Automatic detection of building typology using deep learning methods on street level images

dc.citation.journalTitleBuilding and Environment
dc.contributor.authorDuque, J.
dc.contributor.authorGonzalez, D.
dc.contributor.authorRueda Plata, Diego
dc.contributor.authorAcevedo, A.
dc.contributor.authorRamos, R.
dc.contributor.authorBetancourt, A.
dc.contributor.authorGarcía, S.
dc.contributor.researchgroupMecánica Aplicadaspa
dc.date.accessioned2020-03-10
dc.date.accessioned2021-04-16T20:10:43Z
dc.date.available2021-04-16T20:10:43Z
dc.date.issued2020-03-20
dc.date.submitted2020-01-18
dc.description.abstractAn exposure model is a key component for assessing potential human and economic losses from natural disasters. An exposure model consists of a spatially disaggregated description of the infrastructure and population of a region under study. Depending on the size of the settlement area, developing such models can be a costly and time-consuming task. In this paper we use a manually annotated dataset consisting of approximately 10,000 photos acquired at street level in the urban area of Medellín to explore the potential for using a convolutional neural network (CNN) to automatically detect building materials and types of lateral-load resisting systems, which are attributes that define a building's structural typology (which is a key issue in exposure models for seismic risk assessment). The results of the developed model achieved a precision of 93% and a recall of 95% when identifying nonductile buildings, which are the buildings most likely to be damaged in an earthquake. Identifying fine-grained material typology is more difficult, because many visual clues are physically hidden, but our model matches expert level performances, achieving a recall of 85% and accuracy scores ranging from 60% to 82% on the three most common building typologies, which account for 91% of the total building population in Medellín. Overall, this study shows that a CNN can make a substantial contribution to developing cost-effective exposure models. © 2020 Elsevier Ltdeng
dc.identifierhttps://eafit.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=11808
dc.identifier.doi10.1016/j.buildenv.2020.106805
dc.identifier.issn03601323
dc.identifier.issn1873684X
dc.identifier.otherWOS;000550146500008
dc.identifier.otherSCOPUS;2-s2.0-85084182110
dc.identifier.urihttp://hdl.handle.net/10784/29226
dc.language.isoengeng
dc.publisherPERGAMON-ELSEVIER SCIENCE LTD
dc.publisher.departmentUniversidad EAFIT. Departamento de Ingeniería Mecánicaspa
dc.relation.urihttps://www.sciencedirect.com/science/article/abs/pii/S0360132320301633
dc.rightsPERGAMON-ELSEVIER SCIENCE LTD
dc.sourceBuilding and Environment
dc.subject.keywordBuildingseng
dc.subject.keywordConvolutional neural networkseng
dc.subject.keywordCost effectivenesseng
dc.subject.keywordDisasterseng
dc.subject.keywordLearning systemseng
dc.subject.keywordLosseseng
dc.subject.keywordRisk assessmenteng
dc.subject.keywordAutomatic Detectioneng
dc.subject.keywordBuilding populationeng
dc.subject.keywordBuilding typologieseng
dc.subject.keywordFine-grained materialeng
dc.subject.keywordLateral load resisting systemseng
dc.subject.keywordSeismic risk assessmenteng
dc.subject.keywordStructural typologieseng
dc.subject.keywordTime-consuming taskseng
dc.subject.keywordDeep learningeng
dc.subject.keywordartificial neural networkeng
dc.subject.keywordautomationeng
dc.subject.keywordconstruction materialeng
dc.subject.keyworddata seteng
dc.subject.keywordnatural disastereng
dc.subject.keywordurban areaeng
dc.titleAutomatic detection of building typology using deep learning methods on street level imageseng
dc.typeinfo:eu-repo/semantics/articleeng
dc.typearticleeng
dc.typeinfo:eu-repo/semantics/publishedVersioneng
dc.typepublishedVersioneng
dc.type.localArtículospa

Archivos

Colecciones