The C-TEP will incorporate research into new ways to extract information from Earth Observation (EO) data archives, in particular using semantic information. The objective is to give practical answers to high-level questions such as “What is the turbidity level of water on estuaries of the Black Sea during winter?” To answer this question, C-TEP shall extract data from turbidity maps obtained by satellite data during the winter period. However, this query would require a list of all the river estuaries of the Black Sea and their geographic location. The Knowledge Discovery in Database (KDD), a software tool developed by DLR (Germany), is expected to help create this type of list. The tool works on a database of radar or optical satellite imagery, and extracts small image features based on visual similarity. The operator provides a semantic annotation of a feature of interest (such as an estuary) and the software learns to recognise similar features. Once the learning is completed using a few examples, the tool automatically extracts semantic features from the complete image database and produces a list which can be further used for automatic processing.