Machine learning is one of the keys in the development of modern Earth Observation satellite missions. Model training requires precise Earth simulation as basis of applications such as vegetation monitoring, prospecting for minerals, soil use and climate change studies, among others.
SENTIFLEX (https://ipl.uv.es/sentiflex) project led by Dr. Jochem Verrelst generates data observation and predictive models for the Earth Observation Missions program from the European Space Agency, in the scope of two future missions: CHIME and FLEXE.
These scenes calculate geometry projection, shadows, cloud presence in the scene as well as cloud shadow projections. A single scene with 10m spatial resolution and 2480 spectral bands can give an output of 300GB in around an hour time.
These vast amount of information will later be used to develop machine learning retrieval algorithms, and in a second phase of the project the generated data will be reduced.