Semantic Building Models

The semantic building models are structured according to the OGC CityGML standard.

Level of Detail 3

Based on the official LOD2 building models and additional MLS point clouds, the buildings in LOD3 with detailed facades were modeled manually according to this guideline. The SketchUp projects as well as the CityGML 2.0 models are available for download per individual building.

Building Models in LOD3 — Semantic LOD3 building models manually created based on point clouds

The individual building models can be cloned with the following command:

git clone --depth 1 git@gitlab.lrz.de:tum-gis/tum2twin-datasets.git
cd tum2twin-datasets/citygml/lod3-building-datasets

Exemplary 3D building model as a CAD model in SketchUp (more SketchUp projects)

Level of Detail 2 + Textures

Derived from the above-mentioned dataset with manually added terrestrial-acquired textures (can be treated as ground-truth). They can be downloaded as CityGML 2.0 datasets for the TUM Campus and as SketchUp projects. The individual building models can be downloaded here and with the following command:

git clone --depth 1 git@gitlab.lrz.de:tum-gis/tum2twin-datasets.git
cd tum2twin-datasets/citygml/lod2-textured-building-datasets

We also store the deprecated initial textured models (used e.g., in Scan2LoD3 paper) under this folder

Level of Detail 2

The building models in Level of Detail (LOD) 2 are based on the official cadastre with stable object identifiers and an absolute accuracy ranging between 1-3cm. They can be downloaded as CityGML 2.0 datasets for the whole of Bavaria from the open data portal of the Bavarian State Mapping Agency.

Building Models in LOD2 — Semantic LOD2 building models provided by the Bavarian State Mapping Agency

The tiles, 690_5334 and 690_5336, can be downloaded with the following command:

curl --compressed --remote-name-all "https://download1.bayernwolke.de/a/lod2/citygml/{690_5334,690_5336}.gml"

HD Map Semantic Vegetation Models