摘要 |
PURPOSE: An apparatus and a method for classifying billet surface defects using an intelligent classifier with an optimized learning function are provided to automatically classify the type of billet surface defects by performing an intelligent inference process using feature vectors of the defects. CONSTITUTION: An apparatus for classifying billet surface defects comprises a defect detector(10), a defect feature vector input device(20), a fuzzifier(30), a fuzzy inference device(40), a defuzzifier(50), a defect classifier(70), and an optimized learning device(60). The fuzzifier converts feature vector information inputted by the defect feature vector input device into a fuzzy set. The fuzzy inference device reasons fuzzy output using the fuzzy rule. The defuzzifier converts the output into numerical information. The defect fuzzifier classifies billet surface defects according to the numerical information. The optimized learning device optimizes the center value required for the fuzzy rule.
|