Image processing device, image processing method, and program

Abstract

To inspect a printed matter with high accuracy even when defects come close to each other.SOLUTION: An information processing device concerning the present invention comprises: acquisition means for acquiring first image data representing a reference image which is a target print result and second image data representing an object image which is an inspection object; and processing means for performing correction for more emphasizing a second subregion in line with a first subregion different from the reference image in the object image than a difference image representing difference between the reference image and the object image by using the first image data and the second image data to inspect the object image.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to an image processing technology for inspecting printed matter. Background Art

[0002] In some cases, it is required to guarantee the quality of printed matter output by a printing apparatus by inspecting whether there are defects such as stains or color missing in the printed matter. Visual inspection, which visually inspects whether there are defects in printed matter, requires a lot of cost, so inspection systems that perform automatic inspection have been developed. In such an inspection system, the presence or absence of a defect is determined based on a difference image representing a difference between a reference image serving as a pre-registered inspection standard for printed matter and an inspection target image obtained by scanning the printed matter to be inspected. Patent Document 1 discloses determining the presence or absence of a defect in printed matter based on the magnitude of variation in pixel values in the aforementioned difference image. Prior Art Literature Patent Literature

[0003] Patent Document 1 Japanese Unexamined Patent Application Publication No. 2019-158757 Summary of the Invention Problems to be Solved by the Invention

[0004] However, according to the technology disclosed in Patent Document 1, the magnitude of variation in pixel values in a difference image is acquired by using a difference of pixel values at a predetermined interval, etc. Therefore, when defects are adjacent to each other at the predetermined interval, the determination accuracy may decrease.

[0005] Accordingly, an object of the present invention is to inspect printed matter with high accuracy even when defects are adjacent to each other. Means for Solving the Problems

[0006] The image processing apparatus according to the present invention includes: acquisition means for acquiring first image data representing a reference image which is the target printing result and second image data representing a target image which is the object to be inspected; generation means for generating a difference image which represents the difference between the reference image and the target image; and processing means for correcting the difference image using a first filter which has a lower degree of emphasis when applied to both the first partial region and a second partial region adjacent to the first partial region than when applied to the first partial region, and a second filter which has a higher degree of emphasis when applied to both the first partial region and the second partial region than when applied to the first partial region, and for inspecting the target image based on the corrected difference image. The processing means performs the correction using the sum of the result of applying the first filter to the difference image and the result of applying the second filter to the difference image. It is characterized by the following: [Effects of the Invention]

[0007] Even when defects are in close proximity, printed materials can be inspected with high precision. [Brief explanation of the drawing]

[0008] [Figure 1] A diagram showing the configuration of a printing system including an image processing device. [Figure 2] A block diagram showing the configuration of an image processing device. [Figure 3] A flowchart illustrating the processes performed by an image processing unit. [Figure 4] A diagram showing examples of reference images and images to be examined. [Figure 5] A diagram showing an example of detection sensitivity. [Figure 6] A diagram showing an example of a difference image. [Figure 7] A diagram showing an example of a defect-enhanced image. [Figure 8] A diagram showing an example of a defect map. [Figure 9] A flowchart illustrating the process of highlighting difference images. [Figure 10] A diagram illustrating an example of an enhancement filter. [Figure 11] A diagram showing an example of test sensitivity. [Figure 12] A diagram showing an example of a defect-enhanced image. [Figure 13] A diagram showing an example of a defect map. [Figure 14] A flowchart illustrating the process of highlighting difference images. [Figure 15] A flowchart illustrating the process of highlighting difference images. [Figure 16] A diagram showing an example of a defect-enhanced image. [Modes for carrying out the invention]

[0009] The following embodiment will be described with reference to the drawings. Note that the following embodiment is not necessarily limiting to the present invention. Furthermore, not all combinations of features described in this embodiment are essential to the solution of the present invention.

[0010] [First Embodiment] In this embodiment, a predetermined sub-region corresponding to a defect in the target image (a region different from the reference image, or a region not present in the reference image) is emphasized compared to the difference image between the reference image, which is the target print result, and the target image, which is the object to be inspected. Therefore, a defect-enhanced image is generated by applying a correction process to the difference image so that the sub-region is emphasized. Based on this enhanced image, the defect region is extracted, and the printed material corresponding to the target image is inspected. The enhancement process uses an enhancement filter that performs image processing to enhance the sub-region corresponding to a single defect (first sub-region), and an enhancement filter that corresponds to a nearby defect, which is a sub-region located next to the first sub-region in the target image (second sub-region). Using these filters, the difference image is enhanced so that the difference corresponding to the defect is emphasized. Note that the inspection in this embodiment is performed based on the absolute value of the difference between the reference image and the target image. Furthermore, the nearby defects may be adjacent to the first sub-region, or they may be spaced apart by a predetermined width, for example, the width of the first sub-region.

[0011] [Printing System Configuration] FIG. 1 is a diagram showing a configuration example of an entire printing system that includes an image processing apparatus 100 and performs output and inspection of printed matter. The printing system according to the present embodiment includes the image processing apparatus 100, a printing server 180, and a printing apparatus 190. The printing server 180 generates a print job of a document to be printed and inputs the print job to the printing apparatus 190. The printing apparatus 190 forms an image on a recording medium (printing paper) based on the print job input from the printing server 180. The printing apparatus 190 has a paper feeding unit 191, and a user sets printing paper in the paper feeding unit 191 in advance. When a print job is input, the printing apparatus 190 conveys the printing paper set in the paper feeding unit 191 along a conveyance path 192, forms an image on the surface (one side or both sides) thereof, and sends the printing paper to the image processing apparatus 100. Although the printing apparatus 190 according to the present embodiment is an electrophotographic printing apparatus, it may be an offset printing type or inkjet type printing apparatus.

[0012] The image processing apparatus 100 performs defect inspection on a printed matter to be inspected that has been subjected to printing. The printed matter to be inspected obtained by the printing apparatus 190 forming an image on printing paper is conveyed along the conveyance path 192 and inspected by the image processing apparatus 100. The image processing apparatus 100 functions as an inspection processing apparatus. The image processing apparatus 100 includes a CPU 101, a RAM 102, and a ROM 103. The image processing apparatus 100 also includes a storage device 104, an image reading device 105, a printing interface (I / F) 106, a general-purpose interface (I / F) 107, a user interface (UI) panel 108, and a main bus 109. Furthermore, the image processing apparatus 100 includes a printed matter conveyance path 110 connected to the conveyance path 192 of the printing apparatus 190, an output tray 111 for printed products that have passed inspection, and an output tray 112 for printed matter in which a defect is found and that has failed inspection. Note that the classification of printed matter is not limited to two types of acceptable and unacceptable, and may be configured to classify printed matter more finely. In addition, in the printing system, the storage device 104, the image reading device 105, the UI panel 108, the conveyance path 110, the output tray 111, and the output tray 112 may be provided outside the image processing apparatus 100.

[0013] The CPU 101 is a processor that comprehensively controls each part of the image processing apparatus 100. The RAM 102 functions as a main memory, a work area and the like for the CPU 101. The ROM 103 stores a program group executed by the CPU 101. The storage device 104 stores applications executed by the CPU 101, data used for image processing, and the like. The image reading device 105 is a scanner, which reads one or both sides of a printed matter sent from the printing device 190 on the conveyance path 110, and acquires the read result as image data.

[0014] The printing I / F 106 is connected to the printing device 190, and is an interface for synchronizing the processing timing of printed matters between the image processing apparatus 100 and the printing device 190 and notifying each other of operating statuses. The general-purpose I / F 107 is a serial bus interface such as USB, IEEE1394 or the like, which enables a user to carry data such as logs. The UI panel 108 is a display device such as a liquid crystal display, and functions as a user interface for informing the user of the current status and settings of the image processing apparatus 100. In addition, the UI panel 108 may be provided with an input device such as a touch panel or buttons, and can receive instructions from the user regarding inspection and the like. Note that the input device may be provided separately from the UI panel 108, such as a mouse or a keyboard. The main bus 109 is a transmission path that connects each module of the image processing apparatus 100.

[0015] The image processing apparatus 100 conveys the printed matter sent from the printing device 190 through the conveyance path 110, and performs the inspection processing described below based on the image data of the printed matter read by the image reading device 105. If the printed matter passes the inspection, it is conveyed to a passed output tray 111; if the printed matter fails the inspection, it is conveyed to a failed output tray 112. Accordingly, only printed matters that are confirmed to meet quality standards can be collected in the output tray 111 as printed matters for delivery.

[0016] [Functional Configuration of Image Processing Apparatus] Figure 2 shows the functional configuration of the image processing device 100. The image processing device 100 includes a reference image setting unit 201, a detection sensitivity setting unit 202, an image acquisition unit 203, a difference image generation unit 204, a difference image enhancement unit 205, and an inspection processing unit 206.

[0017] The reference image setting unit 201 sets an image represented by image data recorded in RAM 102 or storage device 104 as the reference image. The detection sensitivity setting unit 202 sets the detection sensitivity for each defect type based on instructions from the user obtained via the UI panel 108. The image acquisition unit 203 acquires inspection target image data representing the inspection target image, obtained by the image reading device 105 reading printed material on the transport path 110. The acquired inspection target image data is stored in RAM 102 or storage device 104. The difference image generation unit 204 generates a difference image based on the reference image and the inspection target image. The difference image enhancement unit 205 generates a defect-enhanced image by enhancing the difference image using enhancement filters corresponding to individual defects and enhancement filters corresponding to nearby defects, which are set according to the detection sensitivity. The inspection processing unit 206 performs inspection processing on the inspection target image based on the defect-enhanced image and outputs the inspection result.

[0018] [Processing performed by the image processing unit] Figure 3 is a flowchart showing the processes performed by the image processing device 100. Hereafter, each step (process) will be represented by the letter S before it.

[0019] In S301, the reference image setting unit 201 sets a reference image. An example of a reference image is shown in Figure 4(a). The reference image data (first image data) representing the reference image is assumed to be created in advance based on an image output by the printing device 190, or a scanned image obtained by reading a printed document printed by the printing device 190, and recorded in the RAM 102 or storage device 104.

[0020] In S302, the detection sensitivity setting unit 202 sets the detection sensitivity for each defect type based on instructions from the user obtained via the UI panel 108. In this embodiment, the defect type is limited to vertical linear defects (hereinafter also referred to as vertical linear defects). In this embodiment, the detection sensitivity setting unit 203 sets the detection sensitivity for vertical linear defects in two stages, High and Low, with the narrower the vertical linear defect being detected as the detection target, the higher the detection sensitivity. An example of defects to be detected according to the detection sensitivity is shown in Figure 5(a). In this embodiment, when the detection sensitivity is set to High (Figure 5(b)), vertical linear defects with a width of 0.3 mm or more are detected, and when the detection sensitivity is set to Low (Figure 5(c)), vertical linear defects with a width of 0.5 mm or more are detected.

[0021] In S303, the image acquisition unit 203 acquires inspection target image data (second image data) representing the image to be inspected. An example of an inspection target image in this embodiment is shown in Figure 4(b). The inspection target image shown in Figure 4(b) is assumed to contain vertical line-shaped defects, i.e., vertical streak-shaped defects.

[0022] In S304, the difference image generation unit 204 calculates the absolute value of the difference in brightness values ​​between corresponding pixels in the reference image and the image to be inspected, based on the reference image and the image to be inspected, and generates a difference image in which each pixel has the absolute value of the difference in brightness values. An example of a difference image in this embodiment is shown in Figure 6(a). In Figure 6, the difference, i.e., the partial region (first partial region) in which the image to be inspected differs from the reference image, is shown in gray, and the partial region with no difference is shown in black.

[0023] In step S305, the difference image enhancement unit 205 enhances the difference image using enhancement filters corresponding to single defects and adjacent defects, which are set based on the detection sensitivity, to generate a defect-enhanced image. The enhancement process enhances the difference region corresponding to the detected defect. An example of a defect-enhanced image in this embodiment is shown in Figure 7. In the difference image, areas with large differences are represented more white, and areas with small differences are represented more black. When the detection sensitivity is Low, as shown in Figure 7(a), narrow vertical line differences are not enhanced, and only wide vertical line differences are enhanced. When the detection sensitivity is High, as shown in Figure 7(b), both narrow and wide vertical line differences are enhanced. In this embodiment, at least two adjacent vertical line differences are enhanced to the same or greater extent than a single vertical line difference that is not adjacent. Details of the difference image enhancement process will be described later.

[0024] In S306, the inspection processing unit 206 performs inspection processing based on the defect-enhanced image and outputs the inspection result of the image to be inspected. Specifically, pixels in the defect-enhanced image whose pixel value is greater than or equal to a preset threshold are extracted as defect regions and a defect map is generated. An example of a defect map in this embodiment is shown in Figure 8. When the detection sensitivity is Low, thick vertical line differences are extracted as defect regions as shown in Figure 8(a) (regions shown in white in Figure 8). When the detection sensitivity is High, both thin vertical line differences and thick vertical line differences are extracted as defect regions as shown in Figure 8(b). The inspection result of the image to be inspected is considered a pass if there are no defect regions and a fail if there are defect regions.

[0025] In S307, the inspection processing unit 206 determines whether to terminate the process based on print information from the printer 190 and user instructions obtained via the UI panel 108. For example, the process is terminated when the inspection of the number of printed pages or the number of pages instructed by the user is complete. If it is determined that the process should not be terminated, the process proceeds to S303.

[0026] [Difference Image Enhancement Processing] Figure 9 is a flowchart showing the process of enhancing the difference image. In this process, a defect-enhanced image is generated by correcting the difference image using enhancement filters for single defects and enhancement filters for nearby defects, which are set based on the detection sensitivity. First-order differential filters and second-order differential filters can be used as enhancement filters. In this embodiment, a LoG (Laplacian of Gaussian) filter, a type of second-order differential filter, is used. The details of the process of enhancing the difference image will be described below.

[0027] In S901, the difference image enhancement processing unit 205 sets an enhancement filter f(x, σ1) corresponding to a single defect. First, for a linear defect to be detected with the detection sensitivity set in S302, if its width is a, the parameter σ1 is set based on this width a. In this embodiment, σ1 is half the width of the linear defect to be detected, as shown in equation (1).

[0028]

number

[0029] Then, according to equation (2), we calculate the enhancement filter f(x, σ1) corresponding to a single defect.

[0030]

number

[0031] By applying the enhancement filter f(x, σ1) calculated as described above to the difference image, it is possible to enhance the difference corresponding to a linear defect of width a.

[0032] However, when linear defects of width a are close together, f(x, σ1) may not adequately emphasize the differences corresponding to the defects. Therefore, in the next step, an enhancement filter is set to emphasize the differences corresponding to adjacent linear defects of width a.

[0033] In S902, the difference image enhancement processing unit 205 sets an enhancement filter f'(x, σ2) corresponding to nearby defects. First, based on the parameter σ1 of the enhancement filter corresponding to a single defect, the parameter σ2 is set according to equation (3).

[0034]

number

[0035] Here, α is set to a value greater than 1. For example, α = 4 is preferred. Then, the enhancement filter f(x, σ²) is calculated according to equation (2). After that, the enhancement filter f'(x, σ²) is calculated according to equation (4).

[0036]

number

[0037] Here, f'(x, σ2) is calculated by correcting f(x, σ2) so that the sum of its coefficients is equal to the sum of the coefficients of f(x, σ1) in the positive range Ω of the coefficients of f(x, σ1). By applying filtering to the difference image using the enhancement filter f'(x, σ2) calculated in this way, it is possible to enhance the difference corresponding to nearby defects of width a. Furthermore, in the example of this embodiment, by setting α such that the positive range of the coefficients of f'(x, σ2) includes the negative range of the coefficients of f(x, σ1), nearby defects that cannot be sufficiently enhanced by f(x, σ1) can be enhanced by f'(x, σ2).

[0038] Furthermore, the coefficient of f(x, σ1) does not necessarily have to be in the negative range; it can be 0. Moreover, if the positive range of the coefficient of f'(x, σ2) is wider than the positive range of the coefficient of f(x, σ1), nearby defects can be emphasized.

[0039] In this embodiment, an example of the enhancement filter when the detection sensitivity is set to High is shown in Figure 10(a), and an example of the enhancement filter when the detection sensitivity is set to Low is shown in Figure 10(b). The higher the detection sensitivity, the smaller the width of the linear defect to be detected, and therefore the smaller the size of the enhancement filter.

[0040] In S903, the difference image enhancement processing unit 205 enhances the difference image using enhancement filters corresponding to single defects and enhancement filters corresponding to nearby defects to generate a defect-enhanced image. In this embodiment, the defect-enhanced image D′ is calculated according to equation (5).

[0041]

number

[0042] D'1 is the result of enhancing the difference image D with an enhancement filter f(x, σ1) corresponding to a single defect, and the difference corresponding to the single defect is enhanced. D'2 is the result of enhancing the difference image D with an enhancement filter f'(x, σ2) corresponding to a nearby defect, and the difference corresponding to the nearby defect is enhanced. The defect-enhanced image D' is obtained by taking the maximum value between D'1 and D'2 for each pixel, and the difference corresponding to both a single defect and a nearby defect is enhanced.

[0043] [Effects of the First Embodiment] As described above, the image processing device in this embodiment sets a reference image, which is the target print result. It sets the detection sensitivity for defects. It acquires image data representing the image to be inspected. It generates a difference image representing the difference between the reference image and the image to be inspected. It enhances the difference image using enhancement filters corresponding to individual defects and enhancement filters corresponding to nearby defects, which are set based on the inspection sensitivity, to generate a defect-enhanced image. It inspects the image to be inspected based on the detection sensitivity and the defect-enhanced image. This makes it possible to enhance nearby defects to the same or greater extent than individual defects. Therefore, even when defects are close together, the printed material can be inspected with high accuracy.

[0044] [Differentiation] In this embodiment, the detection sensitivity setting unit 202 sets the detection sensitivity in two stages, but the method of setting the detection sensitivity is not limited to this. For example, the detection sensitivity may be set in three or more stages. Also, the detection sensitivity for longitudinal defects may be fixed. Furthermore, the width of the longitudinal defects to be detected is not limited to the example shown in this embodiment; it is sufficient that the width becomes narrower as the inspection sensitivity increases.

[0045] Furthermore, in this embodiment, the difference image enhancement processing unit 205 sets an enhancement filter that is larger the wider the linear defect to be detected in S901 and S902. However, it is also acceptable to correct the enhancement filter to be smaller than the predetermined size so that it does not become too large. Figure 10(c) shows an example in which the enhancement filter is corrected so that it does not become larger than the predetermined size when the detection sensitivity is Low. In this case, the negative coefficients included within the predetermined size are corrected by limiting them with a threshold or the like so that the sum of the negative coefficients is equal to the sum of the positive coefficients. By limiting the correction in this way, it is possible to reduce the filter size while enhancing single defects and nearby defects, thereby reducing computational costs.

[0046] Furthermore, in this implementation, the difference image enhancement processing unit 205 calculates and sets the enhancement filter in S901 and S902. However, it is also acceptable to pre-calculate and record the enhancement filter corresponding to the inspection sensitivity, and then read it out and set it in S901 and S902.

[0047] Furthermore, in this implementation, the difference image enhancement processing unit 205 calculates and sets two enhancement filters in S901 and S902, but it is also acceptable to use three or more enhancement filters. For example, when using three enhancement filters, σ3 can be set to a value greater than σ2, and after calculating the enhancement filter f(x, σ3) according to equation (2), the enhancement filter f'(x, σ3) can be calculated according to equation (4).

[0048] Furthermore, in this implementation, the difference image enhancement processing unit 205 sets an enhancement filter using a horizontal second-order differential filter in S901 and S902, but a vertical smoothing filter may also be used as an enhancement filter. That is, a smoothing filter orthogonal to the enhancement filter is used. As the smoothing filter, mean filters, Gaussian filters, etc., can be used. In this case, the defect-enhanced image D′ is calculated in S903 according to equation (6).

[0049]

number

[0050] Here, g(y) represents the smoothing filter.

[0051] Furthermore, in this embodiment, the difference image enhancement processing unit 205 adopted the maximum values ​​of D'1 and D'2 for each pixel as the defect-enhanced image D' in S903, but the method for generating the defect-enhanced image is not limited to this. For example, the defect-enhanced image D' may also be obtained by taking the sum of D'1 and D'2 for each pixel according to equation (7).

[0052]

number

[0053] Furthermore, in this embodiment, the difference image enhancement processing unit 205 enhances nearby defects to the same or greater extent as single defects by filtering, but nearby defects may also be enhanced to the same or greater extent as single defects by a method other than filtering. For example, the defect to be detected may be used as a template, pattern matching may be performed for each pixel of interest, and the difference of the pixel of interest may be enhanced as the number of differences similar to the defect to be detected in the neighboring region of the pixel of interest increases. Alternatively, the difference of the pixel of interest may be suppressed as the number of differences similar to the defect to be detected decreases in the neighboring region of the pixel of interest.

[0054] Furthermore, in this embodiment, the inspection processing unit 206 extracts defective areas using a preset threshold in the defect-enhanced image, but the threshold may be set according to the inspection sensitivity. Also, areas with small areas may be excluded from the extracted defective areas.

[0055] Furthermore, the inspection processing unit 206 in this embodiment may include the highlighted sub-region described above, extract the region as a defect region, and display the region on the UI panel 108 to notify the user, or transmit it to an external device, such as a printing server 180. In this case, the extracted region may be displayed as a difference image or defect map as described above. Moreover, when displaying, adjacent regions may be highlighted more than the difference image or defect map image. Alternatively, the extracted region may be enclosed in a frame or the like within the target image for display.

[0056] [Second Embodiment] In the first embodiment, the defect type was limited to vertical linear defects. In this embodiment, there are three types of defects: dot-like defects (hereinafter referred to as dot-like defects), vertical linear defects, and horizontal linear defects (hereinafter referred to as horizontal linear defects). The configuration of the printing system and the functional configuration of the image processing apparatus 100 in this embodiment are the same as those in the first embodiment, so their description is omitted. Below, the processes S302, S305, and S306, which differ between this embodiment and the first embodiment, will be mainly described. Components identical to those in the first embodiment will be denoted by the same reference numerals.

[0057] In S302, the detection sensitivity setting unit 202 sets the detection sensitivity for each defect type based on instructions from the user, for example, obtained via the UI panel 108. In this embodiment, there are three types of defects: dot defects, longitudinal line defects, and transverse line defects. In this embodiment, the detection sensitivity setting unit 203 sets the detection sensitivity for each defect type in two stages: High and Low. The higher the detection sensitivity, the smaller the dot defects or the narrower the longitudinal line defects that are detected. An example of defects to be detected according to the detection sensitivity is shown in Figure 11(a). In this embodiment, when the detection sensitivity is set to High, dot defects with a size of 0.3 mm or more or linear defects with a width of 0.3 mm or more are detected. When the detection sensitivity is set to Low, dot defects with a size of 0.5 mm or more or linear defects with a width of 0.5 mm or more are detected. An example of setting the detection sensitivity is shown in Figure 11(b).

[0058] In S305, the difference image enhancement unit 205 enhances the difference image using enhancement filters corresponding to individual defects and enhancement filters corresponding to nearby defects, which are set based on the detection sensitivity for each defect type, to generate a defect-enhanced image. Through the enhancement process, the difference region corresponding to the detected defect is enhanced for each defect type. An example of a reference image in this embodiment is shown in Figure 4(a), an example of an image to be inspected is shown in Figure 4(c), and an example of a difference image is shown in Figure 6(b). The image to be inspected in this embodiment includes point-like defects, vertical line-like defects, and horizontal line-like defects, and the difference image contains differences corresponding to each defect. An example of a defect-enhanced image corresponding to a point-like defect in this embodiment is shown in Figure 12(a). Only individual or nearby point-like differences are enhanced. An example of a defect-enhanced image corresponding to a vertical line-like defect in this embodiment is shown in Figure 12(b). Only individual or nearby vertical line-like differences are enhanced. An example of a defect-enhanced image corresponding to a horizontal line-like defect in this embodiment is shown in Figure 12(c). Only individual or nearby horizontal line-like differences are enhanced. Details of the difference image enhancement process will be described later.

[0059] In S306, the inspection processing unit 206 performs inspection processing based on the defect-enhanced image corresponding to each defect type and outputs the inspection result of the image to be inspected. Specifically, in the defect-enhanced image corresponding to each defect type, pixels whose pixel value is above a preset threshold are extracted as defect regions, and a defect map is generated. An example of a defect map corresponding to point defects in this embodiment is shown in Figure 13(a). Single or adjacent point-like defects are extracted as defect regions. An example of a defect map corresponding to vertical line defects in this embodiment is shown in Figure 13(b). Single or adjacent vertical line defects are extracted as defect regions. An example of a defect map corresponding to horizontal line defects in this embodiment is shown in Figure 13(c). Single or adjacent horizontal line defects are extracted as defect regions. The inspection result of the image to be inspected is considered a pass if there are no defect regions in the defect map corresponding to all defect types, and a fail if there are defect regions in the defect map corresponding to any of the defect types.

[0060] [Difference Image Enhancement Processing] Figure 14 is a flowchart showing the process of enhancing difference images in this embodiment. In this process, enhancement filters corresponding to single defects and enhancement filters corresponding to nearby defects are used, which are set based on the detection sensitivity for each defect type. By enhancing the difference image with these filters, a defect-enhanced image is generated. There are three types of defects in this embodiment: point defects, vertical line defects, and horizontal line defects. When the defect type is a vertical line defect, the process from S1401 to S1403 is the same as S901 to S903 in the first embodiment, so the explanation is omitted. When the defect type is a horizontal line defect, in S1401 and S1402, the enhancement filter corresponding to the horizontal line defect is set by rotating the enhancement filter corresponding to the vertical line defect described in the first embodiment by 90°. Then, in S1403, a defect-enhanced image is generated based on the enhancement filter corresponding to the horizontal line defect. This makes it possible to enhance the difference corresponding to the horizontal line defect. When the defect type is a point defect, it can be handled by using a two-dimensional first-order differential filter or a second-order differential filter as an enhancement filter. In this embodiment, a LoG filter is used. The following describes each process for cases where the defect type is a point-like defect.

[0061] In S1401, the difference image enhancement processing unit 205 sets an enhancement filter f(x, y, σ1) corresponding to a single defect. First, the parameter σ1 is set according to equation (1) based on the size a of the point defect. Then, as shown in equation (8), the enhancement filter f(x, y, σ1) corresponding to the single defect is calculated.

[0062]

number

[0063] By applying the enhancement filter f(x, y, σ1) calculated as described above to the difference image, it is possible to enhance the differences corresponding to point defects of size a.

[0064] However, when point defects of size a are close together, f(x, y, σ1) may not adequately emphasize the differences corresponding to the defects. Therefore, in the next step, an enhancement filter is set to emphasize the differences corresponding to adjacent point defects of size a.

[0065] In S1402, the difference image enhancement processing unit 205 sets an enhancement filter f'(x, y, σ2) corresponding to nearby defects. First, based on the parameter σ1 of the enhancement filter corresponding to a single defect, the parameter σ2 is set according to equation (3). Then, the enhancement filter f(x, y, σ2) is calculated according to equation (8). After that, the enhancement filter f'(x, y, σ2) is calculated according to equation (9).

[0066]

number

[0067] Here, f'(x, y, σ2) is calculated by correcting f(x, y, σ2) so that the sum of its coefficients is equal to the sum of the coefficients of f(x, y, σ1) in the positive range Ω of the coefficients of f(x, y, σ1). By applying a filter to the difference image using the enhancement filter f'(x, y, σ2) calculated in this way, it is possible to enhance the differences corresponding to adjacent point defects of size a. Furthermore, in this embodiment, by setting α such that the positive range of the coefficients of f'(x, y, σ2) includes the negative range of the coefficients of f(x, y, σ1), adjacent defects that cannot be sufficiently enhanced by f(x, y, σ1) can be enhanced by f'(x, y, σ2).

[0068] Furthermore, the coefficient of f(x, y, σ1) does not necessarily have to be in the negative range; it can be 0. Moreover, if the positive range of the coefficient of f'(x, y, σ2) is wider than the positive range of the coefficient of f(x, y, σ1), then nearby defects can be emphasized.

[0069] In S1403, the difference image enhancement processing unit 205 enhances the difference image using enhancement filters corresponding to single defects and enhancement filters corresponding to nearby defects to generate a defect-enhanced image. In this embodiment, the defect-enhanced image D' is calculated according to equation (10).

[0070]

number

[0071] D'1 is the result of enhancing the difference image D with an enhancement filter f(x, y, σ1) corresponding to a single defect, where the difference corresponding to the single defect is enhanced. D'2 is the result of enhancing the difference image D with an enhancement filter f'(x, y, σ2) corresponding to a nearby defect, where the difference corresponding to the nearby defect is enhanced. The defect-enhanced image D' uses the maximum values ​​of D'1 and D'2 for each pixel, where the differences corresponding to both single defects and nearby defects are enhanced. Through the above processing, even when point defects are the target of detection, nearby defects can be enhanced to the same or greater extent than single defects. Therefore, even when defects are close together, printed materials can be inspected with high accuracy.

[0072] In S1404, the difference image enhancement processing unit 205 determines if there are any unprocessed defect types. If there are any unprocessed defect types, the process proceeds to S1401. If all defect types have been processed, the process ends.

[0073] [Effects of the second embodiment] As explained above, a defect-enhanced image is generated by enhancing the difference image using enhancement filters corresponding to individual defects and enhancement filters corresponding to nearby defects, which are set based on the inspection sensitivity for each defect type. This makes it possible to enhance nearby defects to the same or greater extent than individual defects for each different defect type. Therefore, even when defects are close together, the inspection of printed materials can be performed with high accuracy for each different defect type.

[0074] In this embodiment, the defect types are defined as point-like defects, vertical line defects, and horizontal line defects. However, the defect types may be one or more of these, or other defect types may be added. For example, diagonal line defects may be added as linear defects that are tilted at a predetermined angle from the vertical direction. In this case, diagonal line defects can be emphasized by applying an enhancement filter corresponding to vertical line defects at a predetermined angle.

[0075] Furthermore, it is acceptable to set inspection sensitivity levels to avoid inspecting certain defect types. This allows for the selection of inspection types based on the inspection sensitivity level.

[0076] Furthermore, in this embodiment, a two-dimensional filter shown in equation (8) was used as the enhancement filter for point defects, but this may be approximated by a combination of one-dimensional filters. That is, at least one of the enhancement filters for single defects and the enhancement filters for nearby defects can be approximated by a LoG filter based on multiple one-dimensional Gaussian filters according to equation (8), for example.

[0077]

number

[0078] Here, σ- represents a value slightly smaller than σ, and σ+ represents a value slightly larger than σ. By approximating in this way, processing costs can be reduced.

[0079] [Third Embodiment] In the first embodiment, the difference image was enhanced using an enhancement filter for adjacent defects in addition to an enhancement filter for individual defects. Therefore, compared to the case where only the enhancement filter for individual defects was used to enhance the difference image, there was a possibility of over-enhancing the difference in areas that did not contain defects, such as the vicinity of the defect or areas containing normal printing variations such as minute positional shifts, which could cause over-detection. In this embodiment, the difference image enhanced using the enhancement filter for individual defects and the enhancement filter for adjacent defects is corrected based on the difference image smoothed according to the defect type. This suppresses excessive enhancement of the difference. Note that the configuration of the printing system and the image processing apparatus 100 in this embodiment are the same as those in the first embodiment, so their description is omitted. Below, the process of S305, which is different from the first embodiment, will be mainly described. Note that components identical to those in the first embodiment will be described using the same reference numerals.

[0080] Figure 15 is a flowchart showing the process of enhancing the difference image in this embodiment. In this process, the difference image is corrected by using an enhancement filter corresponding to a single defect and an enhancement filter corresponding to a nearby defect, based on the difference image that has been smoothed according to the defect type. Processes S1501 to S1503 are the same as S901 to S903 in the first embodiment, so the explanation is omitted.

[0081] In S1504, a smoothing filter corresponding to the defect type is set. In this embodiment, the defect type is limited to longitudinal linear defects. Here, a vertical smoothing filter is set as the smoothing filter corresponding to longitudinal linear defects. An average value filter, a Gaussian filter, etc., can be used as the smoothing filter. By using a smoothing filter that closely matches the shape of the defect type, the influence of noise can be suppressed.

[0082] In S1505, the difference image is smoothed using a smoothing filter. According to equation (9), the smoothing filter g(y) is applied to the difference image D to calculate the smoothed difference image Ds.

[0083]

number

[0084] In S1506, the enhanced difference image is corrected based on the smoothed difference image. According to equation (10), the corrected defect-enhanced image D'' is calculated based on the defect-enhanced image D′ and the smoothed difference image Ds.

[0085]

number

[0086] Here, k is a constant for adjusting the pixel values ​​of the corrected defect-enhanced image D''. For example, by setting k to the maximum pixel value that the smoothed difference image Ds can take, the pixel values ​​of the corrected defect-enhanced image D'' can be made smaller than the pixel values ​​of the defect-enhanced image D′. An example of the difference image D in this embodiment is shown in Figure 6(a), an example of the defect-enhanced image D′ is shown in Figure 16(a), an example of the smoothed difference image Ds is shown in Figure 16(b), and an example of the corrected defect-enhanced image D'' is shown in Figure 16(c). These are examples of obtaining a defect image in which the difference near the defect has been corrected by correcting the defect-enhanced image in which the difference near the defect has been excessively emphasized using a smoothed difference image. In addition, it is possible to suppress the excessive emphasis of differences that occur in normal printed materials, such as differences that occur near the edge due to minute positional shifts, differences caused by color variations in printed materials, and differences due to noise during scanning.

[0087] [Effects of the third embodiment] As explained above, based on the difference image smoothed according to the defect type, the enhanced difference image is corrected using enhancement filters corresponding to single defects and enhancement filters corresponding to nearby defects. This suppresses excessive enhancement of differences in areas without defects. Therefore, over-detection is suppressed, and the inspection of printed materials can be performed with high accuracy.

[0088] In this embodiment, only vertical linear defects are used as the defect type, but other defect types such as dot defects and horizontal linear defects may also be used. In this case, it is preferable in S1504 to set a smoothing filter with a shape similar to the defect type as the smoothing filter corresponding to the defect type. For example, a circular smoothing filter is preferable for dot defects, and a horizontally elongated smoothing filter is preferable for horizontal linear defects. Multiple defect types may also be used, as in the second embodiment.

[0089] Furthermore, in this embodiment, inspection processing is performed based on the corrected defect-enhanced image D'', but inspection processing may also be performed based on the defect-enhanced image D'' and the smoothed difference image Ds. In this case, S1506 is omitted, and inspection processing is performed in S306 based on the defect-enhanced image D'' and the smoothed difference image Ds. In the inspection processing, pixels that are above a preset threshold th0 in the defect-enhanced image D'' and above a preset threshold th1 in the smoothed difference image Ds are extracted as defect regions, and a defect map is generated. Through the above processing, over-detection of regions that do not contain defects can be suppressed.

[0090] [Other embodiments] The present invention can also be realized by supplying a program that implements one or more of the functions of the above-described embodiments to a system or device via a network or storage medium, and by having one or more processors in the computer of that system or device read and execute the program. It can also be realized by a circuit (e.g., an ASIC) that implements one or more functions. [Explanation of symbols]

[0091] 201 Reference Image Setting Unit 202 Detection Sensitivity Setting Unit 203 Image acquisition unit 204 Difference Image Generation Unit 205 Difference Image Enhancement Section 206 Inspection Processing Unit

Claims

1. An acquisition means for acquiring a first image data representing a reference image which is the target print result, and a second image data representing a target image which is the subject of inspection, A generation means for generating a difference image representing the difference between the reference image and the target image, Processing means for correcting the difference image using a first filter which has a lower degree of emphasis when applied to both the first partial region and a second partial region adjacent to the first partial region than when applied to the first partial region, and a second filter which has a higher degree of emphasis when applied to both the first partial region and the second partial region than when applied to the first partial region, and for inspecting the target image based on the corrected difference image, It has, The processing means performs the correction using the sum of the result of applying the first filter to the difference image and the result of applying the second filter to the difference image. An image processing apparatus characterized by the following:

2. The image processing apparatus according to claim 1, characterized in that the second partial region is adjacent to the first partial region, or is adjacent with a predetermined width between them.

3. An acquisition means for acquiring a first image data representing a reference image which is the target print result, and a second image data representing a target image which is the subject of inspection, A generation means for generating a difference image representing the difference between the reference image and the target image, Processing means for correcting the difference image using a first filter which has a lower degree of emphasis when applied to both the first partial region and a second partial region adjacent to the first partial region than when applied to the first partial region, and a second filter which has a higher degree of emphasis when applied to both the first partial region and the second partial region than when applied to the first partial region, and for inspecting the target image based on the corrected difference image, A first setting means for setting the aforementioned reference image, A second setting means for setting the types of the first and second partial regions, It has, The processing means corrects the difference image, which is more exaggerated than the difference image representing the difference between the reference image and the target image, based on the smoothed difference image obtained by smoothing the difference image according to the type, and then inspects the target image. An image processing apparatus characterized by the following:

4. An acquisition means for acquiring a first image data representing a reference image which is the target print result, and a second image data representing a target image which is the subject of inspection, A generation means for generating a difference image representing the difference between the reference image and the target image, Processing means for correcting the difference image using a first filter which has a lower degree of emphasis when applied to both the first partial region and a second partial region adjacent to the first partial region than when applied to the first partial region, and a second filter which has a higher degree of emphasis when applied to both the first partial region and the second partial region than when applied to the first partial region, and for inspecting the target image based on the corrected difference image, A first setting means for setting the aforementioned reference image, A second setting means for setting the types of the first and second partial regions, It has, The processing means inspects the target image based on a smoothed difference image obtained by smoothing the difference image according to the type, and an enhanced difference image which is more exaggerated than the difference image representing the difference between the reference image and the target image. An image processing apparatus characterized by the following:

5. A program for causing a computer to function as one of the means of an image processing apparatus described in any one of claims 1 to 4.

6. An acquisition process to acquire a first image data representing a reference image which is the target printing result, and a second image data representing a target image which is the subject of inspection, A generation step of generating a difference image representing the difference between the reference image and the target image, A processing step of correcting the difference image using a first filter which has a lower degree of emphasis when applied to both the first partial region and a second partial region adjacent to the first partial region than when applied to the first partial region, and a second filter which has a higher degree of emphasis when applied to both the first partial region and the second partial region than when applied to the first partial region, and then inspecting the target image based on the corrected difference image, It has, The processing step performs the correction using the sum of the result of applying the first filter to the difference image and the result of applying the second filter to the difference image. An image processing method characterized by the following:

Images