Automatic detection of building typology using deep learning methods on street level images
dc.citation.journalTitle | Building and Environment | |
dc.contributor.author | Duque, J. | |
dc.contributor.author | Gonzalez, D. | |
dc.contributor.author | Rueda Plata, Diego | |
dc.contributor.author | Acevedo, A. | |
dc.contributor.author | Ramos, R. | |
dc.contributor.author | Betancourt, A. | |
dc.contributor.author | García, S. | |
dc.contributor.researchgroup | Mecánica Aplicada | spa |
dc.date.accessioned | 2020-03-10 | |
dc.date.accessioned | 2021-04-16T20:10:43Z | |
dc.date.available | 2021-04-16T20:10:43Z | |
dc.date.issued | 2020-03-20 | |
dc.date.submitted | 2020-01-18 | |
dc.description.abstract | An 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 Ltd | eng |
dc.identifier | https://eafit.fundanetsuite.com/Publicaciones/ProdCientif/PublicacionFrw.aspx?id=11808 | |
dc.identifier.doi | 10.1016/j.buildenv.2020.106805 | |
dc.identifier.issn | 03601323 | |
dc.identifier.issn | 1873684X | |
dc.identifier.other | WOS;000550146500008 | |
dc.identifier.other | SCOPUS;2-s2.0-85084182110 | |
dc.identifier.uri | http://hdl.handle.net/10784/29226 | |
dc.language.iso | eng | eng |
dc.publisher | PERGAMON-ELSEVIER SCIENCE LTD | |
dc.publisher.department | Universidad EAFIT. Departamento de Ingeniería Mecánica | spa |
dc.relation.uri | https://www.sciencedirect.com/science/article/abs/pii/S0360132320301633 | |
dc.rights | PERGAMON-ELSEVIER SCIENCE LTD | |
dc.source | Building and Environment | |
dc.subject.keyword | Buildings | eng |
dc.subject.keyword | Convolutional neural networks | eng |
dc.subject.keyword | Cost effectiveness | eng |
dc.subject.keyword | Disasters | eng |
dc.subject.keyword | Learning systems | eng |
dc.subject.keyword | Losses | eng |
dc.subject.keyword | Risk assessment | eng |
dc.subject.keyword | Automatic Detection | eng |
dc.subject.keyword | Building population | eng |
dc.subject.keyword | Building typologies | eng |
dc.subject.keyword | Fine-grained material | eng |
dc.subject.keyword | Lateral load resisting systems | eng |
dc.subject.keyword | Seismic risk assessment | eng |
dc.subject.keyword | Structural typologies | eng |
dc.subject.keyword | Time-consuming tasks | eng |
dc.subject.keyword | Deep learning | eng |
dc.subject.keyword | artificial neural network | eng |
dc.subject.keyword | automation | eng |
dc.subject.keyword | construction material | eng |
dc.subject.keyword | data set | eng |
dc.subject.keyword | natural disaster | eng |
dc.subject.keyword | urban area | eng |
dc.title | Automatic detection of building typology using deep learning methods on street level images | eng |
dc.type | info:eu-repo/semantics/article | eng |
dc.type | article | eng |
dc.type | info:eu-repo/semantics/publishedVersion | eng |
dc.type | publishedVersion | eng |
dc.type.local | Artículo | spa |