
Our results for Continental Europe (https://doi.org/10.5281/zenodo.4724549) show that machine learning can be used to remove canopy and buildings, although some post-processing is still needed to create hydrologically correct DTMs. We will next build such a system for the whole world (compare with FABDEM) and then also derive some key terrain variables that can be used for various projects.
What is the challenge?
On the one hand, there is an increasing number of sources of elevation / altitude data while, on the other hand, users typically require only a single functional elevation service, preferably one that most accurately depicts hydrological relationships in a landscape.
Land relief parameterization follows the higher resolution DTM product. Higher resolution data product leads to higher and more intense computation.
Our solution
Machine learning can be used to remove canopy and buildings, although some post-processing is still needed to create hydrologically correct DTMs. We will next build such a system for the whole world (compare with FABDEM) and then also derive some key terrain variables that can be used for various projects.
Integration of WhiteBoxTool developed by a member in ISG, Prof. John Lindsay. Powered by the programme, the efficiency of connecting hydrological and parallelization improves dramatically.

Who will benefit?
Soil scientists, hydrologists, geomorphologists, geologists, as well as the International Society for Geomorphometry will all benefit from this use case.
Scope
Target Partner Organizations
OEMC Leading Partner
Get in touch!
Type of output
Digital terrain model (global), topographical varibles (derivatives of DTMs).
Technology readiness level
TRL5: Technology validated in relevant environment
Location
Worldwide
Links of interest
- Ensemble Digital Terrain Model (EDTM) of the world
- OpenLandMap ensemble digital terrain model
- A Simply Updatable Cloud-based Ensemble Digital Terrain Model
- A 30 m global map of elevation with forests and buildings removed
- An Ensemble Digital Terrain Model of the world at 30 m spatial resolution (EDTM30)
- Yu-Feng Ho: Using GRASS, SAGA and Whiteboxtool to map global land relief parameterization
- Yu-feng Ho and Johannes Heisig: Accessing Big Vector Data on the Cloud using Arrow Parquet