The Defects Detecting System Of Glassware Based On Computer Vision

Nowadays, the production of product has made great process in the field of production scale, production speed and product quality, which has brought challenge to traditional artificial detection method. After all, it's not competent for artificial examination method to complete such a great job. On the other hand, with the development of science and technique, especially in computer, the detection technique of computer vision has put forward. The detection system, which designed with this technique, won't be influenced by bad environment and subjective factors. It can detect the product fleetly and accurately that the human can't do. It is an indispensable tool in modernization production. The detection technique of computer vision is a production which combine computer image processing with pattern identification theory, it comprises correlative knowledge of different fields which including computer technique, data structure, image processing, pattern identification, software engineering and so on. As to glass bottles, the author concentrate on developing a detection system with excellent result—lower price and better performance. This paper introduces image processing technique including image filtering and image segmentation detailedly and ameliorates the typical median filtering arithmetic. The paper seeks out the featest image segmentation technique by analyzing and comparing a variety of image segmentation technique, so as to found a foundation for distilling and analyzing the blemish. The paper analyses a variety of pattern identification technique and raises a specific scheme for the detection of the glassware blemish. The scheme uses the clustering arithmetic to distill blemish, and identifies the glassware by analyzing the character of blemish. At the same time, this article brought forward a new cluster algorithm based on binary image, the algorithm have greatly improved processing speed of the detection system. This system's hardware is mainly composed of vidicon, image capture card and computer. The image of glass bottle is collected by vidicon, and turned to digital images that can be identified with computer by image capture card, and identify glass bottles'blemish with the software in the computer. The author has finished the software development job of bottles'blemish detection with Visual C++. The detection system has successfully realized the blemish detection of bottles. It reaches the detection precision need in experimental conditions and can detect the blemish of glass bottles, such as crack, black spots, bleb, etc.