roofer performs fully automatic LoD2 building reconstruction from a pointcloud and a building footprint polygon. It was originally developed, and is still used, for the 3DBAG project.
- It reconstructs a 3D building model from a pointcloud and a 2D footprint polygon
- It is a fully automated process, there is no manual modelling required.
- It is possible to tweak the reconstruction parameters to adjust it a little to different input data qualities.
- It can output different Level of Details: LoD1.2, LoD1.3, LoD2.2. See the refined Levef of Details by the 3D geoinformation research group.
- Usable either as a command line application or as a library with C++ and Python bindings.
- With the CLI application the building models are outputted to a CityJSONSequence file.
Documentation is provided at https://3dbag.github.io/roofer/. The source code is available at https://github.com/3DBAG/roofer.
The building reconstruction algorithms underpinning roofer were originally developed within the 3D geoinformation research group at the Technical University of Delft. From 2022 onwards, 3DGI, a spinoff company from the aforementioned research group, joined the development efforts.
- This project has received funding from the European Research Council (ERC):
- 2016-2022 under the European Unions Horizon2020 Research & Innovation Programme (grant agreement no. 677312 UMnD: Urban modelling in higher dimensions).
- 2022-2024 under the Horizon Europe Research & Innovation Programme (grant agreement no. 101068452 3DBAG: detailed 3D Building models Automatically Generated for very large areas).
In 2024 this project has received funding from Kadaster, the Netherlands' Cadastre, Land Registry and Mapping Agency.
Prior to 2024 the building reconstruction algorithms were developed as part of Geoflow and the gfp-building-reconstruction plugin. During the summer of 2024 the code was refactored and the roofer project was born.
The roofer project is licensed under the GPLv3 license.