dtm loader and converter to enable accurate terrain following using DJI drones

!Tested with DJI Mavic 3 Enterprise & DJI Matrice 350 RTK!

Small script that turns a pile of 1-metre German DGM raster tiles into a single, cloud-optimised GeoTIFF in geographic WGS-84 coordinates and ellipsoidal heights. Can be used for any other tiles as well.

What it does

1. Download tiles • Reads the Metalink catalogue .meta4 files • Uses aiohttp + aiofiles to fetch every listed GeoTIFF

2. Mosaic in UTM32N or any other CRS
   • Merges the tiles with Rasterio

3. Vertical datum shift (DHHN2016 (or any other orthometric heights) → WGS-84 ellipsoid)
   • Loads the national geoid model
   • Adds the geoid undulation to every pixel

4. Re-project & COG
   • Calls gdal.Warp to transform the temp raster to EPSG 4326, bilinear resample, LZW-compress and tile.
   • Output: a ready-to-serve GeoTIFF at native ground resolution.

-> Now import this to DJI's remote controller to use it for terrain following!

Dependencies

• GDAL ≥ 3.10 (gdal / libgdal-core)
• Rasterio 1.4
• NumPy ≥ 2.0
• aiohttp 3.12 · aiofiles 24.1
• tqdm 4.67

Installation (Conda)

This package was developed and tested on Windows Server 2016 running python=3.9. We use conda to create a new environment:

user@userpc: /dtm_loader$ conda env create -f environment.yml conda activate dtm_loader