DEVICE FOR RECOGNIZING SHAPE OF CAR PARTS USING PATTERN RECOGNITION ALGORITHM

摘要

PURPOSE: A device for recognizing a shape of car parts using the neural network is provided to automatically classify the shape of a pattern classification object by using a signal detected from optical sensors and a pattern recognition algorithm. CONSTITUTION: Pattern classification object transporters(01-06) transport a pattern recognition object such as wheels(71-73). A light source(31) irradiates light to the pattern classification object. The optical sensors(51-70) detect the light interfered by the pattern classification object by being placed to an opposite side of the light source. The learned neural network(81) is learned by basic pattern data formed based on an optical detection signal pattern of each pattern classification object, and automatically classifies the pattern of the object by comparing the signal detected from the optical sensor with the basic pattern data, and performing deduction.