One of the main obstacles in satellite imagery is the presence of clouds, which limits the capability to make observations and ultimately to take optimal decisions – especially in missions with long revisit times. Moreover, since cloudless conditions depend on the seasons and are not spread evenly throughout the year, periods of low or no data availability can extend for a significant amount of time.
ClearSky Vision addresses these problems by integrating data from various missions (Sentinel-1, Sentinel-2, and more) and using a novel neural network architecture to predict urban and landscape changes underneath cloud cover. This will allow for more consistent and frequent retrieval of parameters and indices that are crucial to improved time-series analysis and anomaly detection, with particular benefits for new and existing solutions in the farming sector.
“ClearSky uses a neural network to make relevant information from satellite data available even during heavy cloud coverage. This increases the reliability of satellite-based recommendations for action in agriculture.” – Jörg Migende, Head of Agricultural Distribution & Head of Digital Farming. BayWa AG
Morten Fjord Pedersen, Malthe Dahl Jensen