EVEN ORGANISMS OF VERY MODEST INTELLIGENCE ROUTINELY CLASSIFY OBJECTS ON THE BASIS OF COLOR. UNFORTUNATELY, INCORPORATING THE SAME CAPABILITY INTO PRACTICAL MACHINE VISION SYSTEMS FOR IMAGE ANALYSIS OR PROCESS CONTROL HAS OFTEN PROVEN BOTH DIFFICULT AND EXPENSIVE. REVIEW OF THE REQUIREMENTS FOR NATURAL COLOR RECOGNITION LEADS DIRECTLY TO A POWERFUL COLOR DISTRIBUTION MATCHING METHOD WHICH OPERATES WELL ON PERSONAL COMPUTERS USING INEXPENSIVE VIDEO EQUIPMENT. THE METHOD IS NOW IN USE FOR SATELLITE IMAGE ANALYSIS AND VARIOUS INDUSTRIAL AND PROCESS CONTROL APPLICATIONS. THE CLASSIFICATIONS GENERALLY COINCIDE WELL WITH THOSE MADE BY A HUMAN INSPECTOR.
