Image information mining for earth observation at ESA

Earth Observation (EO) images are retrieved from archives using attributes like geographical location, time of acquisition, sensor, etc., which do not provide any insight in their information content. Experts apply their knowledge, interpreting the images to extract information, through a human intensive and therefore expensive process. This creates a gap between offer (data) and need (information), which is continuously increasing since more missions and even constellations are being deployed, with broader sensor variety and increasing data rates. This contribution describes the European Space Agency (ESA) initiatives in Image Information Mining (IIM), which seems to be a technology capable to provide a breakthrough to possibly reduce this gap. ESA started since 2000 specific projects in the field and chairs the European Image Information Mining, Coordination Group (IIMCG).