JP2008257394A - Unit, method and program for image processing - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To securely detect a position and direction of a ruled line with high accuracy, even when a spurious image is formed in the vicinity of the ruled line, when detecting the position and direction of the ruled line in an image data having the described ruled line. <P>SOLUTION: An image processing method includes the steps of: extracting a plurality of pixel rows each aligned in a substantially perpendicular direction to the ruled line; calculating an average density of the extracted plurality of pixel rows among the pixels disposed in the ruled line direction; comparing the above average density with a predetermined density threshold in a successive order from one end of the pixel rows; obtaining an interval from the first pixel exceeding the predetermined threshold to the last pixel; if the above interval ranges within a predetermined width threshold, detecting the ruled line position to be from the first pixel to the last pixel by determining that a spurious image does not exist from the first pixel to the last pixel, while if the above interval exceeds the predetermined width threshold, determining that the spurious image exists between the first pixel and the last pixel. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an image processing device, an image processing method, and an image processing program, and more particularly, to an image processing device, an image processing method, and an image processing program for detecting ruled lines from an image including ruled lines.

  Conventionally, OCR (Optical Character Recognition) or the like is known as a technique for detecting a predetermined shape such as a character from image data generated by scanning with a scanner or the like and recognizing the shape. In such a shape recognition technology, the generated image data has an image of dust around the shape to be detected due to dust and dirt adhering to the paper surface to be scanned and the reading surface of the scanner. When (hereinafter referred to as a dust image) is formed, there is a problem that the detection accuracy decreases.

In order to solve such a problem, a character image for one character is cut out from the image data, the number of segments included in the character image, the area of each segment, the distribution of the line width of the character image, the character dictionary and the character image, A technique is known in which a dust image is erased from a character image based on the collation of the character to improve the accuracy of character recognition (for example, Patent Document 1).
JP-A-7-105312

  By the way, as a technique for detecting and using a predetermined shape from image data, there are automatic trimming and automatic upright rotation of image data. In order to perform trimming of a range surrounded by straight lines such as ruled lines, for example, ruled lines are first detected. In this ruled line detection, a part where pixels having a predetermined density or more are continuously formed in a straight line with a predetermined width is detected. However, if a dust image is formed in the vicinity of the ruled line, the width of the ruled line changes, and even if the ruled line detection fails or the ruled line detection is successful, there is a problem with the accuracy of the detected ruled line position and ruled line direction. There will be.

  In the upright rotation, the character recognition as in the technique described in Patent Document 1 is performed, the inclination is obtained from the recognized character, and the rotation process is performed to correct the obtained inclination. However, in general, characters are small in size, and there is a limit in accuracy at an angle obtained from the inclination of the characters. Therefore, in order to perform accurate upright rotation, it is preferable to describe a reference line composed of straight lines such as ruled lines in the image data and perform upright rotation to correct the inclination of the reference line.

  As described above, in order to perform automatic trimming and automatic upright rotation with high accuracy, it is necessary to accurately detect ruled lines, and for this purpose, it is necessary to reduce the influence of dust images. However, with the ruled lines, it is impossible to remove dust such as cutting out a character image for one character, or decomposing the cut out character image into segments to detect a dust image. Is difficult.

  The present invention has been made in view of the above problems, and in detecting the ruled line position and ruled line direction of image data in which ruled lines are described, even when a dust image is formed in the vicinity of the ruled line, Another object of the present invention is to provide an image processing apparatus, an image processing method, and an image processing program capable of accurately and reliably detecting the direction of a ruled line.

  In order to solve the above-described problem, the present invention provides an image processing apparatus for detecting a ruled line position of image data in which a ruled line having a predetermined width is described, and includes a column of pixels lined up in a direction substantially orthogonal to the ruled line. A plurality of columns, and an average density calculating means for calculating an average density between pixels located in a line in the ruled line direction in the extracted plurality of pixels, and the average density sequentially from one end of the pixel column And a ruled line width detecting means for obtaining an interval from the first pixel to the last pixel that exceeds the predetermined density threshold, and if the interval falls within the predetermined width threshold, It is determined that there is no dust image from the last pixel to the last pixel and the first pixel to the last pixel is detected as a ruled line position, while if the interval exceeds a predetermined width threshold, From pixel to the last pixel It is constituted comprising, a dust determination means for determining that the dust image exists.

Further, the above-described image processing apparatus includes various modes such as being implemented in another device or being implemented together with another method. Furthermore, the present invention provides a printing system including the above-described image device, a control method having a process corresponding to the configuration of the above-described device, a program for causing a computer to realize a function corresponding to the above-described device configuration, and a computer reading that records the program It can also be realized as a possible recording medium. The inventions of the printing system, the image transmission method, the image transmission program, and the medium on which the program is recorded also have the above-described operations and effects. Of course, the configurations described in claims 2 to 5 are also applicable to the system, the method, the program, and the recording medium.
Other features of the present invention will become apparent from the description of the present specification and the accompanying drawings.

At least the following matters will become clear from the description of the present specification and the accompanying drawings.
According to the configuration of the main invention, the average density calculating unit extracts a plurality of columns of pixels adjacent to one column in a direction substantially orthogonal to the ruled line, and the ruled line direction in the extracted plurality of pixel columns The average density is calculated between the pixels at the positions aligned with. That is, since the density is averaged among the plurality of pixel columns, the density of the image excluding the ruled lines that can be uniformly included in each pixel column decreases. Accordingly, only the density of the ruled lines becomes apparent in the distribution of the average density, and the discrimination between the dust image and the ruled lines becomes easy.

  Then, the ruled line width detection means compares the average density with a predetermined density threshold in order from one end of the pixel row, and obtains an interval from the first pixel to the last pixel that exceeds the predetermined density threshold. . That is, a predetermined density threshold is set, and the interval between the first pixel and the last pixel having a density equal to or higher than the threshold is set as the width of the ruled line.

  When the interval falls within a predetermined width threshold, the dust determination unit determines that there is no dust image from the first pixel to the last pixel, and determines from the first pixel to the last pixel. While it is detected as a ruled line position, if the interval exceeds a predetermined width threshold value, it is determined that a dust image exists from the first pixel to the last pixel. That is, the preset width threshold value is compared with the width of the ruled line detected by the ruled line width detecting means. Therefore, even if a dust image is included in the image data more than expected and the density of the dust image is not sufficiently reduced compared to the ruled line density even with the average density, the dust image may be mistaken for a ruled line. There is no.

  Here, the column of pixels arranged in a line in a direction substantially orthogonal to the ruled line is one pixel in the ruled line direction and a plurality of pixels arranged in a direction substantially orthogonal to the ruled line. Further, the plurality of pixel columns may be adjacent to each other in the ruled line direction, or a position having a gap in which there is a pixel column that is not extracted between the pixel columns. It may be related. The dust image is an image formed contrary to the intention of the creator of the image data. For example, the dust image is caused by noise components or dust attached to the reading surface when optically reading a document. Thus, it is formed on the read image.

  In such an image processing apparatus, when the dust determination unit determines that a dust image exists from the first pixel to the last pixel, the plurality of pixels used in the calculation of the average density are determined as the plurality of pixels. It is desirable to perform recalculation of the average density by replacing with a column of pixels located in the vicinity of the column. According to such an image processing apparatus, instead of a plurality of pixel rows used for calculating the average density first, a pixel row that is less likely to be affected by the dust image is calculated, and as a result, a ruled line. This is used for detecting the width of the image, and there is a possibility that the influence of the dust image can be eliminated.

  In such an image processing apparatus, when the interval exceeds the predetermined width threshold, the dust determination unit recalculates the average density by increasing the number of pixel columns used for the calculation of the average density. Is desirable. According to such an image processing apparatus, even if a dust image is included in a specific pixel column, the number of samples used for calculating the average density is increased, so that the influence of the dust image can be reduced. I can do it. Therefore, the width of the ruled line detected by the ruled line width detecting means may be accurate.

  In such an image processing apparatus, it is preferable that the ruled line position is approximately in the middle of the first pixel and the last pixel. Since the ruled line width detecting means detects the width of the ruled line using a predetermined density threshold value, it is not clear which position of the ruled line corresponds to both ends thereof. However, the ruled line density is distributed approximately symmetrically across the middle in the width direction. Therefore, it can be expected that the approximate middle of the width of the detected ruled line substantially matches the middle of the actual ruled line.

  In such an image processing apparatus, printing that causes a printing apparatus to print image data having a predetermined color and a predetermined density in a predetermined range in the sub-scanning direction together with ruled lines extending in one or both of the main scanning direction and the sub-scanning direction of the image data. Means, print result acquisition means for acquiring an optically read image as a result of the printing from the optical reading device, and erect rotation and trimming of the optically read image based on the ruled line position of the image data detected by the ruled line detecting means Image correction means for performing image correction comprising either or both of the above, and a density change in the sub-scanning direction is measured from the optically read image after the image correction, and a correction value for eliminating the density change is generated in the printing apparatus. It is desirable to include correction value setting means for setting. In such an image processing apparatus, if the detected ruled line position is used for image correction such as erecting rotation or trimming of the optically read image, it contributes to improvement in accuracy.

Hereinafter, embodiments of the present invention will be described in the following order.
(1) Configuration of printing system:
(2) Density unevenness correction value setting processing:
(3) Modified example of ruled line detection processing:

(1) Configuration of printing system:
FIG. 1 is a block diagram showing a printing system to which a density correction printing control apparatus according to an embodiment of the present invention is applied.
In the figure, the computer 10 corresponds to a personal computer or the like. The computer 10 includes a CPU, ROM, RAM as hardware, an external storage device such as a hard disk (not shown), a scanner 30 (optical reader), and the like, a keyboard, a mouse, etc. as input devices, and a display device. A display is provided. The computer 10 realizes a predetermined function by organically linking software and hardware. In this embodiment, the printer configuration includes a printer driver 10c, an application 10b, a correction value setting program 10a, and the like as shown in FIG.

  The application 10b outputs dot matrix image data to the printer driver 10c. This image data is gradation data in which each RGB color component is expressed with 256 gradations (RGB data). The printer driver 10c that has received RGB data from the application 10b performs resolution conversion processing, color conversion processing, gradation conversion processing, and rasterization processing on the RGB data. In the present embodiment, the resolution conversion process, the color conversion process, and the gradation conversion process are handled by the halftone (H / T) module 10c1, and the rasterization process is performed by the interlace module 10c2.

  When receiving the RGB data, the H / T module 10c1 first performs resolution conversion processing. In the resolution conversion process, the resolution of the color image data handled by the application program is converted to a resolution that can be handled by the printer driver 10c. Next, the H / T module 10c1 performs color conversion processing. The color conversion process is a process of converting RGB gradation values into CMYK (C: cyan, M: magenta, Y: yellow, K: black) gradation values, and each dot data of image data expressed in RGB is converted. Conversion into CMYK dot data (CMYK data). Thereafter, the H / T module 10c1 performs gradation conversion processing on the CMYK data. The gradation conversion process is a halftone process for converting the CMYK gradation value of each dot and expressing it by the distribution of the dots of the ink droplets. The head drive data for attaching the ink at the converted recording density is obtained. Generate.

  When receiving the head drive data, the interlace module 10c2 performs a rasterization process. In the printer 20, an ejection nozzle array (not shown) is mounted as an ink ejection device. In the nozzle array, a plurality of ejection nozzles are arranged in parallel in the sub-scanning direction. used. Therefore, rasterization is performed in which the head drive data arranged in the main scanning direction is rearranged in order so that data to be used simultaneously is buffered by the printer 20. Here, the main scanning direction is the direction in which the print head reciprocates, whereas the paper feed direction is the sub-scanning direction. After the rasterization process, the interlace module 10c2 generates print data to which predetermined information such as image resolution is added, and outputs the print data to the printer 20 via the I / F to perform printing.

  The printer 20 includes at least a CPU, a ROM, and a RAM, and mainly controls the control unit 21, a print head unit 25 that discharges color ink, and a main scanning direction in which the head unit is orthogonal to the paper feed direction of the printing paper. A carriage unit 26 that reciprocates the paper, a paper feed unit 27 that conveys printing paper in the sub-scanning direction, which is the paper feed direction, a unit control circuit 24 that drives the units 25 to 27 in accordance with the control of the control unit 21, and a computer And an interface (I / F) 29 connected to 10 I / Fs.

  When the printer 20 receives print data from the computer 10, the control unit 21 controls each unit (print head unit, carriage unit, paper feed unit) to perform printing based on the print data. In addition to these, the printer 20 monitors the internal conditions of the printer 20 (paper position, ink amount, print head position, etc.) and outputs a detection result to the control unit 28, and the I / O of the computer 10 An I / F 29, a non-volatile memory 23, and the like that are communicably connected to the F15 are provided.

  FIG. 2 is a diagram showing the nozzle arrangement of the print head. The print head unit 25 includes a print head 25a composed of a plurality of nozzles. The print head unit 25 receives the applied voltage data corresponding to the print data from the unit control circuit 24, generates an applied voltage pattern to the piezo element, and drives the piezo element to compress and expand the ink chamber of each nozzle. Then, ink droplets are ejected from the nozzles. This print head is provided with a plurality of nozzles corresponding to each color ink, and is aligned for each color ink. Nozzle rows corresponding to each color ink are aligned at a constant nozzle pitch along the transport direction.

  The carriage unit 26 reciprocates the print head unit 25 in the main scanning direction by a motor. The carriage unit 26 reciprocates in synchronism with the ejection timing so that the dots of the color ink ejected from the print head adhere to a predetermined position in the main scanning direction on the printing paper under the control of the unit control circuit 24. .

  The paper feed unit 27 conveys paper in the sub-scanning direction. The paper feeding unit includes a conveyance roller, and feeds a predetermined amount of paper in synchronization with the reciprocation of the print head unit.

  The printer 20 of the present embodiment described above is described as an ink jet printer that prints by forming ink dots on a print medium by ejecting CMYK (cyan, magenta, yellow, and black) inks from a print head. However, of course, the printing device can use C, M or K light color, Y dark color, red, violet, uncolored ink, etc., or any device that does not use any of CMYK. is there.

  By the way, the printer driver 10c includes a correction module 10c3 that corrects density unevenness that occurs in the main scanning direction. The correction module 10c3 acquires and acquires image data from the H / T module 10c1, and performs processing for correcting density unevenness on the acquired image data according to the correction value 13a stored in the memory 13. Here, correction of density unevenness realized by the correction module will be described.

  As shown in FIG. 2, the nozzles of the print head 25a are aligned in the sub-scanning direction for each color. Each nozzle forms a raster by ejecting ink droplets onto the paper surface while scanning the print head in the main scanning direction. However, the ink discharge directions of the nozzles are not completely the same, and there are variations in the ink discharge directions. For example, the ink ejection direction may be directed upstream in the main scanning direction, or may be directed downstream in the main scanning direction. Therefore, if ink is ejected from each nozzle evenly, density unevenness appears in the printing result due to variations in the ink ejection direction.

  Therefore, when printing with the printer 20, using the fact that the nozzles used for printing each raster in the print data are uniquely determined, the measurement pattern is printed, the density unevenness of the printing result is measured, and the density unevenness is measured. A correction value that eliminates the problem is created and set in the printer. When printing is performed using print data corrected based on the correction value, it is possible to prevent such density unevenness from occurring. Thus, the correction module 10c3 that corrects the image data based on the correction value constitutes an image correction unit.

  This correction value is created by the following density unevenness correction value setting process and set in each printer. The correction value setting process is executed at a factory or the like before shipment of the printer, and a correction value created for each printer is set. Then, when a printer driver or the like is installed by a user who has purchased a printer with a correction value set, it is read from the printer and stored in the computer. Using the stored correction value, the correction module corrects the image data.

(2) Density unevenness correction value setting processing:
Hereinafter, setting of correction values for correcting density unevenness will be described. The correction value is set by causing the printer 20 to print a test pattern in the inspection process of the printer manufacturing factory, reading the test pattern with the scanner 30, and measuring the change in the density of the test pattern in the sub-scanning direction. The correction value is stored in the memory 13 of the printer 20. That is, the correction value stored in the printer 20 reflects the characteristic of density unevenness in each printer 20.

FIG. 3 is a flowchart showing the process of the correction value setting program executed in the inspection process after manufacturing the printer 20.
First, the inspector connects the printer 20 to be inspected to the computer 10 and executes the correction value setting program. A scanner 30 is also connected to the computer 10 in advance, and a correction value is calculated from the printer driver 10c of the printer 20, the scanner driver for controlling the scanner 30, and the image data of the test pattern read by the scanner 30. Then, the correction value setting program 10a for setting the correction value in the printer 20 is installed.

When the process is started, in step S30, the printer 20 is caused to print a test pattern. In step S30, the computer 10 that causes the printer 20 to print a test pattern constitutes a printing unit.
FIG. 4 is an explanatory diagram of a test pattern. This test pattern is printed at a print resolution of, for example, 720 × 720 dpi, and four correction patterns are arranged in order of K, C, M, and Y in order from the left. Each correction pattern is composed of five types of density (gradation value) band patterns, an upper ruled line surrounding the band pattern, a lower ruled line, a left ruled line, and a right ruled line.

  Each of the belt-like patterns is formed as image data that has a constant gradation value in a predetermined range in the sub-scanning direction and in which belt-like patterns having different gradation values are arranged in the main scanning direction. In FIG. 4, the darker density pattern is in order from the left belt-like pattern. Note that when the test pattern is printed, the density correction process based on the correction value by the correction module 10c3 described above is not performed.

  Next, the inspector sets the paper on which the test pattern is printed by the printer 20 to the reading position of the scanner 30 and causes the reading to be performed. Then, the computer 10 acquires image data (optical read image) of the read test pattern via the scanner driver (step S32).

  FIG. 5 is an explanatory diagram of the reading range of the correction pattern. The range of the alternate long and short dash line surrounding the correction pattern is the range where scanning is performed. Parameters SX1, SY1, SW1, and SH1 for specifying this range are set in advance in the scanner driver by the correction value setting program. This range is set wider than the test pattern. If this range is read by the scanner 30, the entire correction pattern is read even if the document on which the test pattern is printed is set slightly on the scanner 30. be able to. By this processing, an image in the reading range in the figure is generated as high-resolution rectangular image data and is captured by the computer 10.

  A computer 10 that acquires an optically read image via the scanner driver constitutes a print result acquisition unit. Then, the following steps S36 to S48 are executed to detect a correction pattern from the read image and calculate and set a correction value. The computer 10 that executes the correction value setting program 10a for calculating and setting the correction value constitutes a correction value setting means.

  Next, the correction value setting program 10a performs the ruled line detection process of FIG. 6 on the cut image to detect the ruled line position, and performs image processing (S38) for rotating the read image upright. The ruled line detection process performed here is performed not only in the process of rotating the read image upright but also in the trimming process and the pattern position detection process.

  In the ruled line detection process shown in FIG. 6, first, ruled line detection is performed from an average density obtained by averaging the densities between pixel columns at predetermined positions in a column of pixels arranged in a line in a direction substantially orthogonal to the ruled line. The result of the ruled line detection process is used to identify the test pattern range and to identify which nozzle has printed each raster printed on the test pattern, and therefore requires high accuracy.

  FIG. 8 is an explanatory diagram of calculation of the average density and ruled line detection using the average density. Of course, the ruled line detection can be performed from the density distribution of any one of the pixel columns arranged in the main scanning direction (aligned in the sub-scanning direction in the trimming process described later). However, the scanned image may include a dust image due to noise, or a dust image having a predetermined size or less may be formed due to the presence of fine dust, dust, or the like attached to the document. In such a case, a normal ruled line can be detected as long as the pixel row does not include a dust image, such as a1, a2, an-1, and an as shown in FIG. If ruled line detection is performed from a pixel column in which a dust image exists such as a3 or pixel column an-2, the ruled line position is shifted, and rotation processing, trimming processing, and pattern position detection processing described later cannot be performed correctly.

  Therefore, attention is paid to the fact that the frequency of occurrence of these dust images in each pixel column is random, while the ruled lines are equally included in any pixel column. That is, when the density is averaged between the pixel columns, randomly generated dust images are averaged and the density is lowered, whereas the ruled line density is maintained at a constant density.

  FIG. 7 is a diagram for explaining a column of pixels used for ruled line detection processing. In the figure, the upper left of the reading range is the origin, the right direction is the x direction, and the lower direction is the y direction. As shown in the drawing, each of the KX1 and KX2 columns includes n pixel columns. The KX3 to KX4 columns are not shown, but are the same as the KX1 to KX2 columns. In step S100, the pixel columns KX1 to KX2 and KX3 to KX4 are extracted. In step S102, the density is set between the pixels arranged in the direction substantially coincident with the ruled line direction (between pixels having the same y coordinate) between the KX1 to KX2 columns and the KX3 to KX4 column. The average is calculated as shown in FIG.

Here, if the density (gradation value) of the pixel at the position (x, y) of the read image is D (x, y), the average density Dave between the pixel columns of the KX1 to KX columns is
It becomes. Similarly, the average density between the pixel columns KX3 to KX4 is
It is represented by Here, KX1 to KX2 and KX3 to KX4 are set so as not to interfere with the vertical ruled lines and the belt-like pattern.

  With the average density calculated in this way, the density of the dust image is lowered. However, the density of the dust image is not completely zero. Therefore, in order to further reduce the influence of the dust image on the ruled line detection, a density threshold is set, and in step S104, a part exceeding the density threshold is detected as a ruled line. That is, if it falls within the density threshold, it is determined that the density of the dust image is not a ruled line. This density threshold is set to be equal to or higher than the density of dust images that can be included in the average density in consideration of the appearance degree of dust images and the like.

  In the comparison with the density threshold value, the correction value setting program compares the average density of the pixel columns with the density threshold value th1 from the top up to KH. Then, the interval KW1 from the first pixel KH1 exceeding the density threshold to the last pixel KH2 exceeding the density threshold is determined as the width of the ruled line. As described above, by appropriately setting the density threshold value, it is possible to more reliably eliminate the influence of the dust image.

  By the way, when there are few dust images or when there is no big dust, the ruled line detection is successful by taking the average density and setting the part exceeding the density threshold as the ruled line. However, when a large number of dust images are included, the density of the dust images does not decrease even when the average density is taken, or the positions used for ruled line detection (KX1 to KX2, KX3 to KX4) are close to the ruled lines. If a large dust image is formed and the width of the ruled line includes the width of the dust image, the ruled line width KW1 may be larger than the actual size.

  Accordingly, as shown in FIG. 9, an upper limit (width threshold value) KW0 of the ruled line width is set in advance, and the ruled line width KW1 is determined in step S106. Is determined to exist. More specifically, the detected ruled line width KW1 and the width threshold value KW0 are compared, and if the ruled line width KW1 is equal to or smaller than the width threshold value KW0, the ruled line detection is successful, and if the ruled line width Kw1 exceeds the width threshold value KW0, the ruled line The correction value setting program is terminated as a detection failure.

  In step S108, the ruled line position is determined from the detected ruled line width. In the present invention, a part exceeding the density threshold is defined as a ruled line width. Therefore, depending on how to set the density threshold, the first position exceeding the density threshold and the last position exceeding the density threshold are different from the positions at both ends of the actual ruled line. Therefore, the middle of the width of the detected ruled line is set as the ruled line position. This makes it possible to always recognize the same position as the ruled line position regardless of what density threshold value is employed. Of course, the ruled line position determination method is not limited to the middle of the detected ruled line width, and for example, the barycentric position of the gradation value included in the average density may be considered as the ruled line position.

Based on the ruled line position detected in the above ruled line detection process, the inclination θ of the correction pattern is detected from the ruled line position included in the read image, and the read image is rotated by rotating the read image according to the inclination θ. Image processing for erecting is performed. Specifically, the inclination θ of the correction pattern is calculated by the following equation, and image data rotation processing is performed based on the calculated inclination θ.
Since this rotation process is performed based on the ruled line position detected by the ruled line detection process, it is possible to perform an accurate rotation process by using an accurate ruled line position without the influence of the dust image.

Next, the correction value setting program performs image processing for trimming unnecessary pixels from the read image.
FIG. 10 is an explanatory diagram of a read image at the time of trimming. First, the correction value setting program includes pixel data of KH1 to KX2 columns, pixel data of KH pixels from the top, and pixel rows of KX3 to KX4 from the read image that has been rotated and erected. Then, pixel data of KH pixels are extracted from the top. Similarly, the correction value setting program includes pixel data of KH1 to KX2 columns from the bottom, and pixel data of KH pixels from the bottom, and pixel columns of KX3 to KX4 from the read image that has been rotated and erected. Then, pixel data of KH pixels are taken out from the bottom. Then, similarly to the ruled line detection process described above, the average density of each pixel column is obtained, and the ruled line position is detected using the density threshold value and the width threshold value.

  Based on the detected ruled line position, the upper ruled line is lower than the detected ruled line position by 1/2 of the ruled line width, and the lower ruled line is lower than the detected ruled line position by 1/2 of the ruled line width. , The nearest pixel boundary is determined as the trimming position. Then, the correction value setting program cuts out the pixels above the trimming position of the upper ruled line and the pixels below the trimming position of the lower ruled line, and performs trimming. The ruled line width for determining the trimming position is set in advance in the correction value setting program.

  This trimming process is also performed based on the ruled line position. Therefore, by detecting the exact ruled line position in the ruled line detection process, the raster position of the print data and the raster position of the read image correspond to each other accurately, and correction by the correction module can be performed correctly.

Next, the correction value setting program converts the resolution of the trimmed read image so that the number of pixels in the Y direction is the same as the number of dot rows constituting the correction pattern.
FIG. 11 is an explanatory diagram of resolution conversion. If the printer 20 ideally forms a correction pattern and the scanner 30 ideally reads the correction pattern, the number of pixels in the Y direction of the read image after trimming should be a predetermined value. However, in actuality, the number of pixels in the Y direction of the read image may not be a predetermined value due to the influence of misalignment during printing or reading.

  Therefore, the correction value setting program performs resolution conversion (reduction processing) on the read image at a magnification of [number of dot columns constituting the correction pattern] / [number of pixels in the Y direction of the read image after trimming]. The number of pixels in the Y direction of the read image after resolution conversion is set to a predetermined value.

  Next, the correction value setting program measures the density of each of the five types of belt-like patterns in each row region. Hereinafter, the measurement of the density of the left band-shaped pattern formed with the gradation values in the first row region will be described. Measurements in other row regions are performed in the same manner. In addition, the measurement of the density of other band-like patterns is performed in the same manner.

  FIG. 12A is an explanatory diagram of a read image when the left ruled line is detected. FIG. 12B is an explanatory diagram of detection of the position of the left ruled line. FIG. 12C is an explanatory diagram of the measurement range of the density of the band-like pattern having a density of 30% in the first row region.

  The correction value setting program extracts pixel data of KX pixels from the left, which are H1 to H2 pixel rows, from the top, from the resolution-converted read image. The parameter KX is determined in advance so that the left ruled line is included in the pixels extracted at this time. Then, the correction value setting program obtains the ruled line position by performing the same process as the ruled line detection process described above.

  Since the shape of the correction pattern is predetermined, a strip-shaped pattern having a density of 30% exists over the width W3 on the right side by X2 from this ruled line position (left ruled line position). Therefore, the correction value setting program extracts pixel data matching the target band pattern position for each area corresponding to the raster with respect to the ruled line position, and the average value of the gradation values of the pixel data included in this area Measure. In this way, the correction value setting program creates the density distribution by measuring the density of the belt-like pattern for each raster.

  In the image processing for detecting the pattern position (position where the pattern is printed), it is important to accurately detect the ruled line position. By detecting the exact ruled line position by the ruled line detection process, the density distribution and the raster position of the print data are correctly associated, and the correction based on the correction value is effective.

  Based on the density distribution of the pixel data measured in this way, a correction value for correcting unevenness of the density distribution is obtained and stored in the memory 23 of the printer 20. Of course, before setting a correction value in the printer 20, the obtained correction value may be applied to print a test pattern to confirm the effectiveness of the correction value.

  As described above, the computer 10 that executes the ruled line detection process in steps S100 to S102 constitutes the average density calculation means, and the computer 10 that executes the ruled line detection process in steps S104 to S108 constitutes the dust determination means.

(3) Modified example of ruled line detection processing:
By the way, in the ruled line detection described above, ruled line detection fails if there is a large dust image in the vicinity of the ruled line in a portion of a plurality of pixel columns for calculating the average density. At this time, there may be no dust image in the vicinity of the ruled line included in the pixel column that is not the target of the average density calculation. Therefore, when the ruled line detection from the average density fails, the average density calculation may be recalculated using the neighboring ruled line.

  FIG. 13 illustrates a modified example of the ruled line detection process. In the drawing, the vicinity of the upper end of the KX1 to KX2 rows is shown. Taking ruled line detection in the erecting rotation process as an example, when ruled line detection from the average density of the pixel columns KX1 to KX2 fails, the pixels in the K5 to K1 (KX5 <KX1) columns that are neighboring pixel columns are detected. The average density is calculated from the pixel columns and the pixel columns KX2 to KX6 (KX2 <KX6), and the ruled line detection process is executed again. Of course, the KX5 to KX1 rows and the KX2 to KX6 rows are set so as not to interfere with the vertical ruled lines and the belt-like patterns.

  Further, not only the neighboring pixel columns but also the pixel columns K5 to K1 and K2 to K6 may be copied and replaced with the pixel column parts KX1 to KX2. That is, the ruled line may be complemented by extending the neighboring ruled line to the ruled line detection position. By this complementation, even if a large dust image is included in the vicinity of the ruled line in columns KX1 to KX2, this dust image is deleted. If the ruled line detection process is performed again on the pixel columns KX1 to KX2 after the ruled line is complemented, there is no influence of the dust image before the ruled line extension.

  Further, as another aspect of using a neighboring pixel column when the ruled line detection from the average density fails, the number of pixel columns used for calculating the average density is increased and the average density is recalculated again. It may be configured to detect the ruled line width and determine the dust. In the example of FIG. 13, when the ruled line detection from the average density of the KX1 to KX2 columns fails, the ruled line detection is tried from the average density of the KX5 to KX6 columns. That is, by increasing the number of sample bases used for calculating the average density, the density of a large dust image generated locally can be reduced. Therefore, even in this mode, it is possible to detect a ruled line position that is not easily affected by a dust image.

  As described above, when ruled line detection fails in the ruled line detection process, a dust image formed locally by re-detecting the ruled line using a pixel column in the vicinity of the pixel column used for the ruled line detection. It becomes possible to avoid the ruled line detection failure due to.

  Note that the present invention is not limited to the above-described embodiments and modifications, and the configurations disclosed in the above-described embodiments and modifications are mutually replaced, the combinations are changed, known techniques, and the above-described implementations. Configurations in which the configurations disclosed in the embodiments and modifications are mutually replaced or the combinations are changed are also included.

1 is a block configuration diagram of a printing system including a computer as an image processing apparatus. It is a figure which shows the nozzle of a print head. It is a flowchart which shows the process of a correction value setting program. It is explanatory drawing of a test pattern. It is explanatory drawing of the reading range of the correction pattern. It is a flowchart which shows a ruled line detection process. It is a figure explaining the row | line | column of the pixel used for a ruled line detection process. It is explanatory drawing of calculation of an average density | concentration, and ruled line detection using this average density | concentration. It is explanatory drawing of ruled line determination by the threshold value of ruled line width. It is explanatory drawing of the read image at the time of trimming. It is explanatory drawing of resolution conversion. It is explanatory drawing of the read image at the time of detection of a left ruled line. It is a figure explaining the modification of a ruled line detection process.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Computer, 10a ... Correction value setting program, 10b ... Application, 10c ... Printer driver, 10c1 ... Halftone (H / T) module, 10c2 ... Interlace module, 10c3 ... Correction module, 13 ... Memory, 20 ... Printer, DESCRIPTION OF SYMBOLS 21 ... Control part, 23 ... Memory, 24 ... Unit control circuit, 25 ... Print head unit, 25a ... Print head, 26 ... Carriage unit, 27 ... Paper feed unit, 29 ... I / F, 30 ... Scanner

Claims (7)

An image processing apparatus for detecting a ruled line position of image data in which a ruled line of a predetermined width is described,
A plurality of columns of pixels arranged in a line in a direction substantially orthogonal to the ruled line is extracted, and an average density calculation is performed to calculate an average density between pixels located in the ruled line direction in the extracted plurality of pixel columns. Means,
Ruled line width detecting means for comparing the average density with a predetermined density threshold in order from one end of a pixel column, and obtaining an interval from the first pixel to the last pixel exceeding the predetermined density threshold;
If the interval falls within a predetermined width threshold, it is determined that there is no dust image from the first pixel to the last pixel, and the first pixel to the last pixel are detected as ruled line positions, Dust determination means for determining that a dust image exists from the first pixel to the last pixel when the interval exceeds a predetermined width threshold;
An image processing apparatus comprising: When the dust discrimination means determines that a dust image exists from the first pixel to the last pixel,
The image processing apparatus according to claim 1, wherein the average density is recalculated by replacing a plurality of pixel columns used in the calculation of the average density with a column of pixels located in the vicinity of the plurality of pixel columns.   3. The dust determination unit re-calculates the average density by increasing the number of pixel columns used for the calculation of the average density when the interval exceeds the predetermined width threshold. Image processing apparatus.   4. The image processing apparatus according to claim 1, wherein the ruled line position is set approximately between the first pixel and the last pixel. 5. Printing means for printing image data having a predetermined color and a predetermined density in a predetermined range in the sub-scanning direction together with ruled lines extending in either or both of the main scanning direction and the sub-scanning direction of the image data;
Printing result acquisition means for acquiring an optically read image as a result of the printing from the optical reading device;
Image processing means for performing image processing comprising at least one of upright rotation and trimming of the optically read image and detection of the printed position of the image data based on the ruled line position of the image data detected by the ruled line detection means; ,
Correction value setting means for measuring a density change in the sub-scanning direction from the optically read image after image processing by the image processing means, generating a correction value for eliminating the density change, and setting the correction value in the printing apparatus;
An image processing apparatus according to any one of claims 1 to 4, further comprising: An image processing method for detecting a ruled line position of image data in which a ruled line of a predetermined width is described,
A plurality of columns of pixels arranged in a line in a direction substantially orthogonal to the ruled line is extracted, and an average density calculation is performed to calculate an average density between pixels located in the ruled line direction in the extracted plurality of pixels. Process,
A ruled line width detection step of comparing the average density with a predetermined density threshold in order from one end of a pixel row to obtain an interval from the first pixel to the last pixel that exceeds the predetermined density threshold;
If the interval falls within a predetermined width threshold, it is determined that no dust image exists from the first pixel to the last pixel, and the first pixel to the last pixel are detected as ruled line positions, A dust determination step of determining that a dust image exists from the first pixel to the last pixel when the interval exceeds a predetermined width threshold;
An image processing method comprising: An image processing program for detecting a ruled line position of image data in which a ruled line of a predetermined width is described,
A plurality of columns of pixels arranged in a line in a direction substantially orthogonal to the ruled line is extracted, and an average density calculation is performed to calculate an average density between pixels located in the ruled line direction in the extracted plurality of pixel columns. Function and
A ruled line width detection function that compares the average density with a predetermined density threshold in order from one end of the pixel column, and obtains an interval from the first pixel to the last pixel that exceeds the predetermined density threshold;
If the interval falls within a predetermined width threshold, it is determined that there is no dust image from the first pixel to the last pixel, and the first pixel to the last pixel are detected as ruled line positions, A dust determination function for determining that a dust image exists from the first pixel to the last pixel when the interval exceeds a predetermined width threshold;
An image processing program for causing a computer to realize the above.