遺伝的アルゴリズムを用いたカラ-画像による物体位置・姿勢情報の獲得

概要

論文の詳細を見る
This paper is concerned with an effective technology on image processing of object using genetic algorithms. Recently，Industrial robots in factory and welfare robots for human support are widely studied. For these robots works autonomously， it needs to acquire the information on a position and a posture of a object form the image that taken in from visual sensor. However，if it does not know where and how an object exists，it is difficult for robots to work. ln this paper, pattern processing on image in which plurality objects exist is investigated and a new image processing system is proposed by the application of genetic algorithms which is effective in an object search. First，specific color domain and edge domain is extracted from color image. Second，search image is made by laying specific color domain image and edge image. Third，search models corresponding to each fruits used by genetic algorithms is created. Finally，some acquisition experiments are performed for fruits by using genetic algorithms，and interesting results are obtained. Through experimental results，it is shown that new technology proposed here is sufficiently valid.
福井大学工学部の論文