A Study Of Real-time Detection Technology Based On Machine Version For Printing Product Defect

With the rapid development of computer technology and the improvement of automatic production, people become stricter with product appearance. Due to the side impact of printing equipment, printing technology and printing environment, however, it is easy to produce printing products with defects, which will not only have negative impact on the manufacturer's reputation, but also cause resource waste. The traditional printing quality detection method is manual detection with high labor intensity and subjectivity. The inspector, with his own eyes, should examine if there is printing detect on the surface of the products. In mass production, it is obvious that this method cannot meet the demands of modern life.The detection method based on machine vision is a new one which simulates human eye detection mechanism, and can replace manual detection. In machine vision detection,the camera and other hardware, such as optics lens, will catch the surface characteristics of the measured object which will be detected later. At a higher speed,this detection method,with its high detection efficiency and strong objectivity,can continuously work for mass production and realize the automatic control of the printing equipment. Therefore, the use of machine vision detection technology is an inevitable trend in printing detection industry,medicine detection industry and steel detection industry.In this study, the writer simulated the printing detects which often appear in actual production process and tested different detection methods based on machine vision. The writer found that the template matching method based on deviation model can detect at a higher speed and with higher accuracy. In addition, this method can detect the defects in both the printed texts and the printed images. This method is based on the image acquisition theory that line-scan camera is used to catch the printed image in real time. At the same time, the writer also developed a test system to stimulate the actual production process.This system adopts multithread programming technology which take image collection,image process and defect inspection as three individual sub-threads, which operate at the same time to improve the operation efficiency of the system.