抄録
This paper presents development results of grading machine for shoe soles using image processing method. In the traditional inspection method, a template made by acryl or metal is used and inspectors measure it's size by own's eye. So there are measurement errors according to their body condition, working fatigue by iterative inspection and finger's pressure power difference. Also there is individual difference according to skillfullness of inspectors. In order to improve those problems, using image processing method we develop a grading machine for shoe soles. Since the types of shoe soles are several, the shape is diffenent to each type. Therefore we must consider the grading algorithm related with the shape or type. In this paper, we adopt a bar code system in order to recognize the types of shoe soles and a correction parameter corresponding to each type. It's effectiveness is proven through experimental results and the measurement range is about 1mm in real time of 3.5sec for a pair of shoes.
