TRACS - TOOLKIT FOR RECOGNITION AND AUTOMATIC CLASSIFICATIONOF SOLAR FEATURES

AMERINEX ARTIFICIAL INTELLIGENCE (AAI) IS PROPOSING A PROJECT CALLED TRACS (TOOL FOR RECOGNITION AND AUTOMATIC CLASSIFICATION OF SOLAR FEATURES) AS A PHASE I SBIR PROJECT IN SOLAR IMAGE PROCESSING. THE TRACS PROJECT IS DESIGNED TO FACILITATE THE AUTOMATIC RECOGNITION, ANALYSIS, AND TRACKING, OF SOLAR FEATURES IN DIGITAL IMAGES. THIS PROJECT WILL PRODUCE A SUITE OF COMPUTER PROGRAMS THAT WILL AUTOMATE A VARIETY OF ANALYSIS PROCESSES THAT ARE CURRENTLY PERFORMED MANUALLY. THE DEVELOPED ALGORITHMS WILL BE ABLE TO EXTRACT SOLAR FEATURES,SUCH AS FILAMENTS OR SUNSPOTS, FROM IMAGES THAT ARE TAKEN WITH A VARIETY OF WAVELENGTHS, AND WHICH ARE TAKEN UNDER VARYING CONDITIONS (DIFFERENT TELESCOPES, VARYING CLOUD COVER). SIX CLASSES OF ALGORITHMS WILL BE DEVELOPED UNDER THE TRACS PROJECT. IMAGE PREPROCESSING ALGORITHMS WILL CORRECT FOR VARIABLE OBSERVATION CONDITIONS. FEATURE EXTRACTION ALGORITHMS WILL EXTRACT THE DESIRED SOLAR FEATURES, AND FEATURE ANALYSIS ALGORITHMS WILL BE USED TO MEASURE AND IDENTIFY SPECIFIC CHARACTERISTICS OF THESE FEATURES. FEATURE GROUPING ALGORITHMS WILL BE USED TO CONSTRUCT HIERARCHICAL DESCRIPTIONS OF IMAGE EVENTS, AND FEATURE TRACKING/PREDICTION ALGORITHMS WILL PERMIT A DESCRIPTION OF THE MOVEMENT OF THE FEATURES OVER TIME. FINALLY, USER INTERFACE/DATA DISPLAY ALGORITHMS WILL MAKE THESE ALGORITHMSUSEFUL TOOLS FOR SOLAR RESEARCHERS AND FORECASTERS.