JP4507762B2 - Printing inspection device - Google Patents

Description

  The present invention relates to a print inspection apparatus that inspects whether original image data is correctly printed on a sheet.

  As a device for inspecting whether original image data is correctly printed on paper, a processing method in which printed paper is read by a scanner and the read image data is compared with separately stored original image data is often used. It is used. As such a prior art, there exists a thing as shown, for example in patent document 1. FIG.

  Patent Document 2 discloses a method for detecting defects in a print image based on read image data alone, not collation with document image data. In this method, the scanned image is divided into multiple regions, the spatial frequency component is obtained for each region, and the power spectrum of the high frequency component is based on the property that the power spectrum of the high frequency component becomes large if there is no image quality defect. Defects are detected from the magnitude of the spectrum.

  Patent Document 3 discloses a technique for detecting a defect in a printing result in which the same character string is repeatedly printed with high definition. In this method, the spatial frequency in the character string direction is obtained in the read image obtained by reading the print result. If the same character string is repeated, the size and spacing of the characters are almost constant, so that the spectrum value (power value) at a spatial frequency other than the low frequency component has a peak. On the other hand, in the character string of the print defect “collapsed” or “cut”, the peak is almost concentrated in the direct current component region, and the power other than the low frequency region does not increase. Using this difference, the method of Patent Document 3 determines a pattern defect.

Japanese Patent Application Laid-Open No. 11-078183 Japanese Patent Laid-Open No. 10-048149 Japanese Patent Laid-Open No. 08-159985 (particularly the third embodiment)

  There is one problem with collation between a document image and a read image disclosed in Patent Document 1 or the like. This is a problem that the read image is deteriorated more than the printed image printed on the paper due to the spatial frequency characteristics of the reading device such as a scanner. This deterioration is particularly noticeable in the case of an image containing a high spatial frequency component. For example, in recent years, bar codes having a very small size have been used. However, as the bar code becomes smaller, the interval between the bars becomes closer and the spatial frequency becomes higher. In addition to this, there are an increasing number of cases where images with a high spatial frequency such as various code images such as glyph codes and graphic images represented by a collection of thin lines are printed. In the case of an image having a spatial frequency denser than the resolution of the reading apparatus, the deterioration of the reading result becomes significant.

  For example, even if a barcode as shown in FIG. 8A is printed on the paper surface, if it is read by a reading device, the high frequency component becomes dull, and the read image is as shown in FIG. 8B. The image becomes blurred.

  Here, in a field where characters and symbols are mainly printed, such as form printing, the original document image is a monochrome binary image. For this reason, in the collation, the read image is binarized and then compared with the original image. However, since the read image is deteriorated as described above, the binarization result is often different from the print image actually printed on paper. Even if the printed image on the paper is the original image shown in FIG. 9A correctly printed without defects, the binarized result of the deteriorated read image may be as shown in FIG. 9B. is there. In such a case, when both are collated, as shown in FIG. 9C, many differences are generated, and it may be determined that the printing is defective. In this case, a correctly printed one is determined to be defective, and therefore it is a collation error. In FIG. 9C, pixels shown in black are pixels that differ between the binarization results of the original image and the read image.

  Although the case where the read image is binarized and collated has been described above, even when the read image is collated as it is without being binarized, the influence of the spatial frequency characteristics of the reading device appears on the read image. The same is true in that this effect has an unfavorable effect on matching.

  There is a need to reduce collation errors due to image degradation due to the spatial frequency characteristics of such a reader.

  The methods disclosed in Patent Documents 2 and 3 are inspection methods that focus on the spatial frequency of a read image, but are not applicable to a method of collating a document image with a read image. Further, these methods in Patent Documents 2 and 3 focus on the spatial frequency component to be inspected such as a character string, and do not pay attention to the spatial frequency characteristics of the reading device.

The print inspection apparatus according to the present invention includes collation means for collating read image data obtained by reading a print image on printed paper with a reading apparatus against original image data based on the print image. A print inspection apparatus for determining the quality of a printed image on printed paper based on a collation result of the collating means, and collating the original image data by applying a spatial frequency filter indicating a spatial frequency characteristic of the reading apparatus. Filter means for generating original document image data, and the collating means first executes collation between the original image data and the read image data, and the original image data obtained by the collation and the read image data when the degree of difference exceeds the allowable value, matching the said read image data verification document image data generated by said filter means And wherein the door.

In the reference example , the image forming apparatus further includes binarization means for binarizing the read image data and the original document image data for verification, and the verification means binarizes the read image data and the original image data for verification. And then collate.

In another aspect , the print inspection apparatus according to the present invention collates read image data obtained by reading a printed image on printed paper with a reading device against original image data based on the printed image. A print inspection apparatus for determining the quality of a printed image on a printed paper based on a result of the collation by the collating means, wherein a spatial frequency filter indicating a spatial frequency characteristic of the reading apparatus with respect to document image data is provided. Filter means for generating document image data for collation by acting, the collating means first executes collation between the document image data and the read image data, and the document image data obtained by this collation When the degree of difference from the read image data exceeds an allowable value, the degree of difference from the read image data in the original image data exceeds a predetermined value. That extracts a high variance area, the controls before notated filtering means so as to apply a spatial frequency filter only the high difference region in the original image data, only the high difference region obtained as a result is filtered The collation document image data is collated with the read image data.

  The best mode for carrying out the present invention (hereinafter referred to as “embodiment”) will be described below with reference to the drawings.

  First, a configuration of a printing system to which a printing inspection apparatus according to the present invention is applied will be described with reference to FIG.

  In this printing system, the paper feed unit 10 feeds the stored paper to the paper transport path 15 in accordance with an instruction from the print controller 40. The developing unit 20 forms an image of the original image data supplied from the print controller 40 on the paper transported on the paper transport path 15. The fixing unit 25 fixes the document image formed on the paper to the paper. In the present invention, the method of the developing unit 20 and the fixing unit 25 is not particularly limited. Various units such as an electrophotographic system and an inkjet system can be used as the developing unit 20, and various types of systems such as thermal fixing can be used as the fixing unit.

  The image reading device 30 is a device that optically reads printed paper on which an image is printed by the developing unit 20 and the fixing unit 25, and includes a light source 35 for illumination, a sensor 37 for reading reflected light from the paper, and An optical system 36 for imaging the reflected light from the paper on the sensor 37 is provided. In this example, the sensor 37 is a CCD (Charge Coupled Device) type line sensor, and is arranged in such a positional relationship that the photosensor array is perpendicular to the sheet conveyance direction below the image reading device 30. Yes. The image reading device 30 is provided on the paper conveyance path 15 at a position behind the fixing unit 25 and before the branch point 19 with the paper reversing path 17 for duplex printing. The image reading device 30 uses sub-scanning as the printed paper is transported on the lower paper transport path 15 and reads out the detection output of each sensor in the photosensor array in the meantime (main scanning). ) Is repeated to read the image on the top surface of the printed paper.

  The print inspection device 50 collates image data read by the image reading device 30 (referred to as read image data) with document image data supplied from the print controller 40 and inspects the quality of the print result. The quality inspection result of the printing result of each sheet is recorded in a recording device (not shown). Further, based on the inspection result, it can be controlled to discharge the non-defective product and the defective product to different paper discharge trays.

  At the time of such a print inspection, the print inspection apparatus 50 according to the present embodiment performs processing in consideration of the spatial frequency characteristics of the image reading apparatus 30. In other words, in the present embodiment, the original image data used for printing is subjected to a spatial frequency filter corresponding to the spatial frequency characteristic of the image reading device 30, and the resulting original document image data (for collation) The original image data is referred to as read image data. Such spatial frequency filter data is registered in advance in the print inspection apparatus 50. Since the document image data for collation is affected by the same spatial frequency characteristics of the image reading device 30 as the read image data, if the print results are correct (no defects), the images are very similar to each other. . Therefore, if the print result is correct, the collation is successful and the product is correctly determined to be good.

  FIG. 2 is a flowchart showing the processing procedure of the printing system of this embodiment.

  In this procedure, first, a print instruction is issued from the print controller 40 (S10), original image data indicating an original image to be printed is supplied to the developing unit 20, and the original image is printed on the fed paper. (S12). FIG. 3 shows an example of the document image 100. Here, a case where a binary image such as a character or a barcode is used as a document as in the example of FIG. 3 is taken as an example. The original image 100 includes an image of a bar code 110 including a high spatial frequency component. The barcode image shown in FIG. 8A is an enlarged image of the barcode 110 shown in FIG.

  The paper on which the original image is printed is conveyed to the image reading device 30, and the print image on the upper surface of the paper is optically read to generate read image data (S14). The read image data is, for example, an image with 256 gradations.

  In parallel with the process of printing the original image (S12) and reading the print image (S14), the print inspection apparatus 50 receives the original image data supplied from the print controller 40, and the original image data includes the image. By applying a filter indicating the spatial frequency characteristics of the reading device 30, data of a document image for collation is created (S20).

  FIG. 4A shows a spatial frequency filter pattern corresponding to the spatial frequency characteristics of the image reading apparatus. In the illustrated filter pattern, a horizontal (x) axis indicated by an arrow indicates a horizontal spatial frequency, and a vertical (y) axis indicates a vertical spatial frequency. The origin, which is the intersection of the horizontal axis and the vertical axis, is a point where the spatial frequency in both the horizontal direction and the vertical direction becomes 0, and the spatial frequency increases as the distance from the origin increases. In this filter pattern, the white portion (that is, the higher brightness) has a higher MTF (Modulation Transfer Function) value indicating the spatial frequency characteristics of the image, and the black portion has a lower MTF. That is, this filter pattern means that the lower the spatial frequency, the higher the MTF, that is, the higher the contrast between light and dark. FIG. 4C shows a pattern of the spatial frequency filter pattern shown in FIG. 4A as a plotter in a three-dimensional space with the spatial frequency taken on the z axis.

  The spatial frequency characteristics of such an image reading apparatus are determined by the combination of the mounted imaging device (for example, a line sensor or area sensor using a CCD) and an optical system, and this can be obtained by measurement or calculation. Therefore, a spatial frequency filter can be created according to this.

  In S20, the raw document image supplied from the print controller 40 is converted into spatial frequency domain data by two-dimensional Fourier transform by a process such as DFT (Discrete Fourier Transform), and the data of each frequency of the conversion result. On the other hand, by multiplying the MTF value of the corresponding frequency in the above spatial frequency filter, the filter is applied to the document image. Then, the result of this calculation is subjected to inverse Fourier transform to restore the original image data in the spatial region. As a result, document image data for verification is generated. The above process will be described in more detail as follows.

  That is, the sample interval in the x direction and the y direction of the raw document image (= reciprocal of the printing resolution) is Δx and Δy, respectively, and M discrete sample values in the x direction and N discrete sample values in the y direction are obtained. If the image data f (mΔx, nΔy) (where m = 0, 1, 2,... M−1, n = 0, 1, 2,... N−1) is given, The discrete Fourier transform is defined by the following formula 1.

  However, in general, f (mΔx, nΔy) in Formula 1 is replaced with f (m, n) and F (kΔu, lΔu) is replaced with F (k, l), and the following Formula 2 is obtained. Used.

  F (k, l) is a document image that is decomposed into spatial frequency components. When the MTF of the image reading apparatus actually measured in advance is represented by T (k, l), the spatial frequency component F ′ (k, l) of the collation original image is obtained by the following Equation 3.

  Therefore, the document image for verification f ′ (k, l) can be obtained by the inverse discrete Fourier transform expressed by the following mathematical formula 4.

  In the above example, the original image is decomposed into spatial frequency components and then the spatial frequency filter is applied. However, as another simple method, the original original image is not decomposed into such spatial frequency components. It is also possible to create a collation document image by applying digital filter processing such as a local weighted average filter to the data.

  By such a filtering process, a collation document image as shown in FIG. 4B is generated from the binary document image as shown in FIG.

  As described above, when the read image data of the print result and the original document image data for matching corresponding to the original document image are prepared, the print inspection apparatus 50 respectively converts them into appropriate binarization thresholds. Binarization is performed using the values (S16). Each binarization threshold value can be determined according to the content of the image data by a conventionally known method. Then, the binarized read image data and collation image data are collated (S18). In the collation, for example, two binarized images are subtracted in units of pixels. Assume that the binarization result of the collation document image is as shown in FIG. 5A, for example, and collated (difference calculation) with the binarization result of the read image (FIG. 5B). In this case, if printing is performed correctly and there are no defects such as print stains or missing prints, the difference between them is zero over almost all pixels. Therefore, if the number of pixels other than 0 in the difference image is less than the predetermined allowable upper limit, it can be determined that the print result matches the original image, and the print result is classified as non-defective.

  In the case of a pixel corresponding to printing smear, the read image is black (value is 1) but the collation document image is white (value is 0). Therefore, when the collation document image is subtracted from the read image, the pixel is Black (value is 1). Conversely, pixels that are black in the original image but white in the read image, such as missing or missing prints, become black when the read image is subtracted from the original image for verification. Accordingly, by performing the process of subtracting the original document image for comparison from the read image and the reverse calculation process in the collation process of S16, defects such as printing smudges and omissions can be detected.

  In this collation process, it is preferable to use a template matching calculation method such as a known residual sequential test method in order not to determine that a slight misalignment between the read image and the document image is defective. That is, in the template matching method, when the scanned image and the binarized images of the matching document image are matched, the difference between the two images is obtained while gradually shifting the mutual positional relationship, and the number of difference pixels is the smallest. It is determined that the two are matched at the position at the time, and the quality of the printing result is determined based on the number of difference pixels at that time.

  Although the example of the collation method was given above, of course, the collation method is not limited to this.

  As described above, in this embodiment, the spatial frequency filter corresponding to the spatial frequency characteristic of the image reading device 30 is applied to the document image data used for printing, and the filter result is collated with the read image. Therefore, it is possible to cancel the influence of the spatial frequency characteristics of the image reading apparatus at the time of reading. As a result, it is possible to reduce problems such as being determined to be defective despite the fact that printing has been performed correctly.

  Next, a modified example of the processing of the printing system according to the present embodiment will be described with reference to FIG. In the procedure of FIG. 6, steps that perform the same processing as the procedure of FIG. 2 are assigned the same reference numerals, and description thereof is omitted.

  Although the processing method of the above embodiment has an effect of reducing the problem of determining that a high spatial frequency image portion such as a barcode is correctly printed even though it is correctly printed, it requires a certain amount of time for the filtering process. There is a problem that it takes. In particular, when filter processing is implemented by software, the problem of processing time becomes significant. Considering the case of inspecting the printing result of a high-speed printing apparatus, the time required for the inspection process should be as short as possible. In the above embodiment, since the spatial frequency filter is uniformly applied to a portion of a normal character or the like that is not a high spatial frequency, there is no possibility that it will affect collation. Considering this, it is often considered that it is better not to perform filtering on images that are not at high spatial frequencies. For example, when inspecting the print result of a print job in which a page including a high spatial frequency image such as a barcode and a page other than that is mixed, the above filtering process is not performed on a page that does not include the high spatial frequency image. The process is faster when the check is performed. This modification proposes a processing method considering such points.

  In the procedure of this modified example, the procedure up to the point where the document image is printed and read by the image reading device 30 (S10 to S14) is the same as the procedure of FIG. Thereafter, in the procedure of FIG. 2, the scanned image and the separately generated collation document image are binarized and collated, but in this modification, the scanned image is first subjected to the spatial frequency characteristics of the image reading apparatus. Collation with a raw document image that is not considered, and collation with a collation document image is performed only when it is not determined that the two match in this collation.

  That is, after S14, the raw document image and the read image are binarized (S30), and the binarized results are collated using a known collation method such as a residual sequential test method (S32). Then, as a result of this collation, it is determined whether or not the number of difference pixels (that is, pixels having different pixel values between the two binarized images) is equal to or greater than a predetermined allowable value (S34).

  If it is determined in S34 that the difference pixel number is not equal to or greater than the predetermined allowable value, it can be determined that the print result matches the original image at this stage, so the print result is determined to be non-defective and the process is terminated. If a page that does not contain a high spatial frequency image such as a barcode is a good print result, the collation is successful at this stage and the process can be completed.

  On the other hand, in the case of a page including a high spatial frequency image, it is highly likely that the difference pixel number is determined to be greater than or equal to the allowable value in S34. In this case, considering the possibility that such a large difference is caused by the spatial frequency characteristics of the image reading apparatus 30, the process proceeds to S36 and subsequent steps.

  In S36, a spatial document filter corresponding to the spatial frequency characteristics of the image reading device is applied to the raw document image data before binarization to create a collation document image. Then, the document image for collation is binarized (S38), and this is collated with the binarization result of the read image obtained in S30 (S40). The collation process at S40 may be the same as the process at S18 in the procedure of FIG.

  As described above, according to the procedure shown in FIG. 6, only the page including the high spatial frequency image is subjected to the filtering process. Therefore, in the print job in which the page including the high spatial frequency image is mixed with the page not including the high spatial frequency image, Processing time can be shortened. In addition, since pages that do not include a high spatial frequency image are collated without performing filter processing, it is possible to reduce the possibility of introducing errors in collation due to the influence of the filter processing.

  In this modified example, a page including a high spatial frequency image is subjected to a spatial frequency filter on the entire page. However, this also applies to a character or the like that is not a high spatial frequency. In order to avoid this, it is also preferable to extract a region having a high spatial frequency from the page and apply a spatial frequency filter only to this region.

  For this purpose, for example, an operator of the printing system may register in advance a region of a high spatial frequency portion such as a barcode in a page with respect to the print inspection apparatus. There is a problem of becoming. Therefore, for example, the following method can be used as a method for automatically detecting a region of high spatial frequency.

  That is, in this method, the difference value for each pixel of the original image and the read image is obtained in S32 of the procedure of FIG. 6, and a difference image having the absolute value of the difference value of each pixel as the pixel value is generated. For example, in the difference image 200 between the original image including the barcode as shown in FIG. 3 and the binarized result of the read image, if printing is performed correctly, as shown in FIG. Both coincide with each other, the pixel value becomes 0, and the pixels having the pixel value of 1 concentrate on the barcode portion 205. The difference image 200 is projected in the vertical and horizontal directions to generate a vertical projection 210 and a horizontal projection 220. The vertical projection 210 is a histogram obtained by summing the pixel values of each column of the difference image 200 in the vertical direction (this sum indicates the number of difference pixels in the column). The same applies to the horizontal projection 220. In the case of a document image including a high spatial frequency portion such as a barcode, both the vertical projection 210 and the horizontal projection 220 have a large value in a range corresponding to the barcode portion 205 as shown in FIG. Therefore, for example, a range (x1, x2) in which the value of the vertical projection 210 exceeds the predetermined threshold Th1 can be obtained, and this range can be set as the horizontal range of the high spatial frequency portion. More preferably, in consideration of an error, a range obtained by expanding the range (x1, x2) thus obtained on both sides by a predetermined ratio may be set as a filter target range. The range of the filter target in the vertical direction can be similarly obtained from the horizontal projection 220. If the range in the vertical direction and the range in the horizontal direction are determined, a rectangular range obtained by combining them is the target of the filtering process. It should be noted that the threshold values Th1 (for vertical projection) and Th2 (for horizontal projection) used at this time may be determined by experiments or the like.

  In this improved procedure, when the difference pixel number is determined to be equal to or larger than the predetermined value in the determination in S34, the high spatial frequency region to be filtered is specified by the method described in FIG. In S 36, a collation original image is generated by applying a spatial frequency filter corresponding to the spatial frequency characteristic of the reading device only to this region of the original image. Then, the document image for collation is binarized in S38 and collated with the binarization result of the read image in S40.

  By such processing, a collation original image can be created by applying a spatial frequency filter selectively only to the high spatial frequency portion of the original image, and this can be collated with the read image.

  The preferred embodiments of the present invention and the modifications thereof have been described above. In all the examples shown above, the scanned image and the original image are binarized and collated, but as will be apparent to those skilled in the art, the scanned image and the original image are not binarized and collated. In this case, a method of applying a spatial frequency filter corresponding to the spatial frequency characteristic of the reading apparatus to the original image and performing collation is applicable and preferable.

1 is a diagram showing a schematic configuration of a printing system to which a printing inspection apparatus according to the present invention is applied. 6 is a flowchart illustrating a processing procedure of the printing system according to the embodiment. It is a figure which shows an example of the original image containing a barcode. It is a figure which shows the result of having applied this filter to the spatial frequency filter according to the spatial frequency characteristic of an image reader, and the original image of a barcode. It is a figure for demonstrating image collation with the reading image and the original document image for collation. 10 is a flowchart illustrating a modification of the processing procedure of the printing system. It is a figure for demonstrating the determination method of a high spatial frequency area | region. FIG. 6 is a diagram for explaining a difference between a document image and a read image. It is a figure for demonstrating the image collation with a read image and a document image.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Paper feed unit, 15 Paper conveyance path, 20 Developing unit, 25 Fixing unit, 30 Image reader, 40 Print controller, 50 Print inspection apparatus.

Claims (2)

A collating unit that collates read image data obtained by reading a print image on printed paper with a reading device with original image data that is the basis of the print image, and based on a collation result of the collating unit. A print inspection apparatus for determining the quality of a printed image on printed paper,
Filter means for generating document image data for collation by applying a spatial frequency filter indicating the spatial frequency characteristics of the reading device to the document image data;
With
The collation means first performs collation between the document image data and the read image data, and when the degree of difference between the document image data and the read image data obtained by the collation exceeds an allowable value, printing inspection apparatus characterized in that matching verification document image data generated by the filter means and the read image data. A collating unit that collates read image data obtained by reading a print image on printed paper with a reading device with original image data that is the basis of the print image, and based on a collation result of the collating unit. A print inspection apparatus for determining the quality of a printed image on printed paper,
Filter means for generating document image data for collation by applying a spatial frequency filter indicating the spatial frequency characteristics of the reading device to the document image data;
With
The collation means first performs collation between the document image data and the read image data, and when the degree of difference between the document image data and the read image data obtained by the collation exceeds an allowable value, the degree of difference between the read image data is extracted a high difference area exceeding a predetermined value in the original image data, so as to act on the spatial frequency filter only the high difference region of the original image data before notated controls filter means, as a result only the high difference region obtained printing inspection apparatus you characterized in that to match with the read image data of the original image data for collation is filtered.