JP2021125845A - Image processing device, image processing method, and program - Google Patents

Abstract

To provide an image processing technique for preferentially suppressing deterioration of graininess in a specific gradation while reducing density unevenness of an image formed on a recording medium.SOLUTION: An image processing device that rearranges thresholds in a threshold matrix for generating quantization data used in a recording head having a plurality of recording elements includes a generation unit that generates a dot pattern of the target gradation using the threshold matrix, a moving unit that moves dots in the dot pattern between pixels in the same row corresponding to one recording element, and an exchange unit that exchanges a threshold value corresponding to the pixel before the movement and a threshold value corresponding to the pixel after the movement in the threshold matrix on the basis of the graininess of the dot pattern before and after the movement of the dots by the moving unit.SELECTED DRAWING: Figure 9

Description

The present invention relates to an image processing technique for reducing density unevenness and streaks that occur when an image is formed by ejecting ink.

In a recording head used in an inkjet printer, the amount of ink ejected may vary among a plurality of recording elements (nozzles) due to an error in manufacturing or the like. When the amount of ink ejected varies, the formed image tends to have uneven density. Conventionally, a head shading technique is known as a technique for reducing density unevenness. In the head shading, the image data is corrected according to the information regarding the ink ejection amount of each recording element (density characteristic of the recording element). By this correction, the number of ejected ink dots can be increased or decreased, and the density of the formed image can be adjusted.

On the other hand, when head shading is performed, the graininess of the printed image may be deteriorated. Patent Document 1 is a technique for correcting a threshold matrix corrected based on the ejection characteristics of a nozzle based on a threshold value in the corrected threshold matrix in order to achieve both reduction of density unevenness and suppression of graininess deterioration. Is disclosed.

Japanese Unexamined Patent Publication No. 2006-263983

However, the technique of Patent Document 1 cannot preferentially suppress the deterioration of graininess in a specific gradation such as a bright part where graininess is easily noticeable in an image.

An object of the present invention is to provide an image processing technique for preferentially suppressing deterioration of graininess in a specific gradation while reducing density unevenness of an image formed on a recording medium.

In order to achieve the above object, one embodiment of the present invention is an image processing apparatus that rearranges a threshold value in a threshold value matrix for generating quantization data used in a recording head having a plurality of recording elements, and is a threshold value matrix. A generation means for generating a dot pattern of a gradation of interest, a moving means for moving dots in the dot pattern between pixels in the same row corresponding to one recording element, and before and after the movement by the moving means. Based on the granularity of the dot pattern after movement, the threshold matrix is characterized by having an exchange means for exchanging a threshold value corresponding to the pixel before movement and a threshold value corresponding to the pixel after movement. And.

According to the present invention, it is possible to preferentially suppress deterioration of graininess in a specific gradation while reducing density unevenness of an image formed on a recording medium.

The figure which shows the inkjet printer schematically Block diagram showing the configuration of the image formation system Block diagram showing the functional configuration of the image processing device Diagram showing a test image Diagram to illustrate headshading Diagram to illustrate headshading Flowchart showing the process of rearranging the thresholds of the threshold matrix The figure for demonstrating the method of the process of rearranging the threshold of the threshold matrix. A flowchart showing a process of rearranging the thresholds of the threshold matrix in a specific gradation. The figure for demonstrating the method of the process of rearranging the threshold of the threshold matrix in a specific gradation. Flowchart showing the process of rearranging the thresholds of the threshold matrix A flowchart showing the details of the process of rearranging the thresholds of the threshold matrix. A flowchart showing the details of the process of rearranging the thresholds of the threshold matrix. The figure for demonstrating the method of the process of rearranging the threshold of the threshold matrix. Block diagram showing the functional configuration of the rearrangement unit

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. It should be noted that the following embodiments do not limit the present invention, and not all combinations of features described in the following embodiments are necessarily essential for the means for solving the present invention. In addition, various embodiments within the scope of the present invention are also included in the present invention, and some of the following embodiments can be appropriately combined.

<Structure of image forming apparatus>
FIG. 1 is a diagram schematically showing an image forming apparatus 10. The image forming apparatus 10 described here is an inkjet printer. The image forming apparatus 10 includes a recording head 100 in its housing. The recording head 100 is a so-called full-line type recording head, and has four nozzle rows corresponding to each ink color of black, cyan, magenta, and yellow. In each nozzle row, nozzles for ejecting ink are arranged in the X direction at regular intervals. For example, in the case of a nozzle row of 15 inches in length and 1200 dpi, about 18,000 nozzles are arranged in the X direction.

The paper 106 as a recording medium is conveyed in the Y direction by the transfer roller 105 rotating by the driving force of a motor (not shown) while sandwiching the sheet 106 with other rollers (not shown). Then, while the paper 106 is being conveyed, each nozzle of the nozzle rows 101 to 104 ejects ink according to the recording data, so that an image is formed on the paper surface. In the following description, the recording resolution in the paper transport direction (Y direction) is 1200 dpi. The recording resolution in the X direction is 1200 dpi corresponding to the nozzle arrangement density.

An in-line sensor 107 is provided at a position downstream of the recording head 100 in the paper transport direction. The in-line sensor 107 optically reads the color of the image on the printed matter by the optical reading elements arranged at regular intervals in the X direction, and outputs a scanned image expressing the color in the RGB color space. In the following description, the reading resolution in the paper transport direction is 1200 dpi.

<Configuration of image formation system>
FIG. 2 is a block diagram showing a hardware configuration of an image forming system including the image forming apparatus 10 shown in FIG. This image forming system includes an image forming apparatus 10 shown in FIG. 1 and an image processing apparatus 20 as a host apparatus thereof. The image processing device 20 described here is a personal computer (PC).

The image processing device 20 includes a CPU 201, a RAM 202, an HDD 203, a display I / F 204, a keyboard / mouse I / F 205, and a data transfer I / F 206. The CPU 201 executes a predetermined process such as the process described later in FIGS. 7 and 9 according to the program held in the RAM 202 or the HDD 203. The RAM 202 is a volatile storage device that temporarily holds programs and data. The HDD 203 is a non-volatile storage device, and also holds programs and data. The display I / F 204 is an interface for controlling the display of a display (not shown) such as a liquid crystal monitor. The keyboard / mouse I / F205 controls the HID (Human Interface Device) of the keyboard, mouse, and the like. The data transfer I / F 206 is an interface that controls the transmission and reception of data to and from the image forming apparatus 10. USB or LAN is used as the connection method for data transmission / reception.

The image forming apparatus 10 includes a CPU 211, a RAM 212, a ROM 213, a head controller 214, a sensor controller 215, and a data transfer I / F 216. The CPU 211 executes a predetermined process such as controlling the print operation described above in FIG. 1 according to a program held in the RAM 212 or the ROM 213. The RAM 212 is a volatile storage device that temporarily holds programs and data. The ROM 213 is a non-volatile storage device, which also holds programs and data. The head controller 214 controls the ink ejection operation in each nozzle row of the recording head 100 based on, for example, a binary-quantized halftone image (quantized data or recording data) stored in the RAM 212. The sensor controller 215 controls the individual optical reading elements of the inline sensor 107. The data transfer I / F 216 is an interface that controls the transmission and reception of data to and from the image processing device 20.

All image processing including image processing related to the threshold value arrangement described in each embodiment described later is performed by the image processing apparatus 20. However, all or part of the image processing related to the threshold arrangement may be executed by the CPU 211 of the image forming apparatus 10.

<Head shading by image correction; reference example>
As a head shading method, there is a method of applying head shading to an image by using a pixel value conversion table (γ correction table) for correcting an image (hereinafter, also referred to as “head shading by image correction”). Further, as another head shading method, there is a method of applying head shading to a threshold matrix (dither matrix) used for quantization (hereinafter, also referred to as “head shading by correcting the threshold matrix”).

In each of the following embodiments, head shading is performed by correcting the threshold matrix. Here, first, as a reference example or a comparative example, head shading by image correction will be described.

FIG. 3A is a block diagram showing a functional configuration of an image processing device capable of performing head shading by correcting an image. The functional configuration of FIG. 3A can be realized by the image forming system of FIG.

The parameter generation unit 301 generates the parameters used by the print processing unit 302. The print processing unit 302 generates halftone image data used for controlling the ink ejection operation in the image forming apparatus 10.

In the print processing unit 302, the color conversion unit 303 represents an image having a color gamut (each pixel is represented by a set of R, G, and B signals) input from the image processing device 20 and can be reproduced in the RGB color space of the print target. Perform color conversion of the image). Specifically, the RGB image is represented by an image having a reproducible color gamut of the image forming apparatus 10 (each pixel is a set of R, G, and B signals) using a three-dimensional look-up table (3D-LUT). It is converted into an image to be output) and output to the color separation unit 304. The resolution of the image handled by the print processing unit 302 is 1200 dpi, which is the same as the nozzle resolution of the recording head 100 mounted on the image forming apparatus 10. Further, the bit depth of each color component of the image handled by the print processing unit 302, including the input RGB image and the RGB image after color conversion, is 16 bits except for the halftone image output by the quantization unit 306. Suppose there is. That is, 65536 gradations can be expressed by each of the R, G, B signals and the C, M, Y, K signals.

The color separation unit 304 converts the color-converted RGB image into an ink value image corresponding to each ink color used in the image forming apparatus 10 by using 3DLUT, and outputs the image to the shading correction unit 305. When the image forming apparatus 10 uses four color inks of cyan (C), magenta (M), yellow (Y), and black (K), the RGB image input to the color separation unit 304 is cyan and magenta. It is converted into an ink value image consisting of four channels of yellow, yellow, and black. The ink value image composed of the four channels is an image in which each pixel is represented by a set of C, M, Y, and K signals.

The shading correction unit 305 individually performs shading correction for each of the C, M, Y, and K signals for each pixel of the ink value image output by the color separation unit 304, and C, M for each pixel as a result. , Y, K signals are output to the quantization unit 306. The processing performed by the shading correction unit 305 and the processing performed by the quantization unit 306, which will be described later, are similarly applied to the ink value image for each ink color. In the following, the case of applying to an ink value image of black, which is one ink color, will be described as an example, and the other ink colors will be omitted.

Here, when the coordinates in the nozzle row direction are x and the coordinates in the paper transport direction orthogonal to the nozzle row direction are y, the gradation value (pixel value) of the pixels of the coordinates (x, y) is D (x, y). ). At this time, the gradation value of the pixel of the coordinate (x, y) in the ink value image before the shading correction is Dpre (x, y), and the gradation of the pixel of the coordinate (x, y) in the ink value image after the shading correction is The value is expressed as Daft (x, y). The gradation value Daft (x, y) after this correction is different for each nozzle position (that is, for each value of x) with respect to the gradation value Dpre (x, y) before correction (hereinafter, “1D-LUT”. It is generated by applying "pixel value conversion table"). The method of generating the pixel value conversion table will be described later. Then, when the value of x is the same, the same pixel value conversion table is applied to each pixel in the y direction. Here, when the pixel value conversion table at a certain nozzle position x (hereinafter, simply referred to as “position x”) is f_x, the gradation value Daft (x, y) after shading correction is expressed by the following equation (1). ).
Daft (x, y) = f_x (Dpre (x, y)) ... Equation (1)

The quantization unit 306 applies a quantization process according to a known dither method to the ink value image after shading correction according to the above formula (1) for each ink color to turn dots on or off. Generate a halftone image represented by a value. The threshold matrix (dither matrix) used in the dither method is a dot-dispersed threshold matrix having blue noise characteristics. This can be generated by a threshold matrix generation unit 309 by a known method (for example, the method described in JP-A-2000-59626). Here, the threshold value of the position (x, y) corresponding to the pixel position (x, y) in the threshold matrix is set to M0 (x, y). At this time, the pixel value H (x, y) of the halftone image obtained as a result of quantization using the threshold matrix is expressed by the following equation (2).
if (Daft (x, y) ≥ M0 (x, y% Sy))
H (x, y) = 1
else
H (x, y) = 0 ... Equation (2)
Where% is the modulo operator,
Pixel value '1' represents on-dot,
The pixel value '0' represents an off-dot.

The size of the threshold matrix in the x direction is 18,000, which is the same as the number of nozzles, and the size Sy in the y direction is 512. That is, as shown by y% Sy, the threshold M0 (x, y) corresponding to the pixel position (x, y) (hence, the threshold matrix) is repeatedly used in the y direction. Subsequently, a method of generating a pixel value conversion table for each nozzle used in the shading correction unit 305 will be described. Since this method is common to the nozzle rows of each ink color, a black nozzle row 101, which is one ink color, will be described below as an example.

First, as shown in FIG. 4, a test image 401 in which a plurality of uniform patches of the same gradation (9 in this case) are arranged is subjected to a condition in which shading correction is turned off, and only a black nozzle row 101 is used. Record on paper. Then, the recording result (image on the printed matter) is read by the in-line sensor 107, and the scanned image (scanned image) expressed in the RGB color space, which is the reading result, is transferred to the nozzle characteristic acquisition unit 307.

In the parameter generation unit 301 (FIG. 3A), the nozzle characteristic acquisition unit 307 performs a luminance conversion process for obtaining the weight sum of the pixel values (RGB values) of the received scanned image for each pixel. Then, in the scanned image after the luminance conversion, the pixel value of each uniform patch is referred to as a path in which each nozzle records an image, that is, a path in which ink droplets are ejected to form dots (hereinafter, referred to as a "recording path"). The value of the gradation reproduction function L_x (d) for each nozzle is obtained by integrating along (.) And converting the value into brightness.

For example, when obtaining the gradation reproduction function L_513 (d) of the nozzle at position x = 513, the pixel values of each patch are integrated along the paths shown by arrows 402 to 410 in FIG. 4, and the values are converted into brightness. do. As a result, nine values of L_513 (d = 0) to L_513 (d = 65535) can be obtained. Then, by interpolating these nine values, the gradation reproduction function L_513 (d) of the nozzle at the position x = 513 can be obtained. The nozzles at other positions x can be obtained in the same manner.

FIG. 5A is a diagram showing the gradation reproduction function obtained as described above. That is, the curve 502 in FIG. 5A shows an example of the gradation reproduction function L_513 (d) of the nozzle at the position x = 513. The horizontal axis of the graph shows the gradation value before or after the shading correction is performed, and the vertical axis of the graph shows the brightness value obtained by optically reading the image on the printed matter. The vertical axis of the graph may be a density value instead of the lightness value. “Paper white” on the horizontal axis of the graph is a gradation value when the ink injection rate (ratio of ejected ink to a predetermined area; hereinafter the same) is minimized. Further, "solid" on the horizontal axis of the graph is a gradation value when the ink injection rate is maximized.

The pixel value conversion table generation unit 308 in the parameter generation unit 301 (FIG. 3A) has a gradation reproduction function for each nozzle generated by the nozzle characteristic acquisition unit 307 and a predetermined function indicating the target brightness described below. Based on this, the pixel value conversion table f_x for each nozzle is generated.

In FIG. 5A, the straight line 501 indicates a function (hereinafter, referred to as “target brightness function”) L_T (d) indicating the target brightness corresponding to the input gradation value. In the example shown in FIG. 5A, L_T (d) is a linear function connecting the lightness of "paper white" and the lightness of "solid" with a straight line, but it is designed to determine the gradation reproduction characteristics of the printer. Any function can be set as a parameter. Further, the same target brightness function L_T (d) is used for all nozzles.

When the gradation value Daft (x, y) after shading correction at the position x = 513 is obtained by using the target brightness function 501 and the gradation reproduction function 502 as described above, the procedure is as follows. First, according to the target brightness function L_T (d), the target brightness L_T (Dpre (513, y)) corresponding to the gradation value Dpre (513, y) before shading correction is obtained. Then, the gradation value Daft (513, y) after shading correction for reproducing the target brightness L_T (Dpre (513, y)) at the position on the paper surface corresponding to the nozzle at the position x = 513 is obtained. The gradation value Daft (513, y) after shading correction satisfies Daft (513, y) = L_513 (L_T (Dpre (513, y))) ^-1 .

From the above, the pixel value conversion table f_513 at the position x = 513 is ^{realized by a function that combines the inverse function L_513 -1} of the gradation reproduction function L_513 (d) and the target brightness function L_T (d). The same applies to the pixel value conversion table f_x of the nozzles at other positions x.

In the example shown in FIG. 5A, the gradation reproduction function L_513 (d) shown by the curve 502 is always lower than the target brightness function L_T (d) shown by the straight line 501. This indicates that when the shading correction is OFF (not implemented), black streaks are generated at the position on the paper surface corresponding to the nozzle at the position x = 513. However, when the shading correction is turned ON (implemented), the gradation value Dpre (513, y) before the correction is corrected to Daft (513, y) by the shading correction. As a result, the ink ejection rate controlled to be reduced on the paper surface is controlled, and black streaks are reduced.

FIG. 5B is a diagram showing a pixel value conversion table. Specifically, the curve 504 is an example of the pixel value conversion table of the nozzle at position x = 513, which is generated as described above. The horizontal axis of the graph shows the gradation value before conversion, and the vertical axis of the graph shows the gradation value after conversion. The straight line 503 is a 45 degree line. As described above in FIG. 5A, the gradation conversion table f_513 shown by the curve 504 is always below the straight line 503, which is a 45-degree line, in order to perform correction to reduce the amount of ink at the position x = 513. ..

According to the image processing shown above, head shading can be realized by correcting the image, and the streak generated due to the variation in the ejection characteristics of each nozzle can be reduced.

<Head shading by correcting the threshold matrix>
Next, the head shading by the correction of the threshold matrix according to each of the following embodiments will be described. FIG. 3B is a block diagram showing a functional configuration of an image processing apparatus when head shading is performed by correcting a threshold matrix. The functional configuration shown in FIG. 3B can be realized by the image forming system of FIG.

The parameter generation unit 310 generates the parameters used by the print processing unit 311. In addition, the print processing unit 311 generates a halftone image used by the image forming apparatus 10 for controlling the ink ejection operation.

In the print processing unit 311, the color conversion unit 312, the color separation unit 313, and the quantization unit 314 have the same elements as the color conversion unit 303, the color separation unit 304, and the quantization unit 306 described in FIG. 3A. be.

Subsequently, the processing in the parameter generation unit 310 will be described. Since the processing content is common to the nozzle rows of each color, the black nozzle row 101 will be described as an example here.

The threshold matrix generation unit 315 is the same element as the threshold matrix generation unit 309. The threshold matrix generation unit 315 generates a dot-dispersed threshold matrix having blue noise characteristics by a known method. The size of the threshold matrix in the x direction is 18000, which is the same as the number of nozzles, and the size in the y direction is 512.

The head shading unit 316 performs correction, that is, head shading, on the threshold matrix generated by the threshold matrix generation unit 315 (hereinafter, referred to as “original threshold matrix”) according to the ejection characteristics of each nozzle. Here, the threshold value before head shading is expressed as M0 (x, y), and the threshold value after head shading is expressed as M1 (x, y). The threshold value M1 (x, y) after head shading is different for each nozzle position (that is, for each value of x) with respect to the threshold value M0 (x, y) before head shading (hereinafter, “threshold value conversion”). It can be obtained by applying "table". When the value of x is the same, the same threshold conversion table is applied to each threshold in the y direction. Here, when the threshold conversion table at a certain nozzle position x (hereinafter, simply referred to as “position x”) is f_x ^-1 , the threshold M1 (x, y) after head shading is expressed by the following equation (3). Represented by.
M1 (x, y) = f_x ^-1 (M0 (x, y)) ... Equation (3)

As will be described later, the threshold value conversion table is an inverse function of the pixel value conversion table f_x described above. Therefore, the threshold conversion table is expressed as ^{f_x -1.}

The nozzle characteristic acquisition unit 318 has the same elements as the nozzle characteristic acquisition unit 307 shown in FIG. 3A, and generates a gradation reproduction function L_x (d) for each nozzle based on the scanned image of the test image.

The threshold value conversion table generation unit 319 generates a threshold value conversion table for each nozzle based on the gradation reproduction function for each nozzle generated by the nozzle characteristic acquisition unit 318 and a predetermined function indicating the target brightness.

FIG. 6A is a diagram showing a target brightness function. In FIG. 6A, the straight line 505 represents the target brightness function L_T (d) corresponding to the input gradation value, and is the same function as the straight line 501 shown in FIG. 5A. Further, in FIG. 6A, the curve 506 represents the brightness L_513 (d) at the position on the paper surface corresponding to the nozzle at the position x = 513, and is the same curve as the curve 502 shown in FIG. 5A. be. As an example, the threshold value M1 (x, y) after head shading at the position x = 513 can be obtained by the following procedure.

First, according to the gradation reproduction function L_x (d), when the threshold value M0 (513, y) is used as the input value, the brightness L_513 (M0 (513, y)) at the position on the paper corresponding to the nozzle at the position x = 513). Ask for. Then, the target brightness function L_T (d) is reverse-looked up, and the threshold value M1 (513, y) after head shading is set to the threshold value M1 (513, y) = L_T (L_513 (M0 (513, y))) ^-1. Ask as.

That is, the threshold conversion table f_513 ^{-1 at} the position x = 513 is realized by a function obtained by synthesizing the inverse function L_T (d) ^-1 of the target brightness function and the gradation reproduction function L_513 (d). Then, this threshold value conversion table f_x ^-1 is an inverse function of the pixel value conversion table f_x described above. The same applies to the pixel value conversion table f_x ^-1 of the nozzles at other positions x. The threshold conversion table for each nozzle is not limited to the above method, and may be generated by a known method, for example, as disclosed in Japanese Patent Application Laid-Open No. 2009-89080.

In FIG. 6A, the gradation reproduction function L_513 (d) shown by the curve 506 is always lower than the target brightness function L_T (d) shown by the straight line 505. This indicates that when head shading is not performed, black streaks occur at positions on the paper surface corresponding to the nozzles at position x = 513. On the other hand, by performing head shading, the threshold value M0 (513, y) is corrected to M1 (513, y), and the threshold value becomes large. As a result, the quantization unit 314 performs quantization that makes it difficult for dots to be struck. As a result, the amount of ink ejected onto the paper surface is reduced, so that black streaks are reduced.

FIG. 6B is a diagram showing a threshold conversion table. That is, the curve 508 of FIG. 6B shows an example of the threshold conversion table of the nozzle at position x = 513 generated as described above. The horizontal axis of the graph of FIG. 6B shows the threshold value before conversion, and the vertical axis of the graph shows the threshold value after conversion. The straight line 507 is a 45 degree line. As described above, at the position x = 513, the threshold value is converted so that the dots are hard to be hit, so that the gradation conversion table f_513 ^-1 shown by the curve 508 always exceeds the straight line 507 which is the 45 degree line. There is.

Using the threshold conversion table f_x ^-1 generated as described above, the original threshold matrix M0 (x, y) is converted by the equation (3) to generate the threshold matrix M1 (x, y) after head shading. do. In order to suppress the deterioration of graininess caused by the head shading during head shading, the threshold matrix M1 (x, y) after shading correction is used in the rearrangement unit 317 in the parameter generation unit 310 of FIG. 3 (b). ) Is rearranged. The content of the process in the rearrangement unit 317 will be described in each embodiment described later.

As a preparation for this, it is shown that by using the threshold matrix M1 (x, y) after head shading, the same halftone image as in the case of performing head shading by image correction can be obtained. That is, in FIG. 3B, consider a case where the rearrangement unit 317 outputs the data quantized by the quantization unit 314 in a state where the threshold value rearrangement process is not performed as it is. Then, it is shown that the halftone image obtained in this case is the same as the halftone image obtained by the image processing of FIG. 3A.

When the quantization processing unit 314 performs the quantization processing using the threshold matrix M1 (x, y) after head shading, assuming that the gradation value output by the color separation unit 313 is Dpre (x, y). The pixel value H (x, y) of the halftone image is expressed by the following equation (4). The gradation value output by the color separation unit 313 is the same as the gradation value output by the color separation unit 304 in Dpre (x, y).
if (Dpre (x, y) ≥ M1 (x, y% Sy))
H (x, y) = 1
else
H (x, y) = 0 ... Equation (4)

Substituting Eq. (3) into the right-hand side of the conditional expression in the if statement of Eq. (4), Eq. (4) becomes as in Eq. (5).
if (Dpre (x, y) ≥ f_x ^-1 (M0 (x, y% Sy)))
H (x, y) = 1
else
H (x, y) = 0 ... Equation (5)

Further, when the above-mentioned pixel value conversion table f_x is applied to both sides of the conditional expression in the if statement of the equation (5), the equation (5) becomes as in the equation (6).
if (f_x (Dpre (x, y)) ≧ f_x (f_x ^-1 (M0 (x, y% Sy))))
H (x, y) = 1
else
H (x, y) = 0 ... Equation (6)

From the equation (1), the left side of the conditional equation in the if statement of the equation (6) is Daft (x, y). Further, since f_x and f_x ^-1 are combined to form an identity conversion, the right-hand side of the conditional expression in the if statement of the expression (6) is M0 (x, y% Sy). Therefore, the equation (6) has the same result as the equation (2). That is, when the rearrangement unit 317 does not perform the processing in FIG. 3B, the halftone image obtained by the image processing of FIG. 3B is the halftone image obtained by the image processing of FIG. 3A. Will be the same. Therefore, if the threshold matrix is corrected by the image processing of FIG. 3B, processing equivalent to head shading by image correction can be realized, and the sizzle generated due to the variation in the ejection characteristics of each nozzle can be reduced.

Generally, when an image having uniform pixel values is quantized using a threshold matrix having blue noise characteristics, a dot pattern having good graininess can be obtained. On the other hand, when an image containing high-frequency components is quantized using a threshold matrix having blue noise characteristics, so-called moire occurs due to interference between the input image and the threshold matrix (interference between high-frequency components), resulting in image graininess. Deteriorates.

According to the same principle, when head shading by image correction is performed, the graininess of the halftone image is deteriorated. That is, when an image having a uniform pixel value is input to the print processing unit 302 of FIG. 3A, the color separation unit 304 outputs an image having a uniform pixel value. However, since the shading correction unit 305 applies a different pixel value conversion table for each nozzle, the image output by the shading correction unit 305 has a high frequency component in the nozzle row direction. When this image is quantized by the quantization unit 306, the graininess of the halftone image is deteriorated as compared with the case where the shading correction is turned off.

As described above, the head shading by the correction of the threshold matrix is equivalent to the head shading by the correction of the image. Therefore, it can be said that the graininess of the halftone image deteriorates even if the head shading is performed by correcting the threshold matrix.

On the other hand, in each embodiment described below, the threshold of the threshold matrix is rearranged in the rearrangement unit 317 so as to suppress the graininess deterioration of the dot pattern indicated by the data quantized by the threshold matrix. Therefore, the threshold relocation process at a specific gradation, such as a bright portion where graininess is easily noticeable, can be performed with priority over other gradations. That is, the threshold relocation process for the gradation to be prioritized can be performed in a state where the degree of freedom of rearrangement is higher than that of other gradations. As a result, it is possible to preferentially suppress the deterioration of graininess in a specific gradation while reducing the density unevenness of the image formed on the recording medium.

(First Embodiment)
Since the configuration of the image forming system in this embodiment is the same as the configuration shown in FIG. 1 described above, the description thereof will be omitted. Here, the details of the rearrangement unit 317 in the functional configuration of the image processing device 20 shown in FIG. 3B will be described. FIG. 15 is a block diagram showing a functional configuration of the rearrangement unit 317. The rearrangement unit 317 includes a dot pattern generation unit 1501, a graininess evaluation unit 1502, a dot movement unit 1503, and a threshold value exchange unit 1504. The dot pattern generation unit 1501 generates a dot pattern having a specific gradation by using the threshold matrix. The graininess evaluation unit 1502 evaluates the graininess of the dot pattern. The dot moving unit 1503 moves the dots in the generated dot pattern between the pixels in the same row corresponding to one recording element. The threshold value exchange unit 1504 exchanges the threshold value corresponding to the pixel before movement and the threshold value corresponding to the pixel after movement in the threshold matrix based on the graininess of the dot pattern before and after the movement of the dots.

Hereinafter, the processing executed by the image processing apparatus 20 in the present embodiment will be described. FIG. 7 is a flowchart showing a parameter generation process performed by the parameter generation unit 310. Further, FIG. 8 is a diagram for explaining the parameter generation process. Hereinafter, each step (step) is represented by adding S before the reference numeral.

In S601 of FIG. 7, the threshold matrix generation unit 315 generates the original threshold matrix M0 (x, y). 701 in FIG. 8 shows the original threshold matrix M0 (x, y), and the size in the x direction is the same as the number of nozzles in the nozzle row 704. In FIG. 8, the size of the threshold matrix is shown to be smaller than the actual size for simplification of explanation and illustration.

In the next S602, the head shading unit 316 applies ^{the threshold conversion table f_x -1} for each nozzle to the original threshold matrix M0 (x, y). 702 in FIG. 8 shows the threshold matrix M1 (x, y) after head shading obtained by the threshold conversion using this table. When a quantization process is performed on an image having a uniform pixel value using this threshold matrix M1 (x, y), the total sum of dots corresponding to each nozzle changes depending on the nozzle in the halftone image. For example, a nozzle that generates white streaks has a larger total number of dots than other nozzles. On the other hand, in the nozzle that generates black streaks, the total number of dots is smaller than that of other nozzles. As a result, it is possible to reduce the stigma generated due to the variation in the discharge characteristics of each nozzle. However, as described above, the threshold matrix M1 (x, y) has a high frequency component in the nozzle row direction due to head shading, which deteriorates the graininess.

In S603, the rearrangement unit 317 repeatedly replaces the threshold matrix M1 (x, y) after applying the head shading in the same column to rearrange the thresholds. As a result, deterioration of graininess caused by head shading can be suppressed. In FIG. 8, 703 shows the threshold matrix M2 (x, y) after rearrangement. By limiting the substitution of the threshold value within the row corresponding to the same nozzle in this way, the effect of head shading can be maintained.

FIG. 9 is a flowchart showing the threshold relocation process by the relocation unit 317 according to the first embodiment. Further, FIG. 10 is a diagram for explaining the rearrangement process. In the present embodiment, the threshold value of the threshold matrix is rearranged so that the graininess of the dot pattern in one gradation (hereinafter, referred to as “specific gradation”) in which graininess is emphasized is improved. When the gradation value is expressed as 0 to 65535, 0 is paper white, and 65535 is solid, in this embodiment, among the gradations of bright areas where graininess is easily noticeable, 4096 is set as a specific gradation. Set.

In S801, the dot pattern generation unit 1501 reads the threshold matrix M1 (x, y) after applying the head shading. 901 in FIG. 10A shows a threshold matrix after applying head shading, and the size in the x direction is the same as the number of nozzles. In FIG. 10, for the sake of explanation, the size of the threshold matrix is shown to be smaller than the actual size.

In S802, the dot pattern generation unit 1501 reads 4096 set as a specific gradation as the gradation d0 that emphasizes graininess.

In S803, the dot pattern generation unit 1501 is a threshold matrix after applying head shading and a value of a specific gradation set as a gradation d0 that emphasizes graininess, as in the process performed by the quantization unit 314. Compare with. Then, as a result of the quantization, a dot pattern of a specific gradation (in this embodiment, a pattern of "1" and "0") is generated. 902 in FIG. 10B shows a dot pattern having a specific gradation generated by comparing 901 in FIG. 10A with 4096, which is a value of a specific gradation.

In S804, the graininess evaluation unit 1502 evaluates the graininess of the dot pattern 902 at a specific gradation. Specifically, a Gaussian function having a standard deviation of 1.8 pixels and a sum of 1 is convolved with respect to the dot pattern 902, and the difference between the maximum value and the minimum value of the obtained two-dimensional data is used as the graininess evaluation value. .. The graininess evaluation method is not limited to the above method, and a known method such as that disclosed in JP-A-2004-64687 may be used.

In S805, the dot moving unit 1503 randomly selects a column for replacing the threshold value in the dot pattern in the specific gradation, and randomly selects a pixel having dots and a pixel without dots in the column. Replace dots (move dots to each other). 903 in FIG. 10 (c) shows the dot pattern after replacing the dot pattern of 902 in FIG. 10 (b), and in the column of x = 23, the dot (“1”) at the position of y = 2 is It can be seen that it is replaced at the position of y = 5.

In S806, the graininess evaluation unit 1502 evaluates the graininess of the replaced dot pattern 903 by the same method as in S804. In S807, the graininess evaluation unit 1502 compares the graininess evaluation values before and after the dot replacement, and determines whether or not the quality before the dot replacement is improved after the dot replacement.

When the graininess is improved, in S808, the threshold value exchange unit 1504 exchanges the corresponding threshold values (corresponding threshold values before and after the movement of the dots) (moves the threshold values to each other). For example, when the graininess of the dot pattern 903 is improved rather than the graininess of the dot pattern 902, the threshold value t1 and the threshold value t2 of the corresponding pixels in the threshold value matrix 901 before the threshold value is replaced are replaced. 904 in FIG. 10D shows the threshold matrix after replacement.

On the other hand, when it is determined in S807 that the graininess is not improved, in S809, the dot moving unit 1503 returns the dot pattern to the state before the replacement. For example, if the graininess of the dot pattern 903 is not improved with respect to the graininess of the dot pattern 902, the dot pattern 903 is returned to the dot pattern 902.

In S810, the dot pattern generation unit 1501 determines whether or not the processes from S804 to S809 have reached a predetermined number of times, and if not, returns to S804, and if it has reached, ends the process. The predetermined number of times, in the present embodiment was 10 ⁸ times.

According to the first embodiment shown above, the threshold value of the threshold matrix can be rearranged so that the graininess of the dot pattern in a specific gradation that emphasizes graininess is improved. Since the graininess of the dot pattern of other gradations is not considered at the time of rearrangement, the degree of freedom of rearrangement can be maximized. As a result, it is possible to preferentially suppress the deterioration of graininess in a specific gradation while reducing the density unevenness of the image formed on the recording medium.

(Second Embodiment)
The second embodiment relates to a mode in which deterioration of graininess is suppressed in all gradations in addition to executing the rearrangement processing in the specific gradation described in the first embodiment. That is, after the above-mentioned rearrangement in the first embodiment, the threshold matrix is set so that the graininess of the dot patterns on the bright side and the dark part side is improved starting from the dot pattern of the specific gradation related to the rearrangement. Relocate the threshold. This makes it possible to improve the graininess of the dot pattern of all gradations, not just the specific gradation. Further, since the rearrangement process in the specific gradation is performed with priority given to other gradations, the degree of freedom of rearrangement can be maximized in the specific gradation. As a result, deterioration of graininess in a specific gradation can be suppressed. Hereinafter, the rearrangement process in the second embodiment will be described.

FIG. 11 is a flowchart showing a process performed by the rearrangement unit 317 according to the present embodiment. FIG. 14 is an explanatory diagram of the rearrangement method according to the present embodiment.

First, in S1001, the rearrangement unit 317 executes the rearrangement process at the specific gradation described in the first embodiment.

In S1002, the rearrangement unit 317 starts from a dot pattern of a specific gradation, increases the gradation, and repeatedly replaces existing dots newly existing in each gradation in the same nozzle to form a threshold matrix. Rearrange the threshold. This rearrangement is performed so that the graininess evaluation value of the dot pattern at each gradation is improved.

In S1003, the rearrangement unit 317 starts from a dot pattern of a specific gradation, reduces the gradation, and repeatedly replaces dots that are newly absent in each gradation in the same nozzle to form a threshold matrix. Rearrange the threshold. This rearrangement is performed so that the graininess evaluation value of the dot pattern at each gradation is improved.

In the above-mentioned processing, the processing order of S1002 and S1003 may be changed.

<Details of S1002>
The details of the processing of S1002 will be described. FIG. 12 is a flowchart showing details of the process of S1002 of FIG.

In S1201, the dot pattern generation unit 1501 reads the threshold matrix obtained by optimizing the dot pattern at the specific gradation d0 in S1001. In FIG. 14A, 1101 shows an example of the read threshold matrix, and the size in the x direction is the same as the number of nozzles. In FIG. 14, the size of the threshold matrix is shown to be smaller than the actual size for simplification of explanation and illustration. Further, in S1201, the initial value of the gradation d of interest is set to the specific gradation d0.

In S1202, the dot pattern generation unit 1501 increments the focus gradation d. In S1203, the dot pattern generation unit 1501 compares the threshold matrix being rearranged with the gradation d of interest and generates a dot pattern of the gradation of interest, as in S803 described above in the first embodiment. Dot identification). FIG. 14 (c) shows the dot pattern 1103 in the gradation of interest generated by comparing the threshold matrix 1101 shown in FIG. 14 (a) with the gradation of interest d. In FIG. 14, “◆” indicates a dot that exists for the first time in the gradation of interest d. FIG. 14B shows the dot pattern 1102 in the gradation before incrementing in S1202.

In S1204, the graininess evaluation unit 1502 evaluates the graininess of the dot pattern 1103 in the gradation d of interest by the same method as in S804 described above in the first embodiment. In S1205, the dot moving unit 1503 randomly selects a column for replacing the threshold value in the dot pattern 1103 in the gradation of interest. Then, in the column, a pixel (indicated by "◆") in which the dot exists for the first time in the gradation d of interest and a pixel in which the dot does not exist are randomly selected, and the dots are replaced between the pixels. 14 (d) shows the dot pattern 1104 after replacing the dot pattern 1103 of FIG. 14 (c), and in the column of x = 28, the dot at the position of y = 3 is replaced with the position of y = 6. An example is shown. In S1206, the graininess evaluation unit 1502 evaluates the graininess of the replaced dot pattern 1104 by the same method as in S1204.

In S1207, the graininess evaluation unit 1502 determines whether or not the graininess evaluation value is improved before and after the dot replacement. When the graininess is improved, in S1208, the threshold exchange unit 1504 replaces the corresponding threshold. For example, when the graininess of the dot pattern 1104 is better than the graininess of the dot pattern 1103, the threshold value t1 and the threshold value t2 in the threshold value matrix 1101 before the threshold value is replaced are replaced. FIG. 14 (e) shows the threshold matrix 1105 after replacement.

When it is determined in S1207 that the graininess is not improved, in S1209, the dot moving unit 1503 returns the dot pattern to the state before replacement. For example, when the graininess of the dot pattern 1104 is not improved with respect to the graininess of the dot pattern 1103, the dot pattern 1104 is returned to the dot pattern 1103.

In S1210, the dot pattern generation unit 1501 determines whether the processes from S1203 to S1209 have reached a predetermined number of times, and if not, returns to S1203, and if it has reached, proceeds to S1211. The predetermined number is set to 10 ⁸ times. Further, in S1211, it is determined whether or not the gradation d of interest has reached the upper limit, and if it has not reached, the process returns to S1202, and if it has reached, the process ends. The upper limit was 65535.

<Details of S1003>
Next, the details of S1003 will be described. FIG. 13 is a flowchart showing the flow of processing in S1003 of FIG.

First, in S1301, the dot pattern generation unit 1501 reads the threshold matrix during rearrangement obtained by the above-mentioned processing of S1002. FIG. 14 (f) shows the read threshold matrix 1106. Further, in S1301, the initial value of the gradation d of interest is set to the specific gradation d0.

In S1302, the dot pattern generation unit 1501 decrements the gradation d of interest. In S1303, the dot pattern generation unit 1501 compares the threshold matrix being rearranged with the gradation of interest d and generates a dot pattern in the gradation of interest, as in S1203. FIG. 14 (h) shows the dot pattern 1108 of the gradation of interest generated by comparing the threshold matrix 1106 of FIG. 14 (f) with the gradation of interest d. In FIG. 14, “x” indicates a dot that does not exist for the first time in the gradation of interest d. FIG. 14 (g) shows the dot pattern 1107 in the gradation before decrementing in S1302.

In S1304, the graininess evaluation unit 1502 evaluates the graininess of the dot pattern 1108 in the gradation d of interest by the same method as in S1204.

In S1305, the dot moving unit 1503 randomly selects a column for replacing the threshold value in the dot pattern 1108 of the gradation of interest. Then, in the column, a pixel in which the dot does not exist for the first time in the gradation d of interest (indicated by "x") and a pixel in which the dot exists are randomly selected, and the dots are replaced between the pixels. .. FIG. 14 (i) shows the dot pattern 1109 after replacing the dot pattern 1108 of FIG. 14 (h), and the dot at the position of y = 6 is replaced with the position of y = 0 in the column of x = 18. The example is shown. In S1306, the graininess evaluation unit 1502 evaluates the graininess of the replaced dot pattern 1109 by the same method as in S1304.

In S1307, the graininess evaluation unit 1502 determines whether or not the graininess evaluation value is improved before and after the dot replacement. When the graininess is improved, in S1308, the threshold exchange unit 1504 replaces the corresponding threshold. For example, when the graininess of the dot pattern 1109 is better than the graininess of the dot pattern 1108, the threshold value t3 and the threshold value t4 in the threshold value matrix 1106 before the threshold value is replaced are replaced. FIG. 14 (j) shows the threshold matrix 1110 after replacement. When it is determined in S1307 that the graininess is not improved, in S1309, the dot moving unit 1503 returns the dot pattern to the state before the replacement. For example, if the graininess of the dot pattern 1109 is not improved with respect to the graininess of the dot pattern 1108, the dot pattern 1109 is returned to the dot pattern 1108.

In S1310, the dot pattern generation unit 1501 determines whether the processes from S1303 to S1309 have reached a predetermined number of times, and if not, returns to S1303, and if it has reached, proceeds to S1311. The predetermined number is set to 10 ⁸ times. In S1311, it is determined whether or not the gradation d of interest has reached the lower limit, and if it has not reached, the process returns to S1302, and if it has reached, the process ends. The lower limit was set to 0.

According to the rearrangement processing in the second embodiment described above, it is possible to improve the graininess of the dot pattern of all gradations as well as the specific gradation. Further, since the rearrangement process in the specific gradation is performed with priority given to other gradations, the degree of freedom of rearrangement can be maximized in the specific gradation. As a result, it is possible to preferentially suppress the deterioration of graininess in a specific gradation while reducing the density unevenness of the image formed on the recording medium.

When the recording head 100 can eject inks of different sizes or inks having different ink densities, different threshold matrices may be used for each dot size and ink density (recording intensity). In this case, the threshold conversion table for head shading is applied to a plurality of threshold matrices corresponding to each recording intensity. The rearrangement process shown in the first embodiment can be applied to the plurality of threshold matrices after the head shading. In this case, the thresholds of the plurality of threshold matrices are rearranged so that the graininess of the composite dot pattern obtained by synthesizing the dot patterns of each recording intensity is improved. The graininess of the composite dot pattern may be evaluated by weighting each recording intensity. When comparing the graininess evaluation values before and after dot replacement, the graininess evaluation value of this synthetic dot pattern is used. The dot and threshold substitution corresponding to S805 to S809 is performed for each recording intensity. For example, when there are two types of dot sizes, large and small, the small dots are first replaced by the same method as in the first embodiment, and when the graininess evaluation value is improved before and after the replacement, the corresponding threshold value is replaced. At this time, even if the threshold value is replaced, the threshold value for small dots and the threshold value for large dots are simultaneously replaced in order to keep the total sum of dots of each size constant. After the small dots are replaced in this way, the large dots are replaced by the same method as in the first embodiment, and when the graininess evaluation value is improved before and after the replacement, the corresponding threshold value is replaced. Also at this time, the threshold values corresponding to each dot size are simultaneously interlocked and replaced. By repeating such a process a predetermined number of times as in the first embodiment, even when dots of different sizes are used, the graininess improving effect by rearrangement can be obtained at a predetermined gradation. By applying this method to the second embodiment, the graininess improving effect can be obtained in all gradations.

(Other embodiments)
The image recording device to which each embodiment can be applied is not limited to a full-line type inkjet printer. For example, a so-called serial type inkjet printer may be used in which the recording head is scanned in a direction intersecting the paper transport direction to form an image. Further, the nozzle spacing in the nozzle row and the recording resolution in the paper transport direction are not limited to 1200 dpi. Similarly, the spacing between the optical reading elements of the in-line sensor 107 and the reading resolution in the paper transport direction are not limited to 1200 dpi. Further, the method of image recording is not limited to the inkjet method, and can be applied to a printer that uses an LED or a heating element as a recording element. Specifically, it can be applied to an electrophotographic printer that uses an LED array as a light source for exposure, and a sublimation printer that uses a thermal head in which minute heating elements are lined up as a heat source for vaporizing solid ink. be.

Further, the present invention supplies a program that realizes one or more functions of the above-described embodiment to a system or a device via a network or a storage medium, and one or more processors in the computer of the system or the device read and execute the program. It can also be realized by the processing to be performed. It can also be realized by a circuit (for example, ASIC) that realizes one or more functions.

20 Image processing device 317 Relocation unit 1501 Dot pattern generation unit 1503 Dot movement unit 1504 Threshold exchange unit

Claims (10)

An image processing device that rearranges threshold values in a threshold matrix for generating quantization data used in a recording head having a plurality of recording elements.
A generation means for generating a dot pattern of a gradation of interest using a threshold matrix, and
A moving means for moving the dots in the dot pattern between pixels in the same row corresponding to one recording element, and
Based on the graininess of the dot pattern before and after the movement of the dots by the moving means, the threshold value corresponding to the pixel before the movement and the threshold value corresponding to the pixel after the movement are exchanged in the threshold matrix. Exchange means and
An image processing device characterized by having. Further having an evaluation means for calculating the evaluation value of the graininess of the dot pattern,
The claim is characterized in that the exchange means exchanges threshold values based on the evaluation value of graininess of the dot pattern before the movement and the dot pattern after the movement calculated by the evaluation means. Item 1. The image processing apparatus according to item 1. Further, it has a dot identification means for identifying dots that newly exist or do not exist when the gradation of interest is changed, starting from a dot pattern of a predetermined gradation.
The image processing apparatus according to claim 1 or 2, wherein the moving means moves the specified dots. The threshold matrix is a plurality of threshold matrices corresponding to the recording intensities of the plurality of recording elements.
The generation means generates a composite dot pattern of the gradation of interest using the plurality of threshold matrices.
The moving means moves the dots in the composite dot pattern between the pixels in the same row.
The exchange means sets the threshold value corresponding to the pixel before the movement and the pixel after the movement in the plurality of threshold matrices based on the graininess of the dot pattern before and after the movement of the dot by the moving means. The image processing apparatus according to any one of claims 1 to 3, wherein the corresponding threshold value is interlocked and exchanged. The image processing apparatus according to claim 4, wherein the recording intensity is a dot size. The image processing apparatus according to claim 4, wherein the recording intensity is an ink density. The image processing apparatus according to claim 4, wherein the relocation of the threshold value is performed for each recording intensity. The image processing apparatus according to claim 7, wherein the threshold value is rearranged in ascending order of recording intensity. An image processing method for rearranging threshold values in a threshold matrix for generating quantization data used in a recording head having a plurality of recording elements.
A generation step that uses a threshold matrix to generate a dot pattern of the gradation of interest,
A moving step of moving the dots in the dot pattern between pixels in the same row corresponding to one recording element, and
Based on the graininess of the dot pattern before and after the movement of the dots in the movement step, the threshold value corresponding to the pixel before the movement and the threshold value corresponding to the pixel after the movement are exchanged in the threshold matrix. Exchange step and
An image processing method characterized by having. A program for operating a computer as each means of the image processing apparatus according to any one of claims 1 to 8.

Images