oseiskar / CityGMLTools

Convert CityGML models (of Espoo) to OBJ format defined in WGS84 ENU

CityGML conversion tools (Espoo)

The main purpose of this repository is allowing conversions of this 3D model data of the city of Espoo https://kartat.espoo.fi/3D/citymodel_en.html to a commonly understood 3D format and coordinate system, using Open Source tools.

The first implementation converts to Wafefront OBJ format, which can be viewed with, e.g., this online tool: https://3dviewer.net/. The result is in ENU coordinates defined at the given origin.

The code includes some hacks (like --auto-fix-walls) that may not be required with other CityGML datasets. Compatibility with other CityGML datasets is currently unknown. On the other hand, other tools such as https://github.com/tudelft3d/CityGML2OBJs, which did not work with the Espoo dataset, may work with those.

Usage

1. Optional but recommended: virtualenv
   • (once) Set up virtualenv: python3 -m venv venv
   • (every time): Activate: venv/bin/activate
2. (once) Install: pip install -r requirements.txt

Coordinate helpers

To perform bounding box queries, you may need to convert coordinate systems between, WGS84 (the "satellite coordinates" you get from, e.g., Google Maps) and local coordinate systems like EPSG:3879 (that, e.g., move with the tectonic plates).

python coordinates.py --coordinateSystem=EPSG:3879 single from_wgs 60.1757 24.8040

which should print something like (6673664.363125221, 25489121.254538767).

Downloading

List available datasets

python download.py --url=https://kartat.espoo.fi/teklaogcweb/wfs.ashx GetCapabilities

Look at the FeatureType/Name fields in the result XML, something like bldg:building_lod2.

Download a particular dataset

mkdir -p data
python download.py --url=https://kartat.espoo.fi/teklaogcweb/wfs.ashx \
  GetFeature "bldg:building_lod2" \
  --coordinateSystem=EPSG:3879 --latitude=6673664 --longitude=25489121 --radius=200 \
  --maxFeatures=100 > data/dataset.xml

After checking that the data looks correct, try expanding the search radius --radius (in meters) or removing max feature limit (by setting --maxFeatures=0).

Conversions

Example:

python convert.py to_obj 6673664 25489121 \
  --auto-fix-walls \
  --coordinateSystem=EPSG:3879 < data/dataset.xml > data/dataset.obj

If the imported area is large (several kilometers in diameter), consider also using --accurate-enu, which should correctly handle the curvature of the Earth and other effects of similar magnitude, which are less accurately modeled in the default mode, --fast-enu.