An intelligent object recognizer and classification system for astronomical useAn account is given of an image-processing system based on AI concepts, which allows input images produced by the CCT/Transit Instrument to be compared with a standard-object hierarchylike network of prototypes presented within the computer as 'frames'. Each frame contains information concerning either a standard object or the links among such objects. This method, by comparison to conventional, statistically-based pattern recognition systems, classifies data as an astronomer would and thereby lends credibility to its conclusions; it also furnishes a natural avenue for the machine's serendipitous discovery of new classes of objects.
Document ID
19870049111
Acquisition Source
Legacy CDMS
Document Type
Conference Paper
Authors
Bernat, Andrew P. (Texas, University El Paso, United States)
Mcgraw, John T. (Steward Observatory Tucson, AZ, United States)