JPH07190955A - Equipment for deciding extent of smear of printed matter - Google Patents

Abstract

PURPOSE:To indicate the extent of smear numerically by calculating the amount representative of the feature of smear from an image data, deciding the smear of intricate shape such as crease, and further deciding the smear at a patterned part. CONSTITUTION:A printed matter 101 is read out by means of a line sensor 103 which then delivers an image signal to an image input unit 103 where the image signal is subjected to A/D conversion and stored in an image memory 104. A feature amount extracting section 105 indicates the extent of smear numerically based on the image data stored in the memory 104, i.e., calculates an amount representative of the feature of smear. A graph display section 106 displays an approximate line representative of the relationship between the numeric amount of smear and the data thereof. An operator decides the goodness of fit between the data presented on the display section 106 and the approximate line and notifies pass or fail to a decision input section depending on the goodness of fit thus decided. A decision section 107 decides the extent of smear of a printed matter based on a numeric amount of smear extracted at the feature amount extracting section 105.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an apparatus for inspecting a printed image, and more particularly to an apparatus for determining the degree of stain on a dirty printed matter.

[0002]

2. Description of the Related Art Conventionally, in order to determine the degree of stain on printed matter,
A method has been adopted in which a human looks at a printed matter and subjectively judges it. In order to automate these, a method has been devised in which a stain is determined by reading a printed matter to be determined with a sensor and comparing it with preset reference data. Furthermore, in the invention disclosed in Japanese Patent Laid-Open No. 189794/1983, the integrated value of the amount of transmitted light of a printed matter is stored in a memory, binarized, and the number of pixels is counted, and the number of pixels is used to discriminate a genuine note or an unfit note. Are doing Specifically, as shown in FIG. 2, the transmitted light from the light source 201 through the paper sheet 202 is received by the photoelectric converter 203 and stored in the memory 204. Then, the binarizing device 205 binarizes the image data stored in the memory 204. Finally, the discriminating device 206 counts the number of pixels binarized by the binarizing device 205, and if it is less than a certain threshold, it is discriminated as a genuine bill, and if not, it is discriminated.

[0003]

DISCLOSURE OF THE INVENTION Problems to be Solved by the Invention
In the method disclosed in Japanese Unexamined Patent Publication No. 8-189794, the shape of the binarized data is not considered at all because the discrimination is made only by the number of binarized pixels. Therefore, it is impossible to discriminate the stain due to the complexity of the shape such as wrinkles. Further, it is impossible to discriminate the portion where the pattern exists.

Therefore, an object of the present invention is to take into consideration the shape of a dirty image of a dirty printed matter, to determine the stain of a complicated shape such as a wrinkle, and further to determine the stain of a portion having a pattern, and to determine the degree of the stain. The purpose of the present invention is to provide a printed matter stain discriminating apparatus that can be represented by.

[0005]

In order to solve the above-mentioned problems, a printed matter stain degree determining apparatus according to the present invention has means for inputting image data of a printed matter, and a stain characteristic from the image data input by the input means. The present invention is characterized by comprising: means for calculating a stain characteristic amount shown; and means for determining the degree of the stain on the basis of the stain characteristic amount input from the calculating means.

Further, the printed matter stain degree determining apparatus according to the present invention comprises means for inputting the image data of the printed matter, means for dividing the image data input by the input means into a plurality of small areas, and each of the small areas. Means for calculating the pollution feature amount showing the characteristics of pollution, based on the pollution feature amount input from the calculating means, means for determining the degree of the pollution for each of the small areas, and each of the small areas The present invention is characterized by comprising means for averaging the degree of stains to obtain the degree of stains on the entire printed matter, and means for outputting a small amount area having a large degree of stains in each of the small areas.

Further, the printed matter stain degree determining apparatus according to the present invention comprises means for calculating a stain characteristic amount indicating the feature of stain from image data, means for inputting image data of the printed matter for determining the stain, and the input. From the image data input by the means, the means for obtaining the first stain characteristic amount using the calculating means, and the image data of the printed matter without stain,
Means for obtaining a second pollution characteristic amount using the calculation means,
The pollution feature is referred to with reference to a table for associating the pollution feature quantity and the degree of pollution calculated in advance with a unit for calculating the pollution feature quantity using the first pollution feature quantity and the second pollution feature quantity. And a means for obtaining the degree of pollution from the quantity.

[0008]

A piece of the input printed image is divided into small regions each having n pixels, and the box count number is obtained in each small region. This operation is performed a plurality of times by changing the number of pixels in one piece of the small area. A graph is created on the basis of data each constituted by the number of pixels in one piece and the box count number obtained for the number of pixels, and an approximate straight line is obtained. A stain feature amount is calculated from this approximate straight line, and the degree of stain is obtained based on the feature amount. The box count number is calculated from the density difference of pixels in the small area, the number of pixels having the value “1” in the binarized image, or the like.

[0009]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the arrangement of a stain degree determining device for printed matter according to an embodiment of the present invention. In the same figure, the printed matter 101 for determining the stain is read by the line sensor 102, and the image signal is sent to the image input device 103. The image input device 103 A / D converts the sent analog image signal and stores it in the image memory 104. The feature amount determination unit 105 digitizes the degree of contamination from the image data stored in the image memory 104. Feature quantity extraction unit 10
Reference numeral 5 has a graph display unit 106, and the feature amount extraction unit 1
The amount of pollution quantified in 05 and the approximate straight line of the data are displayed in a graph. The operator uses the graph display unit 106
Judge the degree of fit between the data displayed in and the approximate straight line,
When it is determined that the data is sufficiently on the approximate straight line, “OK” is input to the operator determination input unit 108.
If it is determined that it is not the case, "impossible" is input.
The discriminating unit 107 discriminates the degree of stain of the printed matter based on the stain amount digitized by the amount extracting unit 105. For example,
If the fouling amount calculated by the feature amount extraction unit 105 is 0 to 20, there is no fouling, 20 to 40 is a fouling, 40 to 60 is a fouling, 60 to 80 is fouling, and 80 to 100 is a fouling. . Table 1 shows the relationship between the feature amount and the degree of stain.

[0010]

[Table 1]

If "OK" is input in the operator judgment input unit 108, the degree of contamination is output as it is, and if "INHIBIT" is input, reject or judgment, output of data error, etc. is output. To do. Further, although the method of discriminating the stain with the intervention of the operator is described here, the stain can be discriminated without the intervention of the operator. In that case, the graph display unit 10
In 6, the degree of fit of the data and the approximate line of the data is
For example, it is represented by the error between the data shown in FIG. 7 and the approximate straight line. When the data 1 (702), the data 2 (703), the data 3 (704) and the approximate straight line 701 of these data are as shown in the figure, the data 1 (702) to the approximate straight line 70
The sum of the length 705 of the perpendicular drawn from 1, the length 706 of the perpendicular drawn from the data 2 (703) to the approximate straight line 701, and the length 707 of the perpendicular drawn from the data 3 (704) to the approximate straight line 701 are given here. Then, the degree of non-conformity is defined, and the smaller the value, the better the agreement between the data and the approximate straight line.
Specifically, when the approximate straight line is given by y = ax + b, the difference 705 between the data 1 (x1, y1) (702) and the approximate straight line is represented by the difference between y1 and ax1 + b, and similarly the data 2 ( The difference 706 between x2, y2) (703) and the approximate straight line is the difference between y2 and ax2 + b, and the data 3 (x3, y3) (7
04) and the approximation straight line 707 are obtained by the difference between y3 and ax3 + b, and the sum of these differences is regarded as the degree of incompatibility between the data and the approximation straight line. And, for example, this value is 20
In the following cases, the operator input unit 108 is “OK”, and in the case of being larger than that, “NO”.

FIG. 3 is a detailed block diagram of the feature quantity extraction unit 105. In the scale selection unit 301, the image memory 10
A plurality of pixel scale ranges to be analyzed are selected based on the size of the image data stored in 4. For example, when the size of the image data is 100 pixels in the vertical direction and 100 pixels in the horizontal direction, the range of the analysis pixel scale is increased from 10 pixels, which is 1/10 of the image size, to 11 pixels, 12 pixels, and 1 pixel, respectively. Determine up to 20 pixels. When the size of the image is known in advance, the analysis pixel scale range may be fixed. The box count unit 302 uses one pixel scale of the analysis pixel scale range selected by the scale selection unit 301 to convert the image data stored in the image memory 104 into a small area of the size of this pixel scale. To divide. Then, the difference between the maximum density value and the minimum density value of the image data is obtained in each small area, the difference is divided by the value of the pixel scale, and 1 is added. This value is used as the box count number in the small area. FIG. 4 is a diagram showing this box count processing. That is, in a small area of a certain pixel scale r (401), the maximum density value I1
(402) and the minimum density value I3 (404) are obtained, and one obtained by adding 1 to the number of blocks of size r which can be entered between them is defined as the box count number. When the box counts are obtained for all the small areas, the box counts for all the small areas are added, and this value is defined as the box count on the pixel scale used previously. Then, the pixel scale and the box count number are paired and stored in the memory 303. If there is a pixel scale that has not been analyzed yet in the pixel scale range selected by the scale selection unit 301, the loop end determination unit 304 repeats the above processing for that pixel scale. The regression line calculation unit 305 takes a log (logarithm) for each data of the set of pixel scale and box count number stored in the memory 303, represents it on a graph, and obtains a regression line (approximation line) using the least squares method. . The data of the set of pixel scale and box count and the regression line are displayed on the graph display unit 106. In the feature amount data calculation unit 306, the stain amount is digitized from the slope of the approximate straight line calculated by the regression straight line calculation unit 305 and the y intercept of the approximate straight line, that is, the box count number in the smallest pixel scale of the analysis pixel scales. Output. As a method of digitizing this, for example, the sum, product, quotient, or difference of the slope of the approximate straight line and the value of the y intercept can be considered.

The regression line by the method of least squares can be obtained as follows. As an example, data 1 (x1, y1) (802) and data 2 (x2, y2) shown in FIG.
) (803), data 3 (x3, y3) (804) and data 4 (x4, y4) (805), the regression line (8
01) is specifically shown. First, the regression line (801) is assumed to be y = ax + b, and a and b are unknowns. The least squares method is to obtain a and b that minimize the sum of squares of vertical deviations between each data and the straight line y = ax + b. When the sum of squares of the vertical deviation between each data and the straight line is calculated,

[0014]

[Equation 1] Becomes From now on, if Q is differentiated with a and b and it is 0,

[0015]

[Equation 2] And unknowns a and b are obtained.

The above is the description of the structure of the stain degree determination device for printed matter. Here, the stain degree is actually determined based on a part of image data of a certain stained print matter as shown in FIG. Indicates. Here, the pixel scale to be analyzed for the image as shown in FIG. 9 is set from 2 pixels to 4 pixels. First, the box count number is calculated for two pixels on the pixel scale. As shown in FIG. 10, the image is divided into 2 × 2 small regions, and the box count number is obtained for each small region. For example, in the small area 1001, since the maximum density value is 6 and the minimum density value is 1, the difference is 5, which is 2 when divided by the pixel scale 2 and 3 when 1 is added. Similarly, in the small area 1002, the maximum density value is 9, the minimum density value is 4, the difference is 5, and when divided by the pixel scale 2, it is 2 and 1
Is added to give a total of 3. The block count number in each small area is shown in a circle in the center of each small area in FIG. Then, the block count number in the pixel scale 2 is
The sum of the block counts of these small areas is 60, which is 60. Subsequently, the box counts are similarly obtained for the pixel scales of 3 pixels and 4 pixels, and from FIG. 11 and FIG.
12 and 12, respectively. Next, if the log is taken for each of these pixel scale and box count, the set of (pixel scale, box count) is (0.3, 1.78), (0.48, 1.08), (0.6, 1.08)
Becomes When these data sets are plotted on a graph, it becomes as shown in FIG. From the plot points on Fig. 13,
The regression line is obtained by the least squares method. Substituting values into equations (6) and (7), respectively, we obtain a regression line,

[0017]

(3) y = -2.46x + 6.67 (8) is obtained. Here, the degree of matching between the data and the regression line will be described when the operator does not intervene. In that case, the error between the data and the approximate straight line needs to be calculated, so the total error in the vertical direction between the data and the approximate straight line is calculated.

[0018]

[Equation 4] Error ε = (-2.46 × 0.3 + 6.67-1.78) + (-2.46 × 0.48 + 6.67-1.08) + (-2.46 × 0.6 + 6.67-1.08) = 12.7 (9)

It becomes Since this value is smaller than a predetermined threshold value of the degree of nonconformity, for example, 20, the operator determination input unit 108 is “OK”. Then, in the discriminating unit 107, the contamination amount quantified by the characteristic amount extracting unit 105, that is, 2.46 which is the absolute value of the slope of the approximate straight line, and the box count number 60 in 2 pixels which is the smallest pixel scale among the analysis pixel scales are 60. The degree of fouling is determined using 1.78 obtained by taking the log of. Alternatively, the degree of contamination is determined by using 6.67 of the y-intercept of the approximate straight line in FIG. Here, using the former, for example, the product of those values was obtained, and the pollution amount was set to 4.4. This value falls into the “no stain” section of the predetermined stain determination step, and the image analyzed here is determined to have no stain.

The feature quantity extraction unit 105 can also perform processing as shown in FIG. The scale selection unit 501 selects a range of pixel scales to be analyzed based on the size of the image data stored in the image memory 104. As described above, when the size of the image data is 100 pixels in the vertical direction and 100 pixels in the horizontal direction, the range of the analysis pixel scale is from 10 pixels which is one tenth of the size of the image to 20 pixels which is 10 pixels larger than that. It's just like that. When the size of the image is known in advance, the analysis pixel scale range may be fixed. The binarization processing unit 502 binarizes the image data stored in the image memory 104 with a threshold value, for example, half the image quantization value (maximum value).
The image is stored in the image memory 508 in the digitization processing unit 502. The box counting unit 503 uses one pixel scale in the analysis pixel scale range selected by the scale selecting unit 501 to use the image memory 50 in the binarization processing unit 502.
The binary image data stored in 8 is divided into small areas of the size of this pixel scale. Then, if there is 1 pixel in each small area, the small area is 1, otherwise 0.
And FIG. 6 shows this box counting process. That is, if the binary image 601 is an inspection target and there is 1 pixel in a small area of a certain pixel scale r (603), (6
02), the small area is set to 1 and (604) 0 if not. In FIG. 6, a small area having one pixel is indicated by diagonal lines. When the determination has been completed for all the small areas, the number of small areas that have become 1 among these small areas is counted, and this value is set as the box count number in the previously used pixel scale. Then, the pixel scale and the box count are paired and stored in the memory 504. In the loop end determination unit 505, if there is a pixel scale that has not been analyzed yet in the pixel scale range selected by the scale selection unit 501, the above processing is repeated for that pixel scale. The regression line calculation unit 506 obtains a log for each set of pixel scale and box count number data stored in the memory 504, represents the log on a graph, and obtains a regression line (approximate straight line) by the least-squares method. The data of the set of the pixel scale and the box count and the approximate straight line are displayed on the graph display unit 106. In the feature amount data calculation unit 507, the stain amount is digitized from the slope of the approximate straight line calculated by the regression straight line calculation unit 506 and the y intercept of the approximate straight line, that is, the box count of the smallest pixel scale of the analysis pixel scales. Output. As a method of digitizing this, for example, the sum, product, quotient, or difference of the slope of the approximate straight line and the value of the y intercept can be considered.

Here, the stain degree determination based on the process of FIG. 5 will be specifically described. It is assumed that there is a part of the image data of the dirty printed matter as shown in FIG. First, the pixel scale to be analyzed is selected (501), but here, it is set to 2 to 4 pixels. Then, binarization processing is performed. Here, the binarization threshold value is set to 8 and the pixels of 8 levels or higher are set to 1. An image binarized in this way is shown in FIG. In FIG. 14, it is assumed that the shaded pixels represent 1. First, the box count number is calculated for two pixels on the pixel scale. As shown in FIG. 15, the image is divided into small areas of 2 × 2, and one pixel is searched for in each small area. For example, since 1 pixel does not exist in the small area 1501, the small area 1501 becomes 0. Since one pixel exists in the small area 1502, the small area 1502 has one pixel. The judgment in each small area is shown in the center of each small area in FIG. Then, the block count number in the pixel scale 2 is 11 in this case because it is the number of the small areas which becomes 1 in all the small areas. Then 3 pixel scale,
The box counts are similarly obtained for the four pixels, and are 4 and 4 from FIGS. 16 and 17, respectively. Next, if the log is taken for each of these pixel scale and box count number, the set of (pixel scale, box count number) is (0.3, 1.04), (0.48, 0.6),
It becomes (0.6, 0.6). When these data sets are plotted on a graph, it becomes as shown in FIG. A regression line is obtained from each plot point on FIG. 18 by the method of least squares. Substituting values into equations (6) and (7), respectively, we obtain a regression line,

[0022]

## EQU00005 ## y = -2.16x + 2.04 (10) is obtained. Here, the degree of matching between the data and the regression line will be described when the operator does not intervene. In that case, the error between the data and the approximate straight line needs to be calculated. Therefore, the total error in the vertical direction between the data and the approximate straight line is calculated, which in this case is about 0.9. Since this value is smaller than a predetermined threshold value of the incompatibility, for example, 10, the operator determination input unit 108 is “OK”. Then, in the discriminating unit 107, it is the amount of pollution quantified by the characteristic amount extracting unit 105, that is, the absolute value of the slope of the approximate line 2.
16 and the log of the box count number 11 for 2 pixels, which is the smallest pixel scale of the analysis pixel scale
The degree of contamination is determined by using 1.04 obtained. Alternatively, the degree of contamination is determined by using the y intercept of 2.04 of the approximate straight line in FIG. Here, the product of the former values is calculated to determine the amount of pollution.
It was set to 2.2. This value falls into the “no stain” section of the predetermined stain determination step, and the image analyzed here is determined to have no stain.

The feature quantity extraction unit 105 can also perform processing as shown in FIG. The binarization processing unit 1901 binarizes the image data stored in the image memory 104 with a threshold value, for example, half the image quantization value, and stores it in the image memory 1905 in the binarization processing unit 1901. To do. The perimeter measurement unit 1902 measures the perimeter of the binary image stored in the image memory 1905 in the binarization processing unit 1901. The area measuring unit 1903 obtains the area of the binary image stored in the image memory 1905 in the binarization processing unit 1901. Feature amount data calculation unit 1904
Then, based on the numerical values calculated by the perimeter measuring unit 1902 and the area measuring unit 1903, the pollution amount is digitized and output. As a method of digitizing this, for example, a value obtained by squaring the perimeter and dividing by the area, a sum, a product, a quotient, or a difference of each value can be considered.

Here, the stain degree determination based on the processing of FIG. 19 will be specifically described. As described above, it is assumed that there is a part of the image data of the dirty printed matter as shown in FIG. First, binarization processing is performed. Here, the binarization threshold value is set to 8 and the pixels of 8 levels or higher are set to 1. An image binarized in this way is shown in FIG. In FIG. 14, it is assumed that the shaded pixels represent 1. Next, in FIG. 14, the perimeter and area of the pixel with diagonal lines are obtained. When FIG. 14 is rewritten with binary values, it becomes as shown in FIG. 20, but the perimeter of one pixel of the pixels 2001 is 4 in this case, and the area is 1. The pixel 2002 is adjacent to the pixel 2003, and the total length of these two pixels is 6, and the area is 2.
In this way, the total of the perimeters and the areas of FIG. 20 is calculated to be 50 and 14, respectively. continue,
Quantify the degree of pollution from the perimeter and area. Here, the method of dividing the perimeter by squaring and dividing it by the area is used, and if the correction parameter 0.1 is further added, the pollution amount becomes 17.9. This value falls into the “no stain” section of the predetermined stain determination step, and the image analyzed here is determined to have no stain.

In the above description, the method of analyzing the entire surface of the printed matter for which the stain is to be judged at once is explained, but it is also possible to perform only for the specific area of the printed matter for which the stain is judged. In this case, this can be done by designating an area for determining stain in advance, or by allowing an operator to look at the printed matter and designate an area for determining stain. Further, as shown in FIG. 21, the printed matter is divided into small areas, the stain degree determination processing is performed for each small area, and when the stain degree is determined for all the small areas, the stain degree of the printed matter is finally obtained. You can also As a processing block in this case, the image dividing unit 110 is provided in the image memory 104 of FIG. 1, and the input image is divided into, for example, nine small areas. Then, the feature amount is extracted for each small area to determine the stain degree, and the image integrating unit 111 associates each small area of the image with the stain degree. Then, the overall stain determination unit 112
Thus, the stain degree of the entire printed matter that has been input is obtained by the partial stain determining section 113 to obtain a partial stain degree and a region with a large stain degree, and the respective results are output. Specifically,
As shown in FIG. 21, the printed matter is divided into, for example, nine small areas. Then, it is assumed that the contamination degree is obtained for each small area and that the contamination degree is obtained as shown in FIG. In FIG. 21, the central portion (2105) of the printed matter is not stained, but the right end (210
(3, 2106, 2109) is more polluted. Then, as the result of the overall stain determination, for example, no stain is 0,
By quantifying the stains to 5, stain stains to 10, stains to 15, and taking the average of the nine small areas, this printed matter outputs a determination result of "slightly stain", for example. Alternatively, the position of the small area for which the determination result of "stainlessness" or "contamination" is obtained as the partial stain determination result is output. Furthermore, here, after the stain degree is determined for each small area, the final stain on the printed matter is determined. However, for each small area, the stain amount is not numerically expressed and is expressed numerically as a stain amount. It is also possible to use values such as the sum, product, quotient, and difference of the stain amounts of the respective small areas in the final judgment. By this method, it is possible to extract not only the entire printed matter but also the soiled area in the printed matter.

Finally, the determination of the stain degree of the printed matter including the pattern will be described. As the processing block in this case,
This can be realized by adding the reference data comparison unit 109 to FIG. First, the amount of stains described above is calculated in advance for a printed matter including a pattern without stains, and is stored as reference dictionary data 114 in the reference data comparison unit 109. Then, when determining the stain degree of the printed matter including the pattern, the difference between the stain amount of the printed matter calculated by the feature amount extraction section 105 and the data stored in the reference dictionary data 114 is used as the reference data comparison section 109. Asked in
The result is the stain amount of the printed matter including the pattern. Later,
The degree of pollution is determined from this amount of pollution.

[0027]

As described above, according to the present invention, since the stain of a printed matter can be expressed by a small number of characteristic quantities quantitatively representing the degree of stain, the stain can be easily discriminated. Note that the difference from the stain degree determination by the conventional differential processing or the like is that the binary image is composed of a simple straight line in the conventional pixel number measurement, the variance measurement, etc. , Or if they are composed of complicated lines, if the number of pixels to be measured is the same, those two binary images will be determined as the same stain degree. On the other hand, in the present invention, by obtaining the complexity of the binary image, it is possible to capture a degree of complexity such as how many wrinkles or creases are not obtained only by measuring the number of pixels.

Further, according to the present invention, the entire surface of the printed matter is divided into small areas, the stain degree is obtained for each small area, and the stain degree of the printed matter is finally obtained. Therefore, the overall stain of the printed matter is obtained. It is also possible to specify the area that is contaminated. Furthermore, according to the present invention, it is possible to determine the degree of stain on a portion where a printed pattern exists.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a stain degree determination device for printed matter according to an embodiment of the present invention.

FIG. 2 is a schematic configuration diagram of a conventional pollution degree determination device.

FIG. 3 is a first detailed block diagram of a feature quantity extraction unit in FIG.

FIG. 4 is a diagram for explaining box count processing according to the present invention.

5 is a second detailed block diagram of the feature amount extraction unit of FIG. 1. FIG.

FIG. 6 is a diagram for explaining box count processing according to the present invention.

FIG. 7 is a diagram for explaining the fitness of an approximate straight line.

FIG. 8 is a diagram for explaining how to obtain an approximate straight line.

FIG. 9 is a diagram showing a part of the dirty printed matter image data.

FIG. 10 is a diagram showing a box count number when the pixel scale is 2 pixels.

FIG. 11 is a diagram showing a box count number when the pixel scale is 3 pixels.

FIG. 12 is a diagram showing a box count number when the pixel scale is 4 pixels.

FIG. 13 is a diagram showing a relationship between a pixel scale and a box count number.

FIG. 14 is a diagram showing binarized image data of a dirty printed matter.

FIG. 15 is a diagram showing a box count number when the pixel scale is 2 pixels.

FIG. 16 is a diagram showing a box count number when the pixel scale is 3 pixels.

FIG. 17 is a diagram showing a box count number when the pixel scale is 4 pixels.

FIG. 18 is a diagram showing a relationship between a pixel scale and a box count number.

FIG. 19 is a third detailed block diagram of the feature amount extraction unit of FIG. 1.

FIG. 20 is a diagram for explaining calculation of a perimeter and an area of each image in a binary image.

FIG. 21 is a diagram for explaining stain determination for each small amount area of a printed matter image.

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI technical display location G07D 7/00 E

Claims (12)

[Claims] 1. An apparatus for determining stains on a printed matter, means for inputting image data of the printed matter, means for calculating a stain characteristic amount indicating a stain characteristic from the image data input by the input means, And a means for discriminating the degree of the stain based on the stain characteristic amount inputted from the calculating means. 2. The stain feature amount calculation means selects a plurality of pixel scales to be analyzed according to the size of the input image, and a plurality of small scales of the image based on the pixel scales. Means for dividing into areas, a difference between the maximum density value and the minimum density value of the pixels in each of the small areas,
And means for performing an operation using the pixel scale to provide an operation result, means for providing the total result by summing the operation results obtained for each of the small areas, and a pair of the pixel scale and the total result. Means for obtaining a graph based on the data represented respectively, means for obtaining an approximate straight line from the graph, means for obtaining a y-intercept and slope of the approximate straight line, and calculating the stain feature amount from the y-intercept and slope. And a means for determining the degree of contamination from the contamination characteristic amount with reference to a table that associates the contamination characteristic amount and the degree of contamination obtained in advance. The printed matter stain degree determination device according to claim 1, wherein 3. The stain feature amount calculation means, means for binarizing the input image, means for selecting a plurality of pixel scales to be analyzed according to the size of the image, and each pixel A unit that divides the image into a plurality of small regions based on a scale; a unit that determines a specific small area number that indicates the number of small regions having the specific binarized data in the plurality of small regions; A means for obtaining a graph based on data represented by each pair of the pixel scale and the number of the specific small areas; a means for obtaining an approximate straight line from the graph; a means for obtaining a y-intercept and a slope of the approximate straight line; Means for obtaining the pollution characteristic amount from the y-intercept and inclination, and the degree discriminating means refers to a table that correlates the previously obtained contamination characteristic amount and the degree of pollution, and refers to the contamination characteristic amount. From Prints defacement degree determination apparatus according to claim 1, characterized in that it comprises a means for obtaining a degree of serial fouling. 4. The stain feature amount calculation means, means for binarizing the input image to provide a binarized image, and a periphery of a pixel having specific binarized data in the binarized image. A unit for determining a length, a unit for determining an area of a pixel having specific binarized data in the binarized image, and a unit for calculating the stain feature amount from the perimeter and the area, 2. The determination means includes means for obtaining the degree of the pollution from the pollution characteristic amount by referring to a table in which the pollution characteristic amount and the degree of the pollution obtained in advance are associated with each other. Deterioration degree determination device for printed matter. 5. An apparatus for determining stains on a printed matter, means for inputting image data of the printed matter, means for dividing the image data input by the input means into a plurality of small areas, and each of the small areas. Means for calculating a pollution feature amount showing the characteristic of pollution, based on the pollution feature amount for each of the small areas input from the calculating means, means for determining the degree of the pollution for each of the small areas, A means for obtaining the degree of stain on the entire printed matter by averaging the degree of stain on each of the small areas; and a means for outputting a small amount area having a large degree of stain on each of the small areas. An apparatus for determining the degree of stains on printed matter, characterized by. 6. The stain feature amount calculation means selects a plurality of pixel scales to be analyzed according to the size of the divided small areas, and the stain feature quantity calculation means selects a plurality of pixel scales to be analyzed based on each pixel scale. A means for dividing an image into a plurality of microscopic areas; a means for calculating the difference between the maximum density value and the minimum density value of pixels in each of the microscopic areas; and a calculation result using the pixel scale; Means for providing a total result for each of the image scales obtained by summing the calculation results obtained for the region, based on the data respectively represented by a pair of the pixel scale and the corresponding total result A means for obtaining a graph, a means for obtaining an approximate straight line from the graph, and a means for obtaining a y-intercept and a slope of the approximate straight line, and a means for calculating a stain feature amount from the y-intercept and the slope. And the degree determining means includes means for obtaining the degree of contamination from the contamination characteristic amount with reference to a table in which the previously obtained contamination characteristic amount and the degree of contamination are associated with each other. The printed matter stain degree determination device according to claim 5. 7. The stain characteristic amount calculation means includes means for binarizing the input image, and means for selecting a plurality of pixel scales to be analyzed according to the size of the divided small area. A means for dividing the image in the small area into a plurality of minute areas based on the pixel scale; a means for searching for specific binary data in each of the small areas; and a minute area having the specific binary data. Means for obtaining the number of specific microscopic regions indicating the number of regions for each pixel scale, means for obtaining a graph based on data represented by each pair of the pixel scale and the number of specific microscopic regions corresponding thereto, and the graph And a means for obtaining a y-intercept and a slope of the approximate straight line, and a means for obtaining the stain characteristic amount from the y-intercept and the slope. 6. The printed matter pollution degree determining apparatus according to claim 5, further comprising: a unit that obtains the degree of pollution from the contamination characteristic amount with reference to a table that associates the contamination characteristic amount and the degree of contamination. . 8. The stain feature amount calculation means, means for binarizing the input image to provide a binarized image, and a periphery of a pixel having specific binarized data in the binarized image. A degree determining means, a means for obtaining an area of a pixel having specific binarized data in the binarized image, and a means for computing the stain feature amount from the perimeter and the area. 6. The means comprises: means for obtaining the degree of the pollution from the pollution characteristic amount by referring to a table in which the pollution characteristic amount and the degree of the pollution obtained in advance are associated with each other. Deterioration degree detector for printed matter. 9. An apparatus for determining stains on a printed matter, means for calculating a stain characteristic amount indicating a characteristic of stains from image data, and means for inputting first image data for a printed matter on which the stain is determined. From the first image data input by the input means, a means for obtaining a first stain characteristic amount using the calculating means, and a second image data using the calculating means from the second image data of the stain-free printed matter. The means for obtaining the pollution feature amount, the means for calculating the pollution feature amount by using the first pollution feature amount and the second pollution feature amount, and the pre-calculated pollution feature amount and the degree of pollution are associated with each other. An apparatus for determining the degree of stains on a printed matter, comprising: means for obtaining the degree of stains from the stain feature amount with reference to a table. 10. The stain feature amount calculation means selects a plurality of pixel scales to be analyzed according to the size of the input image, and a plurality of small areas of the image based on the pixel scales. And a difference between the maximum density value and the minimum density value of the pixels in each of the small areas,
And means for performing an operation using the pixel scale to provide an operation result, means for providing the total result by summing the operation results obtained for each of the small areas, and a pair of the pixel scale and the total result. Means for obtaining a graph based on the respective data, means for obtaining an approximate straight line from the graph, means for obtaining a y-intercept and slope of the approximate straight line, and means for calculating a stain feature amount from the y-intercept and slope When,
The printed matter stain degree determination device according to claim 9, further comprising: 11. The means for obtaining the stain feature quantity comprises means for binarizing the input image, means for selecting a plurality of pixel scales to be analyzed according to the size of the image, and the pixel Means for dividing the image into a plurality of small areas based on a scale; means for searching for specific binary data in each of the small areas; and a specific small area indicating the number of small areas having the specific binary data. A means for obtaining the number of quantitative ranges, a means for obtaining a graph based on data represented by each pair of the pixel scale and the number of specific small areas, a means for obtaining an approximate straight line from the graph, a y-intercept of the approximate straight line 10. The printed matter pollution degree determination device according to claim 9, further comprising: a means for obtaining a slope and a means for obtaining the stain characteristic amount from the y-intercept and the slope. 12. The stain characteristic amount calculation means, means for binarizing the input image to provide a binarized image, and a periphery of a pixel having specific binarized data in the binarized image. A unit for determining a length, a unit for determining an area of a pixel having specific binary data in the binary image, and a unit for calculating the stain feature amount from the perimeter and the area. The printed matter contamination degree determination device according to claim 9.