JP2011188109A - Document image analysis apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately detect unwanted streaks in an image generated during printing of a document. <P>SOLUTION: An image analysis apparatus includes a step of detecting a luminance gradient direction of each of pixels constituting the image and its strength, based on Y (luminance) data of images of a document, respectively (S105). While switching a pixel to be defined as a pixel of interest sequentially over an entire screen, pixels being present within a predetermined region (e.g., a rectangular region of five pixels respectively in longitudinal and lateral directions) around the pixel of interest are then classified based on the luminance gradient direction of these pixels (S106). When the pixels within the predetermined region are classified eccentrically to a specific direction range group and there is considerable luminance fluctuation in the luminance gradient direction in the pixel of interest, the pixel of interest is determined as an edge pixel (S107). Further, based on a distribution density of pixels determined to correspond to the edge pixel, presence of unwanted streaks in the document is discriminated and moreover, a direction of the unwanted streaks is discriminated based on the luminance gradient direction of the edge pixels (S108). <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an apparatus for analyzing image data read from an original, and more particularly to an apparatus for detecting streaks in an image generated when an original is printed.

  As described in Patent Document 1, a printing apparatus including a line-type inkjet head is known. The line-type inkjet head has a size corresponding to the width of the printing paper, is fixed to the printing apparatus main body, and is a printing paper conveyed in a direction (sub-scanning direction) perpendicular to the line direction (main scanning direction). On the other hand, an image is formed by ejecting ink.

  FIG. 8A is a diagram illustrating an example of the configuration of a line-type inkjet head and ejected ink dots. As shown in this figure, nozzle units 240a and 240b each having a plurality of nozzles arranged obliquely in a predetermined number unit (three in this figure) in the line direction to increase the resolution are bonded together to form one block. 230 is formed. Then, as shown in FIG. 8B, a plurality of the blocks 230 are arranged in the line direction (six blocks of 1 to 6 in the figure) to form a line-type inkjet head 220.

  As shown in FIG. 8A, the nozzle unit 240a and the nozzle unit 240b are bonded to each other with a gap p1 in the line direction. The interval p1 is set to be 1/2 of the nozzle pitch dp, and the nozzle of the nozzle unit 240b is arranged between the nozzle of the nozzle unit 240a and the nozzle in the line direction. Thereby, dots with uniform density are formed at equal intervals by the ink droplets ejected from each nozzle of the nozzle unit 240a and each nozzle of the nozzle unit 240b.

  Since the inkjet head 220 is an industrial product, there is some variation between products. Therefore, as shown in FIG. 8C, there may be a block in which the interval p2 between the nozzle unit 240a and the nozzle unit 240b is wider or narrower than 1/2 of the nozzle pitch dp. In such a block 230, a periodic deviation occurs in the nozzle pitch. That is, two adjacent nozzles form a set, the interval between dots in the set is reduced, and the interval between dots between sets is increased.

  As shown in FIG. 8D, a case where the block 230 is arranged to be inclined with respect to the line direction may occur. Even in such a block 230, a periodic deviation occurs in the nozzle pitch. That is, two adjacent nozzles form a set, the interval between dots in the set is reduced, and the interval between dots between sets is increased.

  Further, as shown in FIG. 8E, even when the interval p1 between the nozzle unit 240a and the nozzle unit 240b is appropriate and the block 230 is appropriately arranged, the ink can be used like the nozzle 241a and the nozzle 241b. When there is a nozzle with a small discharge amount, or when the ink discharge amount is small as a whole block 230, the interval between dots is widened.

  If the dot interval is large, streaks occur in the printing result at that portion. For example, when a solid-painted original image as shown in FIG. 9A is printed using a line-type inkjet head 220 with an interval between dots, as shown in FIG. Will occur. Also, when printing is performed using a printing apparatus other than a line-type inkjet head, a printing result in which unnecessary stripes are generated in one direction for various reasons may be obtained.

JP-A-8-132645

  When the printing result in which such a streak is generated is copied using a copying apparatus, as shown in FIG. 9C, the streak is printed as it is or emphasized, and the copy image quality is increased. Will deteriorate. In particular, when the copying apparatus employs a line-type ink jet head, as shown in FIG. 9D, streaks may further increase. Further, the streak direction can occur in both the vertical and horizontal directions of the image data read due to the relationship between the printing direction and the reading direction. In order to make the streak inconspicuous, it may be possible to uniformly smooth the read image data. However, an overall blurred and deteriorated print result is obtained.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a document image analysis apparatus capable of accurately detecting unnecessary streaks in an image generated during printing of a document. is there.

In order to achieve the above object, the present invention provides:
An apparatus (for example, the copying apparatus 10 in FIG. 1) that analyzes data of an image read from a document and detects streaks in the image generated when the document is printed.
For each pixel constituting the image data, a luminance gradient direction detecting means (for example, S105 in FIG. 2) that detects a luminance gradient direction with respect to surrounding pixels,
Each pixel is classified into a direction range group to which the luminance gradient direction detected by the luminance gradient direction detection unit of each pixel belongs among a plurality of direction range groups obtained by dividing the distribution range of the luminance gradient direction into a plurality of ranges. Classification means to perform (for example, S106 in FIG. 2);
While sequentially switching the position of the pixel of interest in the data, the pixel of interest is in the specific direction based on the classification weight of the pixel existing in the predetermined area centered on the pixel of interest with respect to the specific direction range group. Determination means (for example, S107 in FIG. 2) for determining whether or not the pixel is an edge pixel having a large change in luminance gradient in a specific direction represented by a range group;
Based on the distribution density of each pixel determined to be the edge pixel by the determination means in the data and the specific direction, the presence / absence of a streak in the image generated at the time of printing the document and its direction are determined. Streak determining means (for example, S108 in FIG. 2),
It is characterized by providing.

  According to the present invention, according to the luminance gradient direction of each pixel, each pixel is classified into any one of the direction range groups obtained by dividing the distribution range of the luminance gradient direction into a plurality, and the pixel classification bias with respect to the specific direction range group Based on the degree, edge pixels that are candidates for unnecessary streak edges of the image generated when the document is printed are extracted. Then, based on the distribution density of the extracted edge pixels, the presence or absence of unnecessary streaks in the image of the document is determined, and on the basis of a specific direction corresponding to a specific direction range group in which the pixels are classified biased, The direction of unnecessary stripes in the original image is determined. Therefore, the ability to distinguish the streaks that the image originally has as part of the image from the streaks that occurred during the printing of the original document is enhanced, and the unnecessary streaks of the image that occurred during the printing of the original document can be accurately detected. Can be detected.

In the above invention,
With each pixel as a target pixel, a density value calculating unit (for example, FIG. 2) calculates an average density value between the target pixel and a pixel existing within a predetermined range around the target pixel as a density value corresponding to the target pixel. S108)
The determination unit performs the determination on the target pixel whose density value calculated by the density value calculation unit is equal to or greater than a predetermined density reference value.

  According to the above invention, an edge pixel whose average density value including peripheral pixels is equal to or higher than a predetermined density reference value is extracted from edge pixels that are candidates for unnecessary streak edges in an image, and is extracted. This is used to determine the presence or absence of unnecessary streaks in the image based on the distribution density of edge pixels. Therefore, edge pixels that are isolated from surrounding pixels with a relatively high density (low distribution density), such as halftone dots generated in the image by performing error diffusion processing when the document is printed, By excluding it from the presence / absence determination target, it is possible to detect an unnecessary streak of an image generated at the time of printing a document with higher accuracy.

Furthermore, in the above invention,
When it is determined by the streak determining means that a streak exists in the image, among the edge pixels, the edge pixel in which the direction range group to which the streak direction belongs belongs to the specific direction range group. , Further comprising a smoothing processing means or step (for example, S109 in FIG. 2) for performing the smoothing processing,
It is characterized by that.

  According to the above invention, when it is determined that there is an unnecessary streak in the image generated at the time of printing the document, the smoothing process is performed on the edge pixels constituting the unnecessary streak. Accordingly, an image processed so as to eliminate unnecessary stripes in the image or to be inconspicuous can be obtained by copying a document. For this reason, it is possible to obtain a high-quality copy image from the original image after accurately detecting unnecessary streaks in the image generated when the original is printed.

  According to the present invention, it is possible to accurately detect unnecessary streaks in an image generated when a document is printed.

1 is a block diagram illustrating a configuration of a copying apparatus according to an embodiment of an image forming apparatus according to the present invention. It is a flowchart explaining a copy process. It is a figure for demonstrating the detection of the luminance gradient direction of each pixel. It is a figure for demonstrating the detection of the luminance gradient direction of each pixel. It is a figure for demonstrating the detection of the luminance gradient direction of each pixel. It is a figure for demonstrating the classification | category of each pixel. It is a figure for demonstrating a smoothing process. It is a figure explaining a line type ink jet head. FIG. 5 is a diagram for explaining a document in which unnecessary streaks are generated during printing.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of a copying apparatus according to an embodiment of an image analysis apparatus according to the present invention. 1 includes an image reading unit 110 and a printing mechanism unit 120, and a printing mechanism unit based on image data read from a document by the image reading unit 110. A document can be copied by printing at 120.

  The image reading unit 110 includes a document table on which a document is placed, a line sensor in which light receiving elements such as a CCD are arranged on a line, a light source, a carriage on which the line sensor is mounted, a motor for transporting the carriage, and the like. The printing mechanism unit 120 includes a print engine, a paper supply / discharge mechanism, and the like. It is assumed that the print engine included in the printing mechanism unit 120 in this embodiment employs a line-type inkjet method. The present invention can be applied particularly effectively when a document printed by a line-type ink jet method is copied and printed by a line-type ink jet method. However, the copying apparatus of the present invention is not limited to the line type ink jet method, and various types such as a serial type ink jet method and a laser method can be used.

  Further, the copying apparatus 10 includes a user interface unit 130 and a control unit 140. The user interface unit 130 includes a liquid crystal display device, an operation panel, and the like, and accepts various operations from the user and displays various information related to the copying apparatus 10. The control unit 140 includes a CPU, a memory, an image processing device, a communication processing device, and the like, and controls various processes in the copying apparatus 10 as will be described later.

  As shown in FIG. 1, in this embodiment, the control unit 140 includes a reading control unit 141, a special copy determination unit 142, an image processing unit 143, a print control unit 144, and a storage unit 145. The reading control unit 141 controls image reading processing in the image reading unit 110.

  The special copy determination unit 142 determines whether the document whose image data has been read by the image reading unit 110 is a printed material having a predetermined characteristic, specifically, a printed material in which streaks are generated in one direction.

  Specifically, luminance data is extracted from the read image data, edge pixels satisfying a predetermined condition included in the read image data are extracted based on the extracted luminance data, and a streak is printed on the document according to the extraction result. And whether the streak direction is included is determined. Edge pixel extraction can be performed based on the luminance gradient direction of each pixel and the number of changes. Further, whether or not a streak is included in the document can be determined based on the continuity of edge pixels whose average luminance value including peripheral pixels is equal to or higher than a predetermined density reference value.

  A printed material on which an image of a document has streaks is typically a printed material printed by a line-type ink jet printing apparatus. Such a streak is generated when a document is printed and is essentially unnecessary.

  The image processing unit 143 converts the RGB format image data read by the image reading unit 110 into the CMYK format, and performs necessary image processing, halftone processing, and the like. The image processing unit 143 includes a special copy processing unit 146.

  The special copy processing unit 146 performs processing for making the image data inconspicuous when the special copy determination unit 142 determines that the printed matter has predetermined characteristics. Specifically, for example, when it is determined that vertical streaks are included, the horizontal smoothing process is performed on the extracted luminance data, and for example, it is determined that horizontal streaks are included. In such a case, the read image data is corrected by performing vertical smoothing on the extracted luminance data.

  The print control unit 144 controls printing processing in the printing mechanism unit 120. The storage unit 145 is configured by a memory and stores image data, print data, and the like, and is used as a work area for the CPU.

  Next, the copying process of this embodiment will be described with reference to the flowchart of FIG. In the copying process, first, the image reading unit 110 reads an image of a document placed on the document table as image data in RGB format (S101). Prior to reading, detailed settings for copy processing such as enlargement / reduction, copy quality, number of copies, and the like may be received via the user interface unit 130.

  The read image is stored in the storage unit 145 in the RGB format (S102). Next, the special copy determination unit 142 determines whether or not the read original is a printed matter in which streaks are generated in one direction. As preparation for this, image data in RGB format is converted to YCbCr format (S103). Here, the YCbCr format is a format in which colors are expressed by Y (luminance) and CbCr (color difference). In general, unnecessary streaks that occur during printing are conspicuous in a portion having a high density (a portion having a low luminance). For this reason, in the present embodiment, RGB format image data is converted to YCbCr format, and Y (luminance) is used for determination.

The conversion from the RGB format to the YCbCr format is, for example, the following (formula 1),
Y = (77 × R + 150 × G + 29 × B) / 256
Cb = (− 43 × R− + 85 × G + 128 × B) / 256
Cr = (128 * R-107 * G-21 * B) / 256
(Equation 1)
Can be done according to. Here, a case of 8 bits for each of RGB is shown. However, you may convert into YCbCr format using another method. Further, as long as Y (luminance) can be acquired, other color expression formats may be used. For example, V (lightness) or L ^* (lightness) may be used instead of Y (luminance) after conversion to the HSV format or L ^* a ^* b ^* color system.

  After conversion to the YCbCr format, Y (luminance) data is extracted (S104). That is, the process related to stripes is performed on Y (luminance) data, and is not performed on CbCr (color difference) data. Then, based on the extracted Y (luminance) data, it is determined whether or not an unnecessary streak exists in the document (S105 to S108).

  Specifically, first, a luminance gradient direction and its intensity are detected for each pixel constituting the image (S105). In order to perform this detection, first, with respect to a pixel (Ix, Iy) whose luminance gradient direction is to be detected, the luminance gradient (lateral gradient, vertical direction) in the horizontal direction (x direction) and vertical direction (y direction) of the image. Directional gradient).

Here, the horizontal gradient TZ (Ix, Iy) is Y (luminance) of neighboring pixels (Ix-1, Iy) and (Ix + 1, Iy) on both sides in the horizontal direction (x direction) of the target pixel (Ix, Iy). ) Difference value (absolute value). The vertical gradient TH (Ix, Iy) is the Y (luminance) of the neighboring pixels (ix, Iy−1) and (Ix, Iy + 1) on both sides in the vertical direction (y direction) of the target pixel (Ix, Iy). It is calculated by the difference value (absolute value). This is expressed by the following formula (Formula 2),
TZ (Ix, Iy) = | Y (Ix + 1, Iy) −Y (Ix−1, Iy) |
TH (Ix, Iy) = | Y (Ix, Iy + 1) −Y (Ix, Iy−1) |
(Equation 2)
It becomes.

The luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is a ratio of the vertical gradient TH (Ix, Iy) and the horizontal gradient TZ (Ix, Iy) obtained as described above ( Tan ⁻¹ of TH (Ix, Iy) / TZ (Ix, Iy)) is obtained, and this is detected by multiplying it by 180 ° / π. That is, the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is expressed by the following (formula 3):
D (Ix, Iy) = tan ⁻¹ {TH (Ix, Iy) / TZ (Ix, Iy)}
× (180 ° / π) (Formula 3)
It becomes.

Further, the intensity G of the luminance gradient (Ix, Iy) indicating the magnitude of the luminance change in the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is the vertical gradient TH obtained as described above. It is obtained by the square sum of squares of (Ix, Iy) and the lateral gradient TZ (Ix, Iy). That is, the intensity G (Ix, Iy) of the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is expressed by the following (formula 4):
G (Ix, Iy) = √ {TH (Ix, Iy) 2 + TZ (Ix, Iy) 2}
(Formula 4)
It becomes.

For example, when a horizontal streak as shown in FIG. 3 exists in the image, the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is the luminance value of the black solid portion indicated by shading in the figure. Is 20 and the brightness value of the streak portion indicated by white in the figure is 200, from the above (Equation 3),
D (Ix, Iy) = tan ⁻¹ {(200−20) / (20−20)}
× (180 ° / π)
= 90 °
Thus, “90 °” indicating that a luminance gradient exists in the vertical direction (y direction) of the image is obtained as the luminance gradient direction D (Ix, Iy).

In addition, the intensity G (Ix, Iy) of the luminance gradient in the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) in this case is obtained from the above (Expression 4).
G (Ix, Iy) = √ {(200−20) 2 + (20−20) 2}
= 180
It becomes.

On the other hand, when vertical streaks as shown in FIG. 4 are present in the image, the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is the luminance value of the black solid portion indicated by shading in the figure. Is 20 and the brightness value of the streak portion indicated by white in the figure is 200, from the above (Equation 3),
D (Ix, Iy) = tan ⁻¹ {(20-20) / (200-20)}
× (180 ° / π)
= 0 °
Thus, “0 °” indicating that a luminance gradient exists in the horizontal direction (x direction) of the image is obtained as the luminance gradient direction D (Ix, Iy).

In addition, the intensity G (Ix, Iy) of the luminance gradient in the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) in this case is obtained from the above (Expression 4).
G (Ix, Iy) = √ {(20−20) 2+ (200−20) 2}
= 180
It becomes.

Furthermore, when a streak in the oblique 45 ° direction as shown in FIG. 5 is present in the image, the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) is the black solid portion indicated by shading in the figure. Assuming that the luminance value is 20 and the luminance value of the streak portion indicated by white in the figure is 200, from the above (Equation 3),
D (Ix, Iy) = tan ⁻¹ {(200-20) / (200-20)}
× (180 ° / π)
= 45 °
Thus, “45 °” indicating that there is a luminance gradient in an oblique 45 ° direction of the image is obtained as the luminance gradient direction D (Ix, Iy).

In addition, the intensity G (Ix, Iy) of the luminance gradient in the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy) in this case is obtained from the above (Expression 4).
G (Ix, Iy) = √ {(200−20) 2 + (200−20) 2}
≒ 254
It becomes.

  Next, as shown in FIG. 2, each pixel is classified based on its luminance gradient direction D (Ix, Iy) (S106). In this classification, the position of the pixel of interest (Ix, Iy) shown in FIG. 6 is sequentially switched in the horizontal and vertical directions over the entire screen, and within a predetermined area centered on the pixel of interest (Ix, Iy). This means that each pixel (25 pixels in this figure) existing in the rectangular area (vertical area of 5 pixels vertically and horizontally in this figure) is classified based on the luminance gradient direction D (Ix, Iy) of each pixel. In this classification, the selection range of the luminance gradient direction D (Ix, Iy) is changed from 0 ° to 90 ° into four direction range groups (direction range group 1: 0 to less than 22.5 °, direction range group 2:22. .5 ° to less than 45 °, direction range group 3: 45 ° to less than 67.5 °, direction range group 4: 67.5 ° to 90 °), and any of these four direction range groups In addition, each pixel in the predetermined area is sorted.

  Next, as shown in FIG. 2, by using the classification result of the pixels existing in the predetermined area for each pixel of interest (Ix, Iy), each pixel of interest (Ix, Iy) has an edge of an unnecessary line in the image. It is determined whether it corresponds to a candidate edge pixel (S107). This determination is made in principle depending on whether or not the luminance gradient of each pixel of interest (Ix, Iy) has changed significantly in a specific luminance gradient direction.

  In the above determination, whether or not the pixels in the predetermined area centered on each pixel of interest (Ix, Iy) are classified in a specific direction range group and the luminance at each pixel of interest (Ix, Iy) It is confirmed whether there is a large luminance variation in the gradient direction D (Ix, Iy). In this confirmation, in order to ensure that a specific direction range group has a dominant difference in the number of pixel classifications over other direction range groups, a predetermined threshold Th1 is set for the number of pixel classifications. use. Further, in order to ensure that there is a large luminance variation in the luminance gradient direction D (Ix, Iy) at the target pixel (Ix, Iy), the intensity G of the luminance gradient in the luminance gradient direction D (Ix, Iy) ( A predetermined threshold Th2 is used for Ix, Iy).

  Next, a specific case is considered below. Here, it is assumed that the number of pixels in the predetermined area classified into the direction range group 1 (specific direction range group) is N1, and the number of pixels in the predetermined area classified into the direction range group 4 is N4. .

First, consider a case where the direction range group 1 is set as a specific direction range group and the distribution bias is determined as the distribution weight. In this case, the following (condition 1),
N1> N4 × Th1 and G (Ix, Iy)> Th2 (Condition 1)
Is satisfied, the luminance change in the vertical direction (y direction) of the screen is larger than the luminance change in the horizontal direction (x direction), and the luminance gradient direction D (( This means that the intensity gradient intensity G (Ix, Iy) of Ix, Iy) varies greatly.

  When this (Condition 1) is satisfied, the pixel of interest (Ix, Iy) is determined to be an edge pixel because the luminance gradient change in the vertical direction (y direction) of the screen is large. Then, in the target pixel (Ix, Iy) corresponding to the edge pixel, a value “−1” indicating that the luminance gradient change in the vertical direction (y direction) of the screen is large is a function F ( Ix, Iy). This value “−1” is a representative luminance gradient direction (specific direction) corresponding to the direction range group 1 (0 ° to less than 22.5 °) in which the target pixel (Ix, Iy) corresponding to the edge pixel is classified. ) Indicates the vertical direction (y direction) of the screen.

Next, a case is considered in which the direction range group 4 is set as a specific direction range group and the distribution bias is determined as the distribution weight. In this case, the following (condition 2),
N4> N1 × Th1 and G (Ix, Iy)> Th2 (Condition 2)
Is satisfied, the luminance change in the horizontal direction (x direction) of the screen is larger than the luminance change in the vertical direction (y direction), and the luminance gradient direction D (( This means that the intensity gradient intensity G (Ix, Iy) of Ix, Iy) varies greatly.

  When this (condition 2) is satisfied, the pixel of interest (Ix, Iy) is determined to be an edge pixel because the change in the luminance gradient in the horizontal direction (x direction) of the screen is large. Then, in this target pixel (Ix, Iy) corresponding to the edge pixel, a value of “1” indicating that the luminance gradient change in the horizontal direction (x direction) of the screen is large is a function F (Ix indicating the edge pixel. , Iy). This value “1” indicates a representative luminance gradient direction (specific direction) corresponding to the direction range group 4 (67.5 ° to 90 °) in which the target pixel (Ix, Iy) corresponding to the edge pixel is classified. , Indicating the horizontal direction (x direction) of the screen.

  By the way, among the target pixels (Ix, Iy) that are determined to be edge pixels when the above (Condition 1) and (Condition 2) are satisfied, actually, the image data at the time of printing the document is not included. In some cases, halftone dots generated by performing error diffusion processing are included. That is, the halftone dots may be isolated from the surrounding halftone dots, and in this case, the luminance difference between the halftone dots and the surrounding white portions is regarded as a luminance gradient.

Therefore, it is possible to exclude halftone dots isolated from the surrounding halftone dots from the edge pixels. In that case, the average density value Iarea (Ix, Iy) of the pixel of interest (Ix, Iy) and the pixels existing in the surrounding predetermined range is compared with a predetermined density reference value Th_dens. And the following (Condition 3),
Iarea (Ix, Iy) <Th_dens (Condition 3)
For the pixel of interest (Ix, Iy) for which the above relationship is established, a value of “0”, which means that the pixel is not an edge pixel, regardless of whether the above (Condition 1) or (Condition 2) is satisfied, Set as value of function F (Ix, Iy).

  Here, the direction range group 1 (0 ° to less than 22.5 °) in which pixels whose luminance gradient direction is the horizontal direction (x direction) is classified, and the luminance gradient direction is the vertical direction (y direction). Whether or not a number of pixels are classified in the direction range group 4 (67.5 ° to 90 °) in which the pixels are classified more than the other direction range groups (whether the number of pixel classifications is large). confirmed.

  However, in the other direction range group 2 (22.5 ° to less than 45 °) and the direction range group 3 (45 ° to less than 67.5 °), the number of pixels is also more biased than the other direction range group. It is good also as a structure which confirms whether it is classified (whether there are many pixel classification numbers).

  In such a configuration, the numbers N2 and N3 of pixels and threshold values classified into the direction range group 2 (22.5 ° to less than 45 °) and the direction range group 3 (45 ° to less than 67.5 °), respectively. Using the intensity gradient strength G (Ix, Iy) and the threshold value Th2 in Th1 and the luminance gradient direction D (Ix, Iy) of the target pixel (Ix, Iy), the above (condition 1) and (condition 2) ).

  Also at this time, it is possible to exclude halftone dots isolated from the surrounding halftone dots from the edge pixels. In this case, the above-described (conditions) are used by using the average density value Iarea (Ix, Iy) of the pixel of interest (Ix, Iy) and the pixels existing in the surrounding predetermined range and the predetermined density reference value Th_dens. 3) It should be confirmed whether it satisfies. The pixel of interest (Ix, Iy) that satisfies this (Condition 3) is not an edge pixel even if it is determined to be an edge pixel for the direction range group 2 or the direction range group 3. The value “0” to be set may be set as the value of the function F (Ix, Iy).

  Next, as shown in FIG. 2, it is determined whether or not the image read from the document includes unnecessary stripes generated during printing (S108). This determination may be performed automatically as described below, or may be determined by receiving settings from the user via the user interface unit 130.

  In the case of automatic determination, first, based on the distribution density of the pixels determined to correspond to the edge pixels in S107, it is determined whether or not an unnecessary streak exists in the document. When it is determined that an unnecessary streak exists, a representative luminance gradient direction (specific direction) corresponding to the classified direction range group indicated by the value of the function F (Ix, Iy) of each edge pixel is further displayed. ) To determine the direction of existing stripes.

First, as the distribution density of the pixels determined to correspond to the edge pixels, whether or not the number of pixels whose function F (Ix, Iy) is not “0” exceeds the threshold Th_edge_cnt, that is, Condition 4)
Σ {F (Ix, Iy)}> Th_edge_cnt
However, F (Ix, Iy) ≠ 0 (Condition 4)
It is confirmed whether or not is established. If (Condition 4) is satisfied, it is determined that an unnecessary streak exists in the image. If (Condition 4) is not satisfied, it is determined that no unnecessary streak exists in the image.

  Next, when it is determined that an unnecessary streak exists in the image, the total value of the function F (Ix, Iy) of each pixel corresponding to the edge pixel is positive as the luminance gradient direction (specific direction) of the edge pixel. Check whether the value is negative or negative. If the total value is a positive value, the streak direction is the horizontal direction (x direction), and if the total value is a negative value, the streak direction is the vertical direction (y direction).

  In addition, the direction range group 2 (22.5 ° to less than 45 °) and the direction range group 3 (45 ° to less than 67.5 °) have a larger number of pixels than the other direction range groups. In this configuration, the numerical value range to which the total value of the function F (Ix, Iy) of each pixel corresponding to the edge pixel belongs can be divided into smaller ranges than positive and negative. Good. As a result, the direction of unnecessary stripes can be set in the direction range group 2 (22.5 ° to less than 45 °) or the direction range group 3 (45 ° to 67 °) other than the horizontal direction (x direction) and the vertical direction (y direction). It is also possible to discriminate the representative luminance gradient directions (specific directions) respectively corresponding to less than .5 °).

  As a result, when it is determined that the image read from the document includes unnecessary streaks that occur during printing (S108: Yes), the content is smoothed according to the direction of the unnecessary streaks. The special copy processing unit 146 performs the processing on the Y (luminance) data (S109). This smoothing process may be performed on all the pixels of the image data. Or you may carry out limitedly with respect to the edge pixel which makes the direction orthogonal to the direction of an unnecessary streak the brightness | luminance gradient direction D (Ix, Iy).

  Then, the smoothing process in the case where the unnecessary stripe direction is the horizontal direction (x direction) is as shown in FIG. FIG. 7A is a diagram illustrating an example of the smoothing process when the direction of unnecessary stripes is the horizontal direction (x direction). In the example of this figure, the luminance Yb of the target pixel b is smoothed by setting Yb = Y_smooth = (Ya + Yc) / 2 using the luminance Ya of the pixel a adjacent in the vertical direction and the luminance Yc of the pixel c. Like to do.

  Further, the smoothing process in the case where the unnecessary streak direction is the vertical direction (y direction) is as shown in FIG. FIG. 7B is a diagram illustrating an example of the smoothing process when the direction of unnecessary stripes is the vertical direction (y direction). In the example of this figure, the luminance Yb of the target pixel b is smoothed by setting Yh = Y_smooth = (Ya + Yc) / 2 using the luminance Ya of the pixel a adjacent in the left-right direction and the luminance Yc of the pixel c. Like to do.

  Note that the specific method of the smoothing process may be another smoothing method different from the method described above.

  As described above, in the present embodiment, smoothing is performed in one direction perpendicular to the streak, thereby preventing excessive image blur and preventing the streak from conspicuous.

  On the other hand, when it is determined that the read original is not an image including unnecessary streaks that occur during printing (S108: No), the process for Y (luminance) is not performed.

In either case, the image data converted into the YCbCr format is reconverted into the RGB format (S110). At this time, if the Y (luminance) data is changed by the smoothing, the Y (luminance) data after the smoothing is converted into the RGB format. The conversion from the YCbCr format to the RGB format is, for example, the following (formula 5)
R = (256 × Y + 359 × Cr) / 256
G = (256 * Y-88 * Cb-183 * Cr) / 256
B = (256 × Y + 454 × Cb) / 256
(Equation 5 above)
Can be done according to. However, you may convert into RGB format using another method.

  Thereafter, conventional image processing such as conversion to CMYK format and halftone processing is performed (S111), and printing is executed (S112). As a result, it is possible to obtain a print result in which stripes are not conspicuous in the copying process of an original in which unnecessary stripes are generated during printing.

  In the above example, the presence or absence of unnecessary streaks in the original printed with the line-type inkjet head printing device was assumed, but unnecessary streaks were also generated in the printing device using the serial-type inkjet head. A print result may be obtained. For example, when printing is performed on a printing paper having a small dot gain, a gap is likely to be generated between dots to be printed, and thus streaks may occur even in a printing apparatus using a serial type ink jet head. . Such a situation can be dealt with, for example, by changing the threshold value or the like. Specifically, the threshold value such as Th_edge_cnt, which can be experimentally determined as described above, is experimentally adjusted using a document printed by a printing apparatus using a serial type inkjet head. A value suitable for a printing apparatus using an inkjet head can be set.

  In the above example, threshold values used for detection of edge pixels that are streak edge candidates, determination of the presence or absence of streaks, and the like have been set in advance by experiments during development. You may enable it to adjust these threshold values based on the instruction | indication from a user. This is because there is a possibility that the setting value at the time of development may not be optimal depending on the use environment of the user, individual differences among devices, and the like. An instruction from the user can be received by a menu operation via the user interface unit 130.

  For example, when it is determined that a streak is generated even though the streak is not noticeable and the image is smoothed and the sharpness of the print result tends to be lost, based on an instruction from the user, The threshold value can be adjusted so as to reduce the sensitivity of the streak determination. Specifically, the threshold values Th1 and Th2 used for edge determination are increased. Alternatively, the threshold value Th_edge_cnt used for determining the presence or absence of streaks may be increased.

  On the other hand, when it is determined that no streak is generated despite the occurrence of a streak and the image is not smoothed and the streak tends to be noticeable, the streak is determined based on an instruction from the user. The threshold value can be adjusted to increase the sensitivity of determination. Specifically, the threshold values Th1 and Th2 used for edge determination are reduced. Alternatively, the threshold Th_edge_cnt used for determining the presence or absence of streaks may be reduced.

  Further, in the above example, after the unnecessary streaks of the image generated during the printing of the document and its direction are discriminated, the edge pixels are smoothed so that the unnecessary streaks disappear or become inconspicuous. The configuration for performing is described. However, the configuration for smoothing the edge pixels may be omitted.

  In the present embodiment, the luminance gradient direction D (Ix, Iy) of 0 ° to 90 ° related to the pixels constituting the image is divided into 22.5 °, 45 °, 67.5 ° and divided into four. did. However, the angle value that becomes a break when dividing into four is not limited to the pattern of the present embodiment. Moreover, the width of each angle range of each of the four divided direction range groups may be different for each direction range group. Furthermore, the number of divisions in the luminance gradient direction D (Ix, Iy) is not limited to four, and may be three or less or five or more.

DESCRIPTION OF SYMBOLS 10 Copy apparatus 110 Image reading part 120 Printing mechanism part 130 User interface part 140 Control part 141 Reading control part 142 Special copy determination part 143 Image processing part 144 Print control part 145 Storage part 146 Special copying process part 220 Inkjet head 230 Block 240a Nozzle unit 240b Nozzle unit 241a Nozzle 241b Nozzle

Claims (3)

An apparatus that analyzes an image read from a document and detects streaks in the image generated when the document is printed,
For each pixel constituting the image, a luminance gradient direction detecting means for detecting a luminance gradient direction with respect to surrounding pixels, and
Classification means for classifying each pixel into a direction range group to which the brightness gradient direction detected by the brightness gradient direction detection means belongs, among a plurality of direction range groups obtained by dividing the brightness gradient direction distribution range into a plurality of directions. ,
Each pixel corresponds to the specific direction range group based on the classification weight indicating the degree of deviation that the pixels existing in the predetermined area centered on the pixel are classified into the specific direction range group. A determination means for determining whether or not the edge pixel has a large change in luminance gradient in a specific direction;
A streak discriminating unit that discriminates the presence / absence of a streak in the image and its direction based on the distribution density of each pixel determined to be the edge pixel by the determining unit and the specific direction;
An apparatus for analyzing an image of a document. A density value calculating unit that calculates each pixel as a target pixel, and calculates an average density value of the target pixel and pixels existing within a predetermined range around the target pixel as a density value corresponding to the target pixel;
The determination unit performs the determination based on the target pixel whose density value calculated by the density value calculation unit is equal to or greater than a predetermined density reference value.
The document image analyzing apparatus according to claim 1, wherein: When it is determined by the streak determining means that a streak exists in the image, among the edge pixels, the edge pixel in which the direction range group to which the streak direction belongs belongs to the specific direction range group. , Further comprising a smoothing processing means for performing a smoothing process,
3. The document image analyzing apparatus according to claim 1, wherein the document image analyzing apparatus is an image analyzing apparatus.