JPH08323963A - Method and device for forming threshold value image in printed matter-inspection device - Google Patents

Abstract

PURPOSE: To realize a threshold value image forming method by which the stabilized inspection can be carried out by using a reference image or a prior reference image, and finding a corresponding threshold value of a central pixel for respective pixels from proximity pixels in the given rectangular range with the pixel to be inspected as its center. CONSTITUTION: A pattern of a traveling printed-matter 4 is input into a signal transmitting section 20 by utilizing the output of a rotary encorder 6 provided with a high-luminance and high-frequency fluorescent light 12 and synchronized with the movement of a three-plate type color line-sensor camera 11 and the printed matter 4, and then output of the pattern is transmitted to a processing section 30 through an optical fiber 8. The input signal is shading corrected in the processing section 30, and the corrected signal is stored as monochromatic data in a reference memory and inspection image memory, and its position is corrected and the print defect inspection is carried out by a defect adjustment section. Also for the image of defect sensing threshold value formed by a learning section, the inspection is carried out by finding a corresponding threshold value of central pixel for respective pixels out of the proximity pixels in the given rectangular range in which the pixel to be inspected is included as its center. Thus the inspection performance of a profile portion of print pattern, can be retained and the stabilized inspection can be carried out.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for optically monitoring the surface condition of a printed material during running, and comparing the captured inspection image with a reference image to create a threshold image for inspection. Regarding

[0002]

2. Description of the Related Art Conventionally, printed matter inspection has been performed manually by various methods. For example, depending on the printing machine and the type of printed matter, the printed matter after printing has been appropriately extracted and visually inspected. For printed matter that runs continuously at a constant speed, strobe lighting synchronized with the printing speed is used.
Attempts have been made to determine a running printed matter as a still image.

On the other hand, in recent years, a system for inspecting a printed matter online by using a line sensor has been developed. That is, the printed paper surface that continuously runs is scanned by an optical sensor, is captured as image data, and is compared with reference data that has been captured in advance to inspect whether the paper surface is defective.

At this time, in creating the threshold image, the brightness value of the pixel is referred to in the vertical direction and in the horizontal direction independently, and the threshold value of each target pixel is calculated.

[0005]

As described above, conventionally, there has been no method for calculating a threshold value by considering diagonal components by using vertical and horizontal windows. Therefore, due to the positional deviation error of the captured image, image distortion, change in luminance value, etc.,
There is a problem that the inspection performance is poor and stable inspection cannot be performed.

The present invention has been made to solve the above-mentioned conventional problems, and a threshold image forming method in a printed matter inspection apparatus that enables stable inspection while maintaining the inspection performance of the contour portion of the printed pattern to the maximum. And to provide a device.

[0007]

SUMMARY OF THE INVENTION The present invention provides a threshold image creating method in a printed matter inspection apparatus for optically monitoring the surface state of a printed matter during traveling and comparing the captured inspection image with a reference image to inspect. Using a reference image or a reference image before that, by determining an adaptive threshold value of the central pixel for each pixel from neighboring pixels within a predetermined rectangular area centered on the pixel to be inspected, It has achieved its purpose.

According to the present invention, the size of the rectangular range is determined by a value that allows for the vertical and horizontal positional displacement in the captured inspection image and the size of expansion and contraction. It has achieved its purpose.

According to the present invention, the threshold value is calculated by using a value obtained by subtracting the minimum value from the maximum value of the brightness values of the pixels included in the rectangular range. It has been achieved.

According to the present invention, the threshold value is calculated as the maximum absolute value of the values obtained by subtracting the brightness value of the central pixel of the rectangular range from the brightness value of each pixel included in the rectangular range. By doing so, the above-mentioned object is achieved in the same manner.

According to the present invention, the threshold value is further multiplied by a predetermined constant α, and a value obtained by adding the predetermined constant β is used as a final threshold value to create a threshold image. The above object is achieved.

The present invention further provides a threshold image creating device in a printed matter inspection device for optically inspecting a surface state of a printed matter during running and comparing the taken inspection image with a reference image to obtain a reference image or before. By using the image serving as the reference, the threshold image generation unit that obtains the adaptive threshold value of the central pixel for each pixel from the neighboring pixels within a predetermined rectangular range centered on the pixel to be inspected Similarly, the above-mentioned object is achieved.

[0013]

According to the present invention, in the printed matter inspection device, the surface state of the printed matter during traveling is optically monitored, and the captured inspection image is compared with the reference image or the reference image before the inspection, and the inspection is performed. In doing so, a threshold image is created using vertical and horizontal windows.

That is, using the reference image or the reference image before that, the adaptive threshold value of the central pixel is calculated for each pixel from the neighboring pixels within a predetermined rectangular area centered on the pixel to be inspected. ing. Since the vertical direction causes a relatively large distortion due to the running of the original fabric, for example, by using a rectangular range in which the vertical direction is larger than the horizontal direction, while maintaining the maximum inspection performance of the contour portion of the printed pattern, A stable inspection can be performed.

Further, when the size of the rectangular range is determined based on the vertical and horizontal positional deviation in the captured inspection image and the value considering the size of expansion and contraction, the positional deviation of the inspection image and It can handle distortion properly,
Accurate inspection is possible.

When the threshold value is calculated using a value obtained by subtracting the minimum value from the maximum value of the brightness values of the pixels included in the rectangular range, the reference image and the reference image are calculated. When the so-called learning images for creating are different, it is possible to obtain a good effect by updating the reference image one after another and using it to obtain the threshold image.

When the threshold value is calculated as the maximum absolute value of the values obtained by subtracting the brightness value of the central pixel of the rectangular range from the brightness value of each pixel included in the rectangular range, This is effective when the reference image and the learning image are equal and the edge may be displaced to the position of each pixel in the rectangular range.

Further, when the threshold value is further multiplied by a predetermined constant α and a threshold image is created with the value obtained by adding the predetermined constant β as the final threshold value, It is possible to obtain a threshold image in consideration of fluctuation and noise level.

[0019]

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

The overall structure of the printed matter inspection apparatus according to this embodiment is shown in FIG.

As shown in FIG. 1, the printed matter inspection apparatus of the present embodiment has an input unit 10 for inputting a pattern printed on a printed matter 4 which is guided by guide rollers 2 and printed as image data, and an input image. A / D conversion of data,
A signal transmission unit 20 that transmits to the processing unit 30 using the optical fiber 8, a processing unit 30 that determines a printing defect, an operation unit 4 that displays a defective printing image on the screen, and operates the apparatus.
It consists of 0 and.

The input unit 10 includes a three-plate color line sensor camera (hereinafter simply referred to as camera) 11 for image input,
A high-intensity high-frequency fluorescent lamp 12 as a light source device.

The synchronization input timing of the camera 11 is created by the signal transmission section 20. That is, the signal transmission unit 2
0 is a rotary encoder 6 synchronized with the movement of the printed matter 4.
Is converted and output to the camera 11.

The high-intensity high-frequency fluorescent lamp 12 used as the light source has a frequency of, for example, 30 KHz.
It can be considered that the input image of 1 is almost continuous lighting.

The structure of the signal transmission unit 20 is shown in FIG.

In order to minimize the influence of noise on the analog signal (image data) output from the camera 11 to the signal transmission section 20, the transmission paths of the R, G and B analog signals are made as short as possible. It is desirable that the signal transmission unit 20 is installed near the camera 11.

The analog signal input to the signal transmission unit 20 is removed by the speed correction unit 21 from the influence of the image due to the fluctuation of the original (printed material) speed, and is converted into a digital signal by the A / D conversion unit 22. It is converted into an optical signal in the electro-optical conversion unit 23 and transmitted to the processing unit 30 through the optical fiber 8 having high noise resistance.

Further, as described above, the timing of the synchronization input is obtained from the rotary encoder 6 installed in the printing press. The signal from the rotary encoder 6 is used to shift the position of the shifter 24 and the timing controller 2 to find the beginning of the pattern.
5 controls the signal for the camera, and the camera 11
Output to.

The structure of the processing unit 30 is shown in FIG.

The image data signal received from the signal transmission unit 20 is first converted by the preprocessing unit 31 into a signal form required for the subsequent processing.

The pre-processing unit 31 converts the optical signal input from the optical fiber 8 into an electric signal, performs shading correction to correct the variation of each image of the camera 11 and uneven illumination of the light source, and converts it into color data RGB. Each is multiplied by a coefficient, added, and converted into monochrome data.

The color data shading-corrected by the preprocessing unit 31 is stored in the full-color memory 32 for later display.
Is stored in

Further, the monochrome data (luminance data) converted by the preprocessing unit 31 is stored in the reference image memory 34 or the inspection image memory 33, and the defective printing is inspected by using this.

The learning section 36 creates a threshold image for defect detection and selects data for position correction. That is, the learning unit 36 has the function of the threshold image creating device. Further, the position correction unit 35 corrects the position deviation of the reference image inspection image by the method described later based on the data from the learning unit 36, and finally, the defect determination unit 37 compares the reference image with the inspection image. Then, the defective portion is determined.

The operation unit 40 operates this apparatus, and when a defect occurs, displays an inspection image including the defect on the monitor 41 and warns that there is a printing defect with a buzzer.
A defective printed matter is removed by a detector, and if the defective printed matter cannot be removed, labeling is performed on the defective portion so that the defective portion can be recognized.

The threshold image creating process will be described in detail below.

FIG. 4 to FIG. 8 show a process of the defect determination processing for comparing the inspection image and the reference image and generating different portions as defect candidate images.

FIG. 4 shows the inspection image 100.
The inspection image 100 includes a print pattern 110 and a defective portion 120, and the arrangement of the luminance values of pixels on a certain line 130 passing through the defective portion 120 can be represented as a graph 140 in the lower part of FIG.

FIG. 5 shows a reference image 200.
The reference image 200 includes only the print pattern 210,
The arrangement of the brightness values of the pixels on the line 220 at the same position as the inspection image 100 is as shown by the graph 230 in the lower part of FIG. If the inspection image 100 is displaced with respect to the reference image 200, the inspection image 1 is corrected after the displacement is corrected.
00 and the reference image 200, the absolute value of the difference between the brightness values is obtained for each pixel, and a difference image 300 as shown in FIG. 6 is created.

As shown in FIG. 6, the contour portion 310 of the print patterns 110 and 210 remains in the difference image 300 as a positional deviation error. Further, since the defect part 120 is not included in the reference image 200, it is left as it is as the defect part 320 in the difference image 300. In the difference image 300, the line 33 at the same position as the line 130 in the inspection image 100
The arrangement of the luminance values of the pixels on 0 is the graph 340 in the lower part of FIG.
Is represented by

Since the contour portion 310 of the printed patterns 110 and 210 is not a defect, the threshold image 4 shown in FIG.
00 is prepared. This threshold image 400 is the reference image 20.
It is created from a reference image input at 0 or before.

The value of each pixel in the threshold image 400 must exceed the positional deviation error occurring in the contour portion of the print pattern and the noise component contained in the image.
Further, the inspection image 100 includes distortion as compared with the reference image 200 due to expansion and contraction of the printing original fabric and a shift in the image capture timing. Therefore, even if the position shift correction is performed, the position of the outline of the pattern is equal to that of the reference image. It will be shifted compared to. Therefore, the picture outline portion 410 in the threshold image 400 needs to be thicker than the outline portion 310 in the difference image 300.

In the point threshold image 400, the arrangement of the luminance values of the pixels on the line 420 at the same position as the line 130 in the inspection image 100 is represented by the graph 430 in the lower part of FIG.

Each pixel value of the difference image 300 is the threshold image 40.
If it is larger than the value of each pixel existing at the same position of 0, the pixel value of the difference image 300 is left as it is, otherwise, the value of the pixel is set to 0.

The defect candidate image 500 shown in FIG. 8 is created in this way. The defect part 510 is left in the defect candidate image 500. This defect candidate image 50
At 0, the arrangement of the luminance values of the pixels on the line 520 at the same position as the line 130 in the inspection image 100 is represented by the graph 530.

After the defect candidate image 500 is created, the size and the brightness value of the defect portion 510 are measured in the subsequent processing,
A final defect determination is made.

A method of creating the threshold image 400 will be described below.

FIG. 9 shows a 3 × 5 window centered on the pixel h, which is assigned to each pixel h of the reference image or the reference image input before that.

Here, the threshold value of the central pixel h is calculated from the luminance values of the central pixel h and the 14 pixels a to o in the vicinity thereof by the following equation (1).

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)

However, the function Max {} indicates that the maximum value of the listed variables is taken, and the function Min {} indicates that the minimum value of the listed variables is taken.

Further, tij may be calculated using the following equation (2) instead of equation (1).

Tij = Max {| a-h |, | b-h |, | c-h |, | d-h |, | e-h |, | f-h |, | g-h |, | i−h |, | j−h |, | k−h |, | m−h |, | n−h |, | o−h |} (2)

With respect to the threshold value Tij thus obtained, the final threshold value Tij is calculated by the following equation (3) using predetermined constants α and β.
Ask for.

Tij = α · tij + β (3)

Here, the constant α is used to prevent the overall brightness from changing, and is set to a temporarily large value after paper splicing. The constant β is used to hide the entire noise level with an offset, and is called the minimum threshold or tuning threshold.

In this way, the threshold value for each pixel can be obtained in consideration of the overall light amount fluctuation and noise level.

Specific examples are shown in FIGS.

All pixels in the reference image 600 shown in FIG. 10 are calculated by the above-described method. For example, FIG.
If the window centering on the pixel 610 shown in 0 is as shown in FIG. 11, the maximum value of the brightness value in this window is 172 and the minimum value is 21. here,
The higher the value, the higher the brightness.

Applying this to the equation (1), tij = 1
72-21 = 151. When the constants in the equation (3) are α = 1.2 and β = 10, the final threshold Tij of the pixel 710 in the threshold image shown in FIG. 12 is Tij = 1.
It becomes 2 × 151 + 10 = 191.

In the present embodiment, the rectangular range of the neighboring pixels is ± 1 in the horizontal direction and ± 2 in the vertical direction. However, in the vertical direction, the image is expanded or contracted in the vertical direction or the image acquisition timing is shifted. This is because the distortion is large. The apparatus of this embodiment is set so that this rectangular range can be freely selected from the 7 × 7 range.

By changing the size of this rectangular range depending on the characteristics of the printing machine and the type of the original, it is possible to stabilize the inspection while maintaining the maximum detection performance.

[0063]

As described above, according to the present invention,
A rectangular range is set according to the amount of misalignment error and image distortion, and the maximum value that can change the brightness value of the central pixel of this rectangular range is calculated as a threshold value. Stable inspection is possible while maintaining the maximum inspection performance of the part.

[Brief description of drawings]

FIG. 1 is an overall configuration diagram showing an outline of a printed matter inspection apparatus according to an embodiment.

FIG. 2 is a block diagram showing a configuration of a signal transmission unit.

FIG. 3 is a block diagram showing a configuration of a processing unit.

FIG. 4 is an explanatory diagram showing an inspection image.

FIG. 5 is an explanatory diagram showing a reference image.

FIG. 6 is an explanatory diagram showing a difference image.

FIG. 7 is an explanatory diagram showing a threshold image.

FIG. 8 is an explanatory view showing a defect candidate image.

FIG. 9 is an explanatory diagram showing a window.

FIG. 10 is an explanatory diagram showing how a threshold image is created from a reference image.

FIG. 11 is an explanatory diagram showing a specific example of a window.

FIG. 12 is an explanatory diagram showing a specific example of a threshold image.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 ... Input part 20 ... Signal transmission part 30 ... Processing part 36 ... Learning part (threshold image creation device) 40 ... Operation part 100 ... Inspection image 200,600 ... Reference image 400,700 ... Threshold image

Claims (6)

[Claims] 1. A threshold image creating method in a printed matter inspection device for optically inspecting a surface state of a printed matter during traveling and comparing the taken inspection image with a reference image, wherein a reference image or a reference before that is used. Using the image
A threshold image creating method in a printed matter inspection apparatus, wherein an adaptive threshold value of the central pixel is obtained for each of the pixels from neighboring pixels within a predetermined rectangular range centered on the pixel to be inspected. 2. The size of the rectangular range according to claim 1, wherein the size of the rectangular range is determined by a value that allows for the vertical and horizontal positional deviation in the captured inspection image and the size of expansion and contraction. A method for creating a threshold image in a characteristic printed matter inspection apparatus. 3. The method according to claim 1 or 2, wherein the threshold value is calculated by using a value obtained by subtracting a minimum value from a maximum value of luminance values of pixels included in the rectangular range. A method for creating a threshold image in a characteristic printed matter inspection apparatus. 4. The absolute value of the threshold value according to claim 1, wherein the threshold value is a value obtained by subtracting a brightness value of a central pixel of the rectangular range from a brightness value of each pixel included in the rectangular range.
A threshold image creating method in a printed matter inspection apparatus, which is calculated as the maximum value. 5. The threshold image according to claim 3 or 4, wherein the threshold value is further multiplied by a predetermined constant α, and a value obtained by adding a predetermined constant β is used as a final threshold value to create a threshold image. A method for creating a threshold image in a printed matter inspection apparatus, comprising: 6. A threshold image creating apparatus in a printed matter inspection apparatus for optically inspecting a surface state of a printed matter during running, comparing the taken inspection image with a reference image, and comparing the reference image with a reference image before the reference image. Using the image
A threshold image in a printed matter inspection apparatus, which is provided with a threshold image generating unit that obtains an adaptive threshold value of the central pixel for each pixel from neighboring pixels within a predetermined rectangular range centered on the pixel to be inspected. Creation device.