JPH042104B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention relates to an apparatus for quantitatively evaluating the quality of printing in halftone dot printing. In gravure printing, the density contrast is expressed by the size of the halftone dots, but when you enlarge the printed halftone dots, you will notice that some of the circular halftone dots are missing, or that one halftone dot is completely missing. You can find things like that. This is a problem with the printing itself.
This is a phenomenon that involves the problems of compatibility between printing plates and ink, and the problem of compatibility between paper and ink, etc., and is a problem that requires efforts to improve in each field of plate production, ink production, and paper manufacturing. requires appropriate evaluation means for printed halftone dots. The present invention meets these needs.

B. Conventional technology Halftone dots are printed at a fixed density on printing paper, the halftone dots are enlarged using a magnifying projector, an optical microscope, etc., and defective halftone dots and missing halftone dots are visually observed during printing of a fixed area. A method was used in which humans counted the number of defective halftone dots and missing halftone dots.

C. Problems to be solved by the invention The above-mentioned visual inspection method is time-consuming because it is carried out by humans, and requires constant monitoring during the manufacturing process of coated paper for printing and feeding back the results to the process. However, in order to improve the reliability of the inspection results, it is necessary to set as many inspection points as possible from one printing surface and check for defective halftone dots and missing halftone dots in each fixed area including those points. It is necessary to detect and count defects, the observation area becomes larger, and the total number of halftone dots becomes enormous, requiring great mental concentration to pick up defects without overlooking them. Moreover, it is easy for subjectivity to enter into the judgment of whether a single halftone dot is normal or defective, and this subjectivity is shaken by the fatigue caused by mental concentration as mentioned above, and when trying to improve the reliability of the inspection, it is rather difficult to trust from a human perspective. There is a problem that sexual performance is decreasing. The present invention seeks to solve the above-mentioned problems of conventional methods.

D. Means for solving the problem An enlarged image of a certain area of the test surface printed at a certain density is stored in the image memory, and each halftone dot exists on the image memory by data processing on the data in this image memory. Determine the correct coordinate range, evaluate the printed halftone dots using a predetermined halftone evaluation algorithm for each coordinate range, and count the total number of halftone dots determined to be defective (including missing halftone dots). I did it like that.

E. Effect According to the method described above, the operations performed by humans are to place the printed test piece under a microscope, focus it, and adjust the direction of the rows and columns of halftone dots in the X and Y directions of the image memory.
In particular, the latter operation can be automated, and the operations from storing the image of the printing surface in the image memory to producing the result can be completed almost instantly by a computer. What makes this kind of data processing possible is the rows of halftone dots.
The features of the present invention lie in the automation of determining the regular number of halftone dots by automating the detection and counting of rows, and the establishment of a method for determining the coordinate range in which each halftone dot should exist. As long as the coordinate range that should exist can be determined, the halftone dot evaluation algorithm is arbitrary.

F. Embodiment (Summary) FIG. 1 shows an apparatus according to an embodiment of the present invention. 1 is a main body of a microscope equipped with a television camera or an imaging device, 2 is an imaging device, and 3 is a monitor CRT for displaying a microscope image or the contents of an image memory. The sample stage 4 of the microscope is equipped with fine movement devices in the X and y directions and a rotation device around the z axis. 5 is a computer (CPU) that processes data, 6 is an image memory,
7 is a keyboard for setting various parameters and inputting operation commands, and 8 is a CRT for display.

If you look at paper printed with halftone dots at a constant density under a microscope, you will see an array of halftone dots as shown in Figure 2. In the following description, the horizontal arrangement of halftone dots is referred to as a row, and the vertical arrangement is referred to as a column.
On the screen of the monitor CRT 3, bright and dark signals in the AXB range shown in FIG. 2 are stored in the image memory 6. The operator sets the test paper strip printed at a constant density on the sample stage 4, and focuses the microscope while looking at the monitor CRT 3, so that the rows of halftone dots are aligned with the CRT.
Rotate the sample stage 4 so that it is parallel to the horizontal scanning line 3, and rotate the sample stage 4 so that the halftone dot image does not overlap the boundary line between A and B, that is, as shown in Figure 2.
After making fine adjustments in the direction and y direction, when a start command is given from the keyboard after completing this adjustment operation, the CPU 5 stores the image data in the image memory 6, performs the data processing described later, and displays the results. Output to CRT 8 and printer 9. The image memory 6 stores the brightness of the enlarged image of the test paper as 8-bit data, for example.

First, the CPU 5 uses the data in the image memory 6 to binarize the brightness and darkness of the image. This binarization is an operation that discriminates one pixel into two values, ie, whether it is a white part of the test paper or a printed halftone part. That is, a black and white discrimination level is set, and if it is blacker than that level, it is black, and if it is less than that level, it is white. A method for determining this identification level will be described later.

Next, the CPU 5 counts the number of rows and columns of halftone dots in the image stored in the image memory 6, that is, the number of rows and columns of halftone dots included in the AXB range of the image shown in FIG. Determine the total number of halftone dots that should be within the range of xB. The number of halftone dots that should be this is because on the actual printed surface there are places where halftone dots are missing or there are defective halftone dots that cannot be considered as halftone dots, and the actual number of halftone dots is the number of rows and columns. Since it is smaller than the number multiplied by , it is said to be the regular number of halftone dots. The method of detecting and counting rows and columns will be described later.

Thereafter, the CPU 5 sets the coordinate range of the inspection area where each halftone dot should exist. As shown in Figure 3, by setting a square surrounding each halftone dot on the image memory, this can be done by setting the squares surrounding each halftone dot on the image memory, and the x-direction addresses x1, x2, x3... and the y-direction addresses y1, y2 on the two-dimensional memory. ,y3
By determining . . . , the range in which one halftone dot should exist is expressed as (x1 to x2) and (y1 to y2), for example. This range may be set wide so that the halftone dots fit in gently, but it is easier to execute the halftone evaluation algorithm if the range is set as narrowly as possible, and it is possible to set this range to be as narrow as possible so that the halftone dot evaluation algorithm can be executed more easily. It may be set to This setting is performed in conjunction with the row and column detection operations described above.

Finally, the CPU 5 evaluates the printed halftone dots for each halftone dot existing range set as described above, counts the number of defects and missing halftone dots, and calculates the percentage of the total number of normal halftone dots. Calculate and display. The halftone evaluation method is arbitrary. The simplest way of thinking is to count the total number of black pixels from the binarized image data within the halftone dot existing range. This number is maximum for printed correct halftone dots, and decreases if there are missing parts of the halftone dots. Therefore, if the number of black pixels is less than or equal to a preset number, it is determined to be a defective halftone dot.

(Binarization of Image Data) The operation of binarization of the image data of the enlarged image of the print surface is the basis for subsequent data processing. The data stored in the image memory 6 is 8-bit data representing the brightness and darkness of pixels constituting an enlarged image of the print surface. If a brightness histogram is drawn for the entire area of A×B shown in FIG. 2 using this 8-bit data, a shape as shown in FIG. 4 will be obtained. In this histogram, the horizontal axis is the brightness level, which is standardized with the darkest data in the image memory 6 as 0 and the brightest as 100, and the vertical axis corresponds to each standardized brightness level. Total number of pixels, this histogram is CPU
You can display it on RT8 by commanding 5. Two peaks can be seen in this histogram, the peak on the right is the brightness distribution of the white background part of the printing paper, the peak on the left is the brightness distribution of the halftone area, and the valley between both peaks is The corresponding brightness C is taken as a determination level for determining whether a single pixel belongs to a halftone area or a background area of paper. If the judgment level is moved further to the left (toward the black side), the judgment of halftone dot quality becomes more difficult. This determination level may be determined visually by a human by displaying a histogram on the CRT 8, but it may also be automated. To do this, it is sufficient to program an algorithm to find the minimum turning point on the density distribution excluding the whitest and blackest points.

By the way, the histogram shown in Figure 4 is a standard shape, and such a histogram can be obtained within a certain printing density range, and if it is darker than that, the histogram will become a mountain on the left as shown in Figure 5b. becomes larger, and the mountains on the white background become unclear. Also, when the density is low, the peak on the right side becomes large as shown in Figure 5 w, and the peak corresponding to the halftone dot becomes small and low, creating a slight bulge that rests on the foot of the peak on the right side. In either case, the valley appearing in FIG. 4 becomes unclear, making it difficult to set the determination level.

(Detection count of rows and columns of halftone dots) Detection of rows and columns of halftone dots is performed using the above-mentioned binary data. If all addresses in the image memory 6 are scanned in the x direction and the binary data of each address is summed for each scanning line in the x direction, with black as 1 and white as 0, then
A periodic function-type histogram with the y-direction address as an independent variable, as shown by F(y) in FIG. 6, is obtained. Each mountain in this histogram indicates a row of halftone dots. In the same way, all addresses of image memory 6 are y
A histogram F(x) is obtained by scanning in the direction. This histogram can also be displayed on the CRT8. Using a program that detects peaks from these histograms F(y) and F(x), counts the numbers, and multiplies them, calculate the total number of halftone dots that should be in the A×B area shown in Figure 2. be able to.

An example of a program that detects and counts peaks from a histogram F(x), etc. stores data such as a histogram F(x) in memory, searches for the highest and lowest values of the histogram, and calculates the value f, which is 1/2 of that value. and set
The data in the memory is read from the end, and when a value exceeding f is detected for the first time, 1 is added to the count, and then the number of addresses becomes a number equivalent to half the preset row or column pitch. Up to this point, even if the read data value exceeds f, it is not counted, and after the pitch exceeds half the pitch, when a value exceeding f is read for the first time, 1 is added to the count, and the operation is repeated.

(Setting the coordinate range in which halftone dots should exist) The data of the histograms F(x) and F(y) described above are used. F(x) will be explained. First, average the histogram. To do this, for example, data at two addresses on the left and right are added to the data at each address in the memory storing the histogram, and the result is divided by five. In a histogram in which fine irregularities are removed in this way, the shapes of the peaks are almost completely the same.
Histogram Fm(x) averaged in this way
Detect the maximum value and minimum value at
Fm(x) searches for addresses that hold data with a value greater than this reference value. What we are doing here, to explain a part of the histogram Fm(x) shown in Figure 7, is to find the range X that indicates values above level L for each peak of the histogram, and the range that is included in the range of X. Since each address has a group of consecutive address numbers, the address numbers at both ends of each of these address groups are taken as x1-x2, x3-x4, etc. in FIG. 3.
The same algorithm is applied to y1-y2, etc. The coordinate range determined in this manner in which the halftone dot should exist constitutes a square that somewhat cuts into the circular edge of the halftone dot. This method works well if the test paper is set so that the rows and columns of the halftone dots are exactly parallel to the x and y axes of the image memory, but if the parallelism between the two is not very good, the rows and columns of the halftone dots , the amount of bite in the rectangular range varies depending on the position on the row. In such a case, the addresses obtained by adding and subtracting a certain number to the address numbers at both ends of each range obtained as above are x1
Expand the range by setting ~x2, etc. However, if the range is expanded, the area required to evaluate each halftone dot becomes larger, and the time required for evaluation becomes longer. Therefore, the choice is made depending on the purpose: whether to shorten the evaluation time by spending time and carefully setting up the test strips, or simply setting up the test strips and spending more time on the evaluation.

(Evaluation of halftone dots) The halftone dot evaluation method described in the summary section of the example is as follows:
The binarized data of the image is summed up using all pixels within the range where the set halftone dot should exist as specified points, and whether the halftone dot is normal or defective is determined based on whether this is greater than or equal to a predetermined value. . However, evaluation methods are not limited to this. If it is within the area, it can be determined even by specifying a small number of points. For example, when halftone dots are represented by binarized image data, normal halftone dots are circular, and defective halftone dots are circular, as shown in FIG. Therefore, consider a base grid-like line that passes through the center of each halftone dot on the image memory, and draw 9 and 25 points (or even 5 points) included in the normal halftone dots around each intersection of this base grid. It is also possible to designate a point, calculate the sum of the binarized data of that point, and if it is not 9 or 25 (or 5), it is determined to be a defective halftone dot. Alternatively, instead of checking all specified points such as 9 points, 25 points, etc., you can check them one by one, and if a white point (binarized 0) is found, it is determined to be a defective halftone point, and the remaining points can be checked. You can stop. The operation of detecting the center lines of the rows and columns of halftone dots can be performed by taking the midpoint between x1 and x2 in the above-described operation of setting the coordinate range in which halftone dots should exist.

(Automation of test strip setting) In the above-described embodiment, test strip setting is performed manually by visual inspection. To automate this, the operation of obtaining the histograms F(x), F(y), etc. described above can be used. Periodicity in the histogram F(x) etc. appears when the rows and columns of halftone dots become parallel to the x and y directions of the image memory to a certain extent, and when the two become completely parallel, the valleys of the histogram appear. Maximum width. Therefore, a stepping motor is used to slightly move the sample stage of the microscope (Fig. 1, 4) in the x and y directions, and rotate it around the z-axis, and the histogram F(x) is calculated while rotating the sample little by little around the z-axis. seek,
All you have to do is create a program that determines the position where the valley is the widest. In this case, while looking at the monitor CRT (Fig. 1, 3), manually adjust the dot rows so that they are approximately parallel to the x direction on the monitor CRT, and
If you adjust the axis, see which direction to turn around the Z axis to make the rows of halftone dots parallel to the x direction, and instruct the CPU 5 in that direction, the CPU will rotate the Z axis of the sample stage 4. Since there is no need to determine in which direction the rotation around the axis should be performed, the program becomes simpler and the time required until the halftone dot rows become parallel to the x direction is shortened. Once the halftone dot rows have been adjusted parallel to the x direction, find the sum of the image data (binarized data may also be used) at each address in the x axis direction corresponding to y axis address 0 in the image memory 6, and find the minimum value. Move the sample stage slightly in the y-axis direction so that Check and the check result is NO
In this case, check by further finely adjusting the sample stage in the y-axis direction.
Make it YES. Fine adjustment in the x-axis direction is performed in exactly the same way.

G. Effect: According to the present invention, things that humans have traditionally done can be omitted.
It is 100% automated and allows for quick quantitative evaluation of print quality, and highly reliable results are obtained without human subjectivity being involved in the judgment. Furthermore, since the device according to the present invention accumulates and holds image data of enlarged images of printing halftone dots, it has the function of not only determining the quality of printing but also providing basic data for various analyzes such as the suitability of paper for printing. It is something that you have. For example, the following analysis can be performed. In the histogram shown in FIG. 4, the mountain on the stone side represents the distribution of uneven whiteness on the background of the printing paper, and the mountain on the left represents the density distribution of the halftone dots. In an ideal case, the whiteness of the paper is the same across the entire surface, so the mountain on the right becomes a single bright line spectral line,
Since the halftone dots are circles of uniform blackness, the mountain on the left also becomes a single line. The spread of the blackness histogram of halftone dots indicates that there are density differences between each halftone dot and within each halftone dot, and the half-maximum width of the peak on the left correlates with the number of defective halftone dots. This makes it possible to roughly, simply, and quickly determine whether the printing condition is good or bad.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of an example of an apparatus for implementing the present invention, FIG. 2 is an enlarged pattern of halftone dot printing, and FIG. 3 is a diagram explaining the meaning of the coordinate range in which halftone dots should exist. Figure 4 is a histogram showing the distribution of blackness in an enlarged image of halftone printing, Figure 5 is a histogram showing the distribution of blackness when printing density is high and low, Figure 6 is a histogram F(x), A diagram explaining the meaning of F(y), Figure 7 is a histogram F(x)
FIG. 8 is an enlarged view of the binarized halftone dots.

Claims (1)

[Claims] 1. A means for storing image data of an enlarged image of a printing surface in an image memory, and binarizing the data of each pixel in the image memory into two types, one with halftone dots and one without halftone dots, according to a predetermined determination level, In the enlarged image of the printing surface made up of this binarized data, search for the position where the halftone dot should be, specify multiple pixel points within a predetermined area including that position, and The number of halftone dots obtained at a point is compared with a set value, and when the set value is not met, the operation of determining a defective halftone dot is performed for each position where the halftone dot should be, and the halftone dot is determined to be a defective halftone dot. A halftone print evaluation device comprising a data processing device that counts the total number of halftone dots.