Earth observation infrastructure is developing rapidly these years, with increasingly more satellites being send into space, some of these in large swarms of entire microsatellite constellations. Consequently, satellite remote sensing is not limited by a lack of data, but rather our capacity to make best use of all the data we have available already. The key is to build autonomous and intelligent systems on top of satellite imagery data streams, with a view to detect, analyse and process images faster and more accurately than ever before.
Imagine the powerful combination of a living digital library that documents every inch of our planet on a daily basis, and the tools and platforms allowing us to extract meaningful insight from this library at scale. We increasingly apply deep learning-based approaches to serve this purpose, allowing us to build effective and scalable solutions that serves unique insight about the change and dynamics on our planet. We use deep learning to underpin the information needs and requirements within many different sectors, including: