GEOINFORMATION VIA PARALLEL IMAGE ENGINEERINGAND IMAGE INFORMATION MINING

摘要

Earth observation is a fundamental source of informationrnin today's society. A multitude of spacebornernand airborne sensors collect daily an enormous amountrnof information in the form of imagery of the earth. It isrnevident that systematic approaches for informationrnmining in such imagery are required. New sensors andrnimage processing techniques extend the imagerninformation base as well as improve quality and speedrnof availability. In this paper we give information on arnnew, European, high-resolution optical sensor, which isrncurrently implemented for airborne use and willrncontribute to geo-referenced earth observation. Thernproject includes also a new approach for imagernprocessing. Via parallel computing for the purpose ofrnfully automated generation of digital surface models andrnorthomosaics, fundamental geometric and visual geoinformationrnis built up. All of this would be cumbersomernwithout a metadata information system, whichrnallows for archiving and retrieval, dissemination andrnworkflow management. Thus, the paper presents aspectsrnof image generation as well as aspects of imagerninformation mining.