JP2000247056A - Method for detecting incorrect collation when gathering sagnature by gathering machine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To effectively and suitably detect incorrect collation when gathering by imaging a sheet surface of a signature when gathered by a gathering machine by a CCD camera, and image processing information including gray level densities based on the image as it is. SOLUTION: The method for detecting incorrect collation comprises the step of detecting incorrect collating or the like of a signature 3 by using a gathering machine 1 in the case of stacking and gathering the signature 3 in order of pages. In this case, after at least part of a printed surface of the signature to be as a reference is imaged by a CCD camera 6 including gray level densities, its image is recorded as a reference image data group as it is. Meanwhile, after at least part of the printed surface of the signature to be inspected as to the presence or absence of the incorrect collating to the reference image data group is imaged by the camera including the gray level densities, the image is recorded as the inspected image data group as it is, the inspected image data group and the reference data group are image processed to detect the incorrect collating of the signature.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for detecting misalignment when a signature is collated by a collating machine, and more particularly, to at least one printed material printed on the paper surface of the collated signature in the collating machine. The present invention relates to a detection method capable of reliably and appropriately detecting misaligned pages at the time of collating, by photographing a portion with a CCD camera and processing the information including the density of the printed material brought by the captured image as it is.

[0002] This method can be applied to a collating machine that sends signatures by a feeder generally called a vise. In addition to this collating machine, a collating machine that sends signatures by a feeder called a rotary system is used. Can also be applied.

[0003]

2. Description of the Related Art A printed paper is folded, and this is called a signature. Bookbinding of books and magazines is performed by stacking the signatures in page order and collecting them into one book. This work is called collation because the signatures are overlaid, and is usually performed by a collator.

A collating machine used for collating a signature is provided with a collating line. While the signatures are sent from the upstream to the downstream along the collating line, the signatures are stacked in page order. One book or magazine is collected and bound.

That is, a plurality of signature laminating zones are provided along a collating line, signatures are sent from each laminating zone to the collating line, and are laminated in the page order on the collating line. The stacking zone is usually composed of a frame, and signatures corresponding to the same page are accommodated in the frame in a stacked state, and the signatures at the bottom are taken out one by one and sent to a collation line.

[0006] The lowermost signatures from the frame of each lamination zone are taken out by a feeder usually called a vise,
The lower surface of the signature at the bottom is sucked by a vice and pulled out to the guide guide, and the signature is sent from the guide guide to the collation line by a feed piece interlocking with the collation line.

In addition to the method of taking out signatures one by one, a rotary feeder is also used. The rotary feeder is configured to remove the signature from the frame of the laminating zone and feed the signature to the collating line by gripping and rotating the signature with a rotary piece.

[0008] When binding a book or the like using such a collating machine, troubles such as misaligned books cannot be avoided, and if such an accident is overlooked even in a single book, the book is bound in the same lot. It seems that the same accident has also occurred in the book that has been collected, so that work such as collection and disposal is common, losing the credibility of the bookbinding company and causing great damage.

For this reason, usually, a collating machine is provided with a detection device for a pamphlet or the like, thereby detecting the pamphlet or the like.
Collation is performed without disorder. However, with such a detection device and detection method, inaccurately-sorted pages cannot be properly detected, and the detection information itself, such as inaccurate pages, is often a false alarm. There is a need for a drastic reduction, and rather an improvement in productivity.

[0010] In view of the conventional example of a detection device and a monitoring device for mis-ordered pamphlets, Japanese Patent Publication No. 5-78440 and Japanese Patent Publication No. 3-044000 each disclose a patrol monitoring device.

These apparatuses irradiate a measuring point on a printed surface of a signature with light, receive reflected light from the measuring point by a light receiving element such as a photodiode, and receive the reflected light as an electric signal corresponding to the amount of light. , And the difference between the two electric signals is calculated as compared with the reference electric signal for each measurement point in which the electric signal for each measurement point is stored in advance. This difference (absolute value)
Is a device that determines whether or not a page is incorrect based on the size of the page.

However, these devices apply light to the printed surface of the signature only at the measurement point, and convert reflected light at the measurement point into an electric signal and compare it with a previously recorded reference signal. Therefore, the judgment of the presence or absence of the incorrect page is not properly performed.

That is, since the light receiving element is a semiconductor element such as a solar cell or a photodiode, the light energy is converted into electric energy, which is merely recorded on the memory cell as a digital signal. .

However, the information recorded in the memory cell or the like is obtained from the reflected light from the measurement point on the printed surface of the signature, and the information relating to the shape and shading of the image is not sent to the memory cell and is not recorded. Not. For this reason, the image read from the memory cell is significantly different from the image on the printed surface of the signature, and thus it is not possible to determine the incorrect page properly and reliably.

Japanese Patent Publication No. 61-46384 discloses an apparatus for monitoring irregularities, which includes a level failure detector, a failure rate detector, and a non-defective judgment circuit, such as an optical sensor and an AD converter.

In this apparatus, the optical sensor is mounted on a stack of a stacking zone of signatures provided along the side of the collation line, and when the signature is pulled out to the collation line, the direction of the signature is determined. Plural pieces are arranged in a direction orthogonal to each other, or in a direction parallel to the collation line. Therefore, when the signature is pulled out from the lamination zone, the signature printing surface is measured at several places along a straight line perpendicular to the drawing direction, and the output of the reflected light at these several places is converted into a digital signal by an AD converter. Of these digital signals, a reference signal as a reference is recorded in a standard memory, while a detection signal from a signature for detecting a broken page is recorded in each monitoring memory.

In this apparatus, the incorrect page compares the test signal stored in each monitor memory with the reference signal value previously recorded in the standard memory and finds the difference. Even one of the differences falls outside a predetermined level. When a failure occurs, a failure signal is transmitted.

However, this device also measures only on a straight line perpendicular to the direction in which the signature is drawn, and even on the straight line, it measures several measurement points. Not given at all.

Further, since the number of measurement points is limited to a few, deviation between the reference image and the inspection image is likely to occur, and in order to prevent this, the number of optical sensors must be considerably increased. However, it is not preferable because the size becomes large.

[0020]

SUMMARY OF THE INVENTION It is an object of the present invention to solve the above-mentioned drawbacks. More specifically, a conventional apparatus and method for detecting or monitoring a mis-matched document is an image of a printed body printed on the paper of a signature. It is an object of the present invention to solve the drawback that the page cannot be detected reliably because the page is determined based on the voltage value based on the reflected light at the local measurement point without using the shading or the shape as it is.

[0021]

That is, according to the present invention, in a method of detecting a signature misalignment of signatures when stacking and collating the signatures in a page order using a collating machine, a signature to be used as a reference is used. After capturing at least a part of the printing surface including the shade with a CCD camera, the captured image is recorded as it is as a reference image data group, and a fold for inspecting the reference image data group for the presence or absence of a page mismatch. After photographing at least a part of the printing surface of the book including the shade using a CCD camera, the image is recorded as it is as an inspection image data group, and the inspection image data group and the reference image data group are image-processed and folded. The present invention is characterized in that the irregularities of the pages are detected.

Further, in such an image theory, a correlation value between the reference image data group and the inspection image data group is calculated, and a signature signature is detected based on the correlation value.

The configuration and operation of these means will be described more specifically with reference to the drawings.

FIG. 1 is a layout diagram of an example of an apparatus for implementing the method of the present invention.

FIG. 2 is a waveform diagram of an example of an electric signal transmitted by the CCD camera by the apparatus shown in FIG.

FIG. 3 is an explanatory diagram of an example of a grayscale signal obtained by processing the electric signal having the waveform shown in FIG.

FIG. 4 is an explanatory diagram showing an example of digitized signals shown in FIG.

First, in bookbinding of books, magazines, and the like, printed and folded paper is called a signature. Collating is to gather the signatures in a page order by stacking the signatures, and usually collation is performed continuously by a collating machine 1 as shown in FIG.

In the case of using a collating machine, reference numeral 2 in FIG.
Is performed along the collation line shown as
A signature 3 is stacked for each page along. From the lamination zone (not shown) where the signatures are stacked, the guide 4
The signature 3 on the guide 4 is pulled out and the feed piece 5
And collated in page order as shown below. That is, while the signatures 3 flow from the upstream to the downstream of the collation line 2, the signatures are collated and collected in page order, and one book or magazine is bound.

More specifically, the number of signature lamination zones provided along the collation line is usually 18 to 20 (this number is determined according to the number of signature pages to be collated). Each lamination zone is usually constituted by a frame, and signatures corresponding to the same page are accommodated in the frame in a laminated state. A signature is sent from the frame body of each such lamination zone to a collating line, and is sent by a feed piece 5 moving at the same speed in the collating line 2,
While the signatures flow from the upstream to the downstream, the signatures are stacked, and after passing through the last stacking zone, the signatures are stacked in page order to form a book or magazine.

The supply of the signatures 3 from each lamination zone to the guide 4 is usually performed by a feeder (not shown) called a vise, and the signatures are released from the bottom of the frame of each lamination zone by the vise. Each sheet is sucked and pulled out one by one and sent to the collating line 2 via the guide 4. In addition to vise, there are feeders that can take out signatures one by one. An example of this is a rotary feeder. In this case, the signature is gripped by the rotary piece, and the signature is taken out of the frame by rotating the rotary piece, and sent to the collating line 2.

As described above, when the signatures are collated by using the collating machine 1 and bound into one book or the like, the characters, figures, and patterns printed on the paper surface of the signatures sent from the respective laminating zones. A part of the printed body is photographed by the CCD camera 6.
The photographing by the CCD camera 6 is as shown in FIG.
It can be performed on the guide 4 but can also be performed at other places, and can be performed by photographing a part of the printed body of the signature when it is pulled out from the bottom of the frame. .

The image of the CCD camera 6 is transmitted to the image processing device 7 including the shading, and the image processing device 7 sets a reference image data group (hereinafter referred to as reference image data) and an image to be inspected. Image processing is performed as a data (hereinafter referred to as inspection image data) group, and a correlation between these two data groups, that is, a correlation value is calculated. Page is wrong).

The above description is more specifically described as follows.

In starting the collating of the signature by the collating machine, a part of the printed material printed on the paper surface of the signature is irradiated by a light source such as a strobe and photographed by a CCD camera.

That is, the CCD camera 6 is provided with a large number of photodiodes, especially an infinite number of photodiodes.
When light is applied to the array surface, the light received by each photodiode generates a charge whose magnitude is proportional to the brightness (that is, the density of the printed material), and is transmitted by the CCD element in a certain order. In the image processing apparatus 7, the continuous electric signal as shown in FIG. 2 is quantized as shown in FIG.
2, 123 are digitized as shown in FIG. 4, and the photographed image of the printed body is recorded as an image data group while including the shading. By reading the digitized charge distribution, it can be converted into an image, and the converted image is displayed on the display 8.
Will be displayed.

Therefore, the image data sent and recorded by the photographing of the CCD camera 6 is a data group of the image as it is including the density of the print.

The image data group photographed by the CCD camera 6 and recorded in the image processing device 7 includes a printed material on a reference signature which is used as a reference at the time of inspection for the presence or absence of a page, and an inspection compared with the reference signature. Ask about the inspection signature to be done. In the two kinds of image data groups obtained in this way, the correlation between the former reference image data group and the latter inspection image data group is quantified as a relative value, and thereby, the presence or absence of a misalignment is determined.

The data group of the reference image and the inspection image for which the correlation value is calculated by the image processing device 7 is a photograph or image of a printed material on the signature paper surface, and is represented as it is. This is a value that is obtained by digitally converting a charge having a proportional magnitude. When the charge (voltage) shown in FIG. 2 is sent as a print as it is from the CCD camera, it is sampled by an 8-bit 256-layer identification pattern and then converted to a digital display. That is, as shown in FIG. 4, the shades of the printed material are recorded in the memory cells of the image processing apparatus as digital signals of 0 and 1 as they are as signals of 256 layers.

When the data group of the reference image and the inspection image composed of 256 layers is compared and calculated, the image of each printed body of the signature to be inspected and the signature to be inspected are compared as they are including the shading. In the case of collation, misaligned pages can be detected with high accuracy. Note that, when an incorrect page is detected, the entire collation line is stopped, and when the incorrect page is corrected, the collation line is operated again.

That is, when two image data groups are compared by the image processing apparatus in the present invention, the absolute value of each data, that is, the magnitude of the electric charge generated according to the density of the printed material is not compared. , And the correlation between them is calculated, and the irregularity is detected based on the correlation value.

The comparison of the two data groups between the inspection image and the reference image according to the present invention is performed by comparing the data without comparing the absolute values of the data of 256 layers.
The comparison is made including the density of the printed material. In this case, the relative value is calculated by normalizing the image data group without calculating the relative value using a digitized value of the image data group, and calculating the correlation value based on the normalized value. Therefore, the correlation value obtained as a result of the calculation indicates the mutual relationship between the figures drawn by the data group.

When looking at the mutual relationship between the two image data groups, the image data group photographed by the CCD camera 6 is divided by a light and dark level defined by a predetermined standard. 0], it is possible to classify it into [1] on the assumption that there is a charge above the reference gray level.
According to this method, even if an enormous amount of image data is
Since the values are divided into three values, the data amount can be reduced, and the calculation of the correlation value can be performed at high speed.

However, if the method of dividing into two is adopted, the characteristics of the image data group photographed by the CCD camera are lost, and the image which can be read from the divided image data group is the actual photographed image. Will be far away from

In particular, when the printed body of a signature is a photograph, a painting, or the like, a completely separate image is obtained, and even if a correlation value is obtained, the incorrect signature cannot be detected properly.

On the other hand, since the image data group obtained by photographing with the CCD camera contains data indicating shading, it is necessary to calculate the correlation, that is, the correlation value using the data to determine the presence or absence of the incorrect page. preferable. However, when the correlation between the data groups including the shading is calculated, the shading depends on the properties of the paper surface of the signature and the illuminance of the light source at the time of photographing. Therefore, it is preferable to eliminate the influence of these factors. For this reason, it is preferable to normalize the image data so that the two data to be compared have the same change width, and calculate the correlation value based on the normalized data.

That is, in the two data groups of the reference image and the inspection image, a normalization operation is performed to match the change widths of these data. A correlation operation is performed between the two normalized image data to judge whether or not they match.

When the correlation is calculated after normalization as described above, the correlation is calculated by shifting one pixel in the inspection image for checking whether there is a misalignment, and repeatedly from one end of the inspection image to the other.

The correlation value serving as a reference for determining whether the reference image matches the inspection image is obtained as follows.

More specifically, the two image data groups displayed as digital signals of 256 layers (an example is shown in FIG. 4) have the same change width between them, and the data group of the digital display is further changed. Then, the correlation value is calculated. In the calculation of the correlation value, the correlation itself is a property in which the data in the two data groups change at the same time while showing a considerable degree of correlation, and any relational expression showing this property can be used. If this relational expression is stored in the image processing apparatus, it can be easily calculated.

As an example, it is possible to statistically process between two image data groups, obtain a correlation coefficient thereof, and determine the presence or absence of a page mismatch based on this value. In other words, the closer the correlation coefficient is to 1, the higher the correlation, and if the correlation coefficient is 1, the data is consistent. In such a case, it can be determined that there is no misbooking. On the contrary, as the correlation coefficient becomes smaller than 1 and becomes closer to 0, the correlation is lost.

Instead of the statistically familiar correlation coefficient, the correlation operation formula (1) is used in consideration of the autocorrelation and the cross-correlation of each image data of the inspection image and the reference image.
Is set, and this arithmetic expression (1) is input to the image processing apparatus, and the correlation value can be obtained by this expression. That is, since the correlation value according to the equation (1) is expressed as a percentage, it is possible to determine the presence or absence of the incorrect page at the percentage. For example, at 65% or more, a shift or the like occurs between two images, and when this is resolved, it approaches 100%, and in fact, they completely match. Therefore, at this time, it can be determined that there is no incorrect page. When it is 65% or less, the correlation between the two image data is considerably lost, and the correlation ratio is a very low value. At this time, it is determined that there is an incorrect page even if the image is shifted. With such a determination, in an actual collating operation, a misaligned page can be detected with an accuracy close to almost 100%. Correlation value (%) = {(positive / negative of A) × A ² } (B × C)｝ × 100 (1) where A is inspected based on the previously recorded reference image and this reference image The correlation between data between images is shown, and N, M, and I are as described later. A = NΣ (I × M) − (ΣI) × (ΣM) B represents an autocorrelation in the data of the inspection image input for the inspection, and N and I are as described later. B = NΣ (I × I) − (ΣI) × (ΣI) C represents the self-correlation in the data of the previously recorded reference image, and N and M are as described later. C = NΣ (M × M)-(ΣM) × (ΣM) Here, N: the number of data included in each image data group, and also the area occupied by image data. M: the density of each data in the data group of the previously recorded reference image;
It is expressed as the density of 56 hierarchical levels. I: The density of each data of the data group of the inspection image to be inspected as compared with the reference image.

The correlation value calculated by the correlation calculation expression (1) is a value obtained by performing a correlation calculation after adjusting the variation width by normalization between two image data groups. Only the relationship of the form

Further, the correlation value obtained by the correlation operation expression (1) is a degree of the mutual relationship shown between the two image data. Therefore, the larger the correlation value is, the more similar the image data group is, and the meaning is different from the absolute value such as the dimension value, but whether the value is large or small is a problem.

That is, when the correlation value is calculated between the two image data groups by the correlation calculation formula (1), since the calculation is a normalization calculation, the brightness at the time of photographing between the two data groups of the reference image and the inspection image is calculated. Even if there is a change such as the properties of the paper of the signature, the shapes of the figures drawn by the data group are compared without being influenced by those influences, and the correlation of the shapes of the figures is obtained.

To put it plainly, even if the brightness at the time of photographing changes or the paper condition of the signature changes, the image data groups can be compared with the same brightness and the same paper condition.

Further, as described above, when judging the presence or absence of a page misalignment or the like by the correlation operation, the following can be easily performed.

At the time of collation, the presence / absence of an incorrect page is detected, and collation is performed so that there is no incorrect page. At this time, as already explained,
First, a signature as a reference is determined, and a part of a printed body of the signature is photographed by a CCD camera as a reference image.
The image data is input to the image processing apparatus as reference image data, and is recorded as a digital signal of 256 layers or the like.

Next, the signatures to be collated are sent to the collating line (see FIG. 1) in the order of pages, and the printed body of each signature sent is similarly photographed by a CCD camera and sent to an image processing apparatus. . In the image processing apparatus, 2
In order to compare with the 56-layer digitally displayed reference image data group, it is recorded in a memory or a memory cell as a 256-layer digitally displayed inspection image data group.

In the detection and monitoring of the irregularity, the detected image data group recorded in the memory cell or the like is shifted by one pixel in the memory cell, and the correlation value between the previously recorded reference image data group and the previously recorded reference image data group is calculated by a correlation operation expression. The calculation is repeatedly performed from the end of the memory cell in which the inspection image data group is recorded. In this way, the correlation calculation is performed on the entire inspection image data group by shifting one pixel at a time, and the position where the correlation value becomes maximum is determined. This is where the data of the reference image data group matches the data of the inspection image data group. In this way, in the image processing apparatus, a portion where the reference image previously recorded coincides with the memory cell where the inspection image data group sent from the CCD camera is recorded is detected.

The correlation calculation process is not a search for checking the degree of coincidence between two image data groups by one operation, but is a search for searching for the most coincident place by repeating the search. Therefore, the incorrect page can be detected without error.

[0062]

As described above, the present invention relates to a method for detecting a broken page or the like when stacking and collating signatures in a page order using a collating machine. At least a part of the image, including shading, is photographed by a CCD camera, and the photographed image is recorded as it is as a reference image data group, while the signature printing surface for inspecting whether there is a misalignment with the reference image data group. After capturing at least a part of the image including the shading with a CCD camera, the image is recorded as an inspection image data group as it is, and the inspection image data group and the reference image data group are image-processed to check the signature of the signature. It is to detect.

Further, in this image theory, a correlation value between the reference image data group and the inspection image data group is calculated, and a broken signature is detected based on the correlation value.

Therefore, according to the present invention, since the image data is sent by the CCD camera, it includes the shading, and the incorrect page can be detected properly and reliably.

Further, since the comparison of the image data is performed by correlation without taking the absolute value, even if there is a slight shift of the printing surface, the misaligned page is not erroneously detected.

Further, in comparing the image data, the correlation can be determined by normalizing the two image data, so that the printed material such as characters and figures will see the figure drawn by the image data, and the detection accuracy of the incorrect page will be high. Extremely high.

[Brief description of the drawings]

FIG. 1 is a layout view of an example of an apparatus for performing the method of the present invention.

FIG. 2 is a waveform diagram of an example of an electric signal transmitted by a CCD camera by the device shown in FIG. 1;

FIG. 3 is an explanatory diagram of an example of a grayscale signal obtained by processing the electric signal having the waveform shown in FIG. 2;

FIG. 4 is an explanatory diagram showing an example in which the signal shown in FIG. 3 is digitized.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Collating machine 2 Collating line 3 Signature 4 Guidance guide 5 Feeding piece 6 CCD camera 7 Image processing device 8 Display

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] June 24, 1999 (1999.6.2
4)

[Procedure amendment 1]

[Document name to be amended] Drawing

[Correction target item name] All figures

[Correction method] Change

[Correction contents]

FIG.

FIG. 2

FIG. 3

FIG. 4

Claims (6)

[Claims] In a method of detecting signature misalignment or the like when stacking signatures in page order using a collating machine, at least a part of a printed surface of the signature to be a reference is shaded. And then record the photographed image as it is as a reference image data group, and print at least a part of the signature printing surface for checking whether there is a misalignment with the reference image data group. After photographing with the CCD camera including the shading, the image is recorded as it is as an inspection image data group, and the inspection image data group and the reference image data group are subjected to image processing to detect a broken signature. A method for detecting incorrect signatures when a signature is collated by a collator. 2. calculating a correlation value between the reference image data group and the inspection image data group in the image theory;
2. The method according to claim 1, wherein the signature is detected based on the correlation value. 3. The correlation value is a value obtained by calculating based on a relational expression indicating that the reference image data group and the inspection image data group show a considerable degree of mutual relation and change at the same time. 3. The method according to claim 1, wherein
A method for detecting misalignment when a signature is collated by the collating machine described. 4. The relation defined by the mutual relation between the reference image data group and the inspection image data group, the autocorrelation in the inspection image data group, and the autocorrelation in the reference image data group. 4. A method according to claim 3, wherein the signature is detected when the signatures are collated. 5. The falsification of signatures by a collating machine according to claim 1, wherein the reference image data group and the inspection image data group are calculated and normalized in the surface image processing. Detection method. 6. The method according to claim 5, wherein the normalization is performed so as to match a variation width of a data value between the reference image data group and the inspection image data group.
A method for detecting misalignment when a signature is collated by the collating machine described.