\mu }_{i}{)}^{2}+({\mu }_{0j}-{\mu }_{j}{)}^{2}+\\ ({\mu }_{0k}-{\mu }_{k}{)}^{2}]+{P}_{{C}_{1}}\left[\right({\mu }_{1i}-{\mu }_{i}{)}^{2}+\\ ({\mu }_{1j}-{\mu }_{j}{)}^{2}+({\mu }_{1k}-{\mu }_{k}{)}^{2}]\end{array}$
由于离散度矩阵σ(r,s,t)无法度量目标C0和背景C1之间差异大小,引入矩阵σ(r,s,t)的迹$\mathrm{T}\mathrm{r}\left[\sigma \right(r,s,t\left)\right]$作为目标与背景间的离散度量,使得矩阵的迹值最大即为最佳阈值,其计算公式为
$({r}^{\mathrm{*}},{s}^{\mathrm{*}},{t}^{\mathrm{*}})=\underset{r,s,t\in [0,L-1]}{\mathrm{A}\mathrm{r}\mathrm{g}\mathrm{m}\mathrm{a}\mathrm{x}}\left\{\mathrm{T}\mathrm{r}\right[\sigma (r,s,t)\left]\right\}$
式(9)中:Argmax表示返回最大值索引的函数;(r*, s*, t*)为经过三维Otsu算法得到的分割图像目标和背景之间的最佳阈值。
基于三维Otsu算法的焊接缺陷图像分割通过对向量(r,s,t)进行遍历可以得到最佳分割阈值(r*, s*, t*),但引入三维Otsu相应需要遍历的参数成指数级增长,数量级达107,计算复杂度提高,图像分割时间较长。所以,本文研究通过模拟退火(simulated annealing,SA)策略改进PSO(particle swarm optimization)算法对迭代过程进行优化,从而得到更优的图像三维阈值分割模型。
PSO算法初始化随机生成一组粒子,每个粒子代表三维直方图中任一向量(r,s,t),各粒子依据速度和位置更新公式不断迭代,直至找到图像分割的最佳阈值,粒子速度和位置更新公式为
${V}_{\mathrm{i}\mathrm{d}}^{\mathrm{t}+1}=\omega {V}_{\mathrm{i}\mathrm{d}}^{t}+{c}_{1}{r}_{1}({P}_{\mathrm{i}\mathrm{d}}^{t}-{X}_{\mathrm{i}\mathrm{d}}^{t})+{c}_{2}{r}_{2}({P}_{\mathrm{g}\mathrm{d}}^{t}-{X}_{\mathrm{i}\mathrm{d}}^{t})$
${X}_{\mathrm{i}\mathrm{d}}^{\mathrm{t}+1}={X}_{\mathrm{i}\mathrm{d}}^{t}+{V}_{\mathrm{i}\mathrm{d}}^{\mathrm{t}+1}$
式中:t为种群迭代次数;ω为惯性权重;d为粒子维度,d=1,2,…,D;i为粒子序号,i=1,2,…,N(N为种群大小);c1,c2分别为个体学习因子和群体学习因子;r1,r2为区间[0,1]的随机数,用以增加搜索的随机性;${V}_{\mathrm{i}\mathrm{d}}^{t}$为粒子it次迭代的d维速度向量;${X}_{\mathrm{i}\mathrm{d}}^{t}$为粒子it次迭代的d维位置向量;${P}_{\mathrm{i}\mathrm{d}}^{t}$为粒子it次迭代d维的历史最优位置;${P}_{\mathrm{g}\mathrm{d}}^{t}$为种群第t次迭代d维的历史最优位置。
在传统PSO算法中,惯性权重ω往往依赖经验取一常数。较小的ω值,粒子对搜索区域精确局部搜索,容易陷入局部最优,较大的ω值,容易导致算法后期在全局最优解附近产生振荡现象。为使算法快速收敛于全局最优,式(10)中ω采用自适应惯性权重进行取值,公式[14]
$\omega \left(t\right)=({\omega }_{\mathrm{s}}-{\omega }_{\mathrm{e}})\mathrm{t}\mathrm{a}\mathrm{n}\left\{0.875\left[1-{\left(\frac{t}{T}\right)}^{0.6}\right]\right\}+{\omega }_{\mathrm{e}}$
式(12)中:ωs为初始惯性权重,取0.9;ωe为终止惯性权重,取0.4;t为当前迭代次数;T为最大迭代次数。
学习因子c1c2反映粒子在个体和群体之间的信息交换。较大的c1值,粒子趋于局部搜索,易陷入局部最优,较大的c2值会使粒子过早收敛与局部最优。因此,在算法迭代初期,采用较大的c1值和较小的c2值,增加粒子迭代的多样性。随着迭代次数的增加,c1线性递减,c2线性递增,从而粒子收敛于全局最优。学习因子c1c2的计算公式[15]
${c}_{1}={c}_{1\mathrm{s}}+({c}_{1\mathrm{e}}-{c}_{1\mathrm{s}})\frac{t}{T}$
${c}_{2}={c}_{2\mathrm{s}}+({c}_{2\mathrm{e}}-{c}_{2\mathrm{s}})\frac{t}{T}$
式中:c1s为初始个体学习因子,取2.5;c1e为终止个体学习因子,取0.5;c2s为初始群体学习因子,取1;c2e为终止群体学习因子,取2.25。
SA算法是一种基于概率的算法,包含Metropolis准则和退火两部分。算法由初始温度开始,根据退火原理温度不断下降,同时Metropolis准则以一定概率跳出局部最优最终趋于全局最优。对于IPSO(improved particle swarm optimization)算法迭代产生的候选解,若该候选解优于全局最优则更新全局最优值;否则以一定概率判断候选解是否接受。概率p计算公式为
$p=\left\{\begin{array}{ll}\mathrm{e}\mathrm{x}\mathrm{p}\left[\frac{{f}_{\mathrm{p}}\left(k\right)-{f}_{\mathrm{g}}\left(k\right)}{{T}_{k}}\right],& {f}_{\mathrm{p}}\left(k\right)<{f}_{\mathrm{g}}\left(k\right)\\ 1,& {f}_{\mathrm{p}}\left(k\right)\ge {f}_{\mathrm{g}}\left(k\right)\end{array}\right.$
式(15)中:$\mathrm{ }{f}_{\mathrm{p}}\left(k\right)$ 为第k次迭代时个体最优位置对应的适应度值;$\mathrm{ }{f}_{\mathrm{g}}\left(k\right)$为第k次迭代时全局最优位置对应的适应度值; Tk为第k次迭代时系统温度。
通过设置系统的初始温度T1,每次迭代产生的候选解根据式(16)判断是否接受,当接受新解时利用衰减系数α使系统温度衰减,其计算公式[16]
${T}_{k}=\left\{\begin{array}{ll}\frac{{f}_{\mathrm{g}}\left(k\right)}{\mathrm{l}\mathrm{g}5},& k=1\\ \alpha {T}_{k},& k>1\end{array}\right.$
式(16)中:α为温度衰减系数,α=0.95。
针对传统PSO算法在优化过程中容易陷入局部最优和过早收敛现象,本文研究在PSO算法中引入自适应惯性权重和非对称学习因子优化算法参数选择,为增强算法跳出局部最优解的能力,利用SA策略的Metropolis准则来判断是否接受新解,优化目标函数最优解的判定过程,确保IPSO算法能准确得到最优解。
基于SA-IPSO算法优化三维Otsu阈值分割模型的具体实现步骤如下。
步骤1 设置迭代次数和种群大小等参数,随机初始化粒子群中粒子位置和速度。
步骤2 读取图像并建立三维直方图,以式(9)为PSO算法的目标函数。
步骤3 计算初始各粒子的适应度值,并根据式(12)~式(14)、式(16)计算惯性权重、学习因子和初始温度。
步骤4 根据式(10)和式(11)更新粒子的速度和位置得到个体最优和全局最优的候选解。
步骤5 根据式(15)计算接受候选解的概率,更新算法个体和全局最优值。
步骤6 当接受候选解时,以式(16)更新系统温度,否则温度不变。
步骤7 根据系统温度或迭代次数等条件判断算法是否收敛,若未收敛返回步骤3。
步骤8 得到最佳阈值(r*,s*,t*),输出模型分割后图像。
在模型建立过程中,将读取到的图像经过灰度化、去噪等预处理创建其三维直方图,并以类间方差作为算法的目标函数。基于SA-IPSO算法优化三维Otsu阈值分割模型能够有效克服人为选择阈值对图像分割带来的不确定性和无法获得深层次图像信息等问题,利用SA-IPSO算法对阈值向量遍历计算进行优化,能有效提升阈值分割结果精度和效率。
SA-IPSO算法优化三维Otsu阈值分割流程图如图2所示。
不同算法阈值分割后的图像分割效果主观上很难分辨优劣,且不具备说服力。为了客观地评价不同算法阈值分割效果,引入峰值信噪比PSNR(peak signal to noise ratio)和结构相似性SSIM(structural similarity)作为客观评价指标[17]。其意义为:PSNR和SSIM越大,说明阈值分割效果更好。
峰值信噪比(PSNR)是衡量图像失真程度的重要指标,假设原图像f(i,j)大小为M×N,阈值分割后的图像为g(i,j),PSNR计算公式为
$PSNR=10lg\frac{\mathrm{M}\mathrm{N}{L}^{2}}{\stackrel{M}{\sum _{i=1}}\stackrel{N}{\sum _{j=1}}\left[f\right(i,j)-g{(i,j)]}^{2}}$
式(17)中:(i,j)为图像中的任一像素点;L为图像数据类型的最大动态范围,本文中L=255。
结构相似性(SSIM)可以评估分割后图像与原图像之间的相似度,SSIM越高,结构相似性越好。SSIM计算公式为
$\mathrm{S}\mathrm{S}\mathrm{I}\mathrm{M}=\frac{(2{\mu }_{\mathrm{f}}{\mu }_{\mathrm{g}}+{c}_{\mathrm{a}})(2{\sigma }_{f,g}+{c}_{\mathrm{b}})}{({\mu }_{\mathrm{f}}^{2}+{\mu }_{\mathrm{g}}^{2}+{c}_{\mathrm{a}})({\sigma }_{\mathrm{f}}^{2}+{\sigma }_{\mathrm{g}}^{2}+{c}_{\mathrm{b}})}$
${\mu }_{\mathrm{f}}=\frac{1}{MN}\stackrel{M}{\sum _{i=1}}\stackrel{N}{\sum _{j=1}}f(i,j)$
${\sigma }_{\mathrm{f}}=\sqrt{\frac{1}{MN-1}\stackrel{M}{\sum _{i=1}}\stackrel{N}{\sum _{j=1}}\left[f\right(i,j)-{\mu }_{\mathrm{f}}{]}^{2}}$
${\sigma }_{\mathrm{f},\mathrm{g}}=\frac{1}{MN-1}\stackrel{M}{\sum _{i=1}}\stackrel{N}{\sum _{j=1}}\left[f\right(i,j)-{\mu }_{\mathrm{f}}]\left[g\right(i,j)-{\mu }_{\mathrm{g}}]$
式中:${\mu }_{\mathrm{f}}、{\mu }_{\mathrm{g}}$分别为图像f(i,j)和g(i,j)的均值;${\sigma }_{\mathrm{f}}、{\sigma }_{\mathrm{g}}$分别为图像f(i,j)g(i,j)的标准差;${\sigma }_{\mathrm{f},\mathrm{g}}$为两幅图像之间的协方差;${c}_{\mathrm{a}}={\left(0.01L\right)}^{2};{c}_{\mathrm{b}}={\left(0.03L\right)}^{2},$引入其目的是为了消除图像数值类型的影响。
为研究SA-IPSO算法对图像的三维Otsu阈值分割效果,对企业实际生产中的裂纹和气孔两大焊接缺陷进行图像采集。采集到的图像经预处理和图像增强等措施细化图像特征,图像的尺寸大小为754×754像素。实验仿真在Windows 10系统、Inter(R) Core(TM)i5-7200U CPU @2.50 GHz、8 GB内存上运行;编程环境为MATLAB R2021b,并采用HALCON 12.0机器视觉算法包对采集到的图像预处理。SA-IPSO算法参数设置为:种群数目N=50,维度空间d=3,最大迭代次数T=300,Vmin=-2.5,Vmax=2.5,系统最低温度Tmin=10-3
分别采用PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法和SA-IPSO优化三维Otsu算法对裂纹和气孔焊接缺陷进行阈值分割,不同算法分割结果如图3图4所示。
图3可知,PSO优化三维Otsu算法在裂纹起始端和末端都能较好的分割出裂纹图像的特征信息,但由于传统PSO算法的不足,裂纹的不同灰度区域未被分割出来;SA-IPSO优化二维Otsu算法分割结果在裂纹末端处存在明显的失真现象,无法完整分割出裂纹原图特征信息;SA-IPSO优化三维Otsu算法弥补了上述各算法的不足,在确保裂纹图像特征信息完整、不失真的前提下,较好地分割出了裂纹起始端边缘和裂纹内部不同灰度区域。
图4可知,PSO优化三维Otsu算法能够较完整的分割出气孔图像的特征信息,但由于传统PSO算法的不足,较小气孔、边缘处灰度区域未被分割出来;SA-IPSO优化二维Otsu算法分割结果在较小气孔和气孔连接处存在明显的失真现象,分割出的图像气孔特征信息部分丢失;SA-IPSO优化三维Otsu算法分割后的气孔图像,一方面较为完整保留了气孔图像特征信息,另一方面对较小气孔和气孔连接处的分割能较好地反映气孔图像的真实情况。
由于不同算法与模型分割后的焊接缺陷图像差异较小,人为评价具有较大的主观性,引入评价指标PSNR和SSIM客观评价分割效果。表1为3种算法分别对裂纹和气孔焊接缺陷图像分割结果的评价指标PSNR值和SSIM值。由表1中各指标可知,采取本文提出的SA-IPSO优化三维Otsu算法分割后的PSNR值和SSIM值均高于其他算法。表明本文算法计算得到的最佳阈值在图像分割精度上有一定的优势。对裂纹缺陷图像阈值分割时,本文算法较PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法在评价指标PSNR值上分别提高了15.57%、8.38%。对气孔缺陷图像阈值分割时,本文算法较PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法在评价指标SSIM值上分别提高了30.62%、19.39%。
综合分析各算法对裂纹和气孔焊接缺陷阈值分割结果。对比图3(b)图3(d)图4(b)图4(d),在阈值分割模型相同的情况下,SA-IPSO优化算法较传统PSO算法在焊接缺陷边缘灰度区域分割结果上更趋近于原图,说明SA-IPSO算法得到了阈值分割模型的最佳阈值。对比图3(c)图3(d)图4(c)图4(d),在优化算法相同的情况下,三维Otsu模型较二维Otsu模型在焊接缺陷图像目标特征信息上更趋近于原图,说明三维Otsu模型能够分割出图像的深层次特征,避免图像分割结果失真。
为进一步说明不同算法的性能和分割效果,对3种分割算法各运行20次,最佳阈值、最优迭代次数和适应度值分别取20次运行结果的均值。各算法对裂纹和气孔焊接缺陷图像分割优化结果如表2所示。
对于裂纹和气孔焊接缺陷图像,表2分别采用PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法和SA-IPSO优化三维Otsu算法进行优化阈值分割。从优化算法效率方面考虑,由于二维Otsu模型较三维Otsu模型所需迭代的参数少、计算复杂性低,SA-IPSO优化二维Otsu算法达到极值所需的迭代次数最小,能更快的求出算法的最佳阈值。从优化算法计算精度方面考虑,PSO优化三维Otsu算法适应度值最低,SA-IPSO优化三维Otsu算法适应度值最高,算法跳出局部最优,得到的最佳阈值分割精度更高。综合考虑算法效率和精度,SA-IPSO优化三维Otsu算法效率稍低于SA-IPSO优化二维Otsu算法,但其分割精度优于该算法,验证了本文算法的可行性。
PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法和SA-IPSO优化三维Otsu算法对裂纹和气孔焊接缺陷图像分割的收敛曲线分别如图5图6所示。可以看出,本文研究提出的SA-IPSO优化三维Otsu算法较其他算法在分割缺陷图像时能求解得到更高的适应度值,表明本文算法跳出局部最优解能力增强,求解得到算法全局最优,且最佳阈值分割精度更高。与传统PSO算法相比,SA-IPSO算法能更快地达到最优迭代次数,提高了算法求解的收敛速度。综上,本文提出的SA-IPSO优化三维Otsu算法在解决焊接缺陷图像分割问题时,在一定程度上提高了最佳阈值求解的效率,使算法更快达到收敛,同时跳出局部最佳阈值,有效提高了图像阈值分割精度。
为解决传统焊接缺陷检测受主观影响出现漏检、误检等情况,提出了三维Otsu阈值分割算法对焊接缺陷图像进行分割,并运用基于模拟退火策略改进的粒子群优化算法(SA-IPSO)提高分割效率和精度。通过引入像素灰度值、邻域像素平均灰度值和中值像素灰度值三个维度构建三维直方图实现三维Otsu阈值分割。为进一步提高分割算法的性能,对传统的粒子群优化算法引入自适应惯性权重和非对称学习因子并融合模拟退火策略,与三维Otsu算法结合实现焊接缺陷图像的阈值分割。得出以下结论。
(1)三维Otsu模型较二维Otsu模型在焊接缺陷图像目标特征信息上更趋近于原图,说明三维Otsu模型能够分割出图像的深层次特征,避免图像分割结果失真。
(2)对于裂纹缺陷图像,本文算法较PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法在评价指标PSNR值上分别提高了15.57%、8.38%。
(3)对于气孔缺陷图像,本文算法较PSO优化三维Otsu算法、SA-IPSO优化二维Otsu算法在评价指标SSIM值上分别提高了30.62%、19.39%。
(4)本文提出的SA-IPSO优化三维Otsu算法在解决焊接缺陷图像分割问题时,提高了最佳阈值求解的效率,使算法更快达到收敛,同时跳出局部最佳阈值,有效提高了图像阈值分割精度。
后续将从分割速度方面做相应研究并将该方法应用于焊接缺陷定位与分类系统中进行缺陷分割。
  • 国家自然科学基金(52375201)
  • 武汉理工大学产学研科技合作项(20231h0544)
参考文献 引证文献
排序方式:
[1]
陈楠, 贺春, 张楠, 等. X80钢螺旋焊管主要焊接缺陷分析及预控措施[J]. 焊接技术, 2019, 48(7): 91-93.
Chen Nan, He Chun, Zhang Nan, et al. Analysis of main weld defect of X80 spiral welded pipe and pre-control measures[J]. Welding Technology, 2019, 48(7): 91-93.
[2]
朱子洵, 张洁, 汪俊亮. 基于显著性检测的坯布疵点图像自适应分割方法[J]. 华中科技大学学报(自然科学版), 2024, 52(6): 39-47.
Zhu Zixun, Zhang Jie, Wang Junliang. Adaptive segmentation method of grey fabric defect image based on saliency detection[J]. Journal of Huazhong University of Science and Technology(Natural Science Edition), 2024, 52(6): 39-47.
[3]
陈浩然, 范方辉, 牟天. 基于OTSU图像分割算法的碎米检测[J]. 食品研究与开发, 2023, 44(20): 175-183, 8.
Chen Haoran, Fan Fanghui, Mou Tian. Broken rice detection base on OTSU image segmentation algorithm[J]. Food Research and Development, 2023, 44(20): 175-183, 8.
[4]
林宁宁, 高心丹. DeepLab V3+改进的树木图像分割[J]. 计算机工程与设计, 2023, 44(1): 232-239.
Lin Ningning, Gao Xindan. Improved trees image segmentation based on DeepLab V3+[J]. Computer Engineering and Design, 2023, 44(1): 232-239.
[5]
田根源. 改进黏菌算法优化图像熵的图像分割[J]. 蚌埠学院学报, 2023, 12(5): 75-81.
Tian Genyuan. Image segmentation based on improved slime mould algorithm optimizing image entropy[J]. Journal of Bengbu University, 2023, 12(5): 75-81.
[6]
马小晶, 贺航, 王宏伟, 等. 基于改进麻雀搜索算法的最大指数熵分割方法[J]. 科学技术与工程, 2023, 23(16): 6983-6992.
Ma Xiaojing, He Hang, Wang Hongwei, et al. Maximum exponential entropy segmentation method based on improved sparrow search algorithm[J]. Science Technology and Engineering, 2023, 23(16): 6983-6992.
[7]
贺航, 许连杰, 李高源, 等. 基于改进蜉蝣算法优化多阈值图像分割[J]. 科学技术与工程, 2024, 24(12): 5059-5068.
He Hang, Xu Lianjie, Li Gaoyuan, et al. Optimized multi-thresho-ld image segmentation based on improved mayfly algorithm[J]. Science Technology and Engineering, 2024, 24(12): 5059-5068.
[8]
Fan Q, Ma Y, Wang P, et al. Otsu image segmentation based on a fractional order moth-flame optimization algorithm[J]. Fractal and Fractional, 2024, 8(2): 87-111.
[9]
Dong Y, Li M, Zhou M. Multi-threshold image segmentation based on the improved dragonfly algorithm[J]. Mathematics, 2024, 12(6): 854-872.
[10]
肖明尧, 李雄飞. 基于高斯分解的多尺度3D Otsu阈值分割算法[J]. 吉林大学学报(工学版), 2017, 47(1): 255-261.
Xiao Mingyao, Li Xiongfei. Multi-scale 3D Otsu thresholding algorithm based on gaussion decomposition[J]. Journal of Jinlin University(Engineering and Technology Edition), 2017, 47(1): 255-261.
[11]
Neha S, Kumar A B, Vinod I K. Fusion-based contextually selected 3D Otsu thresholding for image segmentation[J]. Multimedia Tools and Applications, 2021, 80(13): 19399-19420.
[12]
曹爽, 贺玉珍, 安建成. 基于改进狼群算法的三维OTSU阈值法[J]. 计算机技术与发展, 2020, 30(4): 94-99.
Cao Shuang, He Yuzhen, An Jiancheng. A three-dimension OTSU threshold algorithm based on improved wolf pack algorithm[J]. Computer Technology and Development, 2020, 30(4): 94-99.
[13]
刘文长, 吴超华, 陈兴安. 双立柱堆垛机立柱挠度模型及其参数优化[J]. 机电工程, 2023, 40(8): 1202-1209.
Liu Wenchang, Wu Chaohua, Chen Xingan. Vertical column deflection model and its parameter optimization of double-column stacker[J]. Journal of Mechanical & Electrical Engineering, 2023, 40(8): 1202-1209.
[14]
于桂芹, 李刘东, 袁永峰. 一种结合自适应惯性权重的混合粒子群算法[J]. 哈尔滨理工大学学报, 2016, 21(3): 49-53.
Yu Guiqin, Li Liudong, Yuan Yongfeng. A hybrid particle swarm optimization algorithm with adaptive inertia weight[J]. Journal of Haebin University of Science and Technology, 2016, 21(3): 49-53.
[15]
吴永红, 曾志高, 邓彬. 基于惯性权重和学习因子动态调整的粒子群算法[J]. 湖南工业大学学报, 2021, 35(1): 91-96.
Wu Yonghong, Zeng Zhigao, Deng Bin. Particle swarm optimization algorithm based on dynamic adjustment of inertial weight and learning factors[J]. Journal of Hunan University of Technology, 2021, 35(1): 91-96.
[16]
闫群民, 马瑞卿, 马永翔, 等. 一种自适应模拟退火粒子群优化算法[J]. 西安电子科技大学学报, 2021, 48(4): 120-127.
Yan Qunmin, Ma Ruiqing, Ma Yongxiang, et al. Adaptive simulated annealing particle swarm optimization algorithm[J]. Journal of Xidian University, 2021, 48(4): 120-127.
[17]
Zhou W, Conrad A B, Rahim H S, et al. Image quality assessment: from error visibility to structural similarity[J]. IEEE Transactions on Image Processing: A Publication of the IEEE Signal Processing Society, 2004, 13(4): 600-612.
2025年第25卷第22期
PDF下载
252
116
引用本文
BibTeX
文章信息
doi: 10.12404/j.issn.1671-1815.2404720
  • 接收时间:2024-06-24
  • 首发时间:2026-02-11
  • 出版时间:2025-08-08
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-06-24
  • 修回日期:2025-04-27
基金
国家自然科学基金(52375201)
武汉理工大学产学研科技合作项(20231h0544)
作者信息
    武汉理工大学机电工程学院, 武汉 430070

通讯作者:

* 吴超华(1980—),男,汉族,湖北孝感人,博士,副教授。研究方向:智能装备系统设计与优化,图像分析与识别。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2404720
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