We generate a Level-of-Detail 1 (LoD1) 3D City Model (buildings and terrain) from OpenStreetMap (osm) contributions with elevation from a raster Digital Elevation Model (DEM).
There are two procesing strategies
| Village/Campus | District [This is on hold] |
|---|---|
| If your Area-of-Interest (aoi) has less than 2 500 buildings, you are welcome to choose village/campus |
Please choose District if your aoi has more than 2 500 buildings |
Input a raster DEM. Script will call for the osm contributions.
Output includes:
i. a topologically correct LoD1 City Model (information rich building models seperate from the ground; but when connected to the terrain form a water-tight surface)*;
ii. one use case of 3D city models. Population estimation and the calculation of Building Volume per Capita; and
iii. an interactive .html which you can navigate and query.
* the goal is a model that conforms to the ISO 19107 standard [connecting and planar surfaces, correct orientation of the surfaces and watertight volumes]. I have not tested this for all possibilities. If the result you achieve is not; you are welcome to raise an issue. I depend on you to help me improve.
Documentation is available. The announcement is noteworthy.
For a quickstart (and to see the value and usefulness of a 3D City Model):
flowchart LR
A[osm_LoD1_3DCityModel-walkthrough.ipynb] -->B(CityJSONspatialDataScience.ipynb)
B --> D[Population estimation];
B --> E[Building Volume per Capita];
B --> F[Interactive visualization];
This repository is inspired by 3dfier; product of the 3D geoinformation group at TUDelft
Its paper: Ledoux H, Biljecki F, Dukai B, Kumar K, Peters R, Stoter J, and Commandeur T (2021). 3dfier: automatic reconstruction of 3D city models. Journal of Open Source Software, 6(57), 2866. website and github are available.