Image processing system, image processing device, and printing device

Abstract

Detects abnormalities in printed materials. [Solution] The image processing comprises: a reference image acquisition unit that acquires a reference image obtained by optically reading a reference printed material on which an image of a specific color is printed and which includes a normal region and a region different from the normal region; a range determination unit that determines a specific range, which is the range of characteristic values ​​related to the color of the normal region, using the reference image; a target image acquisition unit that acquires a target image obtained by optically reading a target printed material that is different from the reference printed material and on which an image of a specific color is printed; and a detection processing unit that performs detection processing on the target image using a first threshold value based on the specific range to detect an abnormality in the target printed material.

Description

Technical Field

[0001] This specification relates to an image processing system, an image processing apparatus, and a printing apparatus for detecting abnormalities in printed matter.

Background Art

[0002] The inspection system disclosed in Patent Document 1 performs inspection of printed matter printed uniformly at a certain density. The inspection system captures a printed image with a camera or the like and converts the captured image into a grayscale image. The inspection system extracts pixels regarded as color development failures based on the luminance distribution of the grayscale image. The inspection system evaluates the printed matter based on the ratio of pixels regarded as color development failures.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] Thus, there is a need for a technique for detecting abnormalities in printed matter.

[0005] This specification discloses a new technique for detecting abnormalities in printed matter.

Means for Solving the Problems

[0006] The technique disclosed in this specification has been made to solve at least a part of the above problems and can be realized as the following application examples.

[0007] [Application Example 1] An image processing system comprising: a reference image acquisition unit that acquires a reference image obtained by optically reading a reference printed material on which an image of a specific color is printed and which includes a normal region and a region different from the normal region; a range determination unit that determines a specific range which is a range of characteristic values ​​relating to the color of the normal region using the reference image; a target image acquisition unit that acquires a target image obtained by optically reading a target printed material different from the reference printed material on which an image of the specific color is printed; and a detection processing unit that performs detection processing on the target image using a first threshold value based on the specific range to detect an abnormality in the target printed material.

[0008] Generally, printing results in some unevenness, making it difficult to prepare a printed document containing only perfectly normal areas. With the above configuration, if a reference printed document containing both normal and non-normal areas can be prepared, it is possible to detect abnormalities in the target printed document.

[0009] [Application Example 2] An image processing apparatus comprising: a reference image acquisition unit that acquires a reference image obtained by optically reading a reference printed material on which an image of a specific color is printed and which includes a normal region and a region different from the normal region; a range determination unit that determines a specific range which is a range of characteristic values ​​relating to the color of the normal region using the reference image; and a recording unit that records a threshold for anomaly detection based on the specific range in the memory of the printing apparatus.

[0010] According to the above configuration, a threshold for detecting abnormalities can be recorded in the printer's memory using a reference printout that includes both normal and non-normal areas. Therefore, the printer can detect abnormalities in the target printout using the threshold for detecting abnormalities.

[0011] [Application Example 3] A printing apparatus comprising: a memory in which a threshold for detecting abnormalities is recorded; a printing processing unit for printing an image of a specific color; a target image acquisition unit for acquiring a target image obtained by optically reading a target printed material on which the image of the specific color is printed; and a detection processing unit for detecting abnormalities in the target printed material by performing a detection process on the target image using the threshold for detecting abnormalities recorded in the memory, wherein the threshold for detecting abnormalities is a threshold based on a specific range which is a range of characteristic values ​​relating to the color of the normal region, and the specific range is determined using a reference image obtained by optically reading a reference printed material different from the target printed material on which the image of the specific color is printed, which includes the normal region and a region different from the normal region.

[0012] According to the above configuration, an anomaly detection threshold, determined using a reference printout that includes both normal and non-normal areas, is stored in memory. Therefore, the printing device can use this anomaly detection threshold to detect anomalies in the target printout.

[0013] Furthermore, the technologies disclosed herein can be implemented in various forms, for example, in the form of a printing apparatus, an image processing apparatus, an image processing system, an image processing method, a computer program for implementing the functions or methods of these apparatuses, a recording medium on which such computer programs are stored, and so on. [Brief explanation of the drawing]

[0014] [Figure 1] A block diagram showing the configuration of System 1000 in the embodiment. [Figure 2] Flowchart of the threshold determination process in the first embodiment. [Figure 3] Diagram illustrating the sheet and scanned image. [Figure 4] A figure showing an example of an image used in the first embodiment. [Figure 5] Flowchart of the anomaly detection process. [Figure 6] Figure 1 shows the detection results of the first embodiment. [Figure 7] Second figure showing the detection results of the first embodiment. [Figure 8] Flowchart of the threshold determination process of the second embodiment. [Figure 9] Explanatory diagram of the threshold determination process of the second embodiment. [Figure 10] Conceptual diagram of the histogram HG of the luminance of the reference patch image RP. [Figure 11] Figure showing the detection results of the second embodiment.

Modes for Carrying Out the Invention

[0015] A. First Embodiment: A-1: Configuration of System 1000 Next, the embodiments will be described based on the examples. FIG. 1 is a block diagram showing the configuration of the system 1000 of the example. The system 1000 includes a computer 100 as the image processing apparatus of the present embodiment and a multifunction machine 200 as the printing apparatus. The multifunction machine 200 and the computer 100 can be communicably connected.

[0016] The computer 100 is a computer of the manufacturer of the multifunction machine 200, for example, a personal computer. The computer 100 includes a CPU 110 as a controller of the computer 100, a non-volatile storage device 120 such as a hard disk drive, a volatile storage device 130 such as a RAM, an operation unit 140 such as a mouse and a keyboard, a display unit 150 such as a liquid crystal display, and a communication unit 180. The communication unit

[0017] The volatile memory device 130 provides a buffer area 131 for temporarily storing various intermediate data generated when the CPU 110 performs processing. The non-volatile memory device 120 stores a computer program PG1. The computer program PG1 is provided by the manufacturer of the multifunction device 200 in a form downloaded from a server or stored in a DVD-ROM or the like. The CPU 110 executes the threshold determination process described later by executing the computer program PG1.

[0018] The multifunction device 200 includes a CPU 210 as a controller of the multifunction device 200, a non-volatile memory device 220 such as a hard disk drive, a volatile memory device 230 such as a RAM, an operation unit 240 such as buttons and a touch panel for acquiring operations by a user, a display unit 250 such as a liquid crystal display, a printing mechanism 260, a reading mechanism 270, and a communication unit 280. The communication unit 280 includes a wired or wireless interface for communicably connecting to an external device, for example, the computer 100.

[0019] The volatile memory device 230 provides a buffer area 2� for temporarily storing various intermediate data generated when the CPU 210 performs processing. The non-volatile memory device 220 stores a computer program PG2, a threshold table THT, and patch print data PD. In this embodiment, the computer program PG2 is a control program for controlling the multifunction device 200 and can be stored and provided in the non-volatile memory device 220 when the multifunction device 200 is manufactured. The threshold table THT is a table in which the thresholds determined in the threshold determination process described later are recorded. The threshold table THT is stored in the non-volatile memory device 220 by the threshold determination process described later. The patch print data PD is print data for printing a print image including a plurality of patch images described later. The patch print data PD is stored in the non-volatile memory device 220 when the multifunction device 200 is manufactured, like the computer program PG2.

[0020] The CPU 210 executes various processes to control the multifunction printer 200 by running the computer program PG2. For example, the CPU 210 controls the printing mechanism 260 to print an image. The CPU 210 also controls the reading mechanism 270 to generate scanned image data. Furthermore, the CPU 210 performs anomaly detection processing, which will be described later.

[0021] The printing mechanism 260 is an inkjet printing mechanism. The printing mechanism 260 is equipped with a print head having nozzles formed for ejecting ink as a printing material, and prints on a substrate such as paper by ejecting multiple types of ink, such as cyan (C), magenta (M), and yellow (Y) inks from the nozzles. The printing mechanism 260 may also be an electrophotographic printing mechanism that prints images using toner contained in a toner cartridge as the printing material.

[0022] The reading mechanism 270 generates scanned image data by optically reading the original document using an image sensor. The image sensor is a one-dimensional image sensor with a structure in which multiple photoelectric conversion elements such as CCDs and CMOSs ​​are arranged in a row. The generated scanned image data is, for example, RGB image data. RGB image data contains the values ​​of multiple pixels, and the value of each pixel indicates the color of each pixel with a color value (also called an RGB value) of the RGB color system. Each RGB value contains the gradation values ​​(for example, 256 gradations from 0 to 255) of the three color components: red (R), green (G), and blue (B).

[0023] A-2. Operation of System 1000 A-2-1. Threshold determination process The operation of system 1000 will now be described. The computer 100 of system 1000 performs threshold determination processing. Threshold determination processing is the process of determining thresholds THd for anomaly detection, recording these thresholds, and storing the threshold table THT (Figure 1) in the non-volatile storage device 220 of the multifunction printer 200. Threshold determination processing is performed for each model of the multifunction printer 200 during the development process of the multifunction printer 200. Alternatively, threshold determination processing may be performed for each individual multifunction printer 200 during the manufacturing process of the multifunction printer 200.

[0024] Figure 2 is a flowchart of the threshold determination process in the first embodiment. In S105, the computer 100 (CPU 110) creates a reference sheet RS. Specifically, the computer 100 sends a print instruction to the multifunction printer 200 for a print image containing multiple patch images. Upon receiving the print instruction, the multifunction printer 200 (CPU 210) controls the printing mechanism 260 using the patch print data PD (Figure 1) stored in the non-volatile storage device 220, causing the printing mechanism 260 to print the print image. This creates a reference sheet RS with the patch images printed on the paper.

[0025] Figure 3 is an explanatory diagram of the sheet and scanned image. Reference sheet RS shown in Figure 3(A) has multiple patch images RPc, RPm, and RPy printed on it. The cyan patch image RPc is printed using print data that shows a uniform image of cyan at a predetermined density (e.g., maximum density). The magenta patch image RPm is printed using print data that shows a uniform image of magenta at a predetermined density. The yellow patch image RPy is printed using print data that shows a uniform image of yellow at a predetermined density. Hereafter, when multiple patch images RPc, RPm, and RPy are not distinguished, the lowercase alphabet at the end of the code will be omitted, and they will also be referred to as patch image RP. Furthermore, to distinguish the patch image RP of reference sheet RS from the patch image SP of the target sheet SS described later, it will also be called reference patch image RP.

[0026] Each reference patch image RP that is actually printed is not perfectly uniform and contains inconsistencies. These inconsistencies can occur, for example, when ink is repelled (not absorbed) by the printing medium, such as paper. These inconsistencies may include areas with higher density than the target density, as well as areas with lower density than the target density. For this reason, each reference patch image RP includes a normal region PAn and an abnormal region PAs that differs from the normal region, as shown in Figure 3(A). The normal region PAn is a region with less inconsistency and higher uniformity compared to the abnormal region PAs.

[0027] The abnormal region PAs may include the quasi-abnormal region PAs1 and the abnormal region PAs2. The quasi-abnormal region PAs1 is a region that is more uneven than the normal region PAn, and less uneven than the abnormal region PAs2. In other words, the quasi-abnormal region PAs1 is less uniform than the normal region PAn, and more uniform than the abnormal region PAs2. The abnormal region PAs2 is a region that is more uneven than the quasi-abnormal region PAs1. In other words, the abnormal region PAs2 is less uniform than the quasi-abnormal region PAs1.

[0028] Abnormal region PAs2 is an area that should be detected in the abnormality detection process described later, and can also be described as an area with color unevenness that exceeds the standard for image quality. In other words, abnormal region PAs2 is an area with uniformity below the standard.

[0029] Here, the multiple reference patch images RP included in the reference sheet RS are printed using a multifunction printer 200 manufactured to meet quality standards, and therefore typically do not include the abnormal region PAs2, but include the normal region PAn and the quasi-abnormal region PAs1.

[0030] Figure 4 shows an example of an image used in this embodiment. Figure 4(A) shows an example of a reference patch image RP. The reference patch image RP in Figure 4(A) includes the normal region PAn and the quasi-abnormal region PAs1, but does not include the abnormal region PAs2.

[0031] In S110, the computer 100 has the reading mechanism 270 of the multifunction printer 200 read the reference sheet RS and obtains the data of the reference scan image RI that represents the reference sheet RS.

[0032] Specifically, the operator places the reference sheet RS on the document tray (not shown) of the reading mechanism 270 and operates the control panel 240 of the multifunction printer 200 to input a reading command. Upon receiving the reading command, the multifunction printer 200 controls the reading mechanism 270 to optically read the reference sheet RS and generate data for the reference scan image RI. The multifunction printer 200 transmits the generated reference scan image RI data to the computer 100. The computer 100 then acquires the data for the reference scan image RI. Figure 3(A) can also be described as a diagram showing the reference scan image RI.

[0033] In S120, the computer 100 performs a grayscale conversion on the reference scan image RI. A known relationship can be used as the correspondence between color gradation values ​​(RGB values) and grayscale gradation values ​​(for example, the correspondence between RGB values ​​in the RGB color space and luminance values ​​Y in the YCbCr color space). In this embodiment, the RGB values ​​of each pixel in the reference scan image RI are converted to luminance values ​​with 256 gradations from 0 to 255.

[0034] In S125, the computer 100 selects one reference patch image RP of a particular color from among multiple reference patch images RP of the grayscale-converted reference scan image RI. In this embodiment, one reference patch image RP is selected from three reference patch images RPc, RPm, and RPy of cyan, magenta, and yellow.

[0035] In S130, the computer 100 performs texture analysis on the reference patch image RP to generate an analyzed image AI. Texture analysis is an analytical method that calculates an evaluation value indicating the texture of an image. In this embodiment, for each pixel of the reference patch image RP, the computer 100 calculates a matrix M called GLCM (Gray-Level Co-occurrence Matrix) using the brightness of a predetermined range (in this embodiment, a range of 5 pixels × 5 pixels) with the pixel as the center pixel. The number of rows and columns of matrix M are the number of grayscale levels of the pixel, respectively. In this embodiment, the number of grayscale levels of brightness values ​​for each pixel of the reference scan image RI is 256, so a matrix M with 256 rows × 256 columns is calculated. Using the calculated matrix M, the computer 100 calculates an index value (also called the degree of dissimilarity Dv) indicating the dissimilarity of each pixel. The degree of dissimilarity Dv is given by the following equation (1), where P(i, j) is the element in row i and column j of matrix M. The heterogeneity index Dv is negatively correlated with uniformity, as a higher value indicates lower uniformity. Therefore, it can be said that heterogeneity index Dv is an indicator of the degree of uniformity.

[0036]

number

[0037] Figure 4(B) shows the analysis image AI generated using the reference patch image RP in Figure 4(A). In the analysis image AI in Figure 4(B), the non-uniformity Dv is greater in lighter colored areas and smaller in darker colored areas. In other words, in the analysis image AI in Figure 4(B), the uniformity is higher in darker colored areas and lower in lighter colored areas.

[0038] In S135, the computer 100 performs a smoothing process on the analysis image AI to generate a processed analysis image TAI. The smoothing process involves, for example, calculating the average value of the pixels in a predetermined range (in this embodiment, a range of 5 pixels × 5 pixels) with the pixel in question as the center pixel for each pixel in the analysis image AI, and using the calculated average value as the smoothed value of that pixel. For such a smoothing process, for example, the OpenCV function "blur" can be used. Figure 4(C) shows an example of a processed analysis image TAI generated using the analysis image AI in Figure 4(B).

[0039] In S140, the computer 100 determines the maximum value of the pixel values ​​(values ​​obtained by smoothing the heterogeneity Dv) of the processed analysis image TAI as the threshold THa for the analysis image AI. This determines an appropriate threshold THa that lies between the minimum value of heterogeneity Dv in the analysis image AI (the black area of ​​the analysis image AI in Figure 4(B)) and the maximum value of heterogeneity Dv in the analysis image AI (the white area of ​​the analysis image AI in Figure 4(B)).

[0040] In S150, computer 100 binarizes the analysis image AI before the smoothing process is performed using a threshold THa for the analysis image AI to generate a region separation image ASI. Specifically, through the binarization process, pixels with a heterogeneity Dv greater than or equal to the threshold THa are classified as heterogeneous pixels, and pixels with a heterogeneity Dv less than the threshold THa are classified as uniform pixels. Figure 4(D) shows the region separation image ASI obtained by binarizing the analysis image AI in Figure 4(B). The black regions in the region separation image ASI in Figure 4(D) are regions composed of uniform pixels. The white regions in the region separation image ASI in Figure 4(D) are regions composed of heterogeneous pixels, i.e., regions with lower uniformity than the black regions. Computer 100 identifies the black regions as normal regions PAn and the white regions as abnormal regions PAs (in this embodiment, quasi-abnormal regions PAs1). As can be seen from the above explanation, the generation of the region separation image ASI identifies the normal region PAn of the reference patch image RP.

[0041] In S155, computer 100 determines the maximum value Vmax of the luminance range BR of the identified normal region PAn. The luminance range BR is the range in which the luminance of the pixels constituting the normal region PAn is distributed. Figure 4(E) shows the histogram HGn of the normal region PAn. The histogram HGn in Figure 4(E) shows the luminance range BR and its maximum value Vmax.

[0042] In S165, the computer 100 determines whether all color reference patch images RP have been processed. In the example in Figure 3(A), it determines whether all C, M, and Y reference patch images RPc, RPm, and RPy have been processed. If all color reference patch images RP have not been processed (S165: NO), the computer 100 returns to S125 and selects the unprocessed reference patch image RP. If all color reference patch images RP have been processed (S165: YES), the computer 100 proceeds to S170.

[0043] In S170, the computer 100 records the anomaly detection threshold THd in the multifunction device 200. By the time the process proceeds to S170, the maximum value Vmax of the luminance range BR corresponding to all reference patch images RPc, RPm, and RPy has been determined. In this embodiment, the maximum value Vmax of each of the three reference patch images RPc, RPm, and RPy for the three colors C, M, and Y is recorded as the anomaly detection threshold THd for C, M, and Y (Vmax = THd). For example, the computer 100 generates a threshold table THT that records the anomaly detection threshold THd for the three colors C, M, and Y. The computer 100 transmits the generated threshold table THT to the multifunction device 200. The multifunction device 200 stores the received threshold table THT in the non-volatile storage device 220. As a result, the threshold table THT is stored in the non-volatile storage device 220, as shown in Figure 1. Once the threshold table THT is stored in the non-volatile storage device 220, the threshold determination process is terminated.

[0044] A-2-2. Anomaly detection process With the threshold table THT stored in the non-volatile storage device 220 through the threshold determination process described above, the system 1000's multifunction printer 200 is shipped and delivered to the user. The user can then use the multifunction printer 200 to print and scan images.

[0045] Here, due to various causes, an abnormality may occur in which the image printed by the printing mechanism 260 of the multifunction printer 200 exhibits unevenness exceeding the standard. The causes may include, for example, the aging deterioration of the printing mechanism 260 or the ink, specifically nozzle clogging, variations in ejection volume, and abnormalities in the ink's characteristics (viscosity, etc.). The multifunction printer 200 performs an abnormality detection process to detect whether or not such an abnormality exists in the printing mechanism 260 at predetermined timings. The predetermined timing may be, for example, every time a predetermined period of time has elapsed, or every time a predetermined number of pages have been printed. Alternatively, the predetermined timing may be the timing when a user inputs an execution instruction.

[0046] Figure 5 is a flowchart of the anomaly detection process. In S205, the multifunction printer 200 (CPU 210) creates the target sheet SS. Specifically, the multifunction printer 200 controls the printing mechanism 260 using the patch print data PD (Figure 1) stored in the non-volatile storage device 220, similar to the creation of the reference sheet RS described above (S105 in Figure 2), causing the printing mechanism 260 to print the print image. This creates the target sheet SS with the patch image printed on the paper.

[0047] Figure 3(B) shows the target sheet SS. The target sheet SS is printed using the patch print data PD (Figure 1) used to create the reference sheet RS in Figure 3(A). For this reason, the target sheet SS, like the reference sheet RS, has multiple patch images SPc, SPm, and SPy printed on it. The cyan patch image SPc is printed using print data that shows a uniform image of cyan at a predetermined density. The magenta patch image SPm is printed using print data that shows a uniform image of magenta at a predetermined density. The yellow patch image SPy is printed using print data that shows a uniform image of yellow at a predetermined density. In the following, when multiple patch images SPc, SPm, and SPy are not distinguished, the lowercase alphabet at the end of the code is omitted, and they are also referred to as patch image SP. Furthermore, to distinguish the patch image SP of the target sheet SS from the reference patch image RP of the reference sheet RS mentioned above, it will also be called the target patch image SP.

[0048] Each target patch image SP that is actually printed, like the reference patch image RP, is not perfectly uniform and contains inconsistencies. For this reason, each target patch image SP includes normal regions PAn and abnormal regions PAs (Figure 3(B)), just like the reference patch image RP (Figure 3(A)).

[0049] Abnormal regions PAs may include quasi-abnormal regions PAs1 and abnormal regions PAs2. As mentioned above, abnormal regions PAs2 are regions that should be detected by the abnormality detection process and are regions with color unevenness that exceeds the standard for image quality. Quasi-abnormal regions PAs1 are regions with more unevenness compared to normal regions PAn and less unevenness compared to abnormal regions PAs2. Normal regions PAn and quasi-abnormal regions PAs1 are regions with color unevenness below the standard and are not regions that should be detected by the abnormality detection process.

[0050] Figure 4(F) shows an example of a target patch image SP. As mentioned above, the target patch image SP may contain abnormal areas PAs2 due to a malfunction of the multifunction printer 200, etc. The target patch image SP in Figure 4(F) contains abnormal areas PAs2 where the color is excessively faded in parts, as shown in the areas enclosed by the rectangular frames F1 and F2.

[0051] In S210, the multifunction device 200 has the reading mechanism 270 read the target sheet SS and acquires data of the target scanned image SI that represents the target sheet SS.

[0052] Specifically, the user places the target sheet SS on the document tray (not shown) of the reading mechanism 270 and operates the control panel 240 of the multifunction printer 200 to input a reading command. Upon receiving the reading command, the multifunction printer 200 controls the reading mechanism 270 to optically read the target sheet SS and generate data for the target scanned image SI. The multifunction printer 200 stores the generated data for the target scanned image SI in, for example, the buffer area 231. Figure 3(B) can also be described as a diagram showing the target scanned image SI.

[0053] In S220, the multifunction printer 200 performs a grayscale conversion on the target scanned image SI. The grayscale conversion in S220 is the same process as the grayscale conversion in S120 shown in Figure 2. As a result, the RGB values ​​of each pixel in the target scanned image SI are converted to luminance values ​​of 256 levels from 0 to 255.

[0054] In the S230, the multifunction printer 200 selects one target patch image SP of a particular color from among multiple target patch images SP of the target scanned image SI after grayscale conversion. In this embodiment, one target patch image SP is selected from three color patch images SPc, SPm, and SPy: cyan, magenta, and yellow.

[0055] In the S250, the multifunction printer 200 obtains the threshold THd for detecting anomalies in the color of interest from the threshold table THT. For example, if the color of interest is cyan, the threshold THd for detecting anomalies in cyan is obtained. The threshold THd for detecting anomalies in cyan is the threshold THd determined using the cyan reference patch image RPc in the threshold determination process described above.

[0056] In S260, the multifunction printer 200 binarizes the target patch image SP of the color of interest to generate a detection result image DI. Specifically, through the binarization process, pixels with a brightness of THd or higher are classified as abnormal pixels, and pixels with a brightness of less than THd are classified as non-abnormal images. Figure 4(G) shows the detection result image DI obtained by binarizing the target patch image SP in Figure 4(F). The black areas in the detection result image DI in Figure 4(G) are composed of non-abnormal pixels. The white areas in the detection result image DI in Figure 4(G) are composed of abnormal pixels. Therefore, the white areas in the detection result image DI are detected abnormal areas (hereinafter also referred to as detected abnormal areas AA). In the detection result image DI in Figure 4(G), it can be seen that the parts of the target patch image SP in Figure 4(F) that contain excessively light-colored unevenness (for example, the parts within frames F1 and F2) are detected as detected abnormal areas AA.

[0057] In S270, the multifunction printer 200 generates a detection result using the detection result image. For example, the multifunction printer 200 calculates the ratio of the detected abnormal area AA to the entire detection result image DI. If the ratio of the detected abnormal area AA is above a predetermined threshold, the multifunction printer 200 generates a detection result indicating an abnormality. If the ratio of the detected abnormal area AA is below a predetermined threshold, the multifunction printer 200 generates a detection result indicating no abnormality.

[0058] In S280, the multifunction printer 200 determines whether it has processed all target patch images SP for all colors. In the example in Figure 3(B), it determines whether it has processed all target patch images SPc, SPm, and RPy for C, M, and Y. If not all target patch images SP for all colors have been processed (S280: NO), the multifunction printer 200 returns to S230 to select the unprocessed target patch images SP. If all target patch images SP for all colors have been processed (S280: YES), the multifunction printer 200 proceeds to S290.

[0059] In S290, the multifunction printer 200 displays the detection results on its display unit 250. By the time the process proceeds to S290, detection results have been generated for all target patch images SPc, SPm, and SPy. Therefore, the multifunction printer 200 displays the detection results for each of the three target patch images SPc, SPm, and SPy (C, M, and Y) on its display unit 250. Once the detection results are displayed, the anomaly detection process is terminated.

[0060] The user can recognize whether there is a malfunction in the printing mechanism 260 of the multifunction printer 200 by looking at the displayed detection results. If the user recognizes that there is a malfunction in the printing mechanism 260, for example, they can take action to resolve the malfunction. For example, the user may have the multifunction printer 200 perform nozzle flushing, replace parts or ink, or send the multifunction printer 200 to a service center or the like to request repairs.

[0061] According to the first embodiment described above, the computer 100 of the system 1000 acquires data of a reference patch image RP obtained by optically reading a reference sheet RS (S110 in Figure 2). The computer 100 uses the reference patch image RP to determine the maximum value Vmax of the brightness range BR of the reference patch image RP (S135 in Figure 2). The multifunction printer 200 of the system 1000 acquires a target patch image SP obtained by optically reading a target sheet SS (S210 in Figure 5). The multifunction printer 200 uses a threshold THd for anomaly detection based on the brightness range BR to perform detection processing on the target patch image SP of the target sheet SS and detects anomalies in the target patch image SP (S260-S270 in Figure 5).

[0062] Generally, printing can result in unevenness in some areas, making it difficult to prepare a reference sheet RS with a printed reference patch image RP containing only normal areas PAn. According to this embodiment, if a reference sheet RS can be prepared with a printed reference patch image RP containing both normal areas PAn and abnormal areas PAs, an appropriate threshold THd for anomaly detection can be determined. Therefore, anomalies in the target patch image SP of the target sheet SS can be detected.

[0063] Furthermore, according to this embodiment, the computer 100 performs texture analysis on the reference patch image RP to separate the reference patch image RP into a normal region PAn and an abnormal region PAs that is different from the normal region (S130-S150 in Figure 2, Figure 4(D)). The computer 100 determines the maximum value Vmax of the luminance range BR of multiple pixels included in the separated normal region PAn (S155 in Figure 2). Since the texture (e.g., uniformity) is considered to be different between the normal region PAn and the abnormal region PAs, the normal region PAn and the abnormal region PAs can be appropriately separated by performing texture analysis. According to this embodiment, by separating the normal region PAn through texture analysis, the maximum value Vmax of the luminance range BR of the normal region PAn can be appropriately identified.

[0064] Furthermore, according to this embodiment, the computer 100 performs texture analysis to separate the reference patch image RP into a first region (the black region in Figure 4(D)) and a second region (the white region in Figure 4(D)) which has lower uniformity than the first region. The first region is designated as the normal region PAn, and the maximum value Vmax of the luminance range BR is determined (S150, S155 in Figure 2). As a result, the region with less printing unevenness and high uniformity in the reference patch image RP is identified as the normal region PAn, so the normal region PAn of the reference patch image RP can be appropriately identified, and an appropriate threshold THd for anomaly detection can be determined.

[0065] Furthermore, according to this embodiment, the computer 100 performs texture analysis to generate an analysis image AI in which the non-uniformity Dv, an index value indicating the degree of uniformity, is set as the value of each pixel (S120 in Figure 2). The computer 100 performs smoothing on the analysis image AI to generate a processed analysis image TAI (S135 in Figure 2). The computer 100 determines the THa for the analysis image based on the maximum value of the non-uniformity Dv of multiple pixels in the processed analysis image TAI (i.e., the value indicating the lowest uniformity) (S140 in Figure 2). The computer 100 separates the reference patch image RP into normal regions PAn and abnormal regions PAs by binarizing the analysis image AI using the threshold THa (S150 in Figure 2). As a result, an appropriate threshold THa can be determined using the processed analysis image obtained by performing smoothing on the analysis image AI. For example, an appropriate threshold THa can be determined that lies between the minimum value of heterogeneity Dv in the analyzed image AI (the black area in the analyzed image AI in Figure 4(B)) and the maximum value of heterogeneity Dv in the analyzed image AI (the white area in the analyzed image AI in Figure 4(B)). Therefore, the reference patch image RP can be accurately separated into normal regions PAn and abnormal regions PAs.

[0066] Furthermore, in this embodiment, the anomaly detection process detects areas with color unevenness exceeding a certain standard as an anomaly area. Therefore, for example, if excessive color unevenness occurs in the printed image due to the aging of the multifunction printer 200 or the ink, the occurrence of such unevenness can be detected as an anomaly.

[0067] Next, the detection results of the first embodiment and the comparative example are illustrated. Figures 6 and 7 show the detection results of the first embodiment. Figure 6 shows the detection results of two types of cyan images. The two types of cyan are referred to as Cyan 1 and Cyan 2. Cyan 2 has a higher concentration than Cyan 1. Figure 6(A) shows the reference patch image RP of Cyan 1, and Figure 6(B) shows the reference patch image RP of Cyan 1. It has been visually confirmed that each reference patch image RP shown in Figures 6 and 7 includes the normal region PAn and the quasi-abnormal region PAs1, but does not include the abnormal region PAs2.

[0068] Figure 6(C) shows the detection result image for Comparative Example 1. The detection result image for Comparative Example 1 is generated by the following method. First, a difference image is generated by calculating the difference between the brightness of the reference patch image RP and the brightness of the target patch image SP for each pixel. If the brightness of the pixel at coordinate (x, y) of the target patch image SP is Is(x, y), and the brightness of the pixel at coordinate (x, y) of the reference patch image RP is Ir(x, y), then the value Id(x, y) of the pixel at coordinate (x, y) in the difference image is expressed by the following equation (2). Note that the range of brightness Is and Ir is 256 levels from 0 to 255.

[0069]

number

[0070] Then, the detection result image of Comparative Example 1 is generated by binarizing the difference image with a predetermined threshold TH1. Specifically, through the binarization process, pixels in the difference image with a value greater than or equal to the threshold TH1 are classified as abnormal pixels, and pixels with a value less than the threshold TH1 are classified as non-abnormal images. The threshold TH1 value used is 3. In the detection result image of Figure 6(C), the black areas are composed of non-abnormal pixels, and the white areas are composed of abnormal pixels. Therefore, the white areas in the detection result image of Figure 6(C) are the detected abnormal areas.

[0071] Figure 6(D) shows the detection result image for Comparative Example 2. The detection result image for Comparative Example 2 is generated by the following method. First, the binarization threshold TH2 is calculated for the reference patch image RP using the formula described in Patent Document 1 (JP 2005-276083 A) mentioned above. Specifically, the binarization threshold TH2 is calculated by the following formula (3), where Y0 is the mode of the brightness of multiple pixels in the reference patch image RP, and Ymin is the minimum brightness of the reference patch image RP.

[0072]

number

[0073] Then, the target patch image SP is converted to grayscale and binarized at the threshold TH2 to generate the detection result image for Comparative Example 2. Specifically, through the binarization process, pixels in the target patch image SP with a brightness equal to or greater than the threshold TH2 are classified as abnormal pixels, and pixels with a brightness less than the threshold TH2 are classified as non-abnormal images. In the detection result image of Figure 6(D), the black areas are composed of non-abnormal pixels, and the white areas are composed of abnormal pixels. Therefore, the white areas in the detection result image of Figure 6(D) are the detected abnormal areas.

[0074] Figure 6(E) shows the detection result image DI generated by the method of the first embodiment. Compared with the detection result images of Comparative Examples 1 and 2 (Figures 6(C) and (D)), the detection result image DI of the first embodiment in Figure 6(E) accurately detects the abnormal region of the target patch image SP (i.e., the region with excessively pale color unevenness in Figure 6(B)). For example, in the target patch image SP in Figure 6(B), there is no abnormal region (an excessively pale region) in the area indicated by the dashed frames Fa and Fb. However, in Comparative Examples 1 and 2 (Figures 6(C) and (D)), the area corresponding to the dashed frames Fa and Fb in Figure 6(B) is white, indicating a false detection of the abnormal region. In contrast, the detection result image DI of the first embodiment in Figure 6(E) does not exhibit such false detections, and the abnormal region of the target patch image SP in Figure 6(B) is detected generally correctly.

[0075] For example, in the method of Comparative Example 1, if the reference patch image RP contains a quasi-abnormal region with unevenness, even if there is no abnormality in the region in the target patch image SP corresponding to the quasi-abnormal region of the patch image RP, that region may be mistakenly detected as an abnormal region in the difference image. In the method of Comparative Example 2, since a quasi-abnormal region with unevenness is not assumed, the quasi-abnormal region is more likely to be mistakenly detected as an abnormal region. In contrast, in the method of the first embodiment, even if the reference patch image RP or the target patch image SP contains a quasi-abnormal region PAs1, the threshold THd for abnormality detection is appropriately determined based on the normal region PAn, so false detections are suppressed.

[0076] Figures 6(F)-(J) show the reference patch image RP, target patch image SP, detection result image for Comparative Example 1, detection result image for Comparative Example 2, and detection result image DI for the first embodiment, respectively. The reference patch image RP in Figure 6(F) shows noticeable unevenness and contains many quasi-abnormal regions. The target patch image SP in Figure 6(G) also shows noticeable unevenness and contains quasi-abnormal regions as well as abnormal regions with excessively light colors. Furthermore, the detection result image DI for the first embodiment in Figure 6(J) accurately detects the abnormal regions in the target patch image SP compared to the detection result images for Comparative Examples 1 and 2 (Figures 6(H) and (I)). For example, in the target patch image SP in Figure 6(G), there are no abnormal regions (areas with excessively light colors) in the area indicated by the dashed frame Fc. However, in Comparative Examples 1 and 2 (Figures 6(H) and (I)), the area corresponding to the dashed frame Fc in Figure 6(G) contains white areas, resulting in false detection of abnormal regions. In contrast, in the detection result image DI of the first embodiment shown in Figure 6(J), no such false detections occurred, and the abnormal region of the target patch image SP shown in Figure 6(G) was detected generally correctly.

[0077] Figure 7 shows the detection results for magenta and yellow images. Figures 7(A)-(E) show the reference patch image RP, target patch image SP, detection result image for Comparative Example 1, detection result image for Comparative Example 2, and detection result image DI for the first embodiment, respectively. The target patch image SP in Figure 7(B) has little unevenness and does not contain abnormal areas with excessively light color. However, the detection result images for Comparative Examples 1 and 2 (Figures 7(C) and (D)) contain white areas even though the target patch image SP does not contain any abnormal areas. In other words, abnormal areas are falsely detected in the detection result images for Comparative Examples 1 and 2 (Figures 7(C) and (D)). In contrast, the detection result image DI for the first embodiment in Figure 7(E) does not contain white areas and no abnormal areas are detected. Thus, for magenta as well, no false detection occurs in the detection result image DI of the first embodiment.

[0078] Figures 7(F)-(J) show the yellow reference patch image RP, the target patch image SP, the detection result image for Comparative Example 1, the detection result image for Comparative Example 2, and the detection result image DI for the first embodiment, respectively. The target patch image SP in Figure 7(B) shows little unevenness and contains almost no abnormal areas with excessively light color. The detection result image for Comparative Example 1 (Figure 7(C)) contains many white areas despite the target patch image SP containing few abnormal areas. In other words, the detection result image for Comparative Example 1 (Figure 7(C)) shows a false detection of abnormal areas. In contrast, the detection result image for Comparative Example 1 in Figure 7(I) and the detection result image DI for the first embodiment in Figure 7(E) contain few white areas and detect only a few abnormal areas. Thus, for yellow, false detections were noticeable in Comparative Example 1, but almost no false detections occurred in Comparative Example 1 and the first embodiment. For yellow, there was almost no difference in the detection results between Comparative Example 1 and the first embodiment.

[0079] From the above results, it was found that the anomaly detection method of the first embodiment can detect anomaly regions with higher accuracy than comparative examples 1 and 2, particularly in the detection of anomalies in cyan and magenta images.

[0080] As can be seen from the above explanation, the reference sheet RS in this embodiment is an example of a reference printout, and the target sheet SS is an example of a target printout. Also, the reference patch image RP in this embodiment is an example of a reference image, the target patch image SP is an example of a target image, and cyan, magenta, and yellow are examples of specific colors. Furthermore, the anomaly detection threshold THd in this embodiment is an example of a first threshold, and the threshold THa for the analysis image is an example of a second threshold.

[0081] B. Second Example In the second embodiment, the threshold determination process differs from that of the first embodiment. The threshold determination process will be described below.

[0082] Figure 8 is a flowchart of the threshold determination process in the second embodiment. In S305, similar to S105 in Figure 2, the computer 100 creates the reference sheet RS by having the multifunction printer 200 perform printing using the patch print data PD.

[0083] In S310, similar to S110 in Figure 2, the computer 100 has the reading mechanism 270 of the multifunction printer 200 read the reference sheet RS and obtains the data of the reference scan image RI that represents the reference sheet RS.

[0084] In S320, similar to S120 in Figure 2, the computer 100 performs a grayscale conversion on the reference scan image RI. This converts the RGB values ​​of each pixel in the reference scan image RI into 256 luminance values ​​ranging from 0 to 255.

[0085] In S325, similar to S15 in Figure 2, the computer 100 selects one reference patch image RP of a particular color from among multiple reference patch images RP of the grayscale-converted reference scan image RI. For example, one reference patch image RP is selected from three reference patch images RPc, RPm, and RPy of cyan, magenta, and yellow.

[0086] In S330, computer 100 generates a histogram HG of the brightness of the reference patch image RP. Figure 9 is an explanatory diagram of the threshold determination process in the second embodiment. Figure 9(A) is a conceptual diagram of the histogram HG of the reference patch image RP. The histogram HG is a histogram obtained by classifying multiple pixels of the patch image RP into 256 bins. The classification into the 256 bins is performed according to the brightness of each pixel. Each of the 256 bins is represented as bin B(n). n is the bin number assigned to the bin, and is one of the 256 integers in the range of 0 ≤ n ≤ 255. Bin B(n) means the bin into which pixels with brightness n are classified.

[0087] In step S335, the computer 100 sorts the multiple bins B(n) of the histogram HG (256 in this embodiment) in descending order of frequency (i.e., the number of pixels belonging to each bin). Figure 9(B) shows the sorted multiple bins B(n) of the histogram HG in Figure 9(A).

[0088] In S340, the computer 100 obtains the effective pixel ratio ER determined for each color. The effective pixel ratio ER is experimentally predetermined for each ink color, in this embodiment, for cyan, magenta, and yellow, and is recorded in the non-volatile storage device 220. The computer 100 obtains the effective pixel ratio ER of the color of interest by reading it from the non-volatile storage device 220. Here, the effective pixel ratios ERc, ERm, and ERy for cyan, magenta, and yellow are preferably determined to experimentally optimal values, for example, to be different values ​​for each ink color. In this embodiment, the effective pixel ratios ERc, ERm, and ERy for cyan, magenta, and yellow are 97%, 98%, and 85%, respectively.

[0089] In S350, computer 100 selects effective bins until the sum of effective pixels reaches the ratio ER. Effective bins are selected from multiple bins of the histogram HG (in this embodiment, 256 bins B(n)) in descending order of the number of pixels they contain. For example, if the sum T(k-1) of pixels belonging to the (k-1) bins from the 1st to the (k-1)th bins in descending order of the number of pixels is less than the value obtained by multiplying the total number of pixels Pm in the reference patch image RP by the ratio ER (Pm × ER), then {T(k-1) < (Pm × ER)}. If the sum T(k) of pixels belonging to the k bins from the 1st to the kth bins in descending order of the number of pixels is greater than (Pm × ER), then {T(k) > (Pm × ER)}, then the k bins from the 1st to the kth bins are selected as effective bins. In this case, bins with a number of pixels from the (k+1)th bin onward are classified as invalid bins. In the examples in Figures 9(A) and (B), single-hatched bins are valid bins, and double-hatched bins are invalid bins.

[0090] In S355, the computer 100 determines the maximum value Vmax2 of the luminance range BR2 of the effective pixels. The luminance range BR2 of the effective pixels is the range in the histogram HG where the effective bin selected in S350 is located, as shown in Figure 9(A).

[0091] Here, we will explain the meaning of the effective pixel brightness range BR2. Figure 10 is a conceptual diagram of the brightness histogram HG of the reference patch image RP. As mentioned above, the reference patch image RP includes the normal region PAn and the quasi-abnormal region PAs1. For this reason, the histogram HG is considered to be a superposition of the histogram PHn of the normal region PAn (Figure 10) and the histogram PHs of the quasi-abnormal region PAs1 (Figure 10).

[0092] The normal region PAn has less unevenness and higher uniformity than the quasi-abnormal region PAs1. For this reason, the histogram PHn of the normal region PAn has less variance than the histogram PHs of the quasi-abnormal region PAs1. For this reason, it is considered that the high-luminance and low-luminance edges FA of the histogram HG of the reference patch image RP are composed only of the histogram PHs of the quasi-abnormal region PAs1. Therefore, the luminance range BR2 of the histogram HG of the reference patch image RP, excluding the edges FA, is considered to be the luminance range of the normal region PAn. As shown in Figure 9, by appropriately determining the ratio ER of effective pixels, the luminance range corresponding to the luminance range of the normal region PAn can be identified as the luminance range BR2 of the effective pixels. Then, by identifying the maximum value Vmax2 of the luminance range BR2 of the effective pixels, the maximum value of the luminance range of the normal region PAn can be identified, similar to the first embodiment.

[0093] In S365, the computer 100 determines whether all color reference patch images RP have been processed. That is, similar to S165 in Figure 2, it determines whether all C, M, and Y reference patch images RPc, RPm, and RPy have been processed. If all color reference patch images RP have not been processed (S365: NO), the computer 100 returns to S325 and selects the unprocessed reference patch images RP. If all color reference patch images RP have been processed (S365: YES), the computer 100 proceeds to S370.

[0094] In S370, the computer 100 records the anomaly detection threshold THd2 in the multifunction device 200. By the time the process proceeds to S370, the maximum value Vmax2 of the luminance range BR2 corresponding to all reference patch images RPc, RPm, and RPy has been determined. In this embodiment, the maximum value Vmax2 of each of the three reference patch images RPc, RPm, and RPy for the three colors C, M, and Y is recorded as the anomaly detection threshold THd2 for C, M, and Y (Vmax2 = THd2). For example, similar to the first embodiment, the computer 100 stores the threshold table THT, in which the anomaly detection threshold THd2 for the three colors C, M, and Y is recorded, in the non-volatile storage device 220 of the multifunction device 200. Once the threshold table THT is stored in the non-volatile storage device 220, the threshold determination process is terminated.

[0095] The anomaly detection process in the second embodiment is the same as the anomaly detection process in the first embodiment (Figure 5). However, instead of the threshold THd in the first embodiment, the threshold THd2 determined in the threshold determination process of the second embodiment is used as the threshold for anomaly detection.

[0096] Next, we will illustrate the detection results of the second embodiment. Figure 11 shows the detection results of the second embodiment. Figure 11 shows the detection results for two types of cyan (Figures 11(A) and (B)), magenta (Figure 11(C)), and yellow (Figure 11(D)). In the example in Figure 11, a common value of 98% is used for the effective pixel ratios ERc, ERm, and ERY of cyan, magenta, and yellow.

[0097] The detection result image of cyan 1 in Figure 11(A) is a detection result image DI obtained by performing an anomaly detection process (Figure 5) on the target patch image SP in Figure 6(B) using the threshold THd2 determined using the reference patch image RP in Figure 6(A). The detection result image of cyan 2 in Figure 11(B) is a detection result image DI obtained by performing an anomaly detection process (Figure 5) on the target patch image SP in Figure 6(G) using the threshold THd2 determined using the reference patch image RP in Figure 6(F). In the detection result image DI of the second embodiment in Figures 11(A) and (B), it can be seen that anomaly detection equivalent to that of the first embodiment is achieved for the cyan 1 and cyan 2 images.

[0098] The magenta detection result image in Figure 11(C) is a detection result image DI obtained by performing an anomaly detection process (Figure 5) on the target patch image SP in Figure 7(B) using the threshold THd2 determined using the reference patch image RP in Figure 7(A). The yellow detection result image in Figure 11(D) is a detection result image DI obtained by performing an anomaly detection process (Figure 5) on the target patch image SP in Figure 7(G) using the threshold THd2 determined using the reference patch image RP in Figure 7(F). In the detection result image DI of the second embodiment in Figures 11(C) and (D), it can be seen that anomaly detection equivalent to that of the first embodiment is achieved for the magenta and yellow images as well.

[0099] From the above results, it was found that the anomaly detection in the second embodiment, like that of the first embodiment, can detect anomaly regions with higher accuracy than comparative examples 1 and 2, particularly in the detection of anomalies in cyan and magenta images.

[0100] According to the second embodiment described above, the computer 100 generates a histogram HG of the reference patch image RP (S330 in Figure 8), selects some of the bins of the histogram HG as effective bins (S350 in Figure 8), and determines the maximum value Vmax2 of the brightness range BR2 of the multiple pixels belonging to the selected effective bins (S355 in Figure 5). Effective bins are selected in descending order of the number of pixels belonging to the multiple bins, until the ratio of the sum of the number of pixels belonging to the bins to the total number of pixels Pm reaches the ratio ER (S350 in Figure 8). As a result, an appropriate brightness range BR2 can be determined by appropriately setting the ratio ER experimentally, for example. Therefore, since the threshold THd2 for anomaly detection can be appropriately determined based on the brightness range BR2, anomaly regions of the target patch image SP can be detected with high accuracy.

[0101] Furthermore, according to this embodiment, the ratio ER is set to a different value for each color (cyan, magenta, and yellow in this embodiment) present in the reference patch image RP and the target patch image SP. As a result, an appropriate luminance range BR2 can be determined according to the colors of the reference patch image RP and the target patch image SP, and an appropriate anomaly detection threshold THd2 can be determined according to the colors of the reference patch image RP and the target patch image SP.

[0102] B. Variations (1) In the threshold determination process of each of the above embodiments, brightness is used as the characteristic value of the pixel to perform texture analysis of the reference patch image RP and analysis of the histogram HG (Figures 2 and 8). In addition, in the anomaly detection process, brightness is used to perform binarization of the target patch image SP (Figure 5). Alternatively, density may be used as the characteristic value of the pixel. Density is a characteristic value that, unlike brightness, indicates that the color is brighter as the numerical value is small and the color is darker as the numerical value is large. For this reason, for example, it is preferable that the minimum value of the density range in which the densities of multiple pixels constituting the normal region PAan are distributed is used as the threshold for anomaly detection.

[0103] Furthermore, either of the RGB values ​​may be used as the pixel characteristic value. For example, in a cyan image, the R value of the RGB values ​​varies according to the cyan density. For this reason, the R value may be used as the pixel characteristic value in a cyan image. In a magenta image, the G value of the RGB values ​​varies according to the magenta density. For this reason, the G value may be used as the pixel characteristic value in a magenta image. In a yellow image, the B value of the RGB values ​​varies according to the yellow density. For this reason, the B value may be used as the pixel characteristic value in a yellow image. In this case, it is not necessary to perform grayscale conversion on the reference patch image RP or the target patch image SP.

[0104] (2) In the threshold determination process of the first embodiment described above, regions with high uniformity within the reference patch image RP are identified as normal regions PAn by texture analysis. The characteristics calculated by texture analysis are not limited to uniformity, but may also be other characteristics, such as periodicity or granularity (roughness). For example, in a printed image, in regions where there are no defects such as missing dots, crushed dots, or blurred dots, the dots (halftone dots) are arranged regularly and periodically. In contrast, in a printed image, in regions where defects such as missing dots, crushed dots, or blurred dots occur, the periodicity of the dots decreases. For this reason, for example, the degree of periodicity for each region within the reference patch image RP may be calculated by texture analysis, and regions where the periodicity is higher than the threshold may be identified as normal regions PAn.

[0105] (3) In the threshold determination process of the first embodiment described above, the maximum value of the pixel value of the processed analysis image TAI after smoothing is used as the threshold THa for binarizing the analysis image AI (S134, S140 in Figure 2). The method for determining the threshold THa for the analysis image AI is not limited to this. For example, the threshold THa for the analysis image AI may be a predetermined value. Specifically, the value of the non-uniformity Dv corresponding to the lower limit of uniformity required for the normal region PAn may be experimentally determined, and the experimentally determined value may be recorded in advance in the non-volatile storage device 120 as the threshold THa for the analysis image AI.

[0106] (4) In the threshold determination process of the second embodiment described above, the effective pixel ratio ER is a different value for each ink color: cyan, magenta, and yellow. However, the effective pixel ratio ER may be a common value for each color.

[0107] (5) In each of the above embodiments, the computer 100 determines the maximum values ​​of the luminance ranges BR and BR2, in which the luminances of multiple pixels constituting the normal region PAn are distributed, as the anomaly detection thresholds THd and THd2. Alternatively, for example, if it is desired to perform anomaly detection with stricter criteria, the computer 100 may determine a value that is a predetermined amount smaller than the maximum value of the luminance ranges BR and BR2 as the anomaly detection threshold, or it may determine a value that is midway between the median and maximum value of the luminance ranges BR and BR2 as the anomaly detection threshold. Furthermore, if it is desired to perform anomaly detection with slightly looser criteria, the computer 100 may determine a value that is a predetermined amount larger than the maximum value of the luminance ranges BR and BR2 as the anomaly detection threshold. Generally, the computer 100 can determine the anomaly detection thresholds THd and THd2 based on the luminance ranges BR and BR2 of the identified normal region PAn.

[0108] (6) In each of the above embodiments, an anomaly detection process is performed to detect whether or not there are any inconsistencies in the image printed by the multifunction printer 200 that exceed a certain standard. However, the anomaly detection process is not limited to this and may be used for various purposes.

[0109] For example, anomaly detection processing may be performed as an abrasion test of an image printed on the surface of a product or the surface of a product's packaging. In this case, for example, the computer acquires reference image data by reading the printed material with the image using a digital camera or scanner. The computer determines a threshold for anomaly detection by performing the threshold determination processing shown in Figures 2 and 8 on the reference image. Next, the worker performs an abrasion process by rubbing the surface of the printed material using an abrasion test machine or the like. The computer acquires target image data by reading the printed material after the abrasion process using a digital camera or scanner. The computer detects whether or not there is an abnormal area on the surface of the printed material after the abrasion process by performing the anomaly detection processing shown in Figure 5 on the target image. In this case, if peeling occurs in the image printed on the surface of the printed material due to the abrasion process, the area where the peeling has occurred is detected as an abnormal area.

[0110] (7) In each of the above embodiments, threshold determination processing and anomaly detection processing are performed using patch images RP and SP of primary colors of printing materials such as cyan, magenta, and yellow. However, the threshold determination processing and anomaly detection processing may also be performed using patch images of black, red, blue, and green, or patch images of various intermediate colors printed using at least two or more printing materials such as cyan, magenta, and yellow.

[0111] (8) In each of the above embodiments, the threshold table THT is stored in the non-volatile storage device 220 of the multifunction printer 200 when it is manufactured. However, this is not limited to this, and for example, the threshold table THT generated in the threshold determination process in Figures 2 and 8 may be stored in a server, and after the sale of the multifunction printer 200, the threshold table THT may be downloaded from the server to the multifunction printer 200 and stored in the non-volatile storage device 220 of the multifunction printer 200. Alternatively, the multifunction printer 200 may access the server each time it performs an anomaly detection process and temporarily obtain the threshold table THT from the server.

[0112] (9) In the threshold determination process of the first embodiment described above, a method called GLCM is used as the texture analysis. However, other texture analysis methods, such as GLSZMN (Gray Level Size Zone Matrix) or NGTDM (Neighborhood Gray-Tone-Difference Matrix), may also be used.

[0113] (10) The computer 100 in each of the above embodiments may be a so-called cloud server composed of multiple computers that can communicate with each other. Also, in the above embodiments, the anomaly detection process in Figure 5 is executed by the CPU 210 of the multifunction printer 200, but it may also be executed by a computer connected to the multifunction printer 200, such as a personal computer, a smartphone, or a cloud server.

[0114] (11) In each of the above embodiments, a part of the configuration that was implemented by hardware may be replaced with software, or conversely, a part or all of the configuration that was implemented by software may be replaced with hardware.

[0115] Furthermore, if some or all of the functions of the present invention are implemented by a computer program, the program may be provided in the form of a computer-readable recording medium (for example, a non-temporary recording medium). The program may be used while stored on the same or a different recording medium (computer-readable recording medium) as it was provided. "Computer-readable recording medium" is not limited to portable recording media such as memory cards and CD-ROMs, but may also include internal storage devices within a computer such as various ROMs, and external storage devices connected to a computer such as hard disk drives.

[0116] The present invention has been described above based on examples and modifications. However, the embodiments of the invention described above are for the purpose of facilitating understanding of the present invention and do not limit it. The present invention can be modified and improved without departing from its spirit, and equivalents thereof are included. [Explanation of symbols]

[0117] 1000...System, 100...Computer, 110...CPU, 120...Non-volatile memory, 130...Volatile memory, 131...Buffer area, 140...Operation unit, 150...Display unit, 180...Communication unit, 200...Multifunction printer, 210...CPU, 220...Non-volatile memory, 230...Volatile memory, 231...Buffer area, 240...Operation unit, 250...Display unit, 260...Printing mechanism, 270...Reading mechanism, 280...Communication unit, AA...Detected anomaly area, AI...Analysis image, ASI...Area separation Image, BR, BR2…Brightness range, DI…Detection result image, Dv…Heterogeneity, ER…Percentage, HG…Histogram, PAn…Normal area, PAs…Abnormal area, PAs1…Semi-abnormal area, PAs2…Abnormal area, PD…Patch print data, PG1, PG2…Computer program, RI…Reference scan image, RP…Reference patch image, RS…Reference sheet, SI…Target scan image, SP…Target patch image, SS…Target sheet, TAI…Processed analysis image, THT…Threshold table

Claims

1. An image processing system, A reference image acquisition unit acquires a reference image obtained by optically reading a reference printout on which an image of a specific color is printed, which includes a normal area and an area different from the normal area. A range determination unit that determines a specific range which is the range of characteristic values ​​relating to the color of the normal region using the aforementioned reference image, A target image acquisition unit acquires a target image obtained by optically reading a target printed material that is different from the aforementioned reference printed material and has an image of the specified color printed on it. A detection processing unit that performs detection processing on the target image using a first threshold based on the specified range to detect an abnormality in the target printed material, An image processing system equipped with the following features.

2. The image processing system according to claim 1, The range determination unit, Texture analysis is performed on the aforementioned reference image to separate the reference image into the normal region and a region different from the normal region. An image processing system that determines the range of characteristic values ​​of a plurality of pixels included in the separated normal region as the specific range.

3. The image processing system according to claim 2, The range determination unit, Performing the texture analysis described above, the reference image is separated into a first region with high uniformity and a second region with low uniformity. An image processing system that determines the specific range, with the first region being the normal region.

4. The image processing system according to claim 3, The range determination unit, Perform the texture analysis described above to generate an analysis image in which the index value indicating the degree of uniformity is set as the pixel value, The aforementioned analysis image is subjected to a smoothing process to generate a processed analysis image. A second threshold is determined based on the value that indicates the lowest uniformity among the index values ​​of multiple pixels in the processed analysis image. An image processing system that separates the reference image into a first region and a second region by binarizing the analyzed image using the second threshold.

5. The image processing system according to claim 1, The range determination unit, A histogram is generated by classifying the multiple pixels of the reference image into multiple bins according to the characteristic value of each pixel. Select some of the bins from the aforementioned plurality of bins, The range of characteristic values ​​of a plurality of pixels belonging to the selected bins is determined as the specific range. An image processing system in which some of the bins are selected in descending order of the number of pixels to which the plurality of bins belong, until the ratio of the total number of pixels to which the bins belong reaches a specific ratio.

6. The image processing system according to claim 5, An image processing system in which the aforementioned specific percentage is set to a different value for each of the aforementioned specific colors present in the reference print and the target print.

7. The image processing system according to claim 1, The aforementioned specific color includes at least one of cyan and magenta in the image processing system.

8. The image processing system according to claim 1, An image processing system in which the abnormality detected by the aforementioned detection processing unit is a color unevenness that exceeds a standard.

9. An image processing system according to any one of claims 1 to 8, The aforementioned characteristic value is luminance, An image processing system in which the first threshold is the maximum brightness belonging to the specific range which is the range of brightness.

10. An image processing device, A reference image acquisition unit acquires a reference image obtained by optically reading a reference printout on which an image of a specific color is printed, which includes a normal area and an area different from the normal area. A range determination unit that determines a specific range which is the range of characteristic values ​​relating to the color of the normal region using the aforementioned reference image, A recording unit that records an abnormality detection threshold based on the aforementioned specific range in the memory of the printing device, An image processing device equipped with the following features.

11. A printing device, A memory where anomaly detection thresholds are recorded, A printing processing unit that prints images of a specific color, A target image acquisition unit acquires a target image obtained by optically reading a target printed material on which an image of the specified color is printed, A detection processing unit that uses the threshold for anomaly detection recorded in the memory to perform detection processing on the target image and detects anomalies in the target printed material, Equipped with, The threshold for detecting anomalies is a threshold based on a specific range, which is the range of characteristic values ​​related to the color of the normal region. The printing apparatus determines the specified range using a reference image obtained by optically reading a reference print that is different from the target print on which the image of the specified color is printed, and which includes the normal region and the region different from the normal region.

Images