Point Cloud for City Generation Code for City4CFD

Description

A Python pipeline that generates Digital Surface Model (DSMs) and converts them into point clouds for City4CFD modeling. Due to the lack of open, high-resolution global DSM datasets, the generation logic differs by target city (see the figure below).

Repository Structure

├── code/
│   ├── main.py
│   ├── get_dtm.py
│   ├── get_dsm.py
│   └── convert_dsm_to_las.py
├── output_example/
│   ├── london.laz
│   ├── seoul.laz
│   ├── madrid.laz
└── README.md

Prerequisites / Setup

• OpenTopography API Key: Required to download DTM datasets (e.g., COP30, NASADEM) for cities outside of England/Netherlands.
• Libraries: Install the required dependencies using the following command:

  pip install earthengine-api geemap geopandas pandas shapely rasterio osmnx overturemaps numpy laspy lazrs pyproj tqdm requests
  

How to Run

1. Download the /code/ directory
2. Run main.py with the necessary arguments Examples by case:
   • [Case 1] For cities in England/Netherlands (uses Google Earth Engine high-resolution data)

     python main.py --region England --city London --crs EPSG:27700
     

   • [Case 2] For cities outside of England/Netherlands (uses OpenTopography for DTM and OSM/Overture Maps for building data)
     • Save as uncompressed LAS

       python main.py --region "South Korea" --city Seoul --crs EPSG:32652 --dtm_dataset NASADEM --api-key _put_your_api_key_ --no-compress
       

     • Convert with subsampling (every 2nd pixel)

       python main.py --region Spain --city Madrid --crs EPSG:25830 --bbox "[-3.7050, 40.4430, -3.6750, 40.4850]" --dtm_dataset COP30 --api-key _put_your_api_key_ --subsample 2
       

Expected Output

File Name	Description	Resolution
[city_name].tif	Merged DSM	• England: 1m × 1m • Netherlands: 0.5m × 0.5m • Others: 1m × 1m
[city_name]_dtm.tif	DTM (Only for non-England/Netherlands cities)	1m × 1m
[city_name].las / .laz	Point Cloud	-

Limitations

For the cities outside England/Netherlands:

• Missing Heights: If a building footprint lacks height data, it is assigned a default height of 3.0m (assuming a single-story).
• Level of Detail: LoD1

Reference

• https://github.com/slzhang-git/cenergy/blob/main/src/cenergy3/core.py