JP4141082B2 - Printed matter inspection method and apparatus - Google Patents

Description

[0001]
[Technical field to which the invention belongs]
The present invention relates to a printed matter inspection method and apparatus, and more particularly, to a printed matter inspection method and apparatus suitable for application when detecting point-like defects that occur during color printing.
[0002]
[Prior art]
A printed image is input as an inspection image from a web (printed material) printed on a rotary printing press and continuously running, and the inspection image and a reference image created in advance are input. In comparison, printing defects are inspected.
[0003]
As described above, the printed matter inspection apparatus that detects the print defect by comparing the inspection image with the reference image measures the number of pixels equal to or greater than a predetermined level difference within a predetermined rectangular area with respect to the difference image between these two images. Japanese Patent Application Laid-Open No. 9-76470 discloses that a pixel is determined to be defective when the number of pixels exceeds a predetermined value.
[0004]
[Problems to be solved by the invention]
However, it has been clarified that the inspection technique disclosed in the above publication has the following problems when applied to inspection of point-like defects with clear outlines (edges).
[0005]
In the detection process using a monochrome inspection image obtained by imaging a printed matter using a monochrome (black and white) camera, the ability for C (blue), M (red), and Y (yellow) compared to the detection ability for K (black) Is inferior. Therefore, when trying to increase the inspection capability for point-like defects that are clearly present on the inspection image, it is possible to detect excessively subtle changes in the density of the printed pattern that is allowed to occur, especially in offset printing. As a result, the printed pattern is detected as defective.
[0006]
On the other hand, when using an inspection image obtained by imaging using a color camera, if independent detection processing is performed for each of R, G, and B, minute misregistration between printed colors, chromatic aberration of a camera lens, etc. May cause excessive detection of subtle color fluctuations in the pattern edge, and image processing is performed independently for each of R, G, and B. Three times the circuit scale is required.
[0007]
The present invention has been made to solve the above-described conventional problems, and can easily and reliably inspect a point defect that clearly exists together with a pattern on an inspection image obtained by imaging the printed matter. It is an object to provide an inspection method and apparatus.
[0008]
[Means for Solving the Problems]
The present invention relates to a printed matter inspection method for detecting a print defect by comparing an inspection image input for each of R, G, and B with a reference image prepared in advance by imaging a printed matter with a color camera. The edge of the pattern is extracted from each inspection image of G and B, and each edge image of R, G, and B is created, and the edge of the pattern of the edge image is extracted from each of the reference images of R, G, and B A threshold image that masks a portion is created, and for each of the created R, G, and B threshold images, a maximum value is obtained from the pixel values of the corresponding pixels, and the maximum value is determined as the pixel value of the corresponding pixel. Each of the R, G, B edge images by masking each of the R, G, B edge images with the combined threshold image, Print defects based on masked edge images By leaving, it is obtained by solving the above problems.
[0009]
The present invention also provides a printed matter inspection apparatus for detecting a print defect by imaging a printed matter with a color camera and comparing an inspection image input for each of R, G, and B with a reference image created in advance. Means for extracting an edge of a pattern from each inspection image of R, G, B and creating each edge image of R, G, B; and from each of the R, G, B reference images, For each of the generated threshold images for R, G, and B, each of which creates a threshold image for masking the edge portion of the pattern, the maximum value is obtained from the pixel values of the corresponding pixels, and the maximum value is determined. Means for creating one composite threshold image as the pixel value of the pixel to be masked, and masking each of the R, G, B edge images with the composite threshold image, to obtain each masked edge image of R, G, B The means to create and each masked error created By having a means for detecting a print defect based on the di-image is obtained by solving the above problems as well.
[0010]
That is, in the present invention, in each of the R, G, and B threshold images obtained from the reference image, the maximum value is selected from the pixel values of the corresponding pixels and replaced with the pixel values of the corresponding pixels. Since the synthesis threshold image is created, only the area for masking the edge portion of the pattern remains in the synthesis threshold image, and the subtle color fluctuation portion caused by the misregistration between colors is excluded. Therefore, since the edge extracted from each inspection image of R, G, B obtained by imaging with a color camera is masked using this composite threshold image, each masked edge image has Since only the edges of defective portions that do not exist in the reference image remain, it is possible to easily and reliably detect point-like defects isolated on the printed matter without being affected by subtle color fluctuations. It becomes like this.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[0012]
FIG. 1 is a main part configuration diagram showing an outline of a printing defect inspection apparatus according to an embodiment of the present invention.
[0013]
The printing defect inspection apparatus of the present embodiment includes a camera 10, an image input board 12, a misregistration calculation board 14, a first defect detection board 16, and a second defect detection board 18 as functional units. Each functional unit is electrically connected through the respective connectors 12A to 18A.
[0014]
The printing defect inspection apparatus includes a mother board 20. The mother board 20 includes a future expansion bus, a misregistration amount bus, R, G, and B image buses and a monochrome image bus in order from the top. Each data bus is laid, and a plurality of connectors 22A to 22E that can be electrically connected to each data bus are provided.
[0015]
The connector 22 </ b> A is connected to the connector 12 </ b> A of the image input board 12, and when a printed matter is captured by the camera 10, image signals corresponding to the R, G, B, and monochrome image buses are output. Yes.
[0016]
The connector 22B is connected to the displacement calculation board 14 via the connector 14A. The calculation board 14 calculates the displacement amount of the inspection image based on the image signal input from the monochrome image bus, and the calculation result. Is output to the misregistration amount bus. Since a monochrome image originally corresponds to a combination of R, G, and B color images, the displacement amount of the monochrome image and the displacement amount of the color image are the same.
[0017]
The connector 22C is connected to the first defect detection board 16 via the connector 16A. In the first defect detection board 16, R, G, B based on the positional deviation amount input from the positional deviation amount bus. The position of the inspection image input from each image bus is corrected, and a print defect is detected based on the image of each color.
[0018]
On the other hand, the connector 22D is connected to the second defect detection board 18 via its connector 18A, and the second defect detection board 18 similarly uses the above-mentioned monochrome based on the positional deviation amount input from the positional deviation amount bus. The position of the inspection image input from the image bus is corrected, and printing defects are detected based on the monochrome image.
[0019]
The connector 22E is used to connect a new functional unit when the inspection function is expanded in the future, and the expansion bus is also used for the same purpose. For the sake of convenience, the case where there is one expansion connector and one bus is shown here, but it goes without saying that both may be two or more.
[0020]
FIG. 2 is a block diagram showing an outline of a hardware configuration of the line sensor camera (one-dimensional CCD camera) in the present embodiment.
[0021]
The camera 10 of this embodiment receives a one-dimensional CCD sensor 10A that is a light receiving unit and light reception signals (odd number signal / even number signal) that are output separately from odd-numbered and even-numbered pixels (CCD elements) of the sensor 10A. Two amplifiers (Amp) for amplifying, an S / H circuit 10B for alternately synthesizing the odd pixel level of the odd signal and the even pixel level of the even signal for the amplified signal, and the black level of the synthesized signal A black level correction circuit 10C for correcting the above, an amplifier (Amp) that amplifies the correction signal and outputs it as a video signal, and a control device (not shown) for these CCD sensor 10A, threshold circuit 10B, and black level correction circuit 10C. And a timing generation circuit 10D that outputs a signal for timing each signal processing based on an external drive signal input from There. As described above, in the line sensor camera 10 of the present embodiment, electrical processing such as amplification is alternately performed on the odd-numbered and even-numbered pixels in order to increase the processing speed of the received light reception signal.
[0022]
FIG. 3 is a block diagram showing the main configuration of a print defect detection unit provided in the printed matter inspection apparatus according to the present embodiment.
[0023]
For convenience, this print defect detection unit represents the R image and does not clearly indicate the G and B images, but corresponds to the first defect detection board 18 shown in FIG. R, G, and B image signals captured via the input board 12 are input via the image buses. The defect detection unit captures an inspection image memory 30 for storing inspection images sequentially input from the camera 10 via the image bus, and captures a normal print pattern with the camera 10 in advance, and appropriately re-captures the image. And A and B reference image memories 32 for storing reference images to be updated.
[0024]
The defect detection unit performs edge extraction processing on the inspection image read from the inspection image memory 30 to create an edge image, and the reference image read from the reference image memory 32. A threshold generation circuit 36 for creating a threshold image, and two threshold image memories A and B for storing the created threshold image are provided. Further, in each of the R, G, and B threshold images read from each memory 38, the maximum luminance value is selected from the corresponding pixels, and the pixel value of the corresponding pixel is replaced with the maximum value. The RGB maximum value selection circuit 40 that creates one composite threshold image, and the R, G, and B edge images created by the edge image generation circuit 34 using the composite threshold image are masked in units of pixels. And a mask circuit 42 for creating a corresponding masked image. Further, in each of the R, G, B masked edge images created by the mask circuit 42, the maximum value of the luminance value is selected from the corresponding pixels, and the pixel value of the corresponding pixel is determined by the maximum value. An RGB composition circuit 44 that creates a composite edge image by replacement, a binarization circuit 46 that creates a binary image by binarizing the created composite edge image with a predetermined set value, and the created binary image A defective pixel number counting circuit 48 that counts the number of binarized pixels as the number of defective pixels and determines that a defect has occurred when the counted value is equal to or higher than a determination level is provided.
[0025]
Here, the reference image memory 32 and the threshold image memory 38 have two sets of A and B, respectively, while the inspection is performed based on the reference image and the threshold image stored in one memory. This is because a new reference image is inputted and stored by a reference update signal inputted as needed, and a threshold image created based on the new reference image is stored. In the figure, the reference update signal on the threshold image memory 38 side has one screen delay because the first reference update signal switches between writing and reading of the reference image memory A or B, and the read reference image is used as a threshold generation circuit. This means that the threshold image is generated by processing in 36 and the reading of the memory is switched one screen after the writing to the threshold image memory A or B is completed.
[0026]
The print defect detection unit in the present embodiment will be further described in detail. However, although illustration is omitted here, by operating the line sensor camera 10 at a predetermined interval in the horizontal direction, a horizontal line image composed of a predetermined number of pixels becomes, for example, a pitch of 1 line (pixel) / mm. Thus, inspection images arranged in the number of scans in the vertical direction are input.
[0027]
In the edge image generation circuit 34, the inspection image input from the inspection image memory 30 and the reference image input from the reference image memory 32 are respectively shown in FIGS. 4A to 4D. A spatial filter is applied to perform pattern edge extraction. Specifically, when one of the filters is superimposed on the image around the target pixel, the value obtained by multiplying the corresponding pixel value by each coefficient is added to the pixel value of the target pixel. By performing the process of resetting to all pixels, an image from which an edge has been extracted is created.
[0028]
As shown in FIG. 5, the threshold value generation circuit 36 selects a maximum value filter circuit 36A that selects the maximum value of pixel values in an area of m pixels × n pixels, and selects a minimum value of pixel values in the same area. A minimum value filter circuit 36B and a difference circuit 36C for obtaining a difference between the selected maximum value max and minimum value min are configured. For convenience, the case where m = 3 and n = 5 will be described below.
[0029]
In this threshold value generation circuit 36, 3 × with respect to the reference image input from the reference image memory 32, the center pixel is i-th in the horizontal direction and j-th in the vertical direction (hereinafter also referred to as coordinates (i, j)). A window consisting of five rectangular areas is set, and the pixel values A to O of the pixel and its surrounding 14 pixels are obtained from the central pixel of the pixel value (luminance value) H as schematically shown in FIG. A threshold value tij calculated by the following equation (1) is set. However, in this expression, Max {} represents the maximum value from the listed pixel values, and Min {} represents the minimum value from the pixel values.
[0030]
tij = Max {A, B, C, D, E, F, G, H, I, J, K, L,
M, N, O} -Min {A, B, C, D, E, F, G, H,
I, J, K, L, M, N, O} (1)
[0031]
As the threshold value tij set for the center pixel, a method of calculating using the following equation (2) can also be adopted.
[0032]
tij = Max {| A−H |, | B−H |, | C−H |, | D−H |,
| EH |, | FH |, | GH |, | IH |, | JH |,
| K-H |, | L-H |, | M-H |, | N-H |, | O-H |} (2)
[0033]
The threshold processing for the center pixel is performed on all pixels of the reference image except for pixels that cannot be calculated. For all the threshold values tij obtained in this way, a threshold value Tij to be set as a pixel value of the final threshold image is obtained by the following equation (3) using predetermined constants α and β.
[0034]
Tij = α · tij + β (3)
[0035]
In this equation, the constant α is for preventing the overall brightness change from occurring, and the constant β is for hiding the entire noise level with an offset, and is a minimum threshold value or tuning threshold value. Called etc. Thus, by using the above equation (3), it is possible to obtain a threshold image composed of threshold values in consideration of the overall light quantity fluctuation and noise level.
[0036]
Here, the threshold value setting process will be described as a specific example in the case where the pixel value of the central pixel shown in FIG. 7 is 75. The maximum luminance value in this window is 172, and the minimum value Therefore, when these values are applied to the equation (1), tij = 172-21 = 151. If the constants in the expression (3) are α = 1.2 and β = 10, the threshold Tij that is the final pixel value set for the pixel of the threshold image corresponding to the central pixel is Tij = 1.2 ×. 151 + 10 = 191. As described above, the rectangular range of the window set for the center pixel is set to ± 1 in the horizontal direction and ± 2 in the vertical direction. Needless to say, the distortion is not limited to this.
[0037]
The threshold value setting process described in detail above is performed for the whole while shifting each pixel (one line) in the horizontal direction and the vertical direction, whereby a threshold image is created, and the image data is stored in the threshold image memory 38. .
[0038]
In the RGB maximum value selection circuit 40, by setting the maximum pixel value among the corresponding pixels of the R, G, B threshold images input from the threshold image memory 38 to the pixels at the same position, One composite threshold image is generated. In this way, in each threshold image, the maximum value is selected from the pixel values of the corresponding pixels, and the composite threshold image is created by replacing it with the pixel value of the corresponding pixel. Only an area for masking the edge portion of the pattern remains in the image, and a subtle color variation portion caused by a misregistration between colors can be eliminated.
[0039]
In the mask circuit 42, a mask process is performed by a composite threshold image for the edge image obtained from the inspection image. If one pixel of the edge image corresponds to the center pixel of the coordinates (i, j) shown in FIG. 6, the absolute value of the pixel value S of the edge image corresponds to the corresponding composition. This is a process of extracting defective candidate pixels by comparing with the pixel value Tij of the threshold image and setting S = 0 if | S | ≦ Tij, and keeping the absolute value in the opposite case. As described above, since the edge extracted from each of the R, G, and B inspection images is masked using the composite threshold image, each masked edge image has a defective portion that does not exist in the reference image. Therefore, it is possible to reliably detect a point-like defect or the like isolated on the printed matter without being affected by subtle color fluctuations.
[0040]
According to the embodiment described in detail above, as conceptually shown in FIG. 8 and FIG. 9, a point-like defect whose luminance value greatly changes at the edge portion, a minute inter-color misregistration of the printed pattern, It is possible to detect easily and reliably without excessively detecting subtle color fluctuations at the pattern edge portion caused by chromatic aberration of the camera lens.
[0041]
That is, R, G, and B threshold images (B) are created from the color reference image (A) created in advance by the threshold generation circuit 36 and stored in the threshold image memory 38. In the threshold image (B), the edge portion of the pattern is extracted with a wide width by the threshold setting process described above. Next, the RGB maximum value selection circuit 40 selects the maximum value among the luminance values of the corresponding pixels from the generated R, G, and B threshold images, and sets them to the corresponding pixels to synthesize them. A threshold image (C) is created.
[0042]
On the other hand, from the image bus, R, G, and B images (E) corresponding to the inspection image in which the point-like defect (defect in the drawing) shown in FIG. Are stored, the edge image (F) is created by the edge image generation circuit 34 for each inspection image (E). In this edge image (F), the edge of the point defect is extracted together with the edge of the pattern. In the attached drawings, the edge of the point defect is not clearly shown due to the performance of the input device. The image (G) described below is the same for the images (H) and (I) that are somewhat visible.
[0043]
Next, the mask circuit 42 masks each edge image (F) using the combined threshold image (C) input from the RGB maximum value selection circuit 40, thereby R, G, B A masked image (G) corresponding to each is created. At this stage, the edge portion of the pattern in the edge image (F) is masked by the edge region of the composite threshold image (C) and disappears, and only the edge of the point defect remains in the G image. The R, G, and B masked edge images are synthesized by the RGB synthesis circuit 44, the resulting synthesized image is binarized with a predetermined set value, and the number of binarized white pixels is counted. Counting is performed by the circuit 48, and when the number, that is, the area is equal to or higher than a predetermined determination level, it is determined that a defect has occurred.
[0044]
According to the present embodiment described in detail above, each edge image is masked using the composite threshold image, and therefore, it is possible to inspect the subtle color misregistration. Therefore, it is possible to realize a color inspection process capable of sufficiently exhibiting the inspection ability in a stable state without excessively detecting.
[0045]
Although the present invention has been specifically described above, the present invention is not limited to that shown in the above embodiment, and various modifications can be made without departing from the scope of the invention.
[0046]
For example, in the above-described embodiment, the number of pixels or the like is specifically shown as a processing unit when creating a threshold image, but the present invention is not limited to this. Further, the spatial filter used for edge extraction is not limited to that shown in the embodiment.
[0047]
【The invention's effect】
As described above, according to the present invention, it is possible to easily and reliably detect a point-like defect that clearly exists together with a pattern on an inspection image obtained by imaging a printed matter.
[Brief description of the drawings]
FIG. 1 is an explanatory diagram illustrating a schematic configuration of an entire inspection apparatus according to an embodiment of the present invention. FIG. 2 is a circuit diagram illustrating a hardware configuration of a line sensor camera according to the embodiment. FIG. 4 is an explanatory diagram showing an example of a spatial filter used for edge extraction. FIG. 5 is a block diagram showing details of a threshold generation circuit. FIG. 6 is an explanatory diagram conceptually showing a window used for threshold processing. FIG. 7 is an explanatory diagram illustrating a specific example of a pixel value in which a window used for threshold processing is set. FIG. 8 is an image diagram illustrating an operation of the embodiment. FIG. 9 is another image diagram illustrating an operation of the embodiment. ]
30 ... Inspection image memory 32 ... Reference image memory 34 ... Edge image generation circuit 36 ... Threshold generation circuit 38 ... Threshold image memory 40 ... RGB maximum value selection circuit 42 ... Mask circuit 44 ... RGB composition circuit

Claims (10)

In a printed matter inspection method for detecting a print defect by comparing an inspection image input for each of R, G, and B by imaging a printed matter with a color camera and a reference image created in advance.
Extracting pattern edges from the R, G, and B inspection images to create R, G, and B edge images,
A threshold image that masks the edge portion of the pattern of the edge image is created from each of the R, G, and B reference images,
For each of the created R, G, and B threshold images, a maximum value is obtained from the pixel values of the corresponding pixels, and one combined threshold image is created using the maximum value as the pixel value of the corresponding pixel.
Masking each of the R, G, and B edge images with the composite threshold image to create R, G, and B masked edge images,
A printed matter inspection method, comprising: detecting defective printing based on each created masked edge image. In claim 1,
As the threshold image, for the reference image, two pixels having the largest difference in pixel value from a predetermined neighborhood range are selected , and α and β are set as constants from the absolute value t ij of the difference between the two pixel values. The following formula
T ij = α · t ij + β
A printed matter inspection method, wherein T ij obtained by the above is set to the pixel value of the center pixel of coordinates (i, j) . In claim 1,
If the absolute value of the pixel value is less than or equal to the pixel value of the corresponding pixel in the composite threshold image, the mask processing exceeds 0 if the absolute value of the pixel value of each pixel constituting each of the R, G, and B edge images If so, the printed pixel inspection method is performed by setting the original pixel values. In claim 1,
The R, G, B masked edge images are combined to create a combined edge image, the combined edge image is binarized to generate a binary image, and defect occurrence is determined based on the binary image And a printed matter inspection method. In claim 4,
A printed matter inspection method, wherein the composite edge image is created by setting a maximum pixel value among the corresponding pixels of the R, G, and B masked edge images to the same pixel. In a printed matter inspection apparatus that detects a print defect by comparing an inspection image input for each of R, G, and B by imaging a printed matter with a color camera and a reference image created in advance.
Means for extracting an edge of a pattern from each inspection image of R, G, B and creating each edge image of R, G, B;
Means for creating a threshold image for masking an edge portion of the pattern of the edge image from each of the R, G, and B reference images;
Means for determining a maximum value from the pixel values of the corresponding pixels for each of the generated threshold images of R, G, and B, and generating one composite threshold image using the maximum value as the pixel value of the corresponding pixel; ,
Means for masking each of the R, G, and B edge images with the composite threshold image to create each of the R, G, and B masked edge images;
A printed matter inspection apparatus comprising: means for detecting a print defect based on each created masked edge image. In claim 6,
As the threshold image, for the reference image, two pixels having the largest difference in pixel value from a predetermined neighborhood range are selected , and α and β are set as constants from the absolute value t ij of the difference between the two pixel values. The following formula
T ij = α · t ij + β
The printed matter inspection apparatus is characterized in that T ij obtained by the above is set to the pixel value of the center pixel of the coordinates (i, j) . In claim 6,
If the absolute value of the pixel value is less than or equal to the pixel value of the corresponding pixel in the composite threshold image, the mask processing exceeds 0 if the absolute value of the pixel value of each pixel constituting each of the R, G, and B edge images If so, the printed pixel inspection apparatus is characterized in that each original pixel value is set. In claim 6,
Based on the binary image, means for synthesizing the R, G, B masked edge images to create a composite edge image, binarizing the composite edge image to create a binary image, and A printed matter inspection apparatus, further comprising means for determining occurrence of a defect. In claim 9,
The printed material inspection apparatus, wherein the composite edge image is created by setting the maximum pixel value among the corresponding pixels of the masked edge images of R, G, and B to the same pixel.

Images