農産物の画像識別へのファジィ理論の応用

概要

論文の詳細を見る
This paper describes a new technology on image recognition of objects using fuzzy theory. The standard method of pattern recognition is geometrical pattern matching. This is accomplished by comparing an object with a standard pattern, which is effective in cases having little ambiguity, such as industrial products. However, it is generally difficult to recognize natural products, because these objects have essential fuzziness caused by the differences is shapes, sizes and colors. In this paper, pictorial pattern recognition including these kinds of fuzziness is investigated, and a new image recognition system is proposed with the application of fuzzy theory. Through discrimination experiments made for ten categories such as apples, bannas, lemonns and others, satisfactory results are obtained. At result, it is certified that the pattern recognition proposed here is more effective to recognition of resemble objects categories.
一般社団法人日本機械学会の論文
1993-09-25