High resolution snow cover fraction (SCF)

SCF is derived from Sentinel-2 and Landsat data using an openEO variant of the SNOWFLAKES algorithm.

The method first splits the target scene into ranges based on solar incidence angles, which is relevant to inspect pixels with a similar illumination condition.

For each range, representative snow/no-snow pixels are selected.

This selection of pixels is used to train an SVM model, which is then used to estimate snow fraction in all other pixels.

The main entrypoint is: openeo_mountains_snow.collect_training.run_openeo.