We use open EO data such as Sentinel imagery from the Copernicus Programme in combination with cutting-edge Open Source Machine Learning to improve environmental layers in Europe.
We use a variety of free and open source software solutions for geospatial data (FOSS4G) in combination with state-of-the-art Machine Learning Algorithms as implemented in the R and Python programming environments. We run processing within EU-supported High Performance Computing (HPC) / and similar Cloud computing infrastructures.
We base harmonization of geographic data (EU scale) on using the EO Sentinel 1/2 and Landsat 6/7/8 layers at 10, 30 and 60 m: cloud-free mosaics (2000–2020) produced via the DIAS infrastructure.
We combine multiscource data — administrative layers, expert-based land cover maps, forest altases and maps, various type of legacy data — with huge volumes of time-series remote sensing images to produce statistical unbiased predictions of target variables and estimate realistic measures of uncertainty.
We develop tools for automated harmonization/conversion of variables including automated upscaling, downscaling, resolving of edge-problems and cross-border problems, imputing of missing values etc.
Latest post
Latest post, data sets, discussion panels and discussion points.