Article(id=1149789604552864204, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1149768563956826506, articleNumber=null, orderNo=null, doi=10.12404/j.issn.1671-1815.2406113, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1723651200000, receivedDateStr=2024-08-15, revisedDate=1742140800000, revisedDateStr=2025-03-17, acceptedDate=null, acceptedDateStr=null, onlineDate=1752060803782, onlineDateStr=2025-07-09, pubDate=1749312000000, pubDateStr=2025-06-08, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1752060803782, onlineIssueDateStr=2025-07-09, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1752060803782, creator=13701087609, updateTime=1752060803782, updator=13701087609, issue=Issue{id=1149768563956826506, tenantId=1146029695717560320, journalId=1146123166801305609, year='2025', volume='25', issue='16', pageStart='6587', pageEnd='7021', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=0, createTime=1752055787314, creator=13701087609, updateTime=1768456850262, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1218559607937618069, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1149768563956826506, language=EN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1218559607937618070, tenantId=1146029695717560320, journalId=1146123166801305609, issueId=1149768563956826506, language=CN, specialIssueTitle=, coverIllustrator=, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=6898, endPage=6912, ext={EN=ArticleExt(id=1149789604854854093, articleId=1149789604552864204, tenantId=1146029695717560320, journalId=1146123166801305609, language=EN, title=Image Feature Inference and Semantic Completion in Occluded Construction Scenes: A Case Study of Construction Fences, columnId=1156262729917780302, journalTitle=Science Technology and Engineering, columnName=Papers·Architectural Science, runingTitle=null, highlight=null, articleAbstract=

In construction safety inspections, visual obstructions often lead to missing features, resulting in dangerous misjudgments. To improve the efficiency of risk identification in construction, a method for occluded feature inference based on amodal completion technology was proposed, using construction fence detection as a case study. First, a dataset of fence detection images with visual obstructions was created using image synthesis techniques. Then, a combination of YOLOv8 instance segmentation and the amodal segmenter based on boundary uncertainty estimation (ASBU) feature completion network was used to infer the visual features of the occluded parts of the fence. The completed features of the occluded construction fences can be applied to various construction safety monitoring tasks, such as closed-loop detection. The approach was validated using fence images from multiple construction sites, achieving precise feature completion for occluded fences (average intersection ratio mIoU>95.5%). The research results provide a framework for feature inference in occluded construction scenes, which enhances the efficiency of intelligent construction safety supervision.

, correspAuthors=Wei WEI, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=Yu-jie LU, Yu-fan CHEN, Wei WEI), CN=ArticleExt(id=1149789627957080169, articleId=1149789604552864204, tenantId=1146029695717560320, journalId=1146123166801305609, language=CN, title=施工场景遮蔽图像特征推理与语义补全方法:以施工围栏为例, columnId=1156262730517565784, journalTitle=科学技术与工程, columnName=论文·建筑科学, runingTitle=null, highlight=null, articleAbstract=

施工安全巡检易因视野遮挡导致特征缺失,出现危险误判情况。为提升施工风险识别效率,针对特征遮挡问题,以施工围栏检测为例,提出一种基于amodal补全技术的遮挡特征推理方法。首先,利用图像合成技术建立存在视野遮挡的围栏检测数据集。然后,结合YOLOv8实例分割和ASBU(amodal segmenter based on boundary uncertainty estimation)特征补全网络,推测围栏被遮挡部分的视觉特征。遮挡特征补全后的施工围栏可应用于闭合检测等多种施工安全监管任务。研究选择多个施工场地的围栏图像进行案例验证,可实现精准的围栏缺失特征补全(平均交并比mIoU>95.5%)。研究成果为施工场景的遮挡特征推理提供研究范式,有效提升了智能化施工安全监管效率。

, correspAuthors=魏伟, authorNote=null, correspAuthorsNote=
* 魏伟(1997—),男,汉族,山东济南人,博士研究生。研究方向:智能建造。E-mail:
, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=AhQv5i5QURBQXq/kO6mMxw==, magXml=TxSdu+XJA05v2pGD6CCeqg==, pdfUrl=null, pdf=/iXqbPm6RkUY4GAg6/hA0A==, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=RnL+ezvVVhm5wNE00x6zog==, mapNumber=null, authorCompany=null, fund=null, authors=

卢昱杰(1985—),男,汉族,贵州贵阳人,博士,教授。研究方向:智能建造、低碳建造与工程管理。E-mail:

, authorsList=卢昱杰, 陈宇凡, 魏伟)}, authors=[Author(id=1178036589532955262, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=lu6@tongji.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1178036589616841347, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589532955262, language=EN, stringName=Yu-jie LU, firstName=Yu-jie, middleName=null, lastName=LU, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1 College of Civil Engineering, Tongji University, Shanghai 200092, China
2 Key Laboratory of Performance Evolution and Control for Engineering Structures of the Ministry of Education, Tongji University, Shanghai 200092, China
3 Shanghai Research Institute of Intelligent Science and Technology, Tongji University, Shanghai 200092, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1178036589683950212, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589532955262, language=CN, stringName=卢昱杰, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, 2, 3, address=1 同济大学土木工程学院, 上海 200092
2 同济大学工程结构性能演化与控制教育部重点实验室, 上海 200092
3 同济大学上海智能科学与技术研究院, 上海 200092, bio={"content":"

卢昱杰(1985—),男,汉族,贵州贵阳人,博士,教授。研究方向:智能建造、低碳建造与工程管理。E-mail:

"}, bioImg=null, bioContent=

卢昱杰(1985—),男,汉族,贵州贵阳人,博士,教授。研究方向:智能建造、低碳建造与工程管理。E-mail:

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1178036589293879924, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=1, ext=[AuthorCompanyExt(id=1178036589306462837, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Civil Engineering, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589314851446, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 同济大学土木工程学院, 上海 200092)]), AuthorCompany(id=1178036589377766007, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=2, ext=[AuthorCompanyExt(id=1178036589381960312, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589377766007, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Key Laboratory of Performance Evolution and Control for Engineering Structures of the Ministry of Education, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589390348921, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589377766007, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 同济大学工程结构性能演化与控制教育部重点实验室, 上海 200092)]), AuthorCompany(id=1178036589453263482, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=3, ext=[AuthorCompanyExt(id=1178036589461652091, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589453263482, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Shanghai Research Institute of Intelligent Science and Technology, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589465846396, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589453263482, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 同济大学上海智能科学与技术研究院, 上海 200092)])]), Author(id=1178036589751059078, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1178036589826556552, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589751059078, language=EN, stringName=Yu-fan CHEN, firstName=Yu-fan, middleName=null, lastName=CHEN, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 College of Civil Engineering, Tongji University, Shanghai 200092, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1178036589897859721, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589751059078, language=CN, stringName=陈宇凡, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1 同济大学土木工程学院, 上海 200092, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1178036589293879924, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=1, ext=[AuthorCompanyExt(id=1178036589306462837, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Civil Engineering, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589314851446, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 同济大学土木工程学院, 上海 200092)])]), Author(id=1178036589969162891, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=2410045@tongji.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1178036590048854669, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589969162891, language=EN, stringName=Wei WEI, firstName=Wei, middleName=null, lastName=WEI, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1 College of Civil Engineering, Tongji University, Shanghai 200092, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1178036590132740750, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, authorId=1178036589969162891, language=CN, stringName=魏伟, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1 同济大学土木工程学院, 上海 200092, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1178036589293879924, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=1, ext=[AuthorCompanyExt(id=1178036589306462837, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Civil Engineering, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589314851446, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 同济大学土木工程学院, 上海 200092)])])], keywords=[Keyword(id=1178036590275347087, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, orderNo=1, keyword=construction safety supervision), Keyword(id=1178036590350844560, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, orderNo=2, keyword=occluded construction features), Keyword(id=1178036590409564817, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, orderNo=3, keyword=amodal feature completion), Keyword(id=1178036590480867986, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, orderNo=4, keyword=instance segmentation), Keyword(id=1178036590531199635, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, orderNo=5, keyword=computer vision), Keyword(id=1178036590581531284, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, orderNo=1, keyword=施工安全监管), Keyword(id=1178036590640251541, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, orderNo=2, keyword=施工特征遮挡), Keyword(id=1178036590711554710, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, orderNo=3, keyword=amodal特征补全), Keyword(id=1178036590803829399, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, orderNo=4, keyword=实例分割), Keyword(id=1178036590862549656, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, orderNo=5, keyword=计算机视觉)], refs=[Reference(id=1178036594297684679, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html, language=null, rfNumber=[1], rfOrder=0, authorNames=中华人民共和国住房和城乡建设部, journalName=null, refType=null, unstructuredReference=中华人民共和国住房和城乡建设部. 住房和城乡建设部办公厅关于2020年房屋市政工程生产安全事故情况的通报[EB/OL]. (2021-05-21)[2024-06-09]. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html., articleTitle=住房和城乡建设部办公厅关于2020年房屋市政工程生产安全事故情况的通报, refAbstract=null), Reference(id=1178036594368987848, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html, language=null, rfNumber=[1], rfOrder=1, authorNames=Ministry of Housing and Urban Rural Development of the People's Republic of China, journalName=null, refType=null, unstructuredReference=Ministry of Housing and Urban Rural Development of the People's Republic of China. Notice from the General Office of the Ministry of Housing and Urban Rural Developmenton production safety accidents in housing and municipal engineering in 2020[EB/OL]. (2021-05-21)[2024-06-09]. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html., articleTitle=Notice from the General Office of the Ministry of Housing and Urban Rural Developmenton production safety accidents in housing and municipal engineering in 2020, refAbstract=null), Reference(id=1178036594431902409, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2024, volume=null, issue=3, pageStart=55, pageEnd=57, url=null, language=null, rfNumber=[2], rfOrder=2, authorNames=柏云飞, journalName=石河子科技, refType=null, unstructuredReference=柏云飞. 智能传感技术在桥梁结构健康监测中的应用研究[J]. 石河子科技, 2024(3): 55-57., articleTitle=智能传感技术在桥梁结构健康监测中的应用研究, refAbstract=null), Reference(id=1178036594499011274, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2024, volume=null, issue=3, pageStart=55, pageEnd=57, url=null, language=null, rfNumber=[2], rfOrder=3, authorNames=Bai Yunfei, journalName=Shihezi Science and Technology, refType=null, unstructuredReference=Bai Yunfei. Research on the application of intelligent sensing technology in bridge structure health monitoring[J]. Shihezi Science and Technology, 2024(3): 55-57., articleTitle=Research on the application of intelligent sensing technology in bridge structure health monitoring, refAbstract=null), Reference(id=1178036594620646091, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2023, volume=156, issue=null, pageStart=105153, pageEnd=null, url=null, language=null, rfNumber=[3], rfOrder=4, authorNames=Agyemang I O, Zhang X, Adjei-Mensah I, journalName=Automation in Construction, refType=null, unstructuredReference=Agyemang I O, Zhang X, Adjei-Mensah I, et al. Automated vision-based structural health inspection and assessment for post-construction civil infrastructure[J]. Automation in Construction, 2023, 156: 105153., articleTitle=Automated vision-based structural health inspection and assessment for post-construction civil infrastructure, refAbstract=null), Reference(id=1178036594700337868, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2021, volume=20, issue=2, pageStart=692, pageEnd=743, url=null, language=null, rfNumber=[4], rfOrder=5, authorNames=Dong C Z, Catbas F N, journalName=Structural Health Monitoring, refType=null, unstructuredReference=Dong C Z, Catbas F N. A review of computer vision-based structural health monitoring at local and global levels[J]. Structural Health Monitoring, 2021, 20(2): 692-743., articleTitle=A review of computer vision-based structural health monitoring at local and global levels, refAbstract=null), Reference(id=1178036594763252429, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2018, volume=85, issue=null, pageStart=1, pageEnd=9, url=null, language=null, rfNumber=[5], rfOrder=6, authorNames=Fang Q, Li H, Luo X, journalName=Automation in Construction, refType=null, unstructuredReference=Fang Q, Li H, Luo X, et al. Detecting non-hardhat-use by a deep learning method from far-field surveillance videos[J]. Automation in Construction, 2018, 85: 1-9., articleTitle=Detecting non-hardhat-use by a deep learning method from far-field surveillance videos, refAbstract=null), Reference(id=1178036594834555598, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2018, volume=91, issue=null, pageStart=53, pageEnd=61, url=null, language=null, rfNumber=[6], rfOrder=7, authorNames=Fang W, Ding L, Luo H, journalName=Automation in Construction, refType=null, unstructuredReference=Fang W, Ding L, Luo H, et al. Falls from heights: a computer vision-based approach for safety harness detection[J]. Automation in Construction, 2018, 91: 53-61., articleTitle=Falls from heights: a computer vision-based approach for safety harness detection, refAbstract=null), Reference(id=1178036594931024591, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2015, volume=50, issue=null, pageStart=64, pageEnd=71, url=null, language=null, rfNumber=[7], rfOrder=8, authorNames=Dagan D, Isaac S, journalName=Automation in Construction, refType=null, unstructuredReference=Dagan D, Isaac S. Planning safe distances between workers on construction sites[J]. Automation in Construction, 2015, 50: 64-71., articleTitle=Planning safe distances between workers on construction sites, refAbstract=null), Reference(id=1178036594998133456, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2024, volume=55, issue=7, pageStart=780, pageEnd=790, url=null, language=null, rfNumber=[8], rfOrder=9, authorNames=曾拓程, 王佳俊, 钟登华, journalName=水利学报, refType=null, unstructuredReference=曾拓程, 王佳俊, 钟登华, 等. 基于大场景视频监控的坝面施工机械潜在碰撞风险预警方法[J]. 水利学报, 2024, 55(7): 780-790, 801., articleTitle=基于大场景视频监控的坝面施工机械潜在碰撞风险预警方法, refAbstract=null), Reference(id=1178036595077825233, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2024, volume=55, issue=7, pageStart=780, pageEnd=790, url=null, language=null, rfNumber=[8], rfOrder=10, authorNames=Zeng Tuocheng, Wang Jiajun, Zhong Denghua, journalName=Journal of Hydraulic Engineering, refType=null, unstructuredReference=Zeng Tuocheng, Wang Jiajun, Zhong Denghua, et al. Collision risk warning method for dam surface construction machinery based on large-scale video surveillance[J]. Journal of Hydraulic Engineering, 2024, 55(7): 780-790, 801., articleTitle=Collision risk warning method for dam surface construction machinery based on large-scale video surveillance, refAbstract=null), Reference(id=1178036595149128402, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2024, volume=166, issue=null, pageStart=105603, pageEnd=null, url=null, language=null, rfNumber=[9], rfOrder=11, authorNames=Kim J, Kim J, Koo N, journalName=Automation in Construction, refType=null, unstructuredReference=Kim J, Kim J, Koo N, et al. Automating scaffold safety inspections using semantic analysis of 3D point clouds[J]. Automation in Construction, 2024, 166: 105603., articleTitle=Automating scaffold safety inspections using semantic analysis of 3D point clouds, refAbstract=null), Reference(id=1178036595216237267, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2022, volume=22, issue=28, pageStart=12449, pageEnd=12459, url=null, language=null, rfNumber=[10], rfOrder=12, authorNames=余萍, 宋祥宇, journalName=科学技术与工程, refType=null, unstructuredReference=余萍, 宋祥宇. 基于Open CV围栏合规性摆放检测方法研究[J]. 科学技术与工程, 2022, 22(28): 12449-12459., articleTitle=基于Open CV围栏合规性摆放检测方法研究, refAbstract=null), Reference(id=1178036595270763220, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2022, volume=22, issue=28, pageStart=12449, pageEnd=12459, url=null, language=null, rfNumber=[10], rfOrder=13, authorNames=Yu Ping, Song Xiangyu, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Yu Ping, Song Xiangyu. Detection method of fence compliance based on Open CV[J]. Science Technology and Engineering, 2022, 22(28): 12449-12459., articleTitle=Detection method of fence compliance based on Open CV, refAbstract=null), Reference(id=1178036595329483477, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2019, volume=19, issue=13, pageStart=171, pageEnd=176, url=null, language=null, rfNumber=[11], rfOrder=14, authorNames=苏婷婷, 王娜, journalName=科学技术与工程, refType=null, unstructuredReference=苏婷婷, 王娜. 基于图像块相似性和补全生成的人脸复原算法[J]. 科学技术与工程, 2019, 19(13): 171-176., articleTitle=基于图像块相似性和补全生成的人脸复原算法, refAbstract=null), Reference(id=1178036595396592342, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2019, volume=19, issue=13, pageStart=171, pageEnd=176, url=null, language=null, rfNumber=[11], rfOrder=15, authorNames=Su Tingting, Wang Na, journalName=Science Technology and Engineering, refType=null, unstructuredReference=Su Tingting, Wang Na. Human face restoration based on natural patch likelihood and generative image inpainting[J]. Science Technology and Engineering, 2019, 19(13): 171-176., articleTitle=Human face restoration based on natural patch likelihood and generative image inpainting, refAbstract=null), Reference(id=1178036595476284119, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[12], rfOrder=16, authorNames=Radford A, journalName=arXiv Preprint, refType=null, unstructuredReference=Radford A. Unsupervised representation learning with deep convolutional generative adversarial networks[J]. arXiv Preprint, 2015: arXiv: 1511.06434., articleTitle=Unsupervised representation learning with deep convolutional generative adversarial networks, refAbstract=null), Reference(id=1178036595564364504, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2020, volume=119, issue=null, pageStart=103332, pageEnd=null, url=null, language=null, rfNumber=[13], rfOrder=17, authorNames=Angah O, Chen A Y, journalName=Automation in Construction, refType=null, unstructuredReference=Angah O, Chen A Y. Removal of occluding construction workers in job site image data using U-Net based context encoders[J]. Automation in Construction, 2020, 119: 103332., articleTitle=Removal of occluding construction workers in job site image data using U-Net based context encoders, refAbstract=null), Reference(id=1178036595618890457, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2015, volume=null, issue=null, pageStart=234, pageEnd=241, url=null, language=null, rfNumber=[14], rfOrder=18, authorNames=Ronneberger O, Fischer P, Brox T, journalName=Medical Image Computing and Computer-assisted Intervention-MICCAI 2015: 18th International Conference, refType=null, unstructuredReference=Ronneberger O, Fischer P, Brox T. U-net: convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-assisted Intervention-MICCAI 2015: 18th International Conference. Munich: Springer International Publishing, 2015: 234-241., articleTitle=U-net: convolutional networks for biomedical image segmentation, refAbstract=null), Reference(id=1178036595677610714, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2014, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[15], rfOrder=19, authorNames=Mirza M, Osindero S, journalName=arXiv Preprint, refType=null, unstructuredReference=Mirza M, Osindero S. Conditional generative adversarial nets[J]. arXiv Preprint, 2014: arXiv: 1411.1784., articleTitle=Conditional generative adversarial nets, refAbstract=null), Reference(id=1178036595732136667, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2015, volume=110, issue=null, pageStart=188, pageEnd=198, url=null, language=null, rfNumber=[16], rfOrder=20, authorNames=Yan S, Liu Q, journalName=Signal Processing, refType=null, unstructuredReference=Yan S, Liu Q. Inferring occluded features for fast object detection[J]. Signal Processing, 2015, 110: 188-198., articleTitle=Inferring occluded features for fast object detection, refAbstract=null), Reference(id=1178036595795051228, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2022, volume=45, issue=null, pageStart=103352, pageEnd=null, url=null, language=null, rfNumber=[17], rfOrder=21, authorNames=Li Z, Li D, journalName=Journal of Building Engineering, refType=null, unstructuredReference=Li Z, Li D. Action recognition of construction workers under occlusion[J]. Journal of Building Engineering, 2022, 45:103352., articleTitle=Action recognition of construction workers under occlusion, refAbstract=null), Reference(id=1178036595853771485, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2018, volume=null, issue=null, pageStart=5689, pageEnd=5698, url=null, language=null, rfNumber=[18], rfOrder=22, authorNames=Yoon J, Jordon J, Schaar M, journalName=International Conference on Machine Learning, refType=null, unstructuredReference=Yoon J, Jordon J, Schaar M. Gain: missing data imputation using generative adversarial nets[C]// International Conference on Machine Learning. Basel: PMLR, 2018: 5689-5698., articleTitle=Gain: missing data imputation using generative adversarial nets, refAbstract=null), Reference(id=1178036595920880350, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=1979, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[19], rfOrder=23, authorNames=Kanizsa G, Legrenzi P, Bozzi P, journalName=Organization in vision: essays on gestalt perception, refType=null, unstructuredReference=Kanizsa G, Legrenzi P, Bozzi P. Organization in vision: essays on gestalt perception[M]. New York: Praeger, 1979., articleTitle=null, refAbstract=null), Reference(id=1178036595983794911, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2022, volume=null, issue=null, pageStart=032014, pageEnd=null, url=null, language=null, rfNumber=[20], rfOrder=24, authorNames=Gu B, Guo H, Huang Y, journalName=IOP Conference Series, refType=null, unstructuredReference=Gu B, Guo H, Huang Y, et al. Computer vision-based human-machine collision warning system on construction site[C]// IOP Conference Series: Earth and Environmental Science. London: IOP Publishing, 2022: 032014., articleTitle=Computer vision-based human-machine collision warning system on construction site, refAbstract=null), Reference(id=1178036596046709472, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[21], rfOrder=25, authorNames=湖北省住房和城乡建设厅, 湖北省市场监督管理局, journalName=建筑施工现场安全防护设施技术规程: DB 42/T 535—2020, refType=null, unstructuredReference=湖北省住房和城乡建设厅, 湖北省市场监督管理局. 建筑施工现场安全防护设施技术规程: DB 42/T 535—2020[S]. 武汉: 湖北省市场监督管理局, 2020., articleTitle=null, refAbstract=null), Reference(id=1178036596113818337, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2020, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[21], rfOrder=26, authorNames=Hubei Provincial Department of Housing and Urban Rural Development, Hubei Provincial Market Supervision Administration, journalName=Building design standard: DB 42/T 535—2020, refType=null, unstructuredReference=Hubei Provincial Department of Housing and Urban Rural Development, Hubei Provincial Market Supervision Administration. Building design standard: DB 42/T 535—2020[S]. Wuhan: Hubei Provincial Market Supervision Administration, 2020., articleTitle=null, refAbstract=null), Reference(id=1178036596180927202, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2021, volume=null, issue=null, pageStart=7396, pageEnd=7405, url=null, language=null, rfNumber=[22], rfOrder=27, authorNames=Nguyen K, Todorovic S, journalName=Proceedings of the IEEE/CVF International Conference on Computer Vision, refType=null, unstructuredReference=Nguyen K, Todorovic S. A weakly supervised amodal segmenter with boundary uncertainty estimation[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. New York: IEEE, 2021: 7396-7405., articleTitle=A weakly supervised amodal segmenter with boundary uncertainty estimation, refAbstract=null), Reference(id=1178036596243841763, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2017, volume=null, issue=null, pageStart=1301, pageEnd=1310, url=null, language=null, rfNumber=[23], rfOrder=28, authorNames=Dwibedi D, Misra I, Hebert M, journalName=Proceedings of the IEEE International Conference on Computer Vision, refType=null, unstructuredReference=Dwibedi D, Misra I, Hebert M. Cut, paste and learn: surprisingly easy synthesis for instance detection[C]// Proceedings of the IEEE International Conference on Computer Vision. New York: IEEE, 2017: 1301-1310., articleTitle=Cut, paste and learn: surprisingly easy synthesis for instance detection, refAbstract=null), Reference(id=1178036596298367716, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, doi=null, pmid=null, pmcid=null, year=2023, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=[24], rfOrder=29, authorNames=Pérez P, Gangnet M, Blake A, journalName=Poisson image editing, refType=null, unstructuredReference=Pérez P, Gangnet M, Blake A. Poisson image editing[M]. Berlin: Springer International Publishing, 2023., articleTitle=null, refAbstract=null)], funds=[Fund(id=1178036593945363139, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, awardId=2022YFC3801700, language=CN, fundingSource=国家重点研发计划(2022YFC3801700), fundOrder=null, country=null), Fund(id=1178036594012472004, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, awardId=2024-1-ZD-02, language=CN, fundingSource=中央高校基本科研业务费专项资金(2024-1-ZD-02), fundOrder=null, country=null), Fund(id=1178036594071192261, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, awardId=22120250350, language=CN, fundingSource=中央高校基本科研业务费专项资金(22120250350), fundOrder=null, country=null), Fund(id=1178036594134106822, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, awardId=52078374, language=CN, fundingSource=国家自然科学基金(52078374), fundOrder=null, country=null)], companyList=[AuthorCompany(id=1178036589293879924, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=1, ext=[AuthorCompanyExt(id=1178036589306462837, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 College of Civil Engineering, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589314851446, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589293879924, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1 同济大学土木工程学院, 上海 200092)]), AuthorCompany(id=1178036589377766007, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=2, ext=[AuthorCompanyExt(id=1178036589381960312, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589377766007, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 Key Laboratory of Performance Evolution and Control for Engineering Structures of the Ministry of Education, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589390348921, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589377766007, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2 同济大学工程结构性能演化与控制教育部重点实验室, 上海 200092)]), AuthorCompany(id=1178036589453263482, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, xref=3, ext=[AuthorCompanyExt(id=1178036589461652091, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589453263482, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 Shanghai Research Institute of Intelligent Science and Technology, Tongji University, Shanghai 200092, China), AuthorCompanyExt(id=1178036589465846396, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, companyId=1178036589453263482, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3 同济大学上海智能科学与技术研究院, 上海 200092)])], figs=[ArticleFig(id=1178036591026127513, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.1, caption=Research framework, figureFileSmall=ivn+PxtXCdSNNV5EJPXcfA==, figureFileBig=aKsRPajGXZB5weFQELNriw==, tableContent=null), ArticleFig(id=1178036591084847770, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图1, caption=研究框架, figureFileSmall=ivn+PxtXCdSNNV5EJPXcfA==, figureFileBig=aKsRPajGXZB5weFQELNriw==, tableContent=null), ArticleFig(id=1178036591143568027, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.2, caption=Image synthesis and results processing framework, figureFileSmall=nHcD9yisiQhlEAkQJws3KQ==, figureFileBig=gKkT91Afd5US5JwqLE6wcw==, tableContent=null), ArticleFig(id=1178036591202288284, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图2, caption=图像合成与结果处理框架, figureFileSmall=nHcD9yisiQhlEAkQJws3KQ==, figureFileBig=gKkT91Afd5US5JwqLE6wcw==, tableContent=null), ArticleFig(id=1178036591265202845, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.3, caption=Overall process of image synthesis, figureFileSmall=5Bai2zR789JJt7yVVgwyKg==, figureFileBig=30cxGxQE6BiNSlgibK/fiQ==, tableContent=null), ArticleFig(id=1178036591319728798, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图3, caption=图像合成整体流程, figureFileSmall=5Bai2zR789JJt7yVVgwyKg==, figureFileBig=30cxGxQE6BiNSlgibK/fiQ==, tableContent=null), ArticleFig(id=1178036591382643359, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.4, caption=YOLOv8 network architecture, figureFileSmall=G31H1EbvYNqeXL7YciEhng==, figureFileBig=nA1w6LJNUEn1/S642LGG2g==, tableContent=null), ArticleFig(id=1178036591441363616, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图4, caption=YOLOv8网络架构

Pn为第n层特特征图;FPN(feature pyramid networks)为特征金字塔网络;PAN(path aggregation network)为路径聚合网络

, figureFileSmall=G31H1EbvYNqeXL7YciEhng==, figureFileBig=nA1w6LJNUEn1/S642LGG2g==, tableContent=null), ArticleFig(id=1178036591495889569, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.5, caption=ASBU network architecture, figureFileSmall=g9ltUDAOBA4r+dJPjiREQw==, figureFileBig=yo7LLTeTyNq5sdC7bU4zoA==, tableContent=null), ArticleFig(id=1178036591558804130, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图5, caption=ASBU网络架构, figureFileSmall=g9ltUDAOBA4r+dJPjiREQw==, figureFileBig=yo7LLTeTyNq5sdC7bU4zoA==, tableContent=null), ArticleFig(id=1178036591638495907, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.6, caption=Adaptive expansion of fence mask, figureFileSmall=BGia5YNHh+iiM4RYXfk3+w==, figureFileBig=rv6KyvAQ0ZOXd+7qtWdDxg==, tableContent=null), ArticleFig(id=1178036591688827556, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图6, caption=围栏掩码自适应膨胀, figureFileSmall=BGia5YNHh+iiM4RYXfk3+w==, figureFileBig=rv6KyvAQ0ZOXd+7qtWdDxg==, tableContent=null), ArticleFig(id=1178036591764325029, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.7, caption=Fence closure judgment process, figureFileSmall=XPwzhATbCJadV2Piybf1zA==, figureFileBig=WlAtOxzC8Pro4qiv0QD58A==, tableContent=null), ArticleFig(id=1178036591835628198, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图7, caption=围栏闭合判断流程, figureFileSmall=XPwzhATbCJadV2Piybf1zA==, figureFileBig=WlAtOxzC8Pro4qiv0QD58A==, tableContent=null), ArticleFig(id=1178036591894348455, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.8, caption=Fence centerline extraction, figureFileSmall=P41kJZ80Hrw0HwJy6WJESQ==, figureFileBig=Tofgr5q7RU4pCtP4I/JRLA==, tableContent=null), ArticleFig(id=1178036591953068712, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图8, caption=围栏中心线提取

绿色表示检测闭合;红色表示检测未闭合

, figureFileSmall=P41kJZ80Hrw0HwJy6WJESQ==, figureFileBig=Tofgr5q7RU4pCtP4I/JRLA==, tableContent=null), ArticleFig(id=1178036592015983273, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.9, caption=Case study, figureFileSmall=a0Lv43zPwxPweAHB/HuaVg==, figureFileBig=nYcR35grb38emczWLNxLnw==, tableContent=null), ArticleFig(id=1178036592078897834, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图9, caption=研究案例, figureFileSmall=a0Lv43zPwxPweAHB/HuaVg==, figureFileBig=nYcR35grb38emczWLNxLnw==, tableContent=null), ArticleFig(id=1178036592141812395, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.10, caption=Training process of YOLOv8 segmentation model, figureFileSmall=L81E88M3nVAf14O5+U9Ymg==, figureFileBig=+AEaM1sPtVrfoJXeSQvL1g==, tableContent=null), ArticleFig(id=1178036592213115564, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图10, caption=YOLOv8分割模型训练过程

I为迭代次数;Lbox为边界框回归损失;Ldfl为分布式焦点损失;Lcls为分类损失;Lseg为分割损失;P为准确率;R为召回率;mAP0.5为以0.5为IoU时算法取得的平均精度;mAP0.5-0.95为IoU以0.05步长在0.5~0.95区间上的平均mAP;IoU为预测目标位置与目标真实位置的重叠率

, figureFileSmall=L81E88M3nVAf14O5+U9Ymg==, figureFileBig=+AEaM1sPtVrfoJXeSQvL1g==, tableContent=null), ArticleFig(id=1178036592271835821, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Fig.11, caption=ASBU training loss curve, figureFileSmall=qX9fwaZyrlyCcZ/+x3dYeQ==, figureFileBig=iKefIFClNHZvMuUYtJSc8w==, tableContent=null), ArticleFig(id=1178036592334750382, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=图11, caption=ASBU训练loss曲线, figureFileSmall=qX9fwaZyrlyCcZ/+x3dYeQ==, figureFileBig=iKefIFClNHZvMuUYtJSc8w==, tableContent=null), ArticleFig(id=1178036592414442159, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 1, caption=

Dataclassification and synthesis

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036592485745328, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表1, caption=

数据归类与合成

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036592548659889, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 2, caption=

YOLOv8 model training environment configuration

, figureFileSmall=null, figureFileBig=null, tableContent=
名称 设置
操作系统 Windows 11
处理器 12th GenIntel(R) Core(TM) i9-12900H 2.50 GHz
内存/GB 64.0
显卡 NVIDIA GeForce RTX 3070 Ti Laptop GPU
深度学习框架 Pytorch 1.13
运算平台 CUDA 11.7
编程语言 Python 3.9
开发环境 PyCharm Community Edition
), ArticleFig(id=1178036592607380146, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表2, caption=

YOLOv8模型训练环境配置

, figureFileSmall=null, figureFileBig=null, tableContent=
名称 设置
操作系统 Windows 11
处理器 12th GenIntel(R) Core(TM) i9-12900H 2.50 GHz
内存/GB 64.0
显卡 NVIDIA GeForce RTX 3070 Ti Laptop GPU
深度学习框架 Pytorch 1.13
运算平台 CUDA 11.7
编程语言 Python 3.9
开发环境 PyCharm Community Edition
), ArticleFig(id=1178036592674489011, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 3, caption=

YOLOv8 model training parameter settings

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 设置
预训练模型(model) YOLOv8 m-seg
输入图像尺寸(image size) 640×640
训练轮次(epochs) 200
每批训练样本数(batch size) 8
优化器(optimizer) SGD
初始学习率(lr0) 0.01
学习率因子(lrf) 0.01
), ArticleFig(id=1178036592733209268, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表3, caption=

YOLOv8模型训练参数设置

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 设置
预训练模型(model) YOLOv8 m-seg
输入图像尺寸(image size) 640×640
训练轮次(epochs) 200
每批训练样本数(batch size) 8
优化器(optimizer) SGD
初始学习率(lr0) 0.01
学习率因子(lrf) 0.01
), ArticleFig(id=1178036592791929525, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 4, caption=

ASBU model training environment configuration

, figureFileSmall=null, figureFileBig=null, tableContent=
名称 内容
操作系统 Ubuntu 20.04.1 LTS
处理器 Intel(R) Xeon(R) CPU E5-2683 v4 @2.10 GHz
内存/GB 128.0
显卡 NVIDIA Corporation Device 2204 (rev a1)
深度学习框架 Pytorch 1.12
运算平台 CUDA 11.4
编程语言 Python 3.7
开发环境 Visual Studio Code
), ArticleFig(id=1178036592859038390, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表4, caption=

ASBU模型训练环境配置

, figureFileSmall=null, figureFileBig=null, tableContent=
名称 内容
操作系统 Ubuntu 20.04.1 LTS
处理器 Intel(R) Xeon(R) CPU E5-2683 v4 @2.10 GHz
内存/GB 128.0
显卡 NVIDIA Corporation Device 2204 (rev a1)
深度学习框架 Pytorch 1.12
运算平台 CUDA 11.4
编程语言 Python 3.7
开发环境 Visual Studio Code
), ArticleFig(id=1178036592921952951, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 5, caption=

ASBU model training parameter settings

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 设置
骨干网络架构(backbone arch) UNet
迭代次数(iteration) 28 000
输入图像尺寸(image size) 256×256
每批训练样本数(batch size) 32
优化器(optimizer) SGD
初始学习率(lr0) 0.001
学习率因子(lrf) 0.1
), ArticleFig(id=1178036592997450424, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表5, caption=

ASBU模型训练参数设置

, figureFileSmall=null, figureFileBig=null, tableContent=
参数 设置
骨干网络架构(backbone arch) UNet
迭代次数(iteration) 28 000
输入图像尺寸(image size) 256×256
每批训练样本数(batch size) 32
优化器(optimizer) SGD
初始学习率(lr0) 0.001
学习率因子(lrf) 0.1
), ArticleFig(id=1178036593060364985, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 6, caption=

Instance segmentation results based on YOLOv8

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036593131668154, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表6, caption=

基于YOLOv8的实例分割结果

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036593202971323, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 7, caption=

ASBU loss validation results

, figureFileSmall=null, figureFileBig=null, tableContent=
迭代次数 loss值
在线验证 离线验证
0 0.767 20 N.A.
2 000 0.098 38 0.090 45
16 000 0.052 97 0.048 64
24 000 0.048 69 0.045 36
28 000 0.046 39 0.040 15
), ArticleFig(id=1178036593278468796, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表7, caption=

ASBU损失验证结果

, figureFileSmall=null, figureFileBig=null, tableContent=
迭代次数 loss值
在线验证 离线验证
0 0.767 20 N.A.
2 000 0.098 38 0.090 45
16 000 0.052 97 0.048 64
24 000 0.048 69 0.045 36
28 000 0.046 39 0.040 15
), ArticleFig(id=1178036593374937789, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 8, caption=

ASBU completion results

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036593450435262, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表8, caption=

ASBU补全结果

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036593530127039, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 9, caption=

Comparison and analysis of occlusion completion methods

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 算法 精度 原理 特点
基于ASBU的遮挡特征amodal补全(本文) ASBU;
amodal补全
mIoU为95.454%;
像素准确率为97.779%
基于modal掩码补全被遮挡物的amodal掩码 可基于单张静态图像推理并还原遮挡区域的真实特征
基于U-Net的施工现场遮挡物移除[13] U-Net;cGAN 图像平均L1损失为7.39%;
标准差为5.88%
根据图像上下文信息修复被遮挡的背景区域 可实现图像补全,但无法还原遮挡区域的真实特征
遮挡条件下的工人姿态识别[17] GAN 下游任务(姿态识别)准确率为86.62% 基于前后帧上下文预测缺失的骨架特征 适用于视频帧输入,而难以处理单张静态图像
), ArticleFig(id=1178036593605624512, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表9, caption=

遮挡补全方法对比与分析

, figureFileSmall=null, figureFileBig=null, tableContent=
方法 算法 精度 原理 特点
基于ASBU的遮挡特征amodal补全(本文) ASBU;
amodal补全
mIoU为95.454%;
像素准确率为97.779%
基于modal掩码补全被遮挡物的amodal掩码 可基于单张静态图像推理并还原遮挡区域的真实特征
基于U-Net的施工现场遮挡物移除[13] U-Net;cGAN 图像平均L1损失为7.39%;
标准差为5.88%
根据图像上下文信息修复被遮挡的背景区域 可实现图像补全,但无法还原遮挡区域的真实特征
遮挡条件下的工人姿态识别[17] GAN 下游任务(姿态识别)准确率为86.62% 基于前后帧上下文预测缺失的骨架特征 适用于视频帧输入,而难以处理单张静态图像
), ArticleFig(id=1178036593672733377, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=EN, label=Table 10, caption=

Fence closure detection results

, figureFileSmall=null, figureFileBig=null, tableContent=
), ArticleFig(id=1178036593735647938, tenantId=1146029695717560320, journalId=1146123166801305609, articleId=1149789604552864204, language=CN, label=表10, caption=

围栏闭合判断结果

, figureFileSmall=null, figureFileBig=null, tableContent=
)], 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.1671-1815.2406113, detailUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/10.12404/j.issn.1671-1815.2406113, pdfUrlCn=https://castjournals.cast.org.cn/joweb/kxjsygc/CN/PDF/10.12404/j.issn.1671-1815.2406113, pdfUrlEn=https://castjournals.cast.org.cn/joweb/kxjsygc/EN/PDF/10.12404/j.issn.1671-1815.2406113, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
施工场景遮蔽图像特征推理与语义补全方法:以施工围栏为例
收藏切换
PDF下载
卢昱杰 1, 2, 3 , 陈宇凡 1 , 魏伟 1, *
科学技术与工程 | 论文·建筑科学 2025,25(16): 6898-6912
收起
收藏切换
科学技术与工程 | 论文·建筑科学 2025, 25(16): 6898-6912
施工场景遮蔽图像特征推理与语义补全方法:以施工围栏为例
全屏
卢昱杰1, 2, 3 , 陈宇凡1, 魏伟1, *
作者信息
  • 1 同济大学土木工程学院, 上海 200092
  • 2 同济大学工程结构性能演化与控制教育部重点实验室, 上海 200092
  • 3 同济大学上海智能科学与技术研究院, 上海 200092
  • 卢昱杰(1985—),男,汉族,贵州贵阳人,博士,教授。研究方向:智能建造、低碳建造与工程管理。E-mail:

通讯作者:

* 魏伟(1997—),男,汉族,山东济南人,博士研究生。研究方向:智能建造。E-mail:
Image Feature Inference and Semantic Completion in Occluded Construction Scenes: A Case Study of Construction Fences
Yu-jie LU1, 2, 3 , Yu-fan CHEN1, Wei WEI1, *
Affiliations
  • 1 College of Civil Engineering, Tongji University, Shanghai 200092, China
  • 2 Key Laboratory of Performance Evolution and Control for Engineering Structures of the Ministry of Education, Tongji University, Shanghai 200092, China
  • 3 Shanghai Research Institute of Intelligent Science and Technology, Tongji University, Shanghai 200092, China
出版时间: 2025-06-08 doi: 10.12404/j.issn.1671-1815.2406113
文章导航
收藏切换

施工安全巡检易因视野遮挡导致特征缺失,出现危险误判情况。为提升施工风险识别效率,针对特征遮挡问题,以施工围栏检测为例,提出一种基于amodal补全技术的遮挡特征推理方法。首先,利用图像合成技术建立存在视野遮挡的围栏检测数据集。然后,结合YOLOv8实例分割和ASBU(amodal segmenter based on boundary uncertainty estimation)特征补全网络,推测围栏被遮挡部分的视觉特征。遮挡特征补全后的施工围栏可应用于闭合检测等多种施工安全监管任务。研究选择多个施工场地的围栏图像进行案例验证,可实现精准的围栏缺失特征补全(平均交并比mIoU>95.5%)。研究成果为施工场景的遮挡特征推理提供研究范式,有效提升了智能化施工安全监管效率。

施工安全监管  /  施工特征遮挡  /  amodal特征补全  /  实例分割  /  计算机视觉

In construction safety inspections, visual obstructions often lead to missing features, resulting in dangerous misjudgments. To improve the efficiency of risk identification in construction, a method for occluded feature inference based on amodal completion technology was proposed, using construction fence detection as a case study. First, a dataset of fence detection images with visual obstructions was created using image synthesis techniques. Then, a combination of YOLOv8 instance segmentation and the amodal segmenter based on boundary uncertainty estimation (ASBU) feature completion network was used to infer the visual features of the occluded parts of the fence. The completed features of the occluded construction fences can be applied to various construction safety monitoring tasks, such as closed-loop detection. The approach was validated using fence images from multiple construction sites, achieving precise feature completion for occluded fences (average intersection ratio mIoU>95.5%). The research results provide a framework for feature inference in occluded construction scenes, which enhances the efficiency of intelligent construction safety supervision.

construction safety supervision  /  occluded construction features  /  amodal feature completion  /  instance segmentation  /  computer vision
卢昱杰, 陈宇凡, 魏伟. 施工场景遮蔽图像特征推理与语义补全方法:以施工围栏为例. 科学技术与工程, 2025 , 25 (16) : 6898 -6912 . DOI: 10.12404/j.issn.1671-1815.2406113
Yu-jie LU, Yu-fan CHEN, Wei WEI. Image Feature Inference and Semantic Completion in Occluded Construction Scenes: A Case Study of Construction Fences[J]. Science Technology and Engineering, 2025 , 25 (16) : 6898 -6912 . DOI: 10.12404/j.issn.1671-1815.2406113
规范的施工安全管理将有效避免工程危险事故发生,是保障工程顺利进行和项目成功交付的关键因素。近年来,建筑施工事故层出不穷,2020年,全国共发生房屋市政工程生产安全事故689起,其中起重机械伤害事故占事故总数的30.43%,严重威胁到工作人员的生命安全[1]
为了提升施工安全监管效率,计算机视觉技术已逐步应用于智能化施工监控。基于目标检测、目标分类、实例分割、目标跟踪等计算机视觉技术,可用于结构健康监测[2-4]、安全装置佩戴监测[5-6]、人机安全距离监控[7-8]、脚手架安全检查[9]、围栏缺口检测[10]等。然而,由于建筑工地的施工主体繁多,施工监控画面中的关注区域易被其他机械、工人等多种变化且不规则的前景遮挡,且遮挡发生的比例和位置分布随机性强,易造成关键识别特征缺失,降低风险识别精度,致使错误报警频发。近年来,已有学者针对特征补全问题展开了相关研究。
在计算机视觉领域,图像修复(Inpainting)是一种实用的解决方法,该方法旨在根据图像的上下文信息,为图像中缺失部分生成视觉上合理且语义一致的内容。苏婷婷等[11]使用深度卷积对抗生成网络(deep convolutional generative adversarial network,DCGAN)[12]对人脸补全修复,并建立了针对人脸的相关性损失函数以促进DCGAN的优化过程,该方法在50%加噪比例下实现了较完整的图像补全。Angah等[13]提出了一种基于U-Net[14]的上下文编码器模型来去除工人遮挡,并结合条件生成对抗网络(conditional generative adversarial nets,cGAN)[15]完成图像修复。这种方法能够在去除前景遮挡后实现背景的高质量填补,从而获得更清晰的工作现场视图。然而,这类方法的目的是使补全部分与图片整体保持视觉观感一致性,缺乏对遮挡部分特征的推理能力。
除图像修复外,遮挡特征提取与推理也是一种常用的特征补全方法。Yan等[16]利用遮挡部分与未遮挡部分之间的全局结构约束来恢复遮挡特征,以提升汽车或行人在部分遮挡情况下的检测性能。但该方法仅适用于目标处于图像边缘或受拍摄视野限制导致部分特征不可见的情况。还有学者采用深度学习提取特征的方法,Li等[17]针对施工工人遮挡场景,提出了一种基于改进生成对抗插补网络(generative adversarial imputation nets,GAIN)[18]的遮挡修复方法,该方法通过卷积神经网络提取工人可见骨架信息构建特征矩阵,并预测因遮挡产生的特征缺失值。但该方法依赖视频帧之间的上下文信息,且工人骨架特征相对简单,难以应对更复杂场景下的遮挡问题。
上述方法虽然在一定程度上解决了遮挡问题,但它们的特征提取过程主要集中在如图像颜色、纹理等显式特征上,这些特征通常依赖预定义的规则,缺乏对物体整体结构的隐式理解,即这些方法无法结合对象的形状、结构等先验知识来推理,容易在识别或分割任务中产生不连贯结果,且在应对复杂场景时存在局限。
为解决上述问题,提出一种基于amodal补全的遮挡补全方法。在计算机视觉领域,amodal补全指的是在遮挡特征缺失的情况下感知目标整体特征的能力[19],该补全方法涉及modal和amodal两种掩码。前者是指目标可见部分的掩码(未被遮挡部分),而后者是指物体可见和不可见部分掩码(被遮挡部分)的并集。所提出的遮挡补全方法可广泛应用于施工安全管理任务。其中,人机作业是施工危险系数最高,最需进行安全监管的施工任务之一[20]。为防止人机碰撞事故,规范要求在开挖深度超过2 m及以上的基坑周边应安装防护围栏且确保围栏封闭[21]。为避免施工特征遮挡对围栏闭合判断造成影响,以施工围栏闭合检测任务为例,验证了遮挡补全方法在施工安全监管任务中的实际应用效果。首先,基于虚拟数据合成方法生成视野遮挡的围栏数据集。其次,将合成数据集应用于ASBU[22](amodal segmenter based on boundary uncertainty estimation)网络,提取目标可见部分的上下文信息,推理遮挡部分的不可见特征。最终,基于完整的施工目标特征应用于围栏闭合检测等施工安全监管任务。研究方法为施工目标遮挡特征推理研究提供理论借鉴,为复杂建造场景智能化施工安全动态管控提供实践参考。
研究目标是施工场景下物体被遮挡部分的特征补全,并以围栏为例,基于补全结果对围栏闭合状态进行判断,整体的研究内容分为:①基于前背景组合的遮挡数据集合成。将挖掘机作为前景对象,合成到围栏背景图像中,形成一套有真值的挖掘机遮挡围栏数据集;②基于YOLOv8的实例分割。识别场景中围栏及挖掘机的像素尺度信息,得到围栏的可见掩码;③基于ASBU的amodal补全。基于YOLOv8分割的可见部分掩码,输入ASBU预测被遮挡部分,补全形成围栏的完整掩码;④基于几何轮廓提取的围栏闭合判断。通过一系列图像层级的后处理操作,以围栏的amodal掩码为输入获取围栏中心线,并基于中心线连续情况判断实际围栏是否闭合。研究框架如图1所示。
ASBU网络需要输入包含目标可见及被遮挡区域的完整图像以进行特征推理,而现实实拍数据多为单帧静态图像,难以获取被遮挡区域的图像信息。
为了解决上述问题,提出一种图像合成的方法来获取包含遮挡区域真值的虚拟数据集,具体的方法框架如图2所示。
首先对挖掘机图像与围栏图像进行标注,以生成挖掘机前景掩码与未被遮挡的围栏掩码。然后,将挖掘机与围栏图像按照观测视角划分为平视(俯仰角:0°~30°)、轻微俯视(俯仰角:30°~50°)、俯视(俯仰角:50°~80°)3个类别。
进一步地,从图像中分离前景掩码并将其放置于同类观测视角的围栏背景中,使得挖掘机前景遮挡围栏的部分区域。通过以上步骤,既能生成围栏遮挡图像,又可获取未被遮挡的围栏掩码(即围栏amodal掩码)。
根据前景对象在背景图像中的位置,可得到合成图像中前景物体的标注信息。基于背景图像中标注的围栏掩码,在挖掘机合成到图像中后,通过图像位运算,可计算得到围栏的可见部分掩码。经过数据处理,可形成如图2右侧所示的各类掩码数据:挖掘机(前景)掩码、围栏amodal掩码及围栏modal掩码。经过处理的数据可根据具体需求生成amodal及modal数据集。
在图像合成过程中,首先调整前景对象的图像尺寸使其与背景图像在同一个像素尺度,接着将原图像中的前景对象插入到背景图像的随机位置。由于图像合成的前景背景来自不同的图像,因此合成图像多出现像素伪影[23],即前景背景交界处存在明显的分辨率与颜色突变。采用泊松融合图像处理技术[24],通过缩小图像亮度和颜色的梯度差异,使前景背景边界的外观过度更加自然,提升合成图像的外观真实性。泊松图像处理方法整体流程如图3所示。
在图像领域要处理的泊松融合函数如式(1)所示,该公式的目标是最小化融合图像中融合部分的梯度与原图像中前景对象梯度之间的差值。
$\min _{f} \iint_{\Omega}|\nabla f-\boldsymbol{v}|^{2} \text { with }\left.f\right|_{\partial \Omega}=\left.f^{*}\right|_{\partial \Omega}$
式(1)中:f为待求解的结果图像;f*为已知的目标图像;Ω为待融合区域; $\nabla$为梯度算子;v为梯度场,即原图像的梯度。
在实际计算中,图像被表示为一个二维离散网络,对于每一个像素点(i,j),离散化的拉普拉斯算子如式(2)、式(3)所示。
$\Delta f_{i, j}=f_{i+1, j}+f_{i-1, j}+f_{i, j+1}+f_{i, j-1}-4 f_{i, j}$
$\Delta f_{i, j}=\nabla \cdot v_{i, j}$
式中:Δ为拉普拉斯算子;fi,j为像素点(i,j)对应的像素值; $\nabla$为散度运算符;vi,j为原图像在(i,j)位置的梯度。
为了方便计算机求解,将公式转换为矩阵形式有式(4),由于未知量巨大,采用Jacobi迭代法来求解,迭代次数设置为20 000次,每次迭代都会更新图像中每个像素的值,并逐渐逼近最终解,如式(5)所示。
Ax=b
$x_{i}^{(k+1)}=\frac{1}{A_{i, i}}\left(b_{i}-\sum_{j \neq i} A_{i, j} x_{j}^{(k)}\right)$
式中:x为待求解目标图像的像素值;b为包含梯度信息的向量;A为稀疏矩阵,对应像素的邻域关系; x i ( k )为第i个未知量在第k次迭代时的值;Ai,jA矩阵中第i行、第j列的元素;bi表示向量b的第i个元素。
为了合成图像中的挖掘机与围栏背景的整体色调一致,采用颜色直方图匹配算法调整挖掘机的颜色分布,使其更接近背景图像的颜色分布。
通过1.1节方法获得数据集,其中形成的modal数据集可用于训练YOLOv8的实例分割。将所输入的图像调整至统一尺寸输入卷积神经网络提取特征,再通过检测头完成对目标的分类及掩码预测。
基于YOLOv8实例分割网络的结构如图4所示,主要由骨干网络(Backbone)、颈部(Neck)、头部(Head)构成,其中,骨干网络从图像中提取图像特征,颈部作为中间层,对骨干网络提取的特征进行多尺度融合,头部作为模型的最后一层,采用Softmax激活函数,输出目标所在的边界框、类别概率和像素级分割掩码。
基于YOLOv8实例分割获取的挖掘机和围栏的modal掩码,输入ASBU[22]网络进行围栏的amodal补全,其网络架构如图5所示。
ASBU接受以下数据作为训练的输入集:①RGB图像;②被遮挡物的modal掩码;③遮挡物掩码及其与被遮挡物的边界。
在输入的数据集中,ASBU随机选取遮挡物和被遮挡物及它们的modal掩码,并计算被遮挡物被前景遮挡后的掩码。同时,根据遮挡物被遮挡的情况计算遮挡发生的边界掩码,该掩码被估计为遮挡物和被遮挡物的形态学放大掩码的交集区域。
将数据输入UNet后,将输出一个H×W×2的特征图,该特征图的两个通道分别用于amodal分割预测和不确定性预测,其中HW分别为输入图像的高度和宽度。不确定性预测旨在隐式地捕获训练对象实例的“先验形状”,通过最小化预测的不确定性,网络能够捕捉到物体的合理形状[22]
amodal分割预测值由Sigmoid函数输出,不确定性预测值由Softplus函数输出,分别如式(6)、式(7)所示。ASBU采用的损失函数结合了数据误差和不确定性两方面,具体如式(8)、式(9)所示。
通过引入该不确定性加权的损失函数,综合考虑了给定的真值modal掩码(伪amodal掩码)和输出的预测amodal掩码的差异及预测的不确定性。
mi=sigmoid(z)= 1 1 + e - z
ui=softplus(z)=ln(1+ez)
式中:mi∈[0,1]为amodal掩码预测值;i为像素的索引;sigmoid、softplus为神经网络的激活函数;z为函数的输入;uiR+为不确定性预测值。
L= 1 N i = 1 NJ( m C i=0)Li+λJ( m C i=1)Li
Li= 1 2 m i * - m i u i 2 + u i 2
式中: m C i为遮挡物掩码值;J(·)为推断函数; m i *为真实掩码值;λ为用于平衡不同区域的损失权重的常数,设置为5,即遮挡物内部被赋予更多的权重;Li为第i个像素的损失函数。
ASBU算法采用平均交并比(mIoU)和像素准确率(pixel accuracy,PA)进行精度评价。平均交并比[式(10)]用于衡量预测amodal掩码与真值掩码之间的重叠程度;像素准确率[式(11)]的数值结果为预测的amodal掩码中正确预测的像素与真值掩码像素的数量比。
mIoU= m p G T m p G T
PA= m p G T G T
式中:mp为预测掩码;GT为真实掩码; ·为集合中元素的数量,即像素数。
由于围栏间的微小间隙[图6(a)]也会影响其闭合状态判断。提出一套基于围栏掩码特征的形态学像素膨胀方法来解决上述问题,该方法综合考虑图像的俯仰角和围栏的规范间距,将规范间距设定为掩码膨胀尺寸,以实际与像素距离的等比关系为尺度提出像素膨胀估算公式,见式(12)。
为了计算条状掩码的近似宽度,以假想长方形作为条状围栏掩码形状,掩码的周长和面积作为已知量,掩码的宽度和高度作为未知量,得到掩码近似宽度的计算公式,见式(13)。考虑到拍摄俯仰角对高度的透视影响,掩码的近似高度计算公式见式(14)。
Di=λ'd W i H i= λ ' d W i h i c o s θ
Wi= p i - p i 2 - 16 A i 4 , p i 4 A i p i 4 , p i < 4 A i
Hi=hicosθ
式中:Di为膨胀像素;λ'为放大系数,取值为1.1~1.3;Wi为图像中第i个条状掩码的近似宽度;d为规定的最大围栏间距;θ为图像对应的俯仰角;Hi为图像俯仰角下的换算实际高度;Ai为掩码的面积;pi为掩码的周长;hi为现实中围栏的实际高度。
hid的设置参考《建筑施工现场安全防护设施技术规程》(DB42/T 535—2020)的围栏尺寸和立杆间距,但该规范中并没有阐明若围栏断开,断开部分规定的最大尺寸限制,因此对于实际围栏间距和高度未知的情况,可考虑采用标准中的围栏高度和栏杆间距代入式(12)做估算。
经上述步骤,存在微小间隙的离散掩码区域[图6(b)]可膨胀并连接为完整的掩码封闭体[图6(c)],为后续闭合逻辑判断奠定基础。
对膨胀后的围栏掩码使用 OpenCV库的skeletonize算法(也称二值图像细化),提取条状掩码的骨架线(后称中心线),获得简洁的围栏中心线信息。围栏闭合判断的整体流程如图7所示。
基于精炼后的围栏中心线,通过对其连续性的判断来验证围栏是否封闭。完整闭合的中心线应无端点(端点指线段在图像中的起点或终点,它只有1个相邻像素),因此若存在端点,则该端点所在的分支必定为非闭合。
本次研究旨在通过端点检测方法判断围栏的闭合情况。具体步骤为:首先,检索图像中的端点,并使用3×3的卷积核跟踪曲线的延伸方向。当遇到另一端点或曲线分叉点时,停止检索并删除该线条上的像素。此过程重复进行,直至图像中不再存在端点。此时,保留在图像中的即为闭合的中心线;若不存在闭合的中心线,则表明图像中的围栏未闭合。当图像中检测到满足条件的闭合中心线时,将图像中的闭合区域与非闭合区域以颜色区分,如图8所示。
通过网络检索与施工现场实拍的方式共收集挖掘机图像651张,围栏图像286张。为了保证数据多样性与泛化性,图像数据包含不同大小、远近和俯仰角的目标物体。
基于1.1节所述方法初步生成22 468张合成数据。受前景放置位置随机性的影响,挖掘机可能位于语义不合理的空间位置,如漂浮于空中,或与背景中的围栏不存在遮挡关系。因此,通过人工审查的方式最终筛选出7 190张图像数据集。如前所述,所有的图像附带包含modal和amodal的语义信息,因此无需额外的人工标注。数据集的归类与合成效果如表1所示。
合成的数据集中的modal掩码用于训练YOLO模型,如前所述共包含7 190张图像及对应的标注数据,并按照7∶3的比例划分为训练集和验证集。
考虑到对围栏掩码识别的精准性和完整性需求,本实验采用YOLOv8 m-seg作为训练模型,该模型相比于YOLOv8s-seg能够达到更高的分割准确率。实验的软硬件配置信息如表2所示,网络的超参数设置如表3所示。
由于ASBU算法逻辑包含获取挖掘机对象作为围栏前景遮挡进行训练的逻辑,因此将收集后标注的原始数据直接输入ASBU网络进行训练。模型训练的软硬件配置如表4所示,训练超参数设置如表5所示。此外,ASBU的验证集使用1.1节的合成数据,该验证集包含合成RGB图像、每个物体的可见及不可见部分的掩码信息。
研究选取中国多个施工现场的围栏图像进行闭合检测方法验证(图9)。通过手机拍照、监控视频截图等方式共收集测试数据104张,数据收集过程保证拍摄位置与拍摄对象之间的俯仰角度位于20°~40°。测试图像中包含红白相间的塑料围栏[图9(a)]、以及附带安全防护网的围栏[图9(b)]。部分图像存在挖掘机前景遮挡的情况[图9(c)],以测试遮挡情况下的围栏恢复与闭合检测性能。
模型在训练集和验证集上的loss曲线及平均精度(mean average precision, mAP)随训练轮次的变化情况如图10(a)图10(b)所示,结果显示,训练集和验证集的边界框回归、目标回归、分类回归损失均在第150轮次后趋于平稳,未出现过拟合。模型的准确率P与召回率R分别为97.8%和94.4%,围栏分割精度mAP0.5与mAP0.5-0.95(mAP0.5为以0.5为IoU时算法取得的平均精度;mAP0.5-0.95为IoU以0.05步长在0.5~0.95区间上的平均mAP)高达96.3%和92.5%,如图10(c)所示,满足工程现场施工围栏检测的精度要求。
施工围栏实例分割的可视化结果如表6所示,选取的图像覆盖了不同视角的多种围栏,且包含挖掘机前景遮挡的情况。结果显示,YOLOv8算法可准确地预测施工现场图像中的挖掘机及围栏的可见掩码。
模型的损失随迭代次数的变化曲线如图11所示,训练损失在2 000次迭代前急速下降,而后呈现震荡波动的缓慢下降趋势,最终损失值稳定在约0.05。
训练过程中程序每迭代2 000次会进行一次训练过程中的验证,在训练结束后也可以根据对应迭代次数的权重文件在验证集上进行离线验证,在本次实验中,挑选了具有代表性的迭代权重文件进行离线验证,依次选取0、2 000、16 000、24 000、28 000次这些具有代表性迭代次数的权重文件进行离线验证,结果如表7所示。
ASBU算法在验证集的amodal补全结果如表8所示。训练后的ASBU网络可准确预测被挖掘机遮挡部分的围栏掩码,为围栏闭合检测的可靠性提供保障。实验结果表明,不同的遮挡位置和比例将会影响特征推理的算法表现。当遮挡位于目标中部时,算法的缺失特征推理精度较高,可精准补全和连接因遮挡造成的不连续掩码;而当遮挡位置处于目标端部时,端部区域缺乏充足的上下文信息,将在一定程度上影响特征推理精度。
基于训练好的UNet权重模型,选取loss最小的第28 000代权重模型用于验证集测试,试验结果表明,mIoU和像素准确率(pixel accuracy)分别高达95.454%和97.779%。预测的掩码与真实掩码几尽重合,验证了本方法在施工遮挡特征推理方面的精确度与可行性。
本次研究结果与其他施工场景下解决遮挡问题的方法进行对比分析,如表9所示。相比于其他方法,通过隐式理解物体的整体形态,能够基于可见掩码预测物体整体的amodal掩码,该方法能够适用于复杂遮挡场景,且在遮挡补全准确性上具有优势。
在获取了围栏的amodal掩码后,对该结果进行围栏闭合的判断验证,共选取100张实际案例中的图像进行判断。表10为8个示例的判断结果,结果显示,以人为判断结果为基准,所提方法与未进行补全时相比,判断准确率从68%提升到82%,能够准确识别画面中围栏的闭合部分。
围栏判断闭合的精度受多种因素影响,如掩码识别精度、掩码膨胀尺度等。首先,因围栏闭合判断结果依赖于掩码识别的精确度,如modal掩码识别不完全,或预测的amodal掩码未将属于同一围栏的掩码正确连接,则基于该掩码的闭合判断结果的准确性也会下降。其次,不恰当的掩码膨胀尺度也可能影响判断结果,如膨胀像素过小会导致微小的围栏间隙未填补;过大则会导致邻近平行围栏融合为一个整体,造成信息丢失。
(1)工程建造场景的施工环境复杂,施工主体繁多,易导致监控视野遮挡,干扰危险事故判断。为了提升施工安全识别的准确性,以施工围栏识别任务为例,提出一种基于amodal补全的施工遮挡特征推理方法。该方法应用ASBU网络提取围栏可见特征,并利用全局上下文信息推断被机械遮挡的不可见信息,以恢复完整的目标特征。实验证明,该方法可在复杂施工遮挡情况下实现精准的围栏特征推理与补全,精度达到95.5%。以围栏为案例进行研究,但该研究方法同样适用于受机械、工人、物料等多种前景目标遮挡的复杂特征推理。该方法还可预测被遮挡区域的RGB外观色彩特征,进而应用于基于彩色点云重建的三维场景特征还原等任务。
(2)本次研究仍存在一定的局限性,现有特征推理方法为基于单张图片的静态特征推理,特征推理精度与遮挡区域的位置分布有关。位于目标端部的遮挡区域易因上下文信息不充分,导致边缘遮挡特征难以推理与界定。未来研究将结合多视角图像特征匹配技术,通过融合与匹配多角度的目标特征,补充有限的单视角静态特征,提升目标边缘区域的遮挡特征推理精度。
  • 国家重点研发计划(2022YFC3801700)
  • 中央高校基本科研业务费专项资金(2024-1-ZD-02)
  • 中央高校基本科研业务费专项资金(22120250350)
  • 国家自然科学基金(52078374)
参考文献 引证文献
排序方式:
[1]
中华人民共和国住房和城乡建设部. 住房和城乡建设部办公厅关于2020年房屋市政工程生产安全事故情况的通报[EB/OL]. (2021-05-21)[2024-06-09]. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html
Ministry of Housing and Urban Rural Development of the People's Republic of China. Notice from the General Office of the Ministry of Housing and Urban Rural Developmenton production safety accidents in housing and municipal engineering in 2020[EB/OL]. (2021-05-21)[2024-06-09]. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html. https://www.mohurd.gov.cn/gongkai/zc/wjk/art/2022/art_17339_768565.html
[2]
柏云飞. 智能传感技术在桥梁结构健康监测中的应用研究[J]. 石河子科技, 2024(3): 55-57.
Bai Yunfei. Research on the application of intelligent sensing technology in bridge structure health monitoring[J]. Shihezi Science and Technology, 2024(3): 55-57.
[3]
Agyemang I O, Zhang X, Adjei-Mensah I, et al. Automated vision-based structural health inspection and assessment for post-construction civil infrastructure[J]. Automation in Construction, 2023, 156: 105153.
[4]
Dong C Z, Catbas F N. A review of computer vision-based structural health monitoring at local and global levels[J]. Structural Health Monitoring, 2021, 20(2): 692-743.
[5]
Fang Q, Li H, Luo X, et al. Detecting non-hardhat-use by a deep learning method from far-field surveillance videos[J]. Automation in Construction, 2018, 85: 1-9.
[6]
Fang W, Ding L, Luo H, et al. Falls from heights: a computer vision-based approach for safety harness detection[J]. Automation in Construction, 2018, 91: 53-61.
[7]
Dagan D, Isaac S. Planning safe distances between workers on construction sites[J]. Automation in Construction, 2015, 50: 64-71.
[8]
曾拓程, 王佳俊, 钟登华, 等. 基于大场景视频监控的坝面施工机械潜在碰撞风险预警方法[J]. 水利学报, 2024, 55(7): 780-790, 801.
Zeng Tuocheng, Wang Jiajun, Zhong Denghua, et al. Collision risk warning method for dam surface construction machinery based on large-scale video surveillance[J]. Journal of Hydraulic Engineering, 2024, 55(7): 780-790, 801.
[9]
Kim J, Kim J, Koo N, et al. Automating scaffold safety inspections using semantic analysis of 3D point clouds[J]. Automation in Construction, 2024, 166: 105603.
[10]
余萍, 宋祥宇. 基于Open CV围栏合规性摆放检测方法研究[J]. 科学技术与工程, 2022, 22(28): 12449-12459.
Yu Ping, Song Xiangyu. Detection method of fence compliance based on Open CV[J]. Science Technology and Engineering, 2022, 22(28): 12449-12459.
[11]
苏婷婷, 王娜. 基于图像块相似性和补全生成的人脸复原算法[J]. 科学技术与工程, 2019, 19(13): 171-176.
Su Tingting, Wang Na. Human face restoration based on natural patch likelihood and generative image inpainting[J]. Science Technology and Engineering, 2019, 19(13): 171-176.
[12]
Radford A. Unsupervised representation learning with deep convolutional generative adversarial networks[J]. arXiv Preprint, 2015: arXiv: 1511.06434.
[13]
Angah O, Chen A Y. Removal of occluding construction workers in job site image data using U-Net based context encoders[J]. Automation in Construction, 2020, 119: 103332.
[14]
Ronneberger O, Fischer P, Brox T. U-net: convolutional networks for biomedical image segmentation[C]// Medical Image Computing and Computer-assisted Intervention-MICCAI 2015: 18th International Conference. Munich: Springer International Publishing, 2015: 234-241.
[15]
Mirza M, Osindero S. Conditional generative adversarial nets[J]. arXiv Preprint, 2014: arXiv: 1411.1784.
[16]
Yan S, Liu Q. Inferring occluded features for fast object detection[J]. Signal Processing, 2015, 110: 188-198.
[17]
Li Z, Li D. Action recognition of construction workers under occlusion[J]. Journal of Building Engineering, 2022, 45:103352.
[18]
Yoon J, Jordon J, Schaar M. Gain: missing data imputation using generative adversarial nets[C]// International Conference on Machine Learning. Basel: PMLR, 2018: 5689-5698.
[19]
Kanizsa G, Legrenzi P, Bozzi P. Organization in vision: essays on gestalt perception[M]. New York: Praeger, 1979.
[20]
Gu B, Guo H, Huang Y, et al. Computer vision-based human-machine collision warning system on construction site[C]// IOP Conference Series: Earth and Environmental Science. London: IOP Publishing, 2022: 032014.
[21]
湖北省住房和城乡建设厅, 湖北省市场监督管理局. 建筑施工现场安全防护设施技术规程: DB 42/T 535—2020[S]. 武汉: 湖北省市场监督管理局, 2020.
Hubei Provincial Department of Housing and Urban Rural Development, Hubei Provincial Market Supervision Administration. Building design standard: DB 42/T 535—2020[S]. Wuhan: Hubei Provincial Market Supervision Administration, 2020.
[22]
Nguyen K, Todorovic S. A weakly supervised amodal segmenter with boundary uncertainty estimation[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision. New York: IEEE, 2021: 7396-7405.
[23]
Dwibedi D, Misra I, Hebert M. Cut, paste and learn: surprisingly easy synthesis for instance detection[C]// Proceedings of the IEEE International Conference on Computer Vision. New York: IEEE, 2017: 1301-1310.
[24]
Pérez P, Gangnet M, Blake A. Poisson image editing[M]. Berlin: Springer International Publishing, 2023.
2025年第25卷第16期
PDF下载
401
138
引用本文
BibTeX
文章信息
doi: 10.12404/j.issn.1671-1815.2406113
  • 接收时间:2024-08-15
  • 首发时间:2025-07-09
  • 出版时间:2025-06-08
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-08-15
  • 修回日期:2025-03-17
基金
国家重点研发计划(2022YFC3801700)
中央高校基本科研业务费专项资金(2024-1-ZD-02)
中央高校基本科研业务费专项资金(22120250350)
国家自然科学基金(52078374)
作者信息
    1 同济大学土木工程学院, 上海 200092
    2 同济大学工程结构性能演化与控制教育部重点实验室, 上海 200092
    3 同济大学上海智能科学与技术研究院, 上海 200092

通讯作者:

* 魏伟(1997—),男,汉族,山东济南人,博士研究生。研究方向:智能建造。E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/kxjsygc/CN/10.12404/j.issn.1671-1815.2406113
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