JP2000079776A - Incorrect detection method - Google Patents

Abstract

(57) [Summary] [PROBLEMS] To reliably detect an incorrect page even when the print density or paper color differs between a printed matter in which an inspection image is captured and a printed matter in which a reference image is captured. SOLUTION: When detecting a misaligned page based on a reference image A inputted from a normal printed matter in advance and an inspection image E inputted from the printed matter in a collating step, a difference image between the reference image A and the inspection image E is obtained. G is created, the difference image is masked with a mask image D created based on the edge image B obtained from the reference image A to create a difference mask image H, and edges are extracted from the difference mask image. Difference mask edge image I
Is created, and the difference mask edge image is binarized.
The incorrect page is determined based on the digitized image J.

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 a mismatched page, which is suitable for detecting a page mismatch occurring in a collating process in a bookbinding process.

[0002]

2. Description of the Related Art Generally, when performing a bookbinding process with a bookbinding machine,
A bundle of printed materials called signatures, which are individually printed in advance by a printing machine, are sequentially superimposed to form one printed material, and then the bundle is bundled to produce one book.

[0003] In this bookbinding process, it is extremely important to overlap each signature in the correct order.
If the order is out of order, it is necessary to reliably detect and deal with it.

[0004] As a conventional misalignment detection technique, for example, Japanese Patent Publication No. 3-44000 discloses a method in which light is applied to a plurality of measurement points on a reference sheet and an inspection sheet, and reflected light from each measurement point corresponds to each measurement point. The difference between the amount of reflected light from the inspection paper surface and the amount of reflected light from the reference paper surface is obtained for each measurement point, and the arithmetic average of the amount of reflection light from the inspection paper surface for each two adjacent measurement points is obtained. By calculating the arithmetic average of the amount of reflected light on the reference paper, taking the difference between the two arithmetic averages, and comparing the number of differences exceeding a predetermined level among the two types of differences with the total number of the two types of differences. A method for detecting a mis-order is disclosed.

[0005] Also, Japanese Patent Application Laid-Open No. 7-257065 discloses that
The CCD camera captures the reference image and each inspection image, registers a partial area of the reference image as a reference pattern, and determines whether the signature is identical or non-identical depending on whether or not the same pattern as the reference pattern exists in each inspection image. There is disclosed a technique for detecting an incorrect page by judging the number of pages. According to this technique, when a solid portion of one color exists in a wide range in the print content, an appropriate region does not exist as a reference pattern. By comparing, it is determined whether the signatures are identical or non-identical, and the incorrect signature is detected.

[0006]

However, the technology disclosed in Japanese Patent Publication No. 3-44000 discloses the following.
Since the change in density on the printed paper surface is detected, erroneous detection occurs due to a change in print density or a change in the amount of reflected light due to misregistration of the printed book, even though collation is correct. There is a problem that can not be. Further, if the detection performance (sensitivity) is reduced to a level that does not cause erroneous detection in order to avoid this, there is a problem that it is not possible to detect incorrect pages depending on the item.

In the technique disclosed in Japanese Patent Application Laid-Open No. 7-2570065, it is determined whether a signature is identical or non-identical based on whether or not the same pattern as a reference pattern exists in each inspection image. Although mischief is detected, this method has the following two problems. The first point is that, even though the same pattern exists in the reference image and the inspection image, if the print density and the paper color are different between the two, it cannot be determined that the patterns are the same. False detection may occur. The second point is that, although the same pattern does not exist in the reference image and the inspection image, another pattern having a similar pattern, such as a character pattern in a telephone directory, is recognized as the same pattern. In some cases, incorrect pages cannot be detected.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned conventional problems, and even if the print density or the paper color is different between the reference image and the inspection image, the incorrect page is detected. It is an object of the present invention to provide a method of detecting an incorrect page.

[0009]

According to the present invention, there is provided a method for detecting a mismatched page based on a reference image previously inputted from a normal printed matter and an inspection image inputted from the printed matter in a collating step. This problem has been solved by creating a difference image between a reference image and the inspection image, and determining the incorrect page based on the difference image.

That is, in the present invention, the misalignment is determined based on the difference image between the reference image and the inspection image. Therefore, for example, whether or not the same printed matter (signature) is determined based on the edge information of the difference image Can be reliably determined.

[0011]

Embodiments of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing a schematic configuration of a misalignment detection apparatus applied to a misalignment detection method according to an embodiment of the present invention.

This apparatus for detecting a mismatched page includes a camera 10 for capturing a two-dimensional image from a signature (not shown), and an image processing unit 12 for performing various image processing on the image captured by the camera 10. The image processing device 12
Image memory 14, misregistration correction unit 16, edge extraction unit 1
Each of the functional units includes an binarization processing unit 20, an expansion processing unit 22, a mask processing unit 24, a difference processing unit 26, and a determination processing unit 28.

In the present embodiment, the camera 10
A difference image is created between a reference image input from a normal signature (printed matter) in advance and an inspection image input from the same type of signature in the collation process, and it is determined whether or not the signature is incorrect based on the difference image. I do.

That is, in this embodiment, the difference image between the reference image and the inspection image, and the edge (contour line) of the pattern from the reference image
Is determined and detected by using the edge image created by extracting.

In the following, the method for detecting incorrect pages according to the present embodiment will be described.
This will be specifically described with reference to FIG.

First, a picture of a normal signature is captured as a two-dimensional image on a bookbinding machine (not shown) by the camera 10 and stored in the image memory 14 as a reference image A.
The reference image A is read from the image memory 14 to the edge extraction unit 18, and an edge extraction operator such as Laplacian or Sobel is applied to the reference image A to extract a picture edge included in the reference image A. Then, an edge image B is created.

The edge image B shown in FIG. 2 is unclear due to the input from the scanner, but a thin line edge exists in the actual figure. The same applies to a difference image G to be described later and an edge image B 'in FIG. The directions of the edges to be extracted here are desirably all directions, but edges only horizontal or only vertical may be extracted depending on the image quality determined by the input system. Further, as a pre-process of edge extraction, a noise removal process using a smoothing filter, a median filter, or the like may be applied.

Next, the edge image B is binarized in the binarization processing section 20 using means such as a modal method or a discriminant analysis method in order to clarify an edge. An edge image C is stored in the image memory 14.
To be stored. Further, the expansion processing section 22 performs an expansion operation on the binary edge image C to obtain a binary edge expanded image (mask image) D, which is also stored in the image memory 14. The above operations up to the creation of the mask image D are performed before inputting the inspection image.

Next, at the start of the bookbinding work, the camera 10 inputs an image from a signature on the bookbinding machine, takes in the inspection image E, and stores it in the image memory 14. Usually, since the reference image A and the inspection image B are misaligned due to misregistration of the printing book, the misregistration correction unit 16 calculates the misalignment amount by a method such as template pattern matching. Then, a position-corrected image F obtained by correcting the position corresponding to the shift amount is stored in the image memory 14 again.

Next, the difference processing section 26 calculates a difference between pixel values for each corresponding pixel between the reference image A and the position correction image (inspection image) F, and uses the difference value as a pixel value. A difference image G is created. The difference image G also includes a triangular picture edge caused by a displacement of one pixel or less that cannot be handled by the displacement correction processing performed by the displacement correction unit 16.

Therefore, in the mask processing section 24,
In order to remove the edge derived from the edge of the pattern included in the difference image G, the 2
A difference mask image H is created by performing a mask process using the value edge dilated image (edge mask image) D. Here, the mask processing is performed by reversing the edge mask image D and the difference image G
And that the AND processing is performed between the corresponding pixels. Therefore, the difference mask image H is white only when the corresponding pixels in both images are both white, and all other pixels are black.

Next, the edge extraction unit 18 extracts the edges of the masked difference mask image H again to obtain a difference mask edge image I, and the difference mask edge image I is processed by the binarization processing unit 20 again. The image is binarized by the method described above to obtain a differential mask edge binary image J.
For the value image J, the determination processing unit 28 counts the number of white pixels corresponding to the edge, and determines whether or not the value image J is incorrect based on the number.

In principle, if the page is correct, the number of white pixels corresponding to the edge is 0. Therefore, if there is at least one page, it means that the page is incorrect. If the number of pixels is 5% or more of the total number of pixels, the page is determined to be incorrect. In order to avoid extracting edges caused by noise, a noise removal operation using a smoothing filter, a median filter, or the like may be applied at a stage prior to edge extraction to remove noise components in advance.

According to the present embodiment described above, even if the inspection image E has a missing portion such as a circle pattern existing in the reference image A, the rectangular image existing in the inspection image E Even if there is an increased portion of the reference image A like a picture, any of them can be reliably detected as a wrong page. By the way, in the case of the right page, the whole after the difference mask image H is substantially a black image.

In this embodiment, since the positional deviation of the inspection image is corrected with respect to the reference image, even if the pattern to be inspected is misaligned due to misregistration of the printing book, etc., the misaligned page can be accurately detected. Can be detected.

When the edge mask image D is created from the reference image A as in the present embodiment, since the reference image A can be created from a plurality of captured images in advance, the edge mask image with high quality As a result, D can be created, and the detection accuracy of the incorrect page can be improved.

Next, a second embodiment according to the present invention will be described with reference to FIG. 3 corresponding to FIG.

This embodiment is substantially the same as the first embodiment except that a mask image for masking a difference image is created based on an inspection image. Therefore, the same apparatus as that in the first embodiment can be used as the incorrect page detecting apparatus.

In FIG. 3, the reference image A, the inspection image E,
The position correction image F and the difference image G are the same as those in the first embodiment, so the same reference numerals are used, and the other images having the same name shown in FIG. Symbols are shown with dashes (').

In this embodiment, the inspection image E after the position correction
The edge extraction unit 18 based on the position correction image F
In the binarization processing section 20 and the expansion processing section 22, the edge image B ', the binary edge image C', and the binary edge expansion image (edge mask image) D 'are sequentially formed in the same manner as in the first embodiment. At the same time, the mask processing section 24 masks the difference image G using the binary edge dilated image D 'to create a difference mask image H'.

Thereafter, based on the difference mask image H ',
As in the case of the first embodiment, a difference mask edge image I 'and a difference mask edge binary image J' are created, and a misplaced page is determined based on the number of white pixels of the binary image J '.

In the present embodiment, the edge mask image D '
Is generated from one inspection image E (position correction image F), the quality of the edge mask image D ′ is inferior to that of the first embodiment, and in addition, for each inspection, in other words, Since it is necessary to create the mask edge image D 'for each copy, it is disadvantageous as compared with the case where the mask edge image D' is created from the reference image A.

Although the present invention has been specifically described above, the present invention is not limited to the above-described embodiment, and can be variously modified without departing from the gist thereof.

For example, in the above embodiment, the case where the difference image G is masked has been described.

In the above embodiment, the case where the camera 10 is used as an image input means for inputting a two-dimensional image has been described. However, the present invention is not limited to this, and a contact line sensor or the like may be used.

[0037]

As described above, according to the present invention,
Even if the print density or the color of the paper differs between the printed material in which the inspection image has been captured and the printed material in which the reference image has been captured, it is possible to reliably detect incorrect pages.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a schematic configuration of a misaligned page detection device applied to a first embodiment according to the present invention;

FIG. 2 is a flowchart showing the operation of the first embodiment.

FIG. 3 is a flowchart showing the operation of the second embodiment.

[Explanation of symbols]

 10 camera 12 image processing device

Claims (5)

[Claims] 1. A method for detecting a misaligned page based on a reference image previously input from a normal printed material and an inspection image input from a printed material in a collating step, comprising the steps of: A method for detecting a misaligned page, comprising: creating a difference image; and determining a mismatched page based on the difference image. 2. The difference mask image according to claim 1, wherein the difference image is masked with a mask image created based on an edge image obtained from the reference image, and a difference mask image is created based on the difference mask image. Judge. 3. The difference mask image according to claim 1, wherein the difference image is masked with a mask image created based on an edge image obtained from the inspection image, and a difference mask image is created based on the difference mask image. Judge. 4. The method according to claim 2, wherein an edge is extracted from the difference mask image to create a difference mask edge image, and a mismatch is determined based on the difference mask edge image. . 5. The method according to claim 4, wherein the difference mask edge image is binarized, and the page misjudgment is determined based on the binarized image.