Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction
Digital twins are increasingly gaining popularity as a method for simulating intricate natural and urban environments, with the precise segmentation of 3D objects playing an important role. This study focuses on developing a methodology for extracting buildings from textured 3D meshes, employing the...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2024-01-01
|
| Series: | ISPRS International Journal of Geo-Information |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2220-9964/13/1/20 |
| _version_ | 1797343569533468672 |
|---|---|
| author | Frédéric Leroux Mickaël Germain Étienne Clabaut Yacine Bouroubi Tony St-Pierre |
| author_facet | Frédéric Leroux Mickaël Germain Étienne Clabaut Yacine Bouroubi Tony St-Pierre |
| author_sort | Frédéric Leroux |
| collection | DOAJ |
| description | Digital twins are increasingly gaining popularity as a method for simulating intricate natural and urban environments, with the precise segmentation of 3D objects playing an important role. This study focuses on developing a methodology for extracting buildings from textured 3D meshes, employing the PicassoNet-II semantic segmentation architecture. Additionally, we integrate Markov field-based contextual analysis for post-segmentation assessment and cluster analysis algorithms for building instantiation. Training a model to adapt to diverse datasets necessitates a substantial volume of annotated data, encompassing both real data from Quebec City, Canada, and simulated data from Evermotion and Unreal Engine. The experimental results indicate that incorporating simulated data improves segmentation accuracy, especially for under-represented features, and the DBSCAN algorithm proves effective in extracting isolated buildings. We further show that the model is highly sensible for the method of creating 3D meshes. |
| first_indexed | 2024-03-08T10:49:40Z |
| format | Article |
| id | doaj.art-bcc65bb989804e15bb80a0fce2eb1d63 |
| institution | Directory Open Access Journal |
| issn | 2220-9964 |
| language | English |
| last_indexed | 2024-03-08T10:49:40Z |
| publishDate | 2024-01-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | ISPRS International Journal of Geo-Information |
| spelling | doaj.art-bcc65bb989804e15bb80a0fce2eb1d632024-01-26T16:50:08ZengMDPI AGISPRS International Journal of Geo-Information2220-99642024-01-011312010.3390/ijgi13010020Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building ExtractionFrédéric Leroux0Mickaël Germain1Étienne Clabaut2Yacine Bouroubi3Tony St-Pierre4Department of Applied Geomatics, Center for Applications and Research in Remote Sensing (CARTEL), University of Sherbrooke, 2500 Boulevard de l’Université, Sherbrooke, QC J1K 2R1, CanadaDepartment of Applied Geomatics, Center for Applications and Research in Remote Sensing (CARTEL), University of Sherbrooke, 2500 Boulevard de l’Université, Sherbrooke, QC J1K 2R1, CanadaDepartment of Applied Geomatics, Center for Applications and Research in Remote Sensing (CARTEL), University of Sherbrooke, 2500 Boulevard de l’Université, Sherbrooke, QC J1K 2R1, CanadaDepartment of Applied Geomatics, Center for Applications and Research in Remote Sensing (CARTEL), University of Sherbrooke, 2500 Boulevard de l’Université, Sherbrooke, QC J1K 2R1, CanadaXEOS Imaging Inc., 1405 Boulevard du Parc-Technologique, Bureau 110, Quebec City, QC G1P 4P5, CanadaDigital twins are increasingly gaining popularity as a method for simulating intricate natural and urban environments, with the precise segmentation of 3D objects playing an important role. This study focuses on developing a methodology for extracting buildings from textured 3D meshes, employing the PicassoNet-II semantic segmentation architecture. Additionally, we integrate Markov field-based contextual analysis for post-segmentation assessment and cluster analysis algorithms for building instantiation. Training a model to adapt to diverse datasets necessitates a substantial volume of annotated data, encompassing both real data from Quebec City, Canada, and simulated data from Evermotion and Unreal Engine. The experimental results indicate that incorporating simulated data improves segmentation accuracy, especially for under-represented features, and the DBSCAN algorithm proves effective in extracting isolated buildings. We further show that the model is highly sensible for the method of creating 3D meshes.https://www.mdpi.com/2220-9964/13/1/20semantic segmentation3D building segmentationcluster analysis3D city models3D meshdata simulation |
| spellingShingle | Frédéric Leroux Mickaël Germain Étienne Clabaut Yacine Bouroubi Tony St-Pierre Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction ISPRS International Journal of Geo-Information semantic segmentation 3D building segmentation cluster analysis 3D city models 3D mesh data simulation |
| title | Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction |
| title_full | Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction |
| title_fullStr | Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction |
| title_full_unstemmed | Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction |
| title_short | Improving Three-Dimensional Building Segmentation on Three-Dimensional City Models through Simulated Data and Contextual Analysis for Building Extraction |
| title_sort | improving three dimensional building segmentation on three dimensional city models through simulated data and contextual analysis for building extraction |
| topic | semantic segmentation 3D building segmentation cluster analysis 3D city models 3D mesh data simulation |
| url | https://www.mdpi.com/2220-9964/13/1/20 |
| work_keys_str_mv | AT fredericleroux improvingthreedimensionalbuildingsegmentationonthreedimensionalcitymodelsthroughsimulateddataandcontextualanalysisforbuildingextraction AT mickaelgermain improvingthreedimensionalbuildingsegmentationonthreedimensionalcitymodelsthroughsimulateddataandcontextualanalysisforbuildingextraction AT etienneclabaut improvingthreedimensionalbuildingsegmentationonthreedimensionalcitymodelsthroughsimulateddataandcontextualanalysisforbuildingextraction AT yacinebouroubi improvingthreedimensionalbuildingsegmentationonthreedimensionalcitymodelsthroughsimulateddataandcontextualanalysisforbuildingextraction AT tonystpierre improvingthreedimensionalbuildingsegmentationonthreedimensionalcitymodelsthroughsimulateddataandcontextualanalysisforbuildingextraction |