抄録
Today, there is an increased need for quality control in the manufacturing sectors. Particularly, the automated detection of the product surface defect is essential to fulfill the requirement of quality assurance. Generally, use of image processing to detect defects on the product surfaces is based on the pattern matching technique which necessitates preparation of possible geometric patterns of defects prior to such trials. Therefore, this method is unable to deal with exceptional defects of prepared patterns. In this study, we have proposed an algorithm of surface defect detection by the application of wavelet transformation to the image processing. This paper provides the algorithm and results of experiment that show the effects of process parameters on the detection accuracy.
