Research And Integration On Web Inspection System Based On Machine Vision

It is more and more impossible to inspect the paper defects by manual inspection as the increasing web width and machine speed. The coming of web inspection system, which based on machine vision, has carried out the online analysis and control as well as improved the inspection efficiency. This paper does the work as collecting image by intelligent image sensor, simulating online web inspection when producing, developing the experiment system.This paper selects hole, spot and wrinkle, which are representational defects, as the research object, and studies the image processing algorithms adapted to the web inspection system, including image smoothing, threshold division, edge detection, and image mathematical morphology processing. Besides, a new algorithm named dynamic double thresholds division based on image brightness has been brought out according to the feature of paper defects, which can be used in high contrast paper defects inspection. Furthermore, this paper puts forward an image feature extraction method based on Hough transformation which is used to solve the problems coming out from the using of routine image feature extraction methods in web inspection, by analyzing the final Hough transformation space, from which can pick up the wrinkle feature and work out the identification.Based on the algorithms study and according to the practical application desire, this paper designs an experiment web inspection system, including the function of image collecting, image pretreatment, image feature extraction and identification, paper defects online inspection, history data query, by software design platform Delphi, and get the system tested and analyzed. Finally, this paper makes some suggestions for the next work due to the deficiencies existed.