AI4Copernicus – Health Bootstrapping services
This service is focused on probabilistic downscaling (super-resolution) of CAMS model output. Current AC and AQ models output forecasts at relatively low resolution (e.g., ca. 10 to 40 km for the CAMS EU and global domains). This service generates high-resolution (currently ~ 10km) air quality maps from low-resolution (~40 -80km) CAMS model (re)analysis and/or forecast output.
License
Apache License 2.0 (Apache-2.0)
To be used in the scope of AI4Copernicus project.
Main Characteristic
- Super resolution of CAMS output to a better resolution (x4 the original resolution of CAMS datasets).
- Quantify the uncertainity in the generated output datasets using GANs.
- Providing scripts for pretraining and training.
- Code packaged in a Docker image and the service can be used as a Docker container.
Technical Categories
AI services
Business Categories
Earth Observation
Last updated
11.10.2023 - 14:09