JP5811516B2 - Correction value acquisition method, correction value acquisition program, and printing apparatus. - Google Patents

Description

  The present invention relates to a correction value acquisition method, a correction value acquisition program, and a printing apparatus.

  2. Description of the Related Art Printing apparatuses that perform recording by ejecting liquid from nozzles and landing droplets (dots) on a medium are known. When printing is performed using such a printing apparatus, density unevenness (for example, white stripes or black stripes) may occur in the printed image, and the image quality of the printed image may deteriorate.

  When density unevenness occurs, a density correction value is acquired for each dot row (raster line), and the print density for each dot row is corrected based on the acquired density correction value, thereby causing image degradation due to density unevenness. Can be solved (BRS correction). In addition, as a method of acquiring the density correction value, a test pattern formed on a medium (such as a test sheet) is read by a scanner to acquire test pattern image data, and each dot in the acquired test pattern image data is acquired. A method of acquiring a density correction value for each dot row based on the density of a pixel row corresponding to the row is known.

  However, when the image data is acquired by reading the test pattern, if the image data contains foreign matter such as dust, the dust portion is also detected as the density, and an accurate density correction value is obtained. Can be difficult. Therefore, first, a test pattern in a state in which foreign matters such as dust are completely removed is prepared, and with respect to image data obtained by reading the test pattern, the minimum value and the maximum value of the density for each pixel column corresponding to each dot column The threshold value is calculated based on the above. Then, when reading the image data of the test pattern for acquiring the correction value, the pixel having a density exceeding the threshold is determined as dust and removed, thereby improving the density reading accuracy for each dot row. A method for obtaining an accurate density correction value has been proposed. (For example, patent document 1).

JP 2006-305957 A

  In the method of Patent Document 1, since a threshold value is calculated by reading a test pattern that does not have foreign matters such as dust in advance with a scanner, the threshold value is the property of the scanner used when reading the test pattern or the medium on which the test pattern is printed. Will depend on. Therefore, when changing the scanner or printing medium, it is necessary to re-print the test pattern without dust each time and calculate the threshold again, which takes time and effort to obtain an accurate density correction value. It was.

  An object of the present invention is to automatically calculate a threshold value for removing foreign matters such as dust attached to a test pattern, and to obtain an accurate density correction value for performing density correction for each dot row.

The main invention for achieving the above object is: (A) By ejecting ink from nozzle rows arranged in a predetermined direction, dot rows formed along a crossing direction intersecting the predetermined direction are formed in the predetermined direction. A lined pattern is formed on the medium, and (B) in the image data obtained by reading the pattern, the average value of the density for the entire pixel corresponding to the pattern and the standard deviation of the density for the entire pixel corresponding to the pattern An upper limit threshold calculated by the sum of a value multiplied by a predetermined coefficient, an average value of density for all pixels corresponding to the pattern in the image data obtained by reading the pattern, and the entire pixel corresponding to the pattern calculating a value obtained by multiplying a predetermined coefficient to the standard deviation of concentration, and the lower limit threshold value calculated by the difference in the for the (C) the pattern For each pixel from the viewing taken image data in a pixel row corresponding to each of the dot rows of the pattern to obtain the concentration compared with the upper threshold and the lower threshold, the lower threshold the concentration and the upper threshold And (D) corresponding to each dot row of the pattern based on the density calculated for each pixel row using the pixel data included in the range of A correction value acquisition method is characterized in that a correction value for correcting a density to be acquired is acquired.

  Other features of the present invention will become apparent from the description of the present specification and the accompanying drawings.

It is a figure which shows the external appearance structure of a recording system. 1 is a block diagram illustrating a configuration of a printer. FIG. 3A is a schematic cross-sectional view illustrating the configuration of the printer of this embodiment. FIG. 3B is a schematic top view illustrating the configuration of the printer according to the present embodiment. It is a figure which shows arrangement | positioning of the head in a head unit. It is a figure explaining the printing method which the printer of this embodiment performs. FIG. 6A is a schematic sectional view of the scanner 150. FIG. 6B is a top view of the scanner 150 with the upper lid 151 removed. FIG. 7A is a diagram for explaining a state of a raster line when dots are ideally formed. FIG. 7B is a diagram for explaining a raster line when density unevenness occurs. It is a flowchart for the conventional correction value H calculation. It is a figure which shows the test pattern for density correction in a prior art example. 6 is a diagram illustrating a result of reading a cyan correction pattern by a scanner 150. FIG. 11A and 11B are diagrams illustrating a specific method for calculating the density unevenness correction value H. FIG. It is a figure which shows the correction value table regarding cyan. It is a figure which shows a mode that the correction value H corresponding to each gradation value is calculated regarding the xth row | line area | region of cyan. FIG. 14A is a graph showing the gradation value of a row area when dust is attached to the row area of the test pattern. FIG. 14B is a graph showing the gradation value of a row area when a missing dot has occurred in the row area of the test pattern. It is a figure explaining the method of specifying a defective pixel using a predetermined threshold value. It is a figure which shows the flow for calculating the correction value H in 1st Embodiment. It is a figure which shows the flow for calculating the threshold value in 1st Embodiment. It is a figure which shows the flow for calculating the gradation value in 1st Embodiment. FIG. 19A is a diagram for explaining a gradation value calculation process in a case where a foreign object flag is not set for a pixel row to be measured. FIG. 19B is a diagram for explaining a gradation value calculation process when a foreign object flag is set in a pixel row to be measured. It is a figure which shows the flow for calculating the correction value H in a comparative example. It is a figure which shows the flow for calculating the threshold value in a comparative example. It is a figure which shows the flow for calculating the correction value H in 2nd Embodiment. It is a figure which shows the flow for calculating the threshold value in 2nd Embodiment.

  At least the following matters will become clear from the description of the present specification and the accompanying drawings.

(A) By ejecting ink from nozzle rows arranged in a predetermined direction, a pattern in which dot rows formed along an intersecting direction intersecting the predetermined direction are arranged in the predetermined direction is formed on the medium (B) ) In the image data obtained by reading the pattern, a predetermined threshold value is calculated using the standard deviation of the density for the entire pixel corresponding to the pattern, and (C) a pixel row corresponding to each dot row of the pattern is configured. Among the pixels to be used, the density of each pixel column is calculated using data of pixels whose density is within the threshold range, and (D) the pattern is calculated based on the density calculated for each pixel column. The correction value acquisition method characterized by acquiring the correction value corresponding to each dot row.
According to such a correction value acquisition method, a threshold value for removing foreign matters such as dust attached to the test pattern is automatically calculated, and an accurate density correction value for performing density correction for each dot row is acquired. be able to.

In this correction value acquisition method, it is preferable that the threshold value is calculated from a standard deviation of density for all pixels corresponding to the pattern and an average value of density for all pixels corresponding to the pattern.
According to such a correction value acquisition method, even when dust adheres along the intersecting direction on the test pattern, such dust can be accurately detected and an accurate density correction value can be calculated. it can.

In this correction value acquisition method, from the standard deviation of the density for the entire pixel corresponding to the pattern and the average value of the density for the pixel row corresponding to each dot row of the pattern, for each pixel row It is preferable that the threshold value is calculated.
According to such a correction value acquisition method, an accurate density correction value can be calculated even when the actual read density varies in the same density region portion in the test pattern.

In this correction value acquisition method, it is preferable to form the pattern using a plurality of heads having the nozzle row.
According to such a correction value acquisition method, an accurate density correction value can be calculated even when printing is performed using a printing apparatus having a plurality of heads.

Further, (A) correction values corresponding to each dot row for dot rows that are stored in the memory and formed along the intersecting direction intersecting the predetermined direction by the ink ejected from the nozzle rows arranged in the predetermined direction. (B) A pixel corresponding to the pattern in the image data obtained by reading the pattern, wherein (B) a pattern in which the dot rows are arranged in the predetermined direction is formed on the medium. A predetermined threshold value is calculated using the standard deviation of the density for the whole, and (D) among the pixels constituting the pixel row corresponding to each dot row of the pattern, the pixels whose density is within the threshold value range Using the data, a density is calculated for each pixel column, and (E) a dot pattern for each dot column of the pattern is calculated based on the density calculated for each pixel column. Obtaining a correction value for, it becomes clear that the correction value acquisition program characterized.
According to such a correction value acquisition program, a threshold for automatically removing foreign matters such as dust attached to the test pattern is automatically calculated, and an accurate density correction value for performing density correction for each dot row is acquired. be able to.

In this correction value acquisition program, it is preferable that the threshold value is calculated from the standard deviation of the density for the entire pixel corresponding to the pattern and the average value of the density for the entire pixel corresponding to the pattern.
According to such a correction value acquisition program, an accurate density correction value can be calculated even when the actual read density varies in the same density region portion in the test pattern.

In this correction value acquisition program, from the standard deviation of the density for the entire pixel corresponding to the pattern and the average value of the density for the pixel column corresponding to each dot column of the pattern, for each pixel column It is desirable to calculate the threshold value.
According to such a correction value acquisition program, even when dust adheres along the intersecting direction on the test pattern, such dust can be detected accurately and an accurate density correction value can be calculated. it can.

(A) a head unit having nozzle rows arranged in a predetermined direction and forming dots by ejecting ink onto the medium; and (B) a control unit for controlling the head unit, the medium In the image data obtained by forming a pattern in which dot rows along the intersecting direction intersecting the predetermined direction are arranged in the predetermined direction on the image data obtained by reading the formed pattern, a density standard for the entire pixels corresponding to the pattern A predetermined threshold value is calculated using a deviation, and among the pixels constituting the pixel row corresponding to each dot row of the pattern, the pixel data whose density is within the threshold range is used for each pixel row. And a control unit, characterized in that a correction value corresponding to each dot row of the pattern is acquired based on the density calculated for each pixel row. Printing device will become apparent that.
According to such a printing apparatus, it is possible to automatically calculate a threshold value for removing foreign matters such as dust attached to the test pattern, and to obtain an accurate density correction value for performing density correction for each dot row. it can.

=== Basic Configuration of Recording System ===
<Overall configuration>
FIG. 1 is an explanatory diagram showing an external configuration of a recording system. The recording system according to the present embodiment includes a printer 1, a computer 110, a display device 120, an input device 130, a recording / reproducing device 140, and a scanner 150. The printer 1 is a printing apparatus that records an image on a medium such as paper, cloth, or film. The computer 110 is communicably connected to the printer 1 and outputs print data corresponding to the image to be printed to the printer 1 in order to cause the printer 1 to print an image.

  A printer driver is installed in the computer 110. The printer driver is a program for causing the display device 120 to display a user interface and converting image data output from the application program into print data. This printer driver is recorded on a recording medium (computer-readable recording medium) such as a flexible disk FD or a CD-ROM. Alternatively, the printer driver can be downloaded to the computer 110 via the Internet. In addition, this program is comprised from the code | cord | chord for implement | achieving various functions.

<Configuration of Printer 1>
As a form of a printing apparatus for carrying out the invention, an ink jet printer (printer 1) will be described as an example.
FIG. 2 is a block diagram illustrating the configuration of the printer 1. 3A is a schematic cross-sectional view of the printer 1, and FIG. 3B is a schematic top view of the printer 1.
The printer 1 includes a transport unit 20, a drive unit 30, a head unit 40, a detector group 50, and a controller 60. The controller 60 controls each unit based on print data received from the computer 110 that is a printing apparatus control unit, and prints an image on a medium. The situation in the printer 1 is monitored by the detector group 50, and the detector group 50 outputs the detection result to the controller 60. The controller 60 controls each unit based on the detection result output from the detector group 50.

<Transport unit 20>
The transport unit 20 is for transporting the medium S (for example, paper) from the upstream side to the downstream side in a predetermined direction (hereinafter referred to as a transport direction). A roll-shaped medium S before printing is supplied to a printing area by a conveyance roller 21 driven by a conveyance motor (not shown), and then the printed medium S is wound into a roll shape by a winding mechanism. Cut to length and discharge. The operation of the carry motor is controlled by the controller 60 on the printer side. In the printing area during printing, the medium S is vacuum-sucked from below, and the medium S is held at a predetermined position.

<Drive unit 30>
The drive unit 30 freely moves the head unit 40 in an X direction corresponding to the transport direction and a Y direction corresponding to the paper width direction of the medium S (a direction orthogonal to the transport direction). The drive unit 30 includes an X axis stage 31 that moves the head unit 40 in the X direction, a Y axis stage 32 that moves the X axis stage 31 in the Y direction, and a motor (not shown) that moves them. ing.

<Head unit 40>
The head unit 40 is for ejecting ink onto the paper S to form an image, and has a plurality of heads 41. On the lower surface of the head 41, a plurality of nozzles which are ink ejecting portions are provided, and each nozzle is provided with an ink chamber containing ink.

  The head unit 40 is provided on the X-axis stage 31. When the X-axis stage 31 moves in the X direction (conveyance direction), the head unit 40 also moves in the X direction. Then, dot lines (raster lines) along the X direction are formed on the medium S by intermittently ejecting ink from the nozzles while the head unit 40 moves in the X direction (conveyance direction). Thereafter, the head unit 40 is moved in the Y direction (paper width direction) by the Y axis stage 32 via the X axis stage 31, and then printing is performed again while the head unit 40 moves in the X direction.

  In this way, an image can be printed on the medium S in the print region by repeating the operation of forming a raster line by the movement of the head unit 40 in the X direction and the movement of the head unit 40 in the Y direction. An operation for printing an image on the medium S supplied to the printing area (image forming operation), and an operation for conveying the medium S in the conveying direction by the conveying unit 20 and supplying a new medium S portion to the printing area (conveying operation). Are alternately repeated to print a large number of images on the continuous medium S.

  FIG. 4 is a diagram showing the arrangement of the plurality of heads 41 in the head unit 40. In practice, the nozzle surface is formed on the lower surface of the head unit 40, but FIG. 4 is a view of the nozzle virtually viewed from the upper surface (the same applies to the following drawings).

  By arranging a large number of nozzles in the Y direction (paper width direction), a large width image can be printed by one movement of the head unit 40 in the X direction (conveyance direction). By doing so, the printing speed can be increased. However, a long head cannot be formed due to manufacturing problems. Therefore, in the printer 1, a plurality of short heads 41 (1) to 41 (n) are arranged in the Y direction. As shown in FIG. 4, the plurality of heads 41 are attached to the base plate BP.

  On the nozzle surface of each head 41, a black nozzle row K for ejecting black ink, a cyan nozzle row C for ejecting cyan ink, a magenta nozzle row M for ejecting magenta ink, and a yellow nozzle row for ejecting yellow ink. Y is formed. Each nozzle row includes 180 nozzles, and the 180 nozzles are aligned at a constant nozzle pitch (180 dpi) in the Y direction. As shown in the drawing, the smaller numbers are assigned in order from the nozzles on the back side in the Y direction (# 1 to # 180).

  Of the two heads adjacent in the Y direction (for example, 41 (1) and 41 (2)), the nozzle # 180 on the front side of the head 41 (1) on the back side and the head 41 (2 on the front side) ) With the innermost nozzle # 1 is also a constant interval (180 dpi). That is, on the lower surface of the head unit 40, the nozzles are arranged at a constant nozzle pitch (180 dpi) in the Y direction. As shown in FIG. 3, in order to set the interval between the end nozzles of different heads 41 to 180 dpi, it is necessary to arrange the heads 41 in a staggered manner due to structural problems of the heads 41. Further, end nozzles of different heads 41 may overlap.

<Detector group 50>
The detector group 50 includes a rotary encoder, a linear encoder (both not shown), and the like. The rotary encoder detects the rotation amount of the transport roller 21 and detects the transport amount of the medium based on the detection result. The linear encoder detects the position of the X axis stage 31 and the Y axis stage 32 in the moving direction.

<Controller 60>
The controller 60 is a control unit (control unit) for controlling the printer. The controller 60 includes an interface unit 61, a CPU 62, a memory 63, and a unit control circuit 64 (FIG. 2).

  The interface unit 61 transmits and receives data between the computer 110 that is an external device and the printer 1. The CPU 62 is an arithmetic processing device for performing overall control of the printer 1. The memory 63 is for securing an area for storing a program of the CPU 62, a work area, and the like, and is configured by a storage element such as a RAM or an EEPROM. Then, the CPU 62 controls each unit such as the transport unit 20 via the unit control circuit 64 in accordance with a program stored in the memory 63.

<Printing operation of printer 1>
FIG. 5 is a diagram illustrating a printing method performed by the printer 1 of the present embodiment. In the drawing, for simplicity of explanation, the number of nozzles arranged in the Y direction (paper width direction) in the head unit 40 is reduced to ten. One operation of forming an image while the head unit 40 moves in the X direction is referred to as “pass”. Here, the printer 1 completes an image in four passes, and forms a raster line of another pass between raster lines (dot rows along the X direction (transport direction)) formed in a pass. By doing so, the print resolution in the Y direction can be made higher than the nozzle pitch (180 dpi), and a high-quality image can be printed.

  Specifically, first, ten raster lines (black circles) are formed while moving the head unit 40 in the X direction in pass 1. Thereafter, the head unit 40 is moved by a predetermined amount f toward the front side in the Y direction by the Y-axis stage 32. Then, ten raster lines (white circles) are formed while moving the head unit 40 in the X direction again in pass 2. At this time, the head unit is moved in the Y direction by a predetermined amount f so that the raster line of pass 2 is formed behind the raster line formed of pass 1 in the transport direction. In this way, an image is completed by repeating the operation of moving the head unit 40 in the X direction to form a lath line and the operation of moving the head unit 40 by a predetermined amount f in the Y direction.

In the printing method shown in FIG. 5, there are unfilled areas between raster lines on the far side and the near side in the paper width direction. Therefore, an area where no gap is generated between raster lines is set to an image width that the printer 1 can print in the X direction.
In addition, for the following description, “pixel region” and “column region” are set. The “pixel area” refers to a rectangular area virtually defined on the medium S, and the size is determined according to the print resolution. One “pixel area” on the medium S corresponds to one “pixel data” on the image data. The “row region” is a region constituted by a plurality of pixel regions arranged in the X direction. The “column region” corresponds to “pixel column data” in which a plurality of pixel data on the image data are arranged along a direction corresponding to the X direction.

  Small numbers are assigned in order from the row area on the far side in the Y direction. For example, in the printing method shown in FIG. 5, the raster line (dotted line) formed in nozzle # 1 in pass 3 is used as the raster line formed in the first row region. The raster line formed in the second row region is formed at the nozzle # 2 in pass 2, and the raster line formed in the third row region is formed at the third nozzle # 3 in pass 1. The raster line formed in the seventh row area is formed in the fourth nozzle # 4 in pass 1, and the raster line formed in the eighth row area is formed in the second nozzle # 2 in pass 4. Is done. For this reason, in the printing method of the present embodiment, even in the row region formed by the same second nozzle # 2, the nozzles that form the raster line in the adjacent row region are not necessarily the same nozzle. I understand that.

<Configuration of Scanner 150>
The scanner 150 is an image input device that reads an image printed by the printer 1 and acquires image data.

FIG. 6A shows a cross-sectional view of the scanner 150. FIG. 6B shows a top view of the scanner 150 with the upper cover 151 removed.
The scanner 150 includes an upper cover 151, a document table glass 152 on which a document 5 (for example, a medium S on which an image is printed by the printer 1) is placed, and the document table 5 through the document table glass 152 while facing the document 5. A reading carriage 153 that moves in the vertical direction, a guide 154 that guides the reading carriage 153 in the sub-scanning direction, a moving mechanism 155 that moves the reading carriage 153, and a scanner controller (not shown) that controls each part in the scanner 150. And. The reading carriage 153 includes an exposure lamp 157 for irradiating the document 5 with light, a line sensor 158 as an example of a reading unit for detecting a line image in the main scanning direction (direction perpendicular to the paper surface in FIG. 6A), and the document. And an optical system 159 for guiding the reflected light from 5 to the line sensor 158. A broken line inside the reading carriage 153 in the drawing indicates a locus of light.

=== Description of Density Unevenness ===
<About banding>
When printing is performed with an ink jet printer, unevenness may occur in the density of the formed raster lines due to variations in the processing accuracy of the nozzle rows that eject ink. When such density unevenness occurs, it appears as if a streak pattern is formed on the printing surface (banding), and the image quality of the printed image deteriorates.
Here, “density unevenness” will be described. In order to simplify the description, the cause of density unevenness occurring in an image printed in a single color will be described.

  FIG. 7A is an explanatory diagram showing a state when dots are ideally formed (when density unevenness does not occur). In the figure, since dots are ideally formed, each dot is accurately formed in a pixel region delimited by a broken line, and raster lines are regularly formed along the row region. In each row region, an image piece having a density corresponding to the coloring of the region is formed. Here, for simplification of explanation, it is assumed that an image having a constant density is printed so that the dot generation rate is 50%.

  Next, FIG. 7B shows an explanatory diagram when density unevenness occurs. In this figure, the raster line formed in the second row region in FIG. 7A is formed closer to the third row region side due to the ink droplets ejected from the nozzle being bent and flying after ejection. Has been. As a result, the density of the second row area is light and the density of the third row area is dark. On the other hand, the ink amount of the ink droplets ejected to the fifth row region is smaller than the specified amount, and the dots formed in the fifth row region are small. As a result, the density of the fifth row region becomes light.

  When an image composed of row regions with different shades is viewed macroscopically, striped density unevenness along the raster line forming direction (X direction in the present embodiment) is visually recognized. The purpose of this embodiment is to correct this density unevenness.

<Density unevenness correction method>
In order to eliminate the striped pattern caused by density unevenness, the tone value indicated by the pixel data corresponding to the row region is corrected so that a dark image piece is formed in the row region that is easily visible. To do. On the contrary, the gradation value indicated by the pixel data corresponding to the row region is corrected so that a dark image piece is formed with respect to the row region that is light and easily visible.

  Here, in FIG. 7B, the reason why the density of the image piece formed in the third row region is high is not due to the influence of the nozzle that forms the raster line in the third row region, but the adjacent second second region. This is due to the influence of nozzles that form raster lines in the row region. For this reason, when a nozzle that forms a raster line in the third row region forms a raster line in another row region, the image piece formed in that row region is not always dark. This is because, as described above, a row region adjacent to a row region assigned to a certain nozzle is not always formed by the same nozzle.

  That is, even if the image pieces are formed by the same nozzle, the density may be different if the nozzles forming the adjacent raster lines are different. In such a case, the density unevenness cannot be suppressed with the correction value simply associated with the nozzle. Therefore, in this embodiment, the density unevenness correction value H is set for each row region (each pixel row data).

  Since density unevenness occurs due to problems such as nozzle processing accuracy, the correction value H for each row region (each pixel row data) is calculated for each printer 1. Hereinafter, a method for calculating the correction value H will be described.

=== Conventional Correction Value Calculation Method ===
First, as a conventional example, a general method for calculating the correction value H will be described.
FIG. 8 shows a flowchart for calculating the correction value H by the conventional method.
By executing the steps S101 to S105 in order using the printer 1, the correction value H for each row region is calculated. The calculated correction value H is stored in the memory 63 of the printer 1, and density correction is performed using the stored correction value H during actual printing. Each step is realized by a correction program installed in the computer 110.

<S101: Print Test Pattern>
FIG. 9 is a diagram showing a density correction test pattern in a conventional example. The test pattern is composed of strip-shaped patterns of three types of density. Each band-like pattern is generated from image data having a constant gradation value, and is formed by arranging a plurality of dot rows along the X direction (conveyance direction) in the Y direction (paper width direction).

  The gradation value for forming the belt-like pattern is called a command gradation value, the command gradation value of the belt-like pattern having a density of 30% is Sa (76), and the command gradation value of the belt-like pattern having a density of 50% is Sb (128). ), The command gradation value of the belt-like pattern having a density of 70% is expressed as Sc (179). It is assumed that a high gradation value indicates a dark density and a low gradation value indicates a light density.

  In the printing method shown in FIG. 5, one belt-like pattern is composed of raster lines formed by four passes using the nozzles of the head unit 40. Then, in order to calculate the correction value H corresponding to each color (black, magenta, cyan, yellow), a correction pattern is formed for each color.

<S102: Reading Test Pattern>
The printed test pattern is set on the scanner 150 by the inspector. Then, the computer 110 causes the scanner 150 to read the test pattern and obtains it as image data. The test pattern is read by reading the line in the main scanning direction by the line sensor 158 while moving the line sensor 158 of the scanner 150 in the sub-scanning direction, thereby acquiring image data.

  The reading resolution at this time is 2880 dpi (X direction) × 2880 dpi (Y direction), and the image data is composed of pixel data of pixels arranged in two dimensions in the X direction and the Y direction. Each pixel data constituting the image data has 256 gradation values.

<S103: Acquisition of gradation value>
For the image data obtained from the test pattern, a read gradation value is acquired for each row region.
FIG. 10 shows the result of reading the cyan correction pattern with the scanner 150. In the graph of FIG. 10, the horizontal axis is the row area number, and the vertical axis is the read gradation value of each row area.

  Hereinafter, the cyan reading result shown in FIG. 10 will be described as an example. Then, after obtaining the reading gradation value of the correction pattern for each color, the pixel column data in the reading data of the scanner 150 (direction corresponding to the X direction on the data) for each band pattern (for each command gradation value) And a single row region (raster line) in the correction pattern have a one-to-one correspondence. Thereafter, the density of each row region is calculated for each strip pattern. An average value of the read gradation values of the pixel data belonging to the pixel column data corresponding to a certain column area is set as the read gradation value of the column area.

  Although each belt-like pattern is uniformly formed with each command gradation value, as shown in FIG. 10, the reading gradation value varies for each row region. For example, in the graph of FIG. 10, the read tone value Cbi of the i column region is lower than the read tone value of the other column region, and the read tone value Cbj of the j column region is the read tone value of the other column region. Higher than. That is, the i-th row region is visually recognized as light, and the j-th row region is viewed as dark. Such a variation in the read gradation value of each row region is the density unevenness of the print image.

<S104: Calculation of Correction Value H>
In order to improve the density unevenness, the variation in the read gradation value for each row region is reduced. That is, the read gradation value of each row region is brought close to a constant value. For this purpose, the average value Cbt of the read gradation values of all the row regions is set as the “target value Cbt” at the same command gradation value (for example, Sb / density 50%). Then, the gradation value indicated by the pixel column data corresponding to each column region is corrected so that the read gradation value of each column region in the command gradation value Sb approaches the target value Cbt.

  Specifically, the gradation value indicated by the pixel column data corresponding to the column region i whose reading gradation value is lower than the target value Cbt in FIG. 10 is corrected to a gradation value darker than the command gradation value Sb. On the other hand, the gradation value indicated by the pixel column data corresponding to the column region j having a reading gradation value higher than the target value Cbt is corrected to a gradation value lighter than the command gradation value Sb. In this way, for the same gradation value, the correction value H for correcting the gradation value of the pixel column data corresponding to each column region is calculated in order to bring the density of all the column regions close to a constant value.

11A and 11B are diagrams illustrating a specific method for calculating the density unevenness correction value H. FIG. First, FIG. 11A shows a state in which a target command tone value (example Sbt) in a command tone value (example Sb) is calculated in an i-row region whose reading tone value is lower than the target value Cbt. The horizontal axis indicates the gradation value, and the vertical axis indicates the read gradation value in the test pattern result. On the graph, the read gradation values (Cai, Cbi, Cci) are plotted against the command gradation values (Sa, Sb, Sc). For example, the target command tone value Sbt for representing the i-th row region with the target value Cbt with respect to the command tone value Sb is calculated by the following equation (linear interpolation based on the straight line BC).
Sbt = Sb + {(Sc−Sb) × (Cbt−Cbi) / (Cci−Cbi)}

Similarly, as shown in FIG. 11B, in a j-row area where the reading gradation value is higher than the target value Cbt, a target command for the j-row area to be represented by the target value Cbt with respect to the command gradation value Sb. The gradation value Sbt is calculated by the following equation (linear interpolation based on the straight line AB).
Sbt = Sa + {(Sb−Sa) × (Cbt−Caj) / (Cbj−Caj)}
Thus, the target command tone value Sbt of each row region with respect to the command tone value Sb is calculated. Then, the cyan correction value Hb for the command gradation value Sb of each row region is calculated by the following equation. Similarly, correction values Ha and Hc for other command gradation values (Sa and Sc) and correction values for other colors are also calculated.
Hb = (Sbt−Sb) / Sb

  FIG. 12 shows a correction value table for cyan. Here, as described above, it is assumed that n raster lines are formed in four passes in the printing method of the present embodiment. For this reason, the correction value H is calculated for each of n column regions in which raster lines are formed in four passes. In the correction value table shown in FIG. 12, correction values H for the first to nth row regions are set. Such a correction value table is also created for other colors. Then, the correction pattern for calculating the correction value H table is stored in the memory 63 of the printer 1 on which the printing is performed.

<Density correction processing>
When actually performing printing, the printer driver requests the printer 1 to transmit the correction value H stored in the memory 63 to the computer 110. The printer driver stores the correction value H transmitted from the printer 1 in a memory in the computer 110.

  FIG. 13 is a diagram illustrating how the correction value H corresponding to each gradation value is calculated for the x-th row region of cyan. The horizontal axis is the gradation value S_in before correction, and the vertical axis is the correction value H_out corresponding to the gradation value S_in before correction. When the printer driver installed in the computer 110 receives a print command from the user, the printer driver generates print data and transmits the print data to the printer 1. First, when the printer driver receives image data from various application software together with a user's print command, the printer driver converts the image data to a resolution corresponding to the print resolution, and performs color conversion according to the ink color YMCK of the printer 1.

  Then, the printer driver performs density correction processing using the correction value H on the 256-gradation data for each color of YMCK. That is, the 256 gradation values (gradation value S_in before correction) of each pixel data constituting the image data are set to the correction value H set for each color and each column area corresponding to the pixel data. Correct by.

If the gradation value S_in before correction is the same as one of the command gradation values Sa, Sb, Sc, the correction value H corresponding to each command gradation value and the correction stored in the memory of the computer 110 The values Ha, Hb, and Hc can be used as they are. For example, if the gradation value S_in before correction is S_in = Sc, the gradation value S_out after correction is obtained by the following equation.
S_out = Sc × (1 + Hc)

When the gradation value S_in before correction is different from the command gradation value, the correction value H_out corresponding to the gradation value S_in before correction is calculated. For example, as shown in FIG. 13, when the gradation value S_in before correction is between the command gradation values Sa and Sb, the linearity of the correction value Ha of the command gradation value Sa and the correction value Hb of the command gradation value Sb. A correction value H_out is calculated by the following equation by interpolation, and a corrected gradation value S_out is calculated.
H_out = Ha + {(Hb−Ha) × (S_in−Sa) / (Sb−Sa)}
S_out = S_in × (1 + H_out)

  When the gradation value S_in before correction is smaller than the command gradation value Sa, the correction value H_out is calculated by linear interpolation between the minimum gradation value 0 and the command gradation value Sa, and the gradation value before correction is calculated. When S_in is larger than the command tone value Sc, the correction value H_out is calculated by linear interpolation between the maximum tone value 255 and the command tone value Sc.

  Thus, the gradation value S_in indicated by the 256 gradation pixel data is corrected by the correction value H set for each color, each column region to which the pixel data belongs, and each gradation value. By doing so, the gradation value S_in of the pixel data corresponding to the row area where the density is visually recognized is corrected to the dark gradation value S_out, and the gradation indicated by the pixel data corresponding to the row area where the density is visually recognized is dark. The value S_in is corrected to a light gradation value S_out. As a result, density unevenness occurring in the printed image can be reduced.

  Then, the printer driver converts the corrected 256-gradation pixel data (S_out) into 4-gradation pixel data corresponding to the types of dots that can be formed by the printer 1 by halftone processing. For example, if a printer is capable of forming three types of dots (large dots, medium dots, and small dots), 8-bit 256-gradation data is converted to 2-bit 4-gradation data by halftone processing. . For example, pixel data indicating “large dot formation” is converted to “11”, pixel data indicating “medium dot formation” is converted to “10”, and pixel data indicating “small dot formation” is converted to “01”. The converted pixel data indicating “no dot” is converted to “00”.

  Finally, the halftone processed matrix image data is rearranged for each data to be transferred to the printer 1 by rasterization processing, and transmitted to the printer 1 as command data together with print data. When the printer 1 receives the print data, the printer 1 performs printing based on the print data.

=== About the case where dust adheres to the test pattern ===
As described above, the test pattern is printed to obtain the correction value H for each row region, and the density is corrected based on the correction value H at the time of printing. Unevenness can be reduced.

  On the other hand, in this method, if there is a defective pixel on the image data obtained by reading the test pattern printed by the scanner 150, the average value of the read gradation values in the row region including the defective pixel cannot be calculated accurately. is there. Here, the defective pixel is a pixel in which the gradation value of a dot to be formed on the pixel is not correctly displayed. For example, when the image has been read, dust has adhered to the test pattern, or the test pattern This occurs when a missing dot has occurred in the dot row constituting the raster line.

  FIG. 14A is a graph showing the read gradation value of the row area including dust when dust is attached to the row area having the test pattern. The horizontal axis of the graph shown on the upper side of FIG. 14A indicates the position of the pixel in the X direction in the row region. The vertical axis indicates the gradation value of the pixel. A solid line graph indicates the read gradation value of each pixel. The dotted line graph shows the average value of the pixel data (read gradation value) of this pixel column. Further, the lower diagram in FIG. 14A shows a pixel column corresponding to the column region and the state of dust contained in the pixel column.

  As shown in FIG. 14A, a pixel including dust in a row area locally shows a very high read tone value, and therefore the average value of the tone values in the entire row area is free of dust. In this case, it becomes higher than the normal average value (≈command gradation value S). In the row area where the average value is increased due to the influence of dust with respect to the command gradation value S, extreme density unevenness occurs only in that row area as compared with other row areas not affected by the dust. As a result, the correction value H larger than necessary is calculated.

FIG. 14B is a graph showing the read gradation value in the row region when dot missing has occurred.
In the row region where the missing dot has occurred, the average value of the pixel data in the row region is lower than the command gradation value S, contrary to the case described above. Therefore, it is recognized that extreme density unevenness has occurred only in that row area compared to other row areas that are not affected by missing dots, and an incorrect correction value H is calculated.

  In this way, if defective pixels (such as a pixel indicating a dust image or a pixel indicating a missing dot image) are present in the column region, the average value of the pixel data in the column region is affected, and the accurate density The correction value H cannot be calculated.

<How to remove the effects of garbage>
In order to remove the influence of dust and the like and obtain an accurate correction value, a predetermined threshold is applied to identify a defective pixel, and an average of gradation values using pixel data other than the defective pixel specified in the column region A method for calculating the value is effective.

  FIG. 15 is a diagram for explaining a method for identifying a defective pixel using a predetermined threshold. The horizontal axis of the graph indicates the position of the pixel in the X direction in the row region. The vertical axis indicates the gradation value of the pixel. A solid line graph indicates a read gradation value of each pixel, and a broken line indicates a threshold value.

  First, a test pattern in which dust or the like is not attached (hereinafter also referred to as a reference test pattern) is prepared, and the reference test pattern is read by the scanner 150 to obtain image data that does not include a defective pixel. . Next, the computer 110 calculates an average value of all the pixels included in the row region, and acquires a predetermined threshold value based on the average value. Then, a threshold range centered on the average value is set as a normal range.

At the time of actual density correction, it is determined whether the read gradation value is included in the normal range for each column area with respect to the image data of the test pattern to be corrected, and pixel data outside this normal range is Identified as a defective pixel. In FIG. 15, the shaded range is outside the normal range, and this portion of the pixel is specified as a defective pixel, and the pixel data is removed. Then, an average value of density (gradation value) is calculated for the column region from the remaining pixel data other than the removed pixel data.
By calculating the density correction value H for each row region based on the calculated average value, it is possible to obtain an accurate density correction value H in which the influence of defective pixels is reduced.

By the way, the calculated threshold value is affected by the medium on which the reference test pattern is printed and the individual performance of the scanner 150 used for reading the reference test pattern (for example, occurrence of deviation at the time of reading). It can be said that the value depends on the threshold calculation environment. That is, when a different scanner other than the scanner 150 is used, or when the printing medium is changed, pixel data of defective pixels may not be accurately removed even when the same threshold value is used.
In such a case, it is necessary to newly calculate the threshold value. To that end, a complete reference test pattern without dust or missing dots must be prepared again.

  In the printer 1 of the present embodiment, it is assumed that the density correction value H is calculated by an operation of the user who performs printing. According to the method described above, every time the user changes the scanner or the medium according to the printing application, it is necessary to prepare a reference test pattern, recalculate an appropriate threshold value, and acquire an accurate density correction value H. That's not realistic.

=== First Embodiment ===
In the first embodiment, an accurate density correction value H is obtained by automatically calculating a threshold value from a printed test pattern (which can be one with dust attached). Since the threshold value can be automatically calculated, the user's trouble as described above can be saved.

FIG. 16 shows a flowchart for calculating the correction value H in the first embodiment.
First, a test pattern is printed (S201) and read by the scanner 150 to obtain image data (S202). Then, a threshold for each row region of the image data is calculated (S203). Based on the calculated threshold value, a defective pixel is identified for each column region, an average of gradation values is obtained using pixel data other than the defective pixel (S204), and a density correction value H for each column region is calculated (S204). S206).

  Hereinafter, each step will be described. Each step is realized by a correction program installed in the computer 110.

<S201: Print Test Pattern>
In the present embodiment, the same pattern as the density correction test pattern shown in FIG. 9 can be used. That is, each of three types of command gradation values (Sa (76) is a command gradation value of a belt-like pattern having a density of 30%, Sb (128) is a command gradation value of a belt-like pattern having a density of 50%, and is a belt having a density of 70%. The command gradation value of the pattern is expressed as Sc (179)). The test pattern is printed for each color of yellow, magenta, cyan, and black, and a correction value for each color is calculated.

<S202: Reading Test Pattern>
As described in (S102), the test pattern printed in (S201) is read by the scanner 150 and acquired as image data. The reading resolution at this time is 2880 dpi (X direction) × 2880 dpi (Y direction), and the image data is composed of pixel data of pixels arranged in two dimensions in the X direction and the Y direction. Each pixel data constituting the image data has 256 gradation values.

  Note that it is also possible to read an image using other means such as a digital camera instead of using the scanner 150 as means for acquiring image data.

<S203: Threshold calculation>
A threshold value for identifying a defective pixel is calculated from image data (pixel data) obtained by reading the test pattern.

FIG. 17 shows a flow for calculating a threshold in the present embodiment. First, read gradation value data is acquired for all pixels for each band-like pattern to be measured (for each density region having the same command gradation value) (S231). For example, for the band-like pattern with 50% density in FIG. 9 (command gradation value Sb = 128), the read gradation values of all the pixels are examined and stored in the memory 63. Then, an average gradation value (average density) Aall for the entire density area is acquired from the read gradation values of all the obtained pixels (S232). Further, the standard deviation σall of the gradation value (density) for the entire density region is calculated from the average gradation value Aall and the read gradation value of each pixel (S233). Then, a threshold value is calculated from Aall and σall (S234). The threshold value is calculated by the following (Formula 1).
Threshold = Aall ± d × σall (Expression 1)
In Equation 1, d is a coefficient, and an optimum value is determined depending on the type of medium to be printed (for example, glossy paper or plain paper).

<S204: Tone value calculation>
For the image data obtained from the test pattern, a read gradation value is obtained for each pixel column corresponding to the column region of the test pattern. At this time, dust is detected using the threshold value calculated in (S203), and the read gradation value (density) of the pixel row is calculated using pixel data other than the pixel where dust is detected.

FIG. 18 shows a flow for calculating a gradation value in this embodiment.
First, for the pixel column corresponding to the predetermined column region to be calculated, the read gradation value (density) of the mth pixel in the pixel column is acquired (S241). The read gradation value of the mth pixel is compared with the threshold value (S242). If the read gradation value is outside the threshold value range, a foreign substance flag is set for the mth pixel (S243), and the threshold value is set. If it is within the range, the read gradation value is stored in the memory 63 as it is. Subsequently, it is determined whether or not the mth pixel is the last pixel in the pixel column (S244). If not, the process proceeds to the next pixel (S245). That is, the read gradation value is obtained for the (m + 1) th pixel. Then, steps (S241) to (S245) are repeated from the first pixel to the last pixel of the pixel column. When the reading gradation values are acquired for all the pixels in the pixel row, the reading gradation values other than the pixels for which the foreign object flag is set are averaged to obtain the gradation value in the pixel row (S246).

  If a large amount of dust adheres to a certain row area on the test pattern, the number of pixels for which the foreign matter flag is set in the corresponding pixel row increases, and the average density of the row area cannot be accurately detected. There is. In such a case, when the foreign substance flag is set for 50% or more of the pixels constituting the pixel column, an error display or warning is issued to stop the calculation of the pixel column density. Thus, it is possible to prevent inaccurate concentration data from being calculated.

19A and 19B are diagrams specifically illustrating the tone value calculation process (S204).
19A and 19B show the read image data of the belt-like pattern, and the right side of FIGS. 19A and 19B shows the read floor for the mth pixel in the measurement target pixel row obtained from the image data. The key value is shown. The number of pixels in the X direction (conveyance direction) of the belt-like pattern is expressed less for the sake of explanation (for 12 pixels). FIG. 19A shows a case where the foreign object flag is not set for the pixel column to be measured, and FIG. 19B shows a case where the foreign object flag is set for the pixel column to be measured.

  In FIG. 19A, since the dust is not included in the pixel row to be measured for the strip pattern (colored portion pixels) (see the left side of FIG. 19A), the read gradations of all the first to twelfth pixels are read. The values are all within the threshold range (see the right side of the figure), and there is no pixel for which the foreign object flag is set. Therefore, the average of the read gradation values of all the first to twelfth pixels is calculated as the gradation value of the pixel column.

  On the other hand, in FIG. 19B, dust is contained in the pixel row to be measured for the band-like pattern (pixels in the colored portion) (see the left side of FIG. 19B), so the seventh to ninth pixels containing dust are included. Among them, the read gradation values of the eighth and ninth pixels are out of the threshold range (see the right side of the figure), and a foreign substance flag is set for the pixel. Therefore, the average of the read gradation values of the first to seventh pixels and the tenth to twelfth pixels for which no foreign object flag is set is calculated as the gradation value of the pixel column.

<S205: Calculation of Correction Value H>
As described in the above (S105), the correction value H for correcting the gradation value of the pixel column data corresponding to each column region is calculated.
The calculated correction value H is stored in the memory 63, and density correction is performed based on the correction value when the user performs printing. The density correction method is the same as in the conventional example.

<Effects of First Embodiment>
According to the method of the first embodiment, it is possible to automatically calculate a threshold value for removing the influence of dust attached to the test pattern.

  In addition, when printing test patterns and reading image data, it is not necessary to prepare a complete test pattern with no dust attached to it, and the user changes each time the printing conditions (medium and scanner to be used) change. It is possible to perform density correction without taking the trouble of reprinting the test pattern and calculating the threshold value.

  In particular, for each strip pattern (each density area) in the printed test pattern, the threshold value is calculated from the average and standard deviation of the density of all the pixels included in the area, so the column area direction (X direction) It is possible to accurately remove the influence of dust adhering along the surface. For example, if dust such as hair adheres along a row area with a test pattern, the read gradation value of the pixel row corresponding to that row region is compared with other pixel rows that are not affected by dust. A high value is calculated. On the other hand, since the threshold value is calculated from the average density and standard deviation of the entire belt-like pattern (each density region) of the test pattern, the density variation of the entire belt-like pattern is small, and the calculated threshold range is also narrow. Become. Therefore, as viewed from the entire belt-like pattern, a pixel column (pixel column including dust) that is slightly high (or low) in density is easily detected.

=== Comparative Example ===
As described above, the printer 1 has a plurality of heads arranged in the Y direction (see FIG. 4), and prints an image by ejecting ink from each head. Since these heads have different ink ejection characteristics due to manufacturing errors, etc., even if printing is performed with the same gradation value, dot rows (raster lines) with different gradation values are formed for each head. There is a case. Therefore, even in the same density region (strip-shaped pattern) in the above test pattern, the density may vary between the dot rows.

  Here, in the first embodiment, the threshold value is calculated using the gradation value of the entire pixel in each density region (band-like pattern) of the test pattern. On the other hand, in a printing apparatus having a plurality of heads such as the printer 1, there may be a large variation in density between the dot rows constituting the belt-like pattern. Therefore, in the first embodiment, the standard deviation σall becomes a large value, The threshold value calculated by the above (Equation 1) is also calculated as a large value. As a result, the range of the threshold value is widened, so that a pixel having a small change in gradation value is difficult to detect, and the dust detection accuracy may be deteriorated accordingly.

  Therefore, as a comparative example, a method for reducing the variation in density and preventing the threshold range from becoming too wide will be described. In this comparative example, a threshold value is calculated by obtaining a density average and a standard deviation for each pixel column in each strip pattern. Then, based on the threshold value calculated for each pixel column, the gradation value (density) of the pixel column is calculated.

  FIG. 20 shows a flowchart for calculating the correction value H in the comparative example. The overall flow is substantially the same as (S201) to (S205) of the first embodiment, and the threshold calculation method of (S303) is different. Hereinafter, the description will focus on (S303).

<S303: Threshold calculation>
FIG. 21 shows a flow for calculating a threshold value in the comparative example. First, read gradation value data is acquired for all pixels for each band-like pattern to be measured (for each density region formed with the same command gradation value) (S331). Then, an average gradation value (average density) Ar is obtained for the r-th pixel column among the obtained reading gradation values of all the pixels (S332). Further, the standard deviation σr of the gradation value (density) is calculated for the r-th pixel column from the average gradation value Ar and the read gradation value of each pixel (S333). Then, using Ar and σr, a threshold value for calculating the gradation value (density) of the r-th pixel column is obtained (S334). That is, a different threshold value is obtained for each pixel column. The threshold value is calculated by the following (Formula 2).
Threshold = Ar ± d × σr (Expression 2)
D in (Expression 2) is a coefficient, and an optimum value is determined according to the type of medium to be printed (for example, glossy paper or plain paper) as in the case of (Expression 1).

  Using the threshold value for the r-th pixel column, the read gradation value of the r-th pixel column is acquired (S304). The calculation method of the gradation value is the same as that in (S204), and a foreign substance flag is set for the gradation value outside the threshold range, and the average value of the gradation values of the pixels for which the foreign substance flag is not set is calculated as the rth value. The gradation value of the second pixel column is used. Thereafter, similarly to (S205), a correction value H for correcting the gradation value of the r-th pixel column is calculated (S305).

<Effect of Comparative Example>
In the comparative example, a threshold value for detecting dust or the like included in the pixel column is calculated for each pixel column corresponding to the column region where the density correction is performed, and thus there is a variation in density between the pixel columns of the printed test pattern. Even if it occurs, the threshold value for each pixel column is calculated as an appropriate value. That is, even when the density variation of the test pattern is large, the influence of such variation is small, so that the problem that the threshold detection range becomes too large and the dust detection accuracy decreases does not occur. Thus, when printing is performed using a printing apparatus that forms an image by ejecting ink from a plurality of heads like the printer 1, it is easy to calculate an accurate density correction value H.

  On the other hand, in the comparative example, the threshold value is calculated by calculating the average gradation value (average density) Ar and the standard deviation σr for each pixel column, so that the amount of calculation is larger than that in the first embodiment. There is a possibility that the load applied to the CPU 62 becomes heavy or the printing speed decreases.

  Further, since the calculation of the threshold value is completed for each pixel column, there is a problem that the influence of dust increases when the number of pixels constituting the pixel column is small. For example, when dust is included in two pixels of a pixel row composed of five pixels, the average gradation value Ar and the standard deviation σr of the pixel row are very high compared to the case where there is no dust. Therefore, it is conceivable that the threshold value is calculated to be larger than necessary and the dust detection accuracy is deteriorated.

=== Second Embodiment ===
In the second embodiment, an appropriate threshold value is calculated while minimizing the influence of density variation even in a printing apparatus that performs printing using a plurality of heads while solving the above-described problems of the comparative example.

  FIG. 22 shows a flowchart for calculating the correction value H in the second embodiment. The overall flow is the same as (S201) to (S205) of the first embodiment and (S301) to (S305) of the comparative example, and the threshold calculation method of (S403) is different. Hereinafter, (S403) will be mainly described.

<S403: Threshold calculation>
FIG. 23 shows a flow for calculating a threshold in the second embodiment. First, read gradation value data is acquired for all pixels for each band-like pattern to be measured (for each density region formed with the same command gradation value) (S431). Then, among the obtained read gradation values of all the pixels, an average gradation value (average density) Ar is obtained for the r-th pixel column (S432). Further, an average gradation value Aall in the density region is calculated from the read gradation values of all the pixels, and the standard deviation σall of the gradation value (density) for each pixel is calculated from Aall and the read gradation value of each pixel. Is calculated (S433). Then, a threshold value for calculating the gradation value (density) of the r-th pixel column is obtained from Ar and σall (S334). The threshold value is calculated by the following (Formula 3).
Threshold = Ar ± d × σall (Expression 3)
D in (Expression 3) is a coefficient, and an optimum value is determined according to the type of medium to be printed (for example, glossy paper or plain paper) as in the case of (Expression 1).

  In other words, in the present embodiment, the threshold value of the r-th pixel column is calculated from the variation of the gradation values of the entire density region and the average of the gradation values of the r-th pixel column (S403). Then, the read gradation value of the r-th pixel column is acquired using the calculated threshold value for the r-th pixel column (S404). The calculation method of the gradation value is the same as that in (S204), and a foreign substance flag is set for the gradation value outside the threshold range, and the average value of the gradation values of the pixels for which the foreign substance flag is not set is calculated as the rth value. The gradation value of the second pixel column is used. Thereafter, similarly to (S205), a correction value H for correcting the gradation value of the r-th pixel column is calculated (S405).

<Effects of Second Embodiment>
Also in the method of the second embodiment, the threshold value for removing the influence of dust attached to the test pattern can be automatically calculated.

  The second embodiment is the same as the comparative example in that a threshold value corresponding to a pixel column is calculated for each pixel column corresponding to a column region where density correction is performed. It is calculated from the standard deviation of the gradation value of the entire pattern and the average gradation value of each pixel column in the strip pattern. That is, the average density serving as a threshold reference is obtained for each pixel column, but the variation in the density (gradation value) that defines the threshold range is obtained from the entire strip pattern of the test pattern.

  As a result, even when the density (gradation value) varies greatly in the belt-like pattern, the density serving as the reference for the threshold is the average density of the pixel column, and therefore is affected by the variation in density between the pixel columns. The threshold value is calculated as an appropriate value for each pixel column. Therefore, even for a printing apparatus that forms an image by ejecting ink from a plurality of heads like the printer 1, the accurate density correction value H is calculated without reducing the dust detection accuracy for each pixel row. can do.

  Although it is necessary to calculate a different threshold value for each pixel column, the standard deviation σall needs to be calculated only once. Therefore, the amount of calculation can be significantly reduced compared to the threshold value calculation method in the comparative example, and the problems that the load on the CPU 62 increases and the printing speed decreases are solved.

  Furthermore, even when the number of pixels in the pixel array is small, the density variation is calculated from the gradation values of the entire belt-like pattern, so that the threshold does not become too large as in the comparative example, and the dust detection accuracy also deteriorates. It becomes difficult to do.

=== Other Embodiments ===
Although a printer or the like as one embodiment has been described, the above embodiment is for facilitating the understanding of the present invention, and is not intended to limit the present invention. The present invention can be changed and improved without departing from the gist thereof, and it is needless to say that the present invention includes equivalents thereof. In particular, the embodiments described below are also included in the present invention.

<About printing devices>
In the above embodiment, the lateral scan type printer 1 has been described as an example. However, the printer may be a so-called line printer with a fixed head, or a serial printer that moves the head 41 together with the carriage. May be.

<Ink used>
In the above-described embodiment, an example of recording using four colors of CMYK inks has been described. However, the present invention is not limited to this. For example, recording may be performed using inks of colors other than CMYK, such as light cyan, light magenta, white, and clear.

<Regarding the arrangement of nozzle rows>
The nozzle rows of the head portion are arranged in the order of KCMY along the transport direction, but the present invention is not limited to this. For example, the order of the nozzle rows may be changed, or the number of nozzle rows for K ink may be greater than the number of nozzle rows for other inks.

<About the printer driver>
The printer driver processing may be performed on the printer side. In that case, a printer is constituted by the printer and the PC on which the driver is installed.

1 printer, 5 manuscripts,
20 transport units, 21 transport rollers,
30 drive unit, 31 X-axis stage, 32 Y-axis stage,
40 head units, 41 heads,
50 detector groups, 60 controllers,
61 interface unit, 62 CPU, 63 memory,
64 unit control circuit, 110 computer,
120 display device, 130 input device, 140 recording / reproducing device,
150 scanner, 151 top cover, 152 platen glass,
153 reading carriage, 154 guide section, 155 moving mechanism,
157 Exposure lamp, 158 line sensor, 159 optical system

Claims (6)

(A) By ejecting ink from nozzle rows arranged in a predetermined direction, a pattern in which dot rows formed along the intersecting direction intersecting the predetermined direction are arranged in the predetermined direction is formed on the medium;
(B) In the image data obtained by reading the pattern, an average density value for all pixels corresponding to the pattern, and a value obtained by multiplying a standard deviation of density for all pixels corresponding to the pattern by a predetermined coefficient, An upper threshold calculated by the sum of
In the image data obtained by reading the pattern, the difference between the average value of the density for the entire pixel corresponding to the pattern and the value obtained by multiplying the standard deviation of the density for the entire pixel corresponding to the pattern by a predetermined coefficient. A lower threshold value to be calculated,
(C) the pattern from the image data read for each pixel constituting the pixel row corresponding to each of the dot rows of the pattern to obtain the concentration compared with the upper threshold and the lower threshold, the concentration of the upper Using the pixel data included in the range between the threshold and the lower threshold , calculate the density for each pixel column,
(D) obtaining a correction value for correcting the density corresponding to each dot row of the pattern, based on the density calculated for each pixel row;
The correction value acquisition method characterized by this. (A) By ejecting ink from nozzle rows arranged in a predetermined direction, a pattern in which dot rows formed along the intersecting direction intersecting the predetermined direction are arranged in the predetermined direction is formed on the medium;
(B) In the image data obtained by reading the pattern, a predetermined coefficient is applied to the average value of the density for the entire pixel corresponding to the pattern and the standard deviation of the density for the pixel row corresponding to each dot row of the pattern. An upper limit threshold calculated by the sum of the multiplied value,
In the image data obtained by reading the pattern, a value obtained by multiplying a predetermined coefficient by an average density value for all pixels corresponding to the pattern and a standard deviation of density for pixel rows corresponding to the respective dot rows of the pattern. And a lower threshold value calculated by the difference between
(C) the pattern from the image data read for each pixel constituting the pixel row corresponding to each of the dot rows of the pattern to obtain the concentration compared with the upper threshold and the lower threshold, the concentration of the upper Using the pixel data included in the range between the threshold and the lower threshold , calculate the density for each pixel column,
(D) obtaining a correction value for correcting the density corresponding to each dot row of the pattern, based on the density calculated for each pixel row;
The correction value acquisition method characterized by this. The correction value acquisition method according to claim 2 ,
A correction value acquisition method, wherein the pattern is formed using a plurality of heads having the nozzle row. (A) Stored in memory,
For the dot rows formed along the intersecting direction intersecting the predetermined direction by the ink ejected from the nozzle rows arranged in the predetermined direction,
A correction value acquisition program for acquiring a correction value corresponding to each dot row,
(B) the dot rows is formed in the predetermined direction to align the pattern of the medium member,
( C ) In the image data obtained by reading the pattern, an average value of the density for the entire pixel corresponding to the pattern, and a value obtained by multiplying the standard deviation of the density for the entire pixel corresponding to the pattern by a predetermined coefficient, An upper threshold calculated by the sum of
In the image data obtained by reading the pattern, the difference between the average value of the density for the entire pixel corresponding to the pattern and the value obtained by multiplying the standard deviation of the density for the entire pixel corresponding to the pattern by a predetermined coefficient. A lower threshold value to be calculated,
(D) the pattern from the image data read for each pixel constituting the pixel row corresponding to each of the dot rows of the pattern to obtain the concentration compared with the upper threshold and the lower threshold, the concentration of the upper Using the pixel data included in the range between the threshold and the lower threshold , calculate the density for each pixel column,
(E) Based on the density calculated for each pixel row, a correction value for correcting the density corresponding to each dot row of the pattern is acquired.
A correction value acquisition program characterized by that. (A) Stored in memory,
For the dot rows formed along the intersecting direction intersecting the predetermined direction by the ink ejected from the nozzle rows arranged in the predetermined direction,
A correction value acquisition program for acquiring a correction value corresponding to each dot row,
(B) the dot rows is formed in the predetermined direction to align the pattern of the medium member,
(C) In the image data obtained by reading the pattern, a predetermined coefficient is applied to the average value of the density for the entire pixel corresponding to the pattern and the standard deviation of the density for the pixel row corresponding to each dot row of the pattern. An upper limit threshold calculated by the sum of the multiplied value,
In the image data obtained by reading the pattern, a value obtained by multiplying a predetermined coefficient by an average density value for all pixels corresponding to the pattern and a standard deviation of density for pixel rows corresponding to the respective dot rows of the pattern. And a lower threshold value calculated by the difference between
(D) the pattern from the image data read for each pixel constituting the pixel row corresponding to each of the dot rows of the pattern to obtain the concentration compared with the upper threshold and the lower threshold, the concentration of the upper Using the pixel data included in the range between the threshold and the lower threshold , calculate the density for each pixel column,
(E) Based on the density calculated for each pixel row, a correction value for correcting the density corresponding to each dot row of the pattern is acquired.
A correction value acquisition program characterized by that. (A) a head portion having nozzle rows arranged in a predetermined direction, and forming dots by ejecting ink onto the medium;
(B) a control unit for controlling the head unit,
Forming a pattern in which dot rows along an intersecting direction intersecting the predetermined direction are arranged in the predetermined direction on the medium;
In the image data obtained by reading the formed pattern, an average value of the density for the entire pixel corresponding to the pattern, and a value obtained by multiplying the standard deviation of the density for the entire pixel corresponding to the pattern by a predetermined coefficient, An upper threshold calculated by the sum of
In the image data obtained by reading the pattern, the difference between the average value of the density for the entire pixel corresponding to the pattern and the value obtained by multiplying the standard deviation of the density for the entire pixel corresponding to the pattern by a predetermined coefficient. The lower threshold value to be calculated,
A density is acquired for each pixel constituting a pixel column corresponding to each dot column of the pattern from the image data obtained by reading the pattern, and the density is compared with the upper limit threshold and the lower limit threshold. Using the pixel data included within the range with the lower threshold , calculate the density for each pixel column,
Obtaining a correction value for correcting the density corresponding to each dot row of the pattern, based on the density calculated for each pixel row;
A control unit, characterized by
A printing apparatus comprising: