RESEARCH ON REAL-TIME DEFECT DETECTION METHODS

摘要

With the rapid development of the printing and dyeing industry,defect detection has become a popular subject among researchers.This project focuses on the research of an online defect detection method which is applied to simple and inexpensive devices,so as to reduce the cost of detection.A hybrid method is proposed,which consists of a feature point based template matching technique where the SURF algorithm is used,k-means algorithm,contour extraction technique and some other approaches,which are used for judging defects.Each of these methods above has not so efficient.In this paper,we propose a new method of combining the known methods and our experimental results show that the proposed methods have good performance and meets industrial requirements of real-time operations.In addition,our results can be used to adjust or fix the printing and dyeing equipment,which make this study more valuable.