JP2006138768A - Inspection method of printed matter - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve shorten treatment time, by reducing the operation quantity in image processing, without lowering the detection accuracy of flaws, in inspecting the flaws of printed matter using image processing. <P>SOLUTION: The inspection method for printed matter is constituted so that the image of the printed matter, to which printing is applied, is photographed and the photographed image is subjected to image processings to detect the flaw of the printed matter. The image processing includes an inclination quantity calculating step of calculating the inclination quantity of the printed matter from the photographed image of the printed matter, an image simplifying step of simplifying the photographed image to form a simplified image, a printing position detecting process of rotating the simplified image, corresponding to the calculated inclination quantity to correct it and using the corrected image to detect a printing position and a mask image processing step of arranging the preliminarily inputted mask pattern at the detected printing position to form a mask arranged image and making the mask arranged image rotate in reverse by the inclination quantity and make a mask image completed. The completed mask image is used to perform the mask process the photographed image. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a method for inspecting a defect such as a foreign matter such as dust, a flaw, or a stain on a printed matter such as a sheet on which characters or designs are printed.

As a method for inspecting a printed sheet body, a method for detecting defects by processing a sheet body by image analysis after masking a printing portion is known, as described in Patent Document 1 below. . However, since the method described in Patent Document 1 uses a long sheet body as an inspection target, no countermeasure is taken for a case where a printed material to be inspected is greatly inclined, and it is effective. It cannot be dealt with.
Japanese Patent No. 2597370

  As an inspection method when the inspection object is tilted, a method using pattern matching is generally known. In pattern matching, a print position and an angle are specified by finding a coincidence point between the print position and the angle while moving and rotating the comparison pattern by trial and error. Therefore, according to the method using the pattern matching, even when the inspection target is tilted, it can be dealt with by trial and error while rotating the comparison pattern according to the tilt. As a result, the tilted inspection target It is possible to create a tilted mask image corresponding to the printed portion.

  As another method when the inspection target is tilted, a method of performing mask processing after rotating a photographed image of the inspection target and correcting the image to be tilted is also conceivable.

  However, in the pattern matching, when the inclination of the inspection target is large, it is necessary to increase the rotation angle of the comparison pattern in trial and error, and there is a problem that the processing time until the proper inclination angle is detected becomes long. is doing.

  In addition, when rotating a captured image to be inspected, there is a problem that it takes a lot of processing time to rotate the entire image. In addition, when rotating the image, it is necessary to interpolate the pixel data, and there is a problem that the detection accuracy is lowered by reducing the contrast of the defect.

  The present invention has been made in view of the problems of the prior art as described above, and it is an object of the present invention to not reduce the detection accuracy of defects when inspecting defects of printed matter using image processing. Another object is to reduce the amount of calculation in the processing to shorten the processing time.

  In order to solve the above-mentioned problems, a printed matter inspection method according to the present invention is a method for inspecting a printed matter by detecting an image of the printed matter by taking an image of the printed matter and performing image processing on the photographed image. In the method, the image processing includes an inclination amount calculating step for calculating an inclination amount of a printed matter from an image of the photographed printed matter, and correcting the image by rotating the image taken according to the calculated inclination amount. A printing position detecting step for detecting a printing position using the image that has been created, and a mask arrangement image is created by arranging a mask pattern that has been input in advance at the detected printing position, and the mask arrangement image is generated by the amount of inclination. And a mask image processing step for completing a mask image by reverse rotation, and masking the photographed image using the completed mask image.

  According to the present invention, since the printing position is detected using the simplified image corrected by the rotation, it is not necessary to detect the inclination angle when detecting the printing position, and this also reduces the amount of calculation in the image processing. It can be greatly reduced. Further, the mask image is masked with the mask image obtained by arranging the mask pattern and rotating the calculated tilt amount in the reverse direction, so that the detection accuracy is not lowered as in the case of rotating the shot image. Defects can be detected. Therefore, according to the printed matter inspection method of the present invention, the processing time can be shortened without reducing the calculation amount in the image processing without reducing the defect detection accuracy.

  In addition, when the captured image is a color image, it further includes an image simplification step for simplifying the captured image to create a simplified image, and in the printing position detection step, instead of the captured image. It is assumed that the simplified image is used.

  In the printing position detection step, the captured image itself is not rotated, but the simplified image created based on the captured image is rotated and corrected, so that the captured image itself is rotated. As compared with the above, the calculation amount in the image processing can be greatly reduced, and the processing time can be shortened by the large reduction in the calculation amount.

  As described above, according to the printed matter inspection method of the present invention, when inspecting a defect of the printed matter using image processing, the calculation amount in the image processing is greatly reduced without reducing the detection accuracy of the defect. The processing time can be shortened.

Hereinafter, an exemplary embodiment of the present invention will be described with reference to the drawings.
FIG. 1 shows a printed sheet 1 on which color printing 2 has been performed as an example of a printed material to be inspected according to the present invention. The shape of the color printing sheet 1 is not particularly limited, and may be a long shape. In the present embodiment, a description will be given of a single-wafer shape in which the angle is not easily displaced and the effect of the present invention is most prominent.

  FIG. 2 is a diagram conceptually showing a configuration example of a defect inspection apparatus used when carrying out the present invention. As shown in FIG. 2, the defect inspection apparatus 10 includes a transport unit 11 (for example, a belt conveyor) for sequentially transporting the color print sheets 1 to be inspected, and an illumination 12 for irradiating the color print sheet 1 with light. An image capturing unit 13 that captures an image of the color print sheet 1 and an image processing unit 14 that determines a defect of the color print sheet 1 by performing image processing on the captured image.

  As the illumination 12, a fluorescent lamp, a halogen lamp, an LED, or the like can be used according to the color printing sheet to be inspected. Also, the illumination method can be appropriately changed depending on the aspect of the color print sheet 1 to be inspected and the type of defect to be detected. For example, if the color printing sheet 1 to be inspected is light-transmitting, it is possible to detect a light-blocking defect such as hair by using transmitted illumination. In addition, when the texture of the color printing sheet 1 is rough, it is preferable to use reflected illumination together, and this can suppress an adverse effect on the image due to the roughness. Further, when the color print sheet to be inspected is not light transmissive, it is preferable to use reflected illumination. In this case, in order to suppress an adverse effect on the image due to the roughness of the texture, for example, light from all angles may be incident using dome-shaped illumination.

  The photographing means 13 receives transmitted light or reflected light from the color printing sheet 1 and converts it into image data. As the photographing means 13, a line sensor, a two-dimensional camera, or the like can be used.

The image processing means 14 performs image processing based on the captured image data and detects defects.
FIG. 3 is a flowchart showing an example of an image processing process in the image processing means 14. The image processing unit 14 in the embodiment includes an image input step 41 for inputting the image data transmitted from the photographing unit 13, a pre-processing step 42 for performing noise removal processing on the obtained image data, and a defect The defect candidate point extraction step 43 for extracting candidate points, the mask processing step 44 for creating a defect detection image by masking the printing unit from the image data from which defect candidate points have been extracted, and the mask processing And a pass / fail judgment step 45 for judging the presence / absence (defective / defective) of the defect using the subsequent defect detection image, and further for creating a mask image used for masking the printing portion in the mask processing step 44 And a mask creating step 46.

  In the preprocessing step 42, noise removal processing, defect enhancement processing, or the like is performed on the input image (that is, the captured image) as necessary. Examples of the noise removal process include an opening process performed on an image with a lot of bright spot noise and a shading correction process performed on an image with shading. Examples of defect enhancement processing include differential processing for enhancing edges and gradation conversion processing for enhancing contrast.

  In the defect candidate point extraction step 43, for example, binarization processing is performed on the image that has undergone the preprocessing step 42 using an appropriate threshold value, and defect candidate points are extracted. As a defect candidate point, in addition to the defect to be detected, a color print portion applied to the print sheet is also detected. Therefore, in the mask processing step 44 described later, this print portion is masked to obtain a defect candidate point. It is necessary to exclude from.

The mask processing step 44 uses, for example, a binary image in which the printing portion is 0 and other portions are 1 as a mask image, and the printing portion is masked by obtaining a theoretical product of the mask image and the defect candidate point image. Then, defect candidate points other than the printing part are extracted.
As another mask processing method, a method of inspecting a mask image region (that is, a printing portion) using different determination criteria and a method of using the mask processing method in combination can be employed. In this case, a threshold value different from that in the defect candidate point extraction step is used, and when extracting the defect candidate points in the printing unit, it is only necessary to mask other than the printing unit with an image obtained by inverting the mask image.

Next, the mask creating process 46 for creating a mask image used in the mask processing process 44 will be described in detail with reference to FIG.
The mask creation process 46 includes (A) an inclination amount calculating step for calculating an inclination amount of the print sheet from the photographed print sheet image, and (B) an image simplification that simplifies the photographed image and creates a simplified image. (C) a printing position detection step for rotating and correcting the simplified image according to the calculated inclination amount, and detecting a printing position using the corrected image; and (D) the detected printing A mask image processing step of creating a mask arrangement image by fixing a mask pattern of a printing unit inputted in advance at a position, and rotating the mask arrangement image in reverse according to the inclination amount to complete a mask image. .

(A) Inclination Amount Calculation Step In the inclination amount calculation step, first, a print sheet region 61 is specified based on a photographed image (input image) 51. When the contrast between the print sheet and the background is sufficiently large, the area 61 of the print sheet can be specified relatively simply by simply binarizing. When the contrast between the print sheet and the background is small, the contrast 61 is corrected by gradation conversion processing or the like, and then binarized to identify the print sheet region 61.

After the area 61 of the print sheet is specified, the amount of inclination of the print sheet is calculated from the linear boundary line 61 ′ constituting the area 61. As a method for calculating the amount of inclination, for example, a method of sampling two or more arbitrary points on the boundary line 61 ′ of the print sheet and calculating an approximate straight line from the sampled points can be employed. If there are two sampled points, and the coordinates are (x1, y1) and (x2, y2), the amount of inclination is calculated as tan ⁻¹ {(y2−y1) / (x2−x1)}. be able to. Further, when there are three or more sampled points, a is obtained as an inclination amount by substituting each coordinate into an approximate straight line equation, that is, ax + by + c = 0, and obtaining each coefficient by the least square method or the like. it can.
The amount of inclination may be calculated from one side of the print sheet, but by calculating from a plurality of sides and averaging them, a more accurate amount of inclination can be obtained. In addition, when calculating the inclination from a plurality of sides, the amount of inclination of the side with the smallest error between the sampling point and the approximate line may be employed.

(B) Image Simplification Step In the image simplification step, the captured color image is simplified and a simplified image 52 is created. As a simplification method, a simplified image 52 is created by extracting at least one of a brightness image, an R image, a G image, or a B image from a color image, and converting the extracted image into a black and white image. The method of doing is mentioned. For example, when the printing color is red, it is preferable to use the B image rather than the lightness image, and a monochrome image (simplified image 52) having a high contrast can be obtained.
It should be noted that other methods may be used as a method for simplifying a color image. Further, when the photographed image is not a color image but a black and white image, the following image simplification step is not performed. In (C) the printing position detection step, the captured image itself may be used instead of the simplified image 52.

(C) Print Position Detection Step In the print position detection step, the simplified image 52 obtained in the (B) image simplification step is rotated by the inclination amount calculated in the (A) inclination amount producing step, and the inclination is obtained. And the print position is detected using the tilt-corrected image 53 with the tilt corrected.
As a printing position detection method, for example, a method such as pattern matching can be employed.

(D) Mask image processing step In the mask image processing step, a mask pattern 62 of the printing portion inputted in advance is arranged at the printing position detected in the (C) printing position detecting step, and the mask arrangement image 54 is displayed. At the same time, the mask arrangement image 54 is reversely rotated in accordance with the inclination amount calculated in the (A) inclination amount calculating step to complete the mask image 55.

The mask pattern 62 is a binary image having the same shape as the printed shape, and a separately prepared image is used, for example, by binarizing the image of the actual print sheet 1.
Further, when there is a defect in the vicinity of the mask, there is a possibility that the defect is mistakenly recognized as being a mask. Therefore, it is preferable to distinguish the defect from the mask by identifying the defect using the print feature amount. As the print feature amount, a shape such as a print line width or area, a color, or the like can be used.

  A mask image 55 completed by reversely rotating the mask pattern 62 of the mask arrangement image 54 becomes an image in which the mask pattern 62 is arranged at the same position as the printing position of the photographed image. Therefore, the mask image 55 is used. Can be masked.

  By using the mask image 55 obtained in this manner and masking the image from which the defect candidate points are detected, it becomes possible to detect a defect existing in a portion other than the printing unit 2. In the printing unit 2 as well, defects other than the printing unit can be masked with an image obtained by inverting the mask image, and then a defect can be detected using an appropriate threshold different from the above.

  As described above, according to the image processing means according to the present embodiment, the color image is simplified to a black and white image, and the simplified image is rotated, so that the color image is directly rotated. The amount of calculation required for rotation can be greatly reduced, and higher-speed processing is possible.

  Specifically, as compared with the case of rotating a 24-bit color image of 8 bits × three colors (RGB), in this embodiment, rotation and creation of an 8-bit monochrome image (simplified image) of 8 bits × single color (black and white) are created. It is only necessary to rotate the 1-bit mask image, and the amount of calculation required for the rotation process can be 9/24.

  Furthermore, according to the present embodiment, it is possible to accurately capture the print shape in order to create a mask image based on the rotation-corrected image. As a result, the print shape, area, width, In addition, since it is possible to accurately compare with feature values such as circularity, it is possible to check whether the print extracted as a mask pattern and the print that should be originally have the same shape, and defects such as missing prints. It can also be inspected.

  Further, according to the present embodiment, since the photographed color image is used for inspection without being rotationally corrected, the defect contrast is not reduced as in the case where the inspection image is rotated.

  In the above-described embodiment, the case where (B) the image simplification step is performed after the (A) inclination amount calculation step in the mask image creation process has been described, but the present invention is not limited to this, and (B) the image After the simplification step, (A) an inclination amount calculation step may be performed.

The top view which showed an example of the structure of the color printing sheet used as the test object of this invention. The conceptual diagram which showed one form of the defect inspection apparatus used in this invention. The figure which showed an example of the image processing process in an image processing means. The flowchart which showed the mask creation procedure in a mask image creation step.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Color printing sheet 2 Printing 10 Defect inspection apparatus 11 Conveying means 12 Illumination 13 Imaging means 14 Image processing means 51 Input image 52 Simplified image 53 Tilt correction image 54 Mask arrangement image 55 Mask image 62 Mask pattern

Claims (2)

A method for inspecting a printed material by detecting a defect of the printed material by photographing an image of a printed material on which printing has been performed and performing image processing on the captured image,
The image processing is
An inclination amount calculating step for calculating the inclination amount of the printed material from the image of the printed material,
A printing position detecting step of rotating and correcting an image captured according to the calculated amount of inclination, and detecting a printing position using the corrected image;
A mask image processing step of creating a mask arrangement image by arranging a mask pattern inputted in advance at the detected printing position, and reversely rotating the mask arrangement image by the inclination amount to complete a mask image;
A method for inspecting a printed matter, wherein the photographed image is masked using the mask image. When the captured image is a color image,
Further comprising an image simplification step for simplifying the captured image to create a simplified image;
2. The printed matter inspection method according to claim 1, wherein, in the printing position detecting step, the simplified image is used instead of the photographed image.

Images