On - Line Inspection And Intelligent Control Of High - Speed Printing Machine

With the development of mechanical manufacturing technology, automatic control technology and electronic technology, printing technology has been developing continuously. The printing industry is moving from the traditional printing to the standard operation process such as PDF/JDF, the whole process is more digital, automated and intelligent. Sheet metal printing is a branch of the printing industry, the amount of ink in real time affects the quality of printing. In high speed printing environment, if we do not monitor on the printing quality timely, when color defect is found in printing, the printing machine has produced a large number of defective products. It not only wasted material, but also reduced production efficiency. So ink online inspection is an indispensable part in the control of printing quality, and in order to cater to the general trend of digital printing, the method of machine vision was used to measure control strip in this paper, in order to detect the amount of ink and make the close loop for metal sheet printing machine. The main contents of this paper are as follows:First of all, the method of machine vision was designed to detect the printing quality, and the overall scheme of ink volume control was introduced, each module was described in detail.Secondly, in view of the special printing situation and the printing request, each kind of hardware for the main machine vision system was analyzed and the best model was selected, whitch was used to build the printing test bed in the production line. In the process of printing, it is inevitable to produce vibration, and the position of the printed matter is not always fixed, which makes the position of the color in the image to be changed. In this paper, some image processing algorithms were illustrated to locate the printed materials, and the advantages and disadvantages of the algorithms were analyzed. Through these algorithms, we can capture the position of control strip and extract the correct color information.Then, according to the characteristics of the CCD camera and the color principle, in this paper, three common algorithms were presented to transform RGB color space to CIE-LAB space for the CCD camera. Its essence is the work of color management. And the advantages and disadvantages of these algorithms were analyzed, laying the foundation for the later model.Finally, two kinds of ink quantity control models are established respectively. The actual data were obtained through the metal sheet printing experiment, and the data were analyzed in detail. The law of color for the different amount of ink was also analyzed. We established the model of the color-ink transformation with the factor of ink vibration. And the inverse operation of the model can realize the quantitative prediction and control of the opening degree of the ink key. In order to take into account the impact of more factors, subtractive clustering-ANFIS algorithm was also used to establish the ink control model, so as to realize the real-time control of ink. In order to solve the problem of less practical training data, the ink transfer model was referred to be as the printing system model, which was used to generate training data and model validation. After verification by the computer, the subtraction clustering ANFIS model was proved to have high precision and good expansibility, and can complete the intelligent adjustment of ink quantity, so the agorithm can be used as reference in practical production.