Application of Computer Microscopic Image Processing in Offset Quality Inspection (Part One)

With the development of society, people have higher and higher requirements for the quality of printed products. How to control the complex image copying process and make the printing quality control data and standardization is a common concern for the majority of producers, operators and consumers. Judging from the international situation, this is also an inevitable trend in the development of printing. Looking at the current status of China's printing industry, the digitization and standardization of print quality control is still far from the international developed countries. Therefore, how to improve and strengthen the digitalization and standardization of printing quality control from the technical and management perspectives is the responsibility that the majority of printing workers cannot shirk; it is also the key to whether or not the company survives in the fierce competition in the future.

The so-called data quality of the printing quality control is through the necessary testing methods, and the necessary control factors for the printing quality are measured and recorded in the form of data and used as the basis for quality control. The so-called standardization is through a large number of data, comprehensively summed up a series of laws that can guide production, such as formulas, curves, charts, data, etc., done in accordance with regulations, there is a standard. At the same time, there are three standards for measuring the degree of data quality and standardization of printing quality control: first, good stability, second, good repeatability, and third, high efficiency.

The purpose of printing is to faithfully reproduce the original. In order to achieve and maintain the high quality of multi-color printing, it is important to measure and control the amount of ink transferred from the yellow, magenta, cyan, and black ink rollers to the paper. The appearance of a halftone dot print depends on two factors: the size of the dot and the density of ink inside the dot. Traditional detection methods based on densitometers or colorimeters can only detect areas around 10 mm2 and cannot obtain the information we require directly from the print screen. Therefore, special patches (signal strips) must be printed on the sheets so that the printing worker can control the printing quality. This method cannot know the ink supply amount and the dot size variation of each monochrome dot. In the use of test strips and signal strips, human eyes are often used for visual inspection or simple test methods are used to obtain the judgment results and serve as a basis for quality control in production. Because the quality of offset printing products is closely related to the complex factors such as equipment, paper, ink, and environment, traditional control methods are bound to be influenced by many subjective factors. Because human subjective observations are not measurable, they cannot achieve good stability, good repeatability, and high efficiency. At the same time, modern printing frequently has a printing speed of nearly 10,000 sheets per hour. When printers find problems and make adjustments based on signal bars, bad printing products have been formed in large numbers. Therefore, finding a correct, reliable and rapid quality control method and developing an instant automatic control system are of very important and far-reaching significance for ensuring stable quality during offset printing, reducing printing costs, and improving printing quality and production efficiency.

Printing quality control is a very delicate job. Although the current test strips and signal strip test methods have proved to be an effective quality control method, this kind of quality control method mainly relies on the operator's long-term accumulation of experience and subjective judgment, so the test results are repeatable and stable. , efficiency and reliability cannot be guaranteed, let alone dataization and standardization of quality control. In recent years, microscopic image technology has made rapid progress in the field of control. The microscopic image technology is used to collect, quantify and process the test strips and signal strips on the proofs and products to obtain accurate test data, thereby laying the foundation for realizing the automatic control of test strips and signal strips in the offset printing production process.

In order to solve this problem, this paper proposes a method to measure and process the signal strips of offset printing products by taking advantage of the advantages of digital technology, such as rapidity, high repeatability, good stability, and free from any subjective factors. Using the microscopic image processing method to obtain a series of indicators consistent with the signal bar control content, so as to achieve real-time digital, standardized control of offset printing product quality, thereby increasing production efficiency and reducing waste. At present, the application of microscopic image processing technology in the printing field is becoming increasingly widespread. For example, the practical PrintVision 9000 and PrintVision 9000NT Automatic Defect Detection System developed by AVT in the United States use microscopic vision technology to solve problems in the printing process, including color change, overprinting, and lapping. , blemishes, etc. are automatically detected and alarmed. The US 3M company RSFISCH used the fast Fourier transform method to identify and classify the dot images generated on the electronic screen, studied the density distribution inside the dots, and developed a micro density meter-based network. Expanding and dot shape testing systems; T.YUASA and H.MISHINA used the two-dimensional Fourier transform to identify several patterns of moiré generated in color printing, and found ways to control the generation of moiré to facilitate better The quality of color reproduction was controlled; Zhao Wei used the edge recognition method to calculate the area of ​​the dot, and used this as a basis to correct the equation of Newcastle. Other studies have used digital image processing techniques to measure ink coverage, average grayscale, dot area, form factor, and overprint accuracy. CHRISTINE BARRATTE and others of SWIDISH NEWSPRINT RESEARCH CENTER have developed a reading system that automatically detects the dot size, density distribution, and other parameters in newspaper printed monochrome prints and has been put into use. With the entry of new technologies, the means and techniques for measuring and controlling printing quality are constantly being updated. The following is a detailed discussion of this technology.

First, the basic principles and methods of microscopic image processing

The so-called microscopic image processing is to use microscope equipment to enlarge the details of images that are difficult for the human eye to distinguish clearly when acquiring image data, so as to obtain a magnified image. Then, a series of image processing methods and methods are used to process the microscopic images to obtain the expected index values, and compared with the traditional test results to find out the correlation between the two, thereby replacing the traditional man-made with digital technology. Unpredictable and controllable means of control.

1. Basic flow of microscopic image processing

2. Basic principles and basic algorithms of image processing

It is not possible to process a general image directly with a computer. Therefore, various types of image signals (photos, graphics, X-ray photos, etc.) should be first converted into digital images, that is, the digitization of images.
A monochromatic image can be viewed as a two-dimensional continuous function whose brightness appears as a continuous function of position, and the digital image in the computer is represented as a matrix. The different processing is to use different algorithms to operate on this image matrix. From the two-dimensional continuous function to the matrix, the function values ​​(brightness) are taken as samples in different spatial positions, and the quantification process of these samples is represented by a set of integer values. This is the sampling and quantification process of the image, that is, the digitization process. The digitized image is divided into many small areas, which are called pixels. The value of each pixel is represented by a luminance value or a gray value. For black and white images, there are only binary information, namely "black" and "white". In a computer, image information is represented by "1", and "0" indicates non-image information.

The original image is input to the computer through the image input device and becomes a discrete digital image, which can be expressed as a matrix:

For a dot microscopic image, it can be explicitly shown in FIG. 3 .

In FIG. 3, the "black" area represents a halftone image, and the "white" area represents a non-graphical part. As can be seen from the figure, the percentage of this dot can be seen as 13/25=50%, and the ratio of length and width is 1:1.

During image processing, such pixel matrixes are calculated and processed. General image processing methods include image smoothing, sharpening, image segmentation, binarization processing, edge tracking, and the like.

Second, offset quality control system

In an offset quality control system, the microscopic image is a test strip on a printed sheet, a proof sheet, or a printing plate. Using image amplification equipment, such as microscopic cameras or micro-cameras, to amplify the measurement and control bar 60 or 100 times, and then use digital equipment to capture binary information of the image to a computer at a certain frequency; use image processing software to image The information is processed to obtain a predetermined set of index values. Through the calculation and comparison, the control information of printing quality is obtained, and then these information are fed back to the printer water supply device, ink supply device, pressure adjustment device, jacket transfer adjustment device and other control devices through the computer, so that the printing conditions can be adjusted in time. And change.

The development of hardware has provided us with great convenience and practicality. At present, the microscopic magnification and digitization of the image can be completed with one device. The computer has become the most common device in the printing factory. For lithographic offset quality control systems, the image processing methods involved are very simple and common. Therefore, the realization of image processing software under the support of high-speed computers, real-time calculation and control is not a problem. A set of practical offset printing quality control system is shown in Figure 4.