| Zipper is an important part of people's daily necessities. Due to the artificial, equipment and other reasons, there are many defacts in zipper teeth and zipper tapes in the production.Those defacts seriously affect the performance and the quality of products. For the lag and instability of results, low efficiency and high cost, the traditional manual inspection has seriously restricted the intelligent and informational development of the enterprise. This article applies machine vision to detection the defacts of metal zipper and develops corresponding methods to inspect the defective zipper teeth and tapes.According to characteristic of high reflection of metal zipper teeth and complex texture of zipper tapes, back optical source and shadowless optical source are selected to build experimental platform.This article analyses the effect of common filter and determines median filter to filter out the noise in the actual working conditions.In the form of the zipper teeth defects, this article discusses the feasibility to detect zipper teeth defects using the distance of zipper teeth root. A method to inspect the teeth defacts based on the distance of zipper teeth root is proposed. After Otsu threshold segmentation, Sobel edge detection, Hough transform for line extraction, the datum line to mesure the distance is determined. This article measures the distance of zipper teeth in the direction of datum line and designs experiment to verify the feasibility of this method.According to the common defects of the zipper tapes, the texture analysis method which is widely used in fabric detection is proposed for the inspection of zipper tapes. This article has an introduction and research on the GLCM for extracting the texture feature, especially on effects of different tectonic factor for the extraction of characteristic quantities. A SVM classifier and an experiment are designed to verify the effectiveness of the method for zipper tapes detection based on GLCM and SVM. |
PDF Full Text Request