An ontology-based modeling of an ocean satellite image retrieval system

The need to access information in large volumes of image data, e.g. big images, large images archives, distributed image repositories, etc, has stimulated the research in the field of Content-based Image Retrieval during last decade [DJLW08, LZLM07, Han08]. Many new concepts have been developed and prototyped. However, the dramatic increase in volume, details, diversity and complexity and the user demand for simultaneous access to multi-domain data urgently requires new approaches for image information mining, multi-domain information management, knowledge management and sharing. In this work we will present the elements of the architecture of a satellite image retrieval system. In order to have a flexible and dynamic image retrieval system, we have designed an architecture which integrates Web technologies with modern pattern recognition systems applied to satellite images [PGC05]. Our proposal aims to develop a satellite image retrieval system in which the kind of image, sensor, satellite, together with the kind of classifiers can be customized by the system user. More concretely, image, sensor and satellite features can be imported/exported and updated, and the algorithms can be imported/exported and refined/improved from the system. In particular, the image retrieval system allows the data and knowledge exchange. In order to equip the system with importing/exporting facilities for data and knowledge exchange, wellknown Web technologies are used: the XML (Extensible Markup Language) [W3C07] for data exchange, and the Web Ontology Language (OWL) for knowledge exchange.