Intelligent image prospecting system for IMAGE/RPI mission

Ivan A. Galkin, Grigori M. Khmyrov, Alexander V. Kozlov, Bodo W. Reinisch, 
and Shing F. Fung, 


The Radio Plasma Imager (RPI) onboard NASAÕs IMAGE spacecraft has acquired over
1.2 million plasmagrams, images of remote sensing of the EarthÕs magnetosphere.
Plasmagram archive is a classic-example dataset posing unrealistic demands of
manual labor in order to analyze each collected image for useful features. We
present an intelligent data prospecting system based on a bio-plausible model of
the pre-attentive vision whose purpose is to draw attention of human analysts to
the most interesting data. Presence of weak signatures in the RPI plasmagrams
makes this dataset an excellent testbed for sensitive image prospecting
techniques. We discuss our progress to date.

________________
World Automation Conference, 2006