Earth Observation by satellite officially started in 1972 with the first Landsat that had a resolution of 60mx80m. In 2002, the first SPOT (Satellite Pour l’Observation de la Terre) was launched, with a 10m resolution. Since the early 2000’s private observation projects for the market from high to very high resolution (sub-metric resolution), e.g. QuickBird, Ikonos, Worldview, Geoeye, Orbview, have pushed the manufacturers towards an improvement of the resolution. In Europe, the instruments have evolved from SPOT (with a resolution of 10m in 1986) to Pleiades in 2009 with a resolution of 0.7m. On-going programs are even below. The evolution has occurred via an increase of spatial resolution, without any significant reduction of field of view. We are currently at a limit of resolution, and improvements will now occur via larger swaths and/or high revisit rates.

On the Low Earth Orbits, the agility of the latest generation of satellites enables the reduction of the revisit time and the acquisition of 3D images, while using state of the art telescopes with “classical” resolution performance.

But in that case the design driver becomes cost and therefore mass and compactness has to be further optimized (the lighter and smaller the spacecraft, the lower the launch cost). This is where a compact and efficient solution benefiting from freeform mirrors and disruptive additive manufacturing structure can be a valuable asset (higher swath with the same instrument volume for example). Secondly, as the Ground Sampling Distance (the projection of the pixel pitch on ground) is going smaller and smaller due to pixel reduction on new detection matrix solution and due to the willingness to increase the imaging resolution on the instruments; the data volume to download becomes a real problem to handle. In order to anticipate and to solve this new challenge, innovative intelligent algorithm for compression and autonomous system for detection and ranging becomes more and more valuable. VIDEO project will assess to test in end to end condition the capacity to detect, recognize and track various element in a wide video scene, as well as compress the useful data in order to minimize the downloaded data to the ground.