Image Information Mining and Understanding
Earth observation data volumes and resolution details have increased significantly over the last decades with Synthetic Aperture Radar (SAR) and optical sensors now acquiring and transmitting to Earth receiving stations several terabytes of data daily. This data acquisition rate is a major challenge for existing data exploitation and dissemination approaches. The challenge is increasingly going to be how to increase the usability of the millions of images being stored in archives for a broader range of end user applications.
Image Information Mining (IIM) is a new field of study that has arisen to seek solutions to automating the mining (extraction) of information from earth observation archives that can lead to knowledge discovery and the creation of actionable intelligence (exploitation). Image information mining is more than just an extension of data mining principles to images: it aims at supporting user understanding. IIM comprises novel concepts and methods to help earth observation data users to access and discover information in large images or image archives, and to rapidly gather information useful for determining courses of action. Interesting applications involve dealing with complicated spatial, structural and temporal relationships among image objects. Thus, new concepts have been introduced on the basis of intensive preprocessing of images to extract relevant features, structures and objects, and to automatically record and analyze their interrelationships, preparatory to learning their behavior and detecting relevant information. The methods are integrated in systems which can be operated using intelligent interfaces able to correlate the information content of the images with the relevant goals of the application. The users have at their disposition tools using semantics for the definition of specific goals.
To reach these goals the following main research activities are carried out:
The developed algorithms are integrated in systems and tools: