JP2687555B2 - Text printing quality control device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION <Industrial field of application> The present invention relates to the print quality of a character printed matter such as a dictionary or a textbook centering on characters printed by a printing machine such as an offset printing machine or a letterpress printing machine. The present invention relates to a character print quality management device for measurement management, and more particularly to a character print quality management device having a blur removing means.

<Prior Art> Generally, in a character printed matter centering on characters printed by a printing machine such as an offset printing machine or a letterpress printing machine, particularly in encyclopedias, dictionaries, textbooks, etc., for all pages of the character printed matter. Maintaining uniform density and thickness of characters is strongly required as the quality of printed characters. In addition, among printing failures that occur during printing, the presence or absence of fog, blurring, and blurring is one of the major factors in evaluating the print quality of characters. Therefore, conventionally, in order to prevent such deterioration of the print quality, the character printed matter during printing is manually extracted at a certain fixed interval, and the character's impression (density, thickness), dulling, The blur is evaluated, and the printing press is adjusted to keep the print quality constant. In view of such a current situation, the character printed matter to be managed is divided into a plurality of pixels by the input means and is taken in as gradation data for each pixel, and the character printed matter is represented from the taken gradation data. A character print quality management device has been proposed which extracts the amount as a gradation histogram and evaluates the print quality of the character printed matter from the gradation histogram.

<Problems to be Solved by the Invention> However, it is physically difficult to obtain an optical system that is ideally free from the influence of blurring in the above-mentioned input means, and this adjustment requires a great deal of labor. for that reason,
There is a drawback that incorrect information is included in the gradation histogram for obtaining the characteristic amount to be measured. In other words, if the optical system is blurred, the gradation measured at the edge of the character part and the paper white part does not change stepwise,
Since the image deterioration phenomenon that causes a gradual change occurs, the outline information of the character also includes the pixel information of the part affected by the blur, and therefore the gradation histogram for obtaining the feature amount of the character group is included. Since it is included as error information, there is a drawback that the calculated feature amount of the character group also has an error. Even if an optical system with no blur is obtained, the sampling interval of the sensor used in the input means does not become 0, so that an erroneous value different from the original density may occur at the edge portion such as the outline portion of the character. The phenomenon of being obtained occurs. In particular, when the feature amount of the character group is obtained, the number of pixels applied to the edge portion is large because the line width of the character is small, and the error information is included in the feature amount of the character group to be obtained in large numbers. . Further, when shape comparison with a gradation histogram obtained from a reference printed matter is performed and the quality of a character group is evaluated, the influence of the edge on the gradation histogram varies greatly depending on the number of pixels applied to the edge. Therefore, when comparing the gradation histograms between a plurality of pages, the number of pixels applied to the edge is different due to the difference in the size and number of characters, so that the shape of the gradation histogram has different effects. This is a major obstacle to obtaining accuracy when performing the shape comparison of the gradation histograms.

Therefore, an error occurs in the feature amount of the character group to be obtained due to an error in the gradation histogram of the character group due to the influence of these edges, and this causes a large problem such as erroneous determination in the quality evaluation of the character group. .

According to the present invention, it is possible to accurately obtain the feature amount of a character group without being affected by blurring, and it is possible to perform the character with higher accuracy even when the shape comparison with the reference gradation histogram is performed. An object is to provide a print quality control device.

<Means for Solving the Problems> The present invention has been made to solve the above problems.
Input means for dividing the character printed matter to be managed into a plurality of pixels and taking in the gradation data for each pixel, and storage means for storing the gradation data for each pixel taken in by the input means. , A secondary differential calculating means for calculating a secondary differential value from the grayscale data stored in the storage means, and a floor for extracting only the grayscale data having a magnitude of the secondary differential value smaller than a certain threshold value. Tone data extracting means, a feature amount extracting means for extracting the feature amount of a character group representing the print quality of a character printed matter as a tone histogram based on the extracted tone data, and the feature amount extracting means. A character mark comprising: a print quality means for evaluating the print quality of a character printed matter from a gradation histogram; and an output display means for outputting and displaying an evaluation result by the print quality evaluation means. It is a quality control equipment.

<Operation> Since the character print quality control device according to the present invention is provided with the second-order differential operation means and the gradation data extraction means, the gradation data of the character print object to be managed, which is taken in from the input means, is the second order. Differentiate, and the value of the second-order differentiated value is used to separate the gradation data having the influence of optical blur from the gradation data having no influence on the basis of a certain threshold value to obtain the optical blur. It is possible to extract only the gradation data that is not affected by.

Then, the characteristic amount of the character printed matter is extracted as a gradation histogram from the gradation data thus extracted, and the print quality of the character printed matter is evaluated from the extracted gradation histogram, and the above-mentioned problems are solved. .

<Example> Hereinafter, one example of the present invention will be described in detail with reference to the drawings.

FIG. 1 is an external view showing a basic hardware configuration example of a character print quality management device according to the present invention, and FIG. 2 is a functional block diagram of the character print quality management device.

First, in FIG. 1, (1) is a character printed matter such as an encyclopedia, a dictionary, or a textbook to be managed, (2) is a flatbed image scanner as an input means, and (3) is an image scanner (2) upper part. Is a transparent mount, (4) is a light-shielding door, and (5) is a switch. Further, (6) is a processor for inputting data from the image scanner (2) through the cable (7), and the processor (6) has a monitor (8), a disk device (9), and a mouse. Equipment (10)
Are respectively connected as shown.

Here, the character printed matter (1) may be printed by a printing machine using an offset printing method or may be printed by a printing machine using a letterpress printing method. Further, the present invention is not limited to the same printing machine, but may be a character printed matter printed by a plurality of printing machines. On the other hand, the image scanner (2) as the input means has a gradation of 8 bits or more and a resolution of 10 per mm.
It is desirable to read in about 30 lines. Regarding the resolution, since the resolution of the human eye is about several tens of μm, if the resolution is too low, the quality control becomes meaningless, and if you input too finely, the amount of data to handle increases, and This is because not only the time required for the arithmetic processing becomes very long, but also the capacity of the memory and the disk for storing the data becomes considerably large. For this reason, in this embodiment, an image scanner using a CCD line sensor having a resolution of 16 lines per mm will be described. The input means is not limited to a flatbed type image scanner, but may be a CCD type camera, a video signal output type camera, a drum type scanner or a flying spot type scanner.

Next, in FIG. 2, (11) scans an entire specific page in the character printed matter (1) to be managed, divides it into a plurality of pixels, and is an optical signal which is a gradation signal for each pixel. An input device (CCD line sensor) for converting the captured image signal as a gray-scale image signal and outputting it, (12) is a LOG circuit that performs appropriate processing on the output signal from this input device (11), (1
3) is a digitizing means for converting an analog signal output from the LOG circuit (12) into a digital signal, and (14) is an image processing means for correcting the output signal from the digitizing means (13). , (15) are storage means comprising a memory for storing the output signal from the image processing means (14), that is, the gradation data. Further, (16) is an extraction processing means for removing the gradation data of the pixel of the character contour portion affected by blur from the gradation data stored in the storage means (15) And a gradation data extraction means for extracting only gradation data of pixels whose magnitude of the second derivative is less than a certain threshold value. Further, (51) represents the print quality of the character printed matter (1) by performing a predetermined calculation / processing based on the gradation data of the pixels extracted by the gradation data extraction means from which the influence of the blur is removed. A feature quantity extracting means for extracting a feature quantity (character density, thickness, blurring, etc.) as a gradation histogram, that is, a density histogram;
The print quality of the character printed matter (1) is evaluated by performing shape comparison between the gradation histogram extracted by the feature amount extraction means and the reference gradation histogram stored in advance in the reference memory (52). It is a processing unit including a print quality evaluation unit. Further, (53) is an auxiliary storage means including a disk for storing the evaluation result by the processing means (51), and (54) is an output display including a monitor for outputting and displaying the evaluation result by the processing means (51). It is a means.

Here, the LOG circuit (12) performs A / D conversion when the grayscale image signal is input as a gradation signal by the CCD line sensor, because the portion with high print density is an important factor in character quality. It is used for the purpose of digitizing the gradation of the high density portion with high accuracy and high resolution.
Further, the image processing means (14), since the signal digitized by the digitizing means (13) may include information error due to optical distortion of the CCD line sensor or variations in photoelectric conversion characteristics, A calibration curve or the like is created in advance, the data is stored in the storage means (14), and the data of the calibration curve is appropriately called to be used for the purpose of correcting the signal value. In this case, as a calibration method, when a flat bed type image scanner is used for the input device (11), linear (1
However, when a TV camera or the like is used, a planar (two-dimensional) correction is performed due to the illumination.

Next, the operation of the character print quality management device configured as described above will be described.

First, the worker puts the character printed matter (1) to be measured on the transparent pedestal (3) above the flatbed type image scanner (2), and the inspection surface of the character printed matter (1) is the pedestal (1). Set so that it touches 3). Next, the worker closes the light blocking door (4) for blocking the outside light and turns on the switch (5), so that the data from the image scanner (2) is transferred to the processor (through the cable (7)). The transfer is started in 6). This data is taken in by the input device (11) as a grayscale image signal for the entire specified page in the character printed matter (1) and is digitized through the LOG circuit (12) and the digitizing means (13). Further image processing means (1
The signal error due to the optical distortion and the variation of the photoelectric conversion characteristic is corrected by 4), and is stored in the storage means (15) as gradation data. Next, in order to remove the optical blur generated in the character outline portion by the input device (11), the storage means (1
From the gradation data stored in 5), the secondary differential value is calculated by the secondary differential value calculating means, and only the gradation data of the pixels whose magnitude of the secondary differential value is less than a certain threshold value is extracted. Is extracted by the gradation data extracting means for performing the predetermined calculation and processing based on the extracted gradation data, and the characteristic amount representing the print quality of the character printed matter (1) is a gradation histogram, that is, a density histogram. To extract. Further, the extracted gradation histogram is stored in the reference memory (52) by the print quality evaluation means of the processing means (51).
The print quality is evaluated by performing a shape comparison with the reference gradation histogram stored in the above. The evaluation result of the print quality is output and displayed to the operator of the printing press by the output display means (54), that is, the monitor (8), and is also stored in the auxiliary storage means (53).

Next, from the gradation data captured by the image scanner (2), the method of removing the influence of blurring, the method of extracting the gradation histogram showing the print quality of the character printed matter (1), and the shape comparison of the gradation histograms are performed. A method of evaluating print quality will be specifically described.

FIG. 3 (a) shows an enlarged part of a character and a cross-sectional situation of an edge in the white part of the paper. If the cross section of the printed characters is an almost ideal step,
The gradation data captured by the input means without blur is also the third
Although the gradation is almost stepwise as shown in FIG. 3B, the gradation data captured by the blurred input means gradually changes as shown in FIG. 3C due to image deterioration due to blurring of the optical system. It becomes curved. Therefore, if an attempt is made to obtain a gradation histogram for directly extracting the feature amount of a character group from the gradation data acquired by the blurred input means, the acquired gradation data will have a blur other than the original density. Since a plurality of pixels having the gradation affected by is also included, the gradation histogram also includes a plurality of erroneous information of the pixels affected by the blur. In particular, since the line width of a character is thin, many pixels are applied to the outline of the character, which has a great influence. That is, when trying to evaluate the print quality from the gradation histogram, the influence of the pixels on the character contour portion appears as a large error, which causes the erroneous evaluation of the print quality.

By the way, in a character printed matter, the number of pixels in the white paper portion is very large, and the number of pixels in the portion affected by blurring is smaller than that. On the other hand, in a character printed matter, the number of pixels in the character portion is small, and thus the influence of blurring becomes large. Information on the character part is important when evaluating the character quality. From these facts, the influence of blurring greatly affects the character portion, so it is important to remove the influence of blurring in the character portion.

In order to eliminate the influence of blur, there is a method to adjust the distance of the lens of the optical system, etc., but it is difficult to obtain an optical system without blur completely, and there are changes due to aging and impact. In that case, there is a big problem such that a complicated adjustment is required again. This adjustment becomes a problem especially when a flat bed type image scanner is used.

Therefore, in the present invention, by adding a means for automatically extracting the gradation data of the pixels not affected by the blur from the gradation data in which the image is deteriorated in the character contour portion including the blur input by the input means, the character contour is added. It was possible to eliminate the effect of blurring in the department.

As a concrete method, a filter by a 3 × 3 Laplacian operator used for extracting an edge portion by an image processing method was used. In this filter, a portion that makes a rapid change is extracted as a large value. That is, when the Laplacian value is small, it can be determined that the pixel is not affected by the edge, and when the Laplacian value is large, it can be determined that the pixel is affected by the edge. This takes advantage of the fact that in printed matter, the density of ink-containing characters and paper white is more stable than that of edges.

FIG. 4 (a) is a diagram showing a cross section of a line of a typical character, and FIG. 4 (b) is a diagram showing absolute values obtained by quadratic differentiation of FIG. 4 (a). That is, the absolute value of the secondary differential value is small in the paper white part and the character part where the density is stable,
Large values appear at the edges. By the way, in FIG. 4, the description is made in one dimension for simplification of the description, but the same applies to the case of a two-dimensional surface, and as an operator for performing the second derivative with respect to the two-dimensional surface, the Laplacian shown in FIG. Used operator. FIG. 5 (a) is a dichotomous operator for eight directions, and FIG. 5 (b) is a dichotomous operator for four directions. Either of these may be used. Next, in order to obtain a threshold value for determining whether it is an edge pixel portion or a stable pixel, 2
A histogram (Fig. 6) of the absolute value obtained by the secondary differentiation was created. As can be seen from this, peaks corresponding to stable pixels with a small secondary differential value and peaks corresponding to edge portions with a large differential value appeared as separate shapes. As a result, the edge portion and the stable portion can be separated, and an appropriate threshold value is obtained, for example, from the second derivative value corresponding to the maximum pixel number of the mountain of the edge portion, and the threshold value = the maximum pixel number Value / 2, and a binary differential value smaller than this value is used as a gradation histogram to obtain the feature quantity for evaluating the quality of the judgment character, with the pixel as a stable pixel with a small gradation difference from the surrounding pixels. Add. At this time, a pixel associated with an edge having a large differential value is not added to the gradation histogram for obtaining the feature amount, and thus a gradation histogram for obtaining the feature amount for evaluating the character quality without being influenced by the edge is obtained. I was able to.

FIG. 7 is a diagram showing the flow of the above-described processing, in which the image captured from the scanner is temporarily stored in the storage circuit A via the LOG circuit or the like. Next, the image of this memory circuit A is
The value obtained by the secondary differentiation is stored in the storage circuit B, the histogram is created from the secondary differentiation value, and the threshold value for discriminating between the edge pixel and the stable pixel is calculated by the threshold value calculation circuit. Next, the threshold value obtained as described above and the secondary differential value stored in the storage circuit B are compared for each pixel, and only the gradation data of the pixel whose secondary differential value is smaller than the threshold value is subjected to the next processing. By providing a comparison / extraction circuit for not storing the gradation data of the pixel having a differential value larger than the threshold value in the means, only stable pixels are displayed in the gradation histogram for obtaining the feature amount in the next processing means. Come on. With these processes, it is possible to restore the histogram for obtaining the feature amount without being affected by the edge, as in the histogram shown in FIG. 3B.

Next, describing the method of extracting and evaluating the feature amount,
First, as the overall gradation, the average value of the character part (31) of the gradation histogram may be used as shown in FIG. 8 (a), the median value of the character part (31), or the character The gradation at the mode (α) of the part (31) may be used.

Next, regarding the situation in the outline portion of the character, when "blurring" is entirely generated in the character printed matter itself, as shown in FIG. 8 (b), the gradation of the character portion (31), that is, the distribution of density. Results in a skirt that spreads toward the low concentration side. This is because the "blurring" of the character causes the peak to shift to a lower density than the original character density, and to take an intermediate value with a gentle outline of the blank sheet. At this time, the character itself is "blurred"
The blurring between the input means and the input means can be distinguished because the outline of the character changes more smoothly than the blur of the input means as shown in FIG. 8 (c).

Also, if “dullness” is occurring overall,
As shown in FIG. 8 (d), the density distribution of the character part (31) has two peaks of density, and a small peak on the lower density side than the original peak is formed.
This is because the printing in the outline portion of the character is not clear, and therefore the density of the intermediate value between the character density and the blank paper density is distributed. That is, by analyzing the distribution of the gradation histogram as shown in the figure, the overall edge condition can be evaluated.

In this case, the spread amount in the character portion (31) may be calculated by calculating the variance in the established statistical phenomenon, or may be calculated by calculating the coefficient of variation or the like.

Next, regarding the thickness of the character, the ratio of the character part to the entire screen is obtained by calculating the ratio of the total number of pixels of the dispersed character part and the number of pixels of the entire screen. May be. On the other hand, when obtaining the feature amount of the character group of the entire screen, if the printed character includes a pattern represented by shading, the feature amount of the character and the pattern are included in the overall feature amount. Since there is a mixture with the feature quantities of, there is an inconvenience. Therefore, the character portion for which the characteristic amount is to be obtained may be manually designated in advance, or the pattern portion may be automatically discriminated and the portion may be excluded. In this case, as a method of manually specifying, the gradation data captured by the image scanner (2) is displayed on the monitor (8), and the portion for which the feature amount is to be acquired is specified by the mouse device (10) or digitizer. There are ways to do it.

The evaluation result of the characteristic amount of the character printed matter obtained as described above is displayed on the monitor (8) to inform the operator of the printing press. At this time, the feature amount of the character group may be graphically displayed. Then, the operator adjusts the printing machine based on the evaluation result of the characteristic amount of the character printed matter displayed on the monitor (8), whereby the variation in the print quality of the character can be suppressed, and the entire character printed matter can be suppressed. The print quality can be improved. Further, by sequentially storing / holding the extracted feature amount in the disk device (9) and managing it, it is possible to prevent variations in print quality due to changes over time. As a result, it is also possible to manage the print quality of the published print by the revision or edition.

<Effects of the Invention> According to the character print quality management apparatus of the present invention, the print quality of a character printed matter such as encyclopedias, dictionaries, and textbooks, which has a particularly high demand for print quality, can be evaluated by various factors. It can be managed automatically and easily without any occurrence.

Further, since the second-order differential calculation means and the gradation data extraction means are provided, the gradation data having an optical blur effect is removed from the taken-in gradation data by these means, and is extracted by this means. There is an effect that the print quality can be evaluated from the generated gradation data that is not affected by the optical blur, and thereby erroneous determination due to a quality evaluation error can be significantly reduced.

Therefore, it is possible to provide a character print quality management device capable of always maintaining the print quality of a character printed matter with high accuracy and stability.

[Brief description of the drawings]

FIG. 1 is a hardware configuration diagram showing an embodiment of a character print quality control device according to the present invention, FIG. 2 is a functional block diagram thereof, and FIG. 3, FIG. 4, FIG. 6 and FIG. FIG. 5 is a diagram for explaining the concept of the present invention for removing gradation data having an influence of blur from the acquired gradation data, FIG. 5 is an explanatory view of a Laplacian operator used at that time, and FIG.
The figure is a diagram for explaining how to evaluate print quality from the extracted gradation histogram. (1) …… Character printed matter, (2) …… Image scanner (3) …… Frame, (4) …… Shade door, (5) …… Switch (6) …… Processor, (7) …… Cable , (8) ...... Monitor (9) ...... Disk device, (10) ...... Mouse device (11) ...... Input device, (12) ...... LOG circuit (13) ...... Numericalization means, (14) ... ... image processing means (15) ... storage means (16) ... extraction processing means (31) ... character part, (51) ... processing means (52) ... reference memory, (53) ... auxiliary storage Means (54) …… Output display means

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Masao Kim 1-5-1, Taitō, Taito-ku, Tokyo Yasushi Takahashi Examiner, Toppan Printing Co., Ltd.

Claims (1)

(57) [Claims] 1. Input means for dividing a character printed matter to be managed into a plurality of pixels to capture as gradation data for each pixel, and gradation data for each pixel captured by the input means. Storage means for storing, secondary differential calculation means for calculating a secondary differential value from the grayscale data stored in this storage means, and only grayscale data whose magnitude of the secondary differential value is smaller than a certain threshold value Gradation data extracting means for extracting the characteristic quantity of the character group representing the print quality of the character printed matter as a gradation histogram based on the extracted gradation data, and the characteristic quantity A print quality evaluation means for evaluating the print quality of the printed matter from the gradation histogram extracted by the extraction means; and an output display means for outputting and displaying the evaluation result by the print quality evaluation means. Character printing quality control apparatus according to claim Rukoto.