JP3638231B2 - Image processing apparatus and recording medium - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a technique for detecting a thin line that is a location where a “line break” phenomenon occurs in image processing, and a technique for preventing the line break phenomenon.
[0002]
[Background]
When the halftone processing of a digital image is performed, a line break phenomenon may occur. For example, as described below, there is a “line break” phenomenon that occurs when the ratio between the halftone dot pitch and the pixel pitch of the digital image is not an integer ratio. 22 and 23 are diagrams for explaining the principle.
[0003]
FIG. 22 shows an example in which a thin line in the vertical direction (longitudinal direction) having a 50% density is formed into halftone dots when the ratio between the halftone dot pitch and the pixel pitch of the digital image is an integer ratio. Here, for example, the dot pitch U1 (= 1/175) as in the case where the number of lines of halftone dots is 175 lines (number of lines per inch) and the pixel pitch is 350 dpi (number of pixels per inch). (Inches)) and pixel pitch U2 (1/350 (inch)) "is shown as a ratio U2 / U1 of 350/175 = 2/1 = 2 (integer). In this case, when each pixel in the original digital image is halftone, the relative position between the unit area of the halftone pattern (hereinafter “unit halftone area DU”) and each pixel PX is Since any part is the same and no spatial change occurs between them, a thin line existing at any pixel position is halftone doted so as to have the same halftone dot area.
[0004]
However, if the ratio between the halftone dot pitch and the pixel pitch of the digital image is not an integer ratio, even if the pixels of the original image have the same density (gradation value) in the halftone processing, Depending on the relative positions of the pixels at the time of halftoning, the result of halftoning (the size of the area at the time of halftoning) varies.
[0005]
For example, in FIG. 23, the halftone dot pitch U1 (= 1/1 /) as in the case where the number of halftone dots is 175 lines (number of lines per inch) and the pixel pitch is 400 dpi (number of pixels per inch). Consider a case where the ratio U3 / U1 of 175 (inch) to pixel pitch U3 (1/400 (inch)) is 400/175 = 2.2, that is, not an integer. In this case, the relative position of each pixel PX of the digital image with respect to the unit dot area DU on the dot pattern gradually shifts and fluctuates with a relatively large period. Therefore, the relative position of the pixel on the halftone dot pattern may exist at the position PA, or may exist at the position PB. When a 1-pixel wide thin line exists at the PA position, the area at the time of halftone dot formation is relatively small (large), and when a 1-pixel wide thin line exists at the PB position, the halftone dot conversion Becomes relatively large (small). Here, when the thin line LA having a width of one pixel exists with a small angle θ with respect to the vertical direction (see FIG. 24A), the relative position of the pixel at the time of halftoning is changed from PA to PB or the like. The thickness of the thin line appears to change according to the periodic fluctuation. Especially in the area where the halftone dot area is small, humans It appears as a “line break” phenomenon in which the thin line appears to be broken. This situation is schematically shown in FIG. 24B (for the sake of convenience, each halftone dot is indicated by a black circle). This line break phenomenon occurs remarkably when the line width is small with respect to the halftone dot pitch.
[0006]
The above is the outline of the “line break” phenomenon that occurs when the ratio between the halftone dot pitch and the pixel pitch of the digital image is not an integer ratio.
[0007]
Further, such a line break phenomenon may be caused by USM (unsharp mask) processing. This USM process is performed in order to avoid a reduction in sharpness that generally occurs during halftone output. The line break phenomenon at this time is accompanied by a decrease in the density level of the fine line due to edge enhancement from both sides. This is because, as shown in FIG. 25, in the fine line portion, peaks due to edge emphasis from both sides of the high density region may overlap and the pixel value may be saturated. This is because the point area ratio is lower than the original one. This area in the thin line appears as “line break” to the human eye as described above. Further, when the thin lines that cause “line breaks” are arranged in parallel at equal intervals, the low density portions of each line are connected to each other, and long-period density fluctuation = moire occurs.
[0008]
[Problems to be solved by the invention]
The line break phenomenon caused by various causes as described above deteriorates the image quality and becomes a major obstacle at the time of image output.
[0009]
Accordingly, in view of the above problems, the present invention provides a technique related to image processing for detecting a thin line that is a source of line breakage, and a technique related to image processing for performing processing for preventing line breakage based on the detection result. Objective.
[0010]
[Means for Solving the Problems]
In order to achieve the above object, an image processing apparatus according to claim 1 is an image processing apparatus that performs processing on an original image, and reduces the original image in a predetermined direction to generate a reduced image. Generating means, thinning processing means for generating a first processed image by thinning a specific density region in a direction orthogonal to the predetermined direction, and the first processing A thickening processing unit that generates a second processed image by performing a thickening process for thickening a specific density region in a direction orthogonal to the predetermined direction, and a difference between the reduced image and the second processed image. And a difference image generating means for generating a certain difference image.
[0011]
The image processing device according to claim 2 is the image processing device according to claim 1, wherein the reduced image generating means averages n pixels along the predetermined direction when the original image is reduced in the predetermined direction. It involves processing.
[0012]
The image processing apparatus according to a third aspect is the image processing apparatus according to the first aspect, wherein the specific density region is a high density region.
[0013]
According to a fourth aspect of the present invention, in the image processing apparatus according to the first aspect, the specific density region is a low density region.
[0014]
An image processing apparatus according to a fifth aspect is the image processing apparatus according to the first aspect, wherein the predetermined directions are two different directions, the reduced image generation means, the thinning processing means, and the thickening. The processing unit and the difference image generation unit perform each processing relating to each of the two directions.
[0015]
An image processing apparatus according to a sixth aspect is the image processing apparatus according to any one of the first to fifth aspects, wherein the difference image is set to a reciprocal multiple of a reduction ratio in the reduced image generating unit. The image processing apparatus further includes enlargement processing means for generating a fine line index indicating the degree of presence or absence of a fine line at each pixel position of the original image by performing enlargement processing that enlarges in a predetermined direction.
[0016]
The image processing device according to claim 7 is the image processing device according to claim 6, wherein a weighting factor for weighting a result of applying a predetermined filter process to the original image to the original image is used as the thin line index. Weight coefficient determination means for determining based on the result, and filter processing means for obtaining a processed image obtained by weighting the original image with the result of applying a predetermined filter process to the original image using the weight coefficient. Features.
[0017]
The image processing apparatus according to claim 8 is the image processing apparatus according to claim 7, wherein the weighting coefficient determining means has a correspondence relationship between the thin line index and the weighting coefficient in advance as a reference table. And
[0018]
The image processing apparatus according to claim 9 is the image processing apparatus according to claim 7, wherein the enlargement processing unit also performs a process of expanding a high density region of the difference image when performing the enlargement process. To generate the fine line index.
[0019]
The recording medium according to claim 10 is a computer-readable recording medium recording a program for causing a computer to function as the image processing apparatus according to any one of claims 1 to 9. To do.
[0020]
DETAILED DESCRIPTION OF THE INVENTION
<A. Configuration>
FIG. 1 is a conceptual diagram showing a hardware configuration of an image processing apparatus 1 according to an embodiment of the present invention. The image processing apparatus 1 includes a CPU 2, a storage unit 3 including a semiconductor memory and a hard disk, a media drive 4 for reading information from various recording media, a display unit 5 including a monitor, an input unit 6 including a keyboard and a mouse, a digital The computer system includes an image input unit 7 for reading an image and an image output unit 8 for outputting a processed image. The CPU 2 is connected to the storage unit 3, the media drive 4, the display unit 5, the input unit 6, the image input unit 7, the image output unit 8, and the like via the bus line BL and the input / output interface IF. The media drive 4 reads information recorded therein from a portable recording medium 9 such as a CD-ROM, DVD (Digital Versatile Disk), or flexible disk. By reading the program from the portable recording medium 9 on which the program is recorded, this computer system has a fine line detection function as described later and a line break prevention function based thereon. Furthermore, the storage unit 3 includes a program storage unit 3a that stores all or part of the read program, an image storage unit 3b that stores various processed images, and the like.
[0021]
FIG. 2 is a functional block diagram illustrating a schematic configuration of the image processing apparatus 1, and FIG. 3 is a more detailed functional block diagram of the image processing apparatus 1. FIG. 3 shows functional blocks when a thin line extending in the vertical direction and the horizontal direction is detected, and a line break prevention process is performed based on the detection result.
[0022]
As shown in FIGS. 2 and 3, the image processing apparatus 1 includes an image input unit 7 that reads a digital image, a fine line detection unit 20 that has a fine line detection function, and a line break that has a line break prevention function based on a detection result. A prevention processing unit 50 and an image output unit 8 that outputs the processed image in halftone dots are provided.
[0023]
In general, the thin line detection unit 20 can detect a thin line extending in a predetermined direction, but here, the thin line detection unit 20 and the vertical direction thin line detection unit 20A that detects a vertical direction thin line are horizontal. It is assumed that a horizontal thin line detection unit 20B that detects a thin line in a direction is included, and a thin line extending in the vertical direction and the horizontal direction is detected (FIG. 3). Here, the “horizontal direction” refers to the row direction in the matrix arrangement of pixels, and the “vertical direction” refers to the column direction in the matrix arrangement. Conversely, it can also be defined.
[0024]
The thin line detection unit 20 (20A, 20B) reduces the read digital image (original image) in a predetermined direction (here, the vertical direction and the horizontal direction) to generate a reduced image generating unit 21 (21A, 20A, 20B). 21B), white thin line extraction unit 23w (23wA, 23wB) and black thin line extraction unit 23b (23bA, 23bB).
[0025]
Here, in this embodiment, a case is considered in which “black” is represented by a gradation value “0” and “white” is represented by a gradation value “255” (in the case of 8-bit expression). For this reason, the black tone region is defined as a “low density region” and the white tone region is defined as a “high density region”. For this reason, the “density” in this embodiment is “white density”. For example, “white line” means “a high density (white) line existing in a low density area (black area)”. Become. In general, an area having a density on the background color (white or black) side of the original image is defined as a low density area.
[0026]
The white thin line extraction unit 23w extracts a thin line (white thin line) expressed as a set of pixels having a large gradation value (high density value), and the black thin line extraction unit 23b has a small gradation value (low level). A thin line (black thin line) expressed as a set of pixels having (density value) is extracted. Here, both black and white fine lines are extracted. However, when only one of the white thin lines or the black thin lines needs to be extracted, only the corresponding one of the extraction units 23w and 23b can be provided. That's fine.
[0027]
The white thin line extraction unit 23w and the black thin line extraction unit 23b respectively have specific density regions (orthogonal directions) in a direction (orthogonal direction) orthogonal to the reduction direction of the reduced image generated in the reduced image generation unit 21 (21A, 21B). A thinning processing unit 25 (25wA, 25bA, 25wB, 25bB) for generating a first processed image by performing a thinning process for thinning a high-density region or a low-density region) A thickening processing unit 27 (27wA, 27bA, 27wB, 27bB) for generating a second processed image by performing a thickening process for thickening a specific density region in a direction orthogonal to the reduction direction, and the reduced image and the second processed image And a difference image generation unit 29 (29wA, 29bA, 29wB, 29bB) for generating a difference image that is a difference from the above.
[0028]
In addition, the fine line detection unit 20 includes a fine line data synthesis unit 31 (31A, 31B) that synthesizes the difference image related to the white fine line and the difference image related to the black fine line, and the high density region expansion unit 33 that performs processing for extending the high density region. (33A, 33B) and the obtained difference image is enlarged in the same direction as the reduction direction (this is enlarged to be the reciprocal of the reduction ratio in the reduced image generation unit 21 to be the same as the number of pixels of the original image). It further includes an enlargement processing unit 35 (35A, 35B) for enlarging. Note that, by the enlargement process in the enlargement processing unit 35, a fine line index that indicates the degree of presence or absence of a fine line at each pixel position of the original image is generated. In addition, the high density area expansion unit 33 performs a process of expanding the high density area of the difference image when performing the enlargement process by the enlargement processing unit 35. The high density area expansion unit 33 and the enlargement processing unit 35 are also performed. And can also be referred to as broadening processing means.
[0029]
Further, the line breakage prevention processing unit 50 includes a vertical horizontal direction thin line data composition unit 51 that synthesizes data related to the above-described vertical and horizontal direction thin line indices, and a weighting coefficient that determines a weighting coefficient based on the obtained thin line indices. A determination unit 53; and a filter processing unit 55 that obtains a processed image by weighting the result of applying a predetermined filter process to the original image based on the weighting coefficient.
[0030]
Hereinafter, in the image processing apparatus 1 having such a schematic configuration, an operation of detecting a thin line extending in the vertical direction and the horizontal direction, and an operation of preventing a line break based on the detection result will be described.
[0031]
<B. Operation>
<B1. Overview>
FIG. 4 is a flowchart showing the processing operation in the image processing apparatus 1. As shown in FIG. 4, first, in step SP10, the original image G to be processed is read using the image input unit 7 or the like. The original image G is obtained as a digital image in which a plurality of pixels are arranged in a matrix in the vertical direction and the horizontal direction, and each of the pixels has a plurality of gradation values (for example, 2 8 = 256 gradations).
[0032]
Next, in step SP20, an operation for detecting a thin line extending in the vertical direction is performed, and in step SP30, an operation for detecting a thin line extending in the horizontal direction is performed. The detection results of these thin lines extending in both directions are synthesized in step SP60. In step SP70, processing for preventing line breakage is performed based on the detection result of the thin line, and in step SP80, the processed image is output.
[0033]
<B2. Vertical thin line detection>
Here, the detail of the operation | movement (step SP20) which detects the thin line extended to a perpendicular direction is demonstrated. FIG. 5 is a flowchart showing the processing procedure.
[0034]
First, in step SP21, reduced image generation processing is performed in which the original image G is reduced to 1 / m in the vertical direction to generate a reduced image.
[0035]
FIG. 6 is a conceptual diagram for explaining this processing. By calculating the average of the gradation values of n pixels (n = 8) arranged in the vertical direction in the original image G, the reduction after the reduction processing is performed. A pixel value of one pixel of the image G0 is calculated. However, n is an integer of 2 or more. In other words, n pixels are reduced to one pixel in the reduction process of step SP21. Here, the scaling factor 1 / m is 1 / n (that is, m = n).
[0036]
<White fine line extraction>
Next, in step SP22, a process of extracting a white thin line as a thin line in which a high density region is connected in a line with a width of about one pixel width is performed. Therefore, the operations of steps SP23 to SP25 are performed on the reduced image G0. FIG. 8 is a diagram for explaining the principle of the white thin line detection operation, which will be described with reference to FIG.
[0037]
In step SP23, the first processed image G1 is generated by performing a thinning process for thinning the high density region in the horizontal direction on the reduced image G0. Specifically, the minimum value of the pixel value of each pixel of interest Pv (i, j) (FIG. 7) in the reduced image G0 and the pixel value of the adjacent pixel Pv (i + 1, j) in the horizontal direction is calculated. The value is set as the pixel value of each pixel Pvi (i, j) in the first processed image G1. That is, Pvi (i, j) = min {Pv (i, j), Pv (i + 1, j)}. Then, the first process image G1 is generated by advancing the above process by sequentially using all the pixels in the reduced image G0 as the target pixel.
[0038]
FIG. 8A shows an example of pixel values of a plurality of pixels arranged in the horizontal direction in the reduced image G0 before processing. This pixel column corresponds to a case where a thin line (white thin line) having a width of one pixel in the horizontal direction and a length of n pixels in the vertical direction exists. When the processing of step SP23 is performed on such an array of pixels Pv, an array of pixels Pvi in FIG. 8B is generated. In FIG. 8B, each pixel value is zero.
[0039]
Next, in step SP24, a second processed image G2 is generated by performing a thickening process to thicken the high density region in the horizontal direction for the first processed image G1. Specifically, the maximum value of the pixel value of each pixel of interest Pvi (i, j) in the first processed image G1 and the pixel value of the adjacent pixel Pvi (i-1, j) on the opposite side in the horizontal direction is calculated. Then, the value is set as the pixel value of each pixel Pvia (i, j) in the second processed image G2. That is, Pvia (i, j) = max {Pvi (i−1, j), Pvi (i, j)}. Then, the second process image G2 is generated by advancing the above process by sequentially setting all the pixels in the first process image G1 as the target pixel.
[0040]
When the high-density region thickening process in step SP24 is performed on the array of pixels Pvi in FIG. 8B, the array of pixels Pvia in FIG. 8C is generated. Here, the second pixel value from the left is restored. It remains zero.
[0041]
In step SP25, a difference image G3 that is a difference between the reduced image G0 and the second processed image G2 is generated. Specifically, the difference value between the pixel value of each pixel of interest Pv (i, j) in the reduced image G0 and the pixel value of the pixel Pvia (i, j) at the corresponding position in the second processed image G2 is calculated as a difference. The pixel value of the pixel in the pixel Mvw (i, j) of the image G3 is used. That is, Mvw (i, j) = Pv (i, j) −Pvia (i, j). However, when the difference value is negative, it is corrected to zero (limit processing), and an image obtained by the difference operation including the correction operation is referred to as a difference image.
[0042]
And the difference image G3 is produced | generated by advancing the said process by making each of all the pixels in the 2nd process image G2 into an attention pixel sequentially.
[0043]
FIG. 8D is a diagram illustrating a result of performing the process of step SP25 on the array of pixels Pv in FIG. 8A and the array of pixels Pvia in FIG. Here, the pixel value of the second pixel from the left has a large value, and it is shown that the degree to which a thin line exists in this portion is high.
[0044]
Here, FIG. 8 is a diagram for explaining the case where the processing of steps SP23 to SP25 is performed on the portion where the thin line to be detected exists, but there is a thin line to be detected with reference to FIG. An example will be given for the case of not. FIG. 9 is a diagram for explaining a case where the same processing is performed when there is a line having a width of two pixels in the horizontal direction and a length of n pixels in the vertical direction. .
[0045]
In this case, when the processing of step SP23 is performed on the array of pixels Pv, the array of pixels Pvi in FIG. 9B is generated, but the pixel value of the second pixel from the left does not become zero. , Value 255. Then, when the high density region thickening process in step SP24 is performed on the array of the pixels Pvi, the array of the pixels Pvia in FIG. 9C is generated. Here, the second pixel value and the third pixel from the left are generated. The value is restored to the original value of the reduced image G0, and the high density region returns to the width of 2 pixels. Accordingly, the pixels Mvw (i, j) of the difference image G3 are all zero, and it is determined that no thin line exists.
[0046]
In this manner, the thinning process and the thickening process are sequentially performed in the horizontal direction that is orthogonal to the reduction direction, and a difference image G3 that is a difference between the image G2 and the reduction image G0 obtained as a result is obtained. It is possible to extract a white thin line expressed as a high density region having a width of one pixel.
[0047]
FIGS. 10A and 10B are diagrams for explaining a case where such processing is performed in a planar manner. FIG. 10A shows a reduced image G0 before processing, and FIG. 10B shows a differential image G3 after processing. . As shown in FIG. 10B, in the lines L2, L3, L5, and L6 that are the pixel arrangement in the horizontal direction of the reduced image G0 before the processing, there is a high density region having a horizontal width of one pixel. Such a region is extracted as a region having a high degree of thin line in the difference image G3. On the other hand, in the lines L1, L4, and L7 of the reduced image G0 before processing, there is a high density region having a horizontal width of two pixels, but in such a region in the difference image G3, there is a thin line. Is low, the value of each pixel of the difference image G3 is zero.
[0048]
<Black fine line extraction>
Next, in step SP26, a process of extracting a black thin line as a thin line in which a low density region is connected in a line shape with a width of about one pixel width is performed. Therefore, the operations of steps SP27 to SP29 are performed on the reduced image G0. In the white thin line extraction processing operation of step SP22 described above, the high density area which is the specific density area is thinned in the horizontal direction and then thickened in the same direction, so that the pixel value is restored as a result. However, in the black thin line extraction processing operation of step SP26, the low density area is selected as the specific density area, and the low density area is thinned in the horizontal direction. By thickening in the same direction, the degree to which a thin line exists is detected based on the difference in the degree to which each pixel value is restored.
[0049]
Specifically, in step SP27, the first processed image G1 is generated by performing a thinning process for thinning the low density region in the horizontal direction on the reduced image G0. Specifically, the “maximum value” between the pixel value of each pixel of interest Pv (i, j) (FIG. 7) in the reduced image G0 and the pixel value of the adjacent pixel Pv (i + 1, j) in the horizontal direction is calculated. The value is set as the pixel value of each pixel Pva (i, j) in the first processed image G1. That is, Pva (i, j) = max {Pv (i, j), Pv (i + 1, j)}.
[0050]
Next, in step SP28, a second processed image G2 is generated by performing a thickening process for thickening the low density region in the horizontal direction on the first processed image G1. Specifically, the “minimum value” between the pixel value of each pixel of interest Pva (i, j) in the first processed image G1 and the pixel value of the adjacent pixel Pva (i−1, j) on the opposite side in the horizontal direction. And the value is set as the pixel value of each pixel Pvai (i, j) in the second processed image G2. That is, Pvai (i, j) = min {Pva (i-1, j), Pva (i, j)}.
[0051]
In step SP29, a difference image G3 that is a difference between the reduced image G0 and the second processed image G2 is generated. Specifically, the difference value between the pixel value of the pixel Pvai (i, j) at the corresponding position in the second processed image G2 and the pixel value of each pixel of interest Pv (i, j) in the reduced image G0 is calculated as a difference. The pixel value of the pixel in the pixel Mvb (i, j) of the image G3 is used. That is, Mvb (i, j) = Pvai (i, j) −Pv (i, j). However, when this difference value is negative, it is corrected to zero.
[0052]
In each of the steps SP27 to SP29, the images G1 to G3 are generated by proceeding with each of the pixels in each of the images G0 to G2 sequentially as the target pixel.
[0053]
FIG. 11A shows an example of pixel values of a plurality of pixels arranged in the horizontal direction in the reduced image G0 before processing. This pixel column corresponds to a case where there is a black thin line having a width of one pixel in the horizontal direction and a length of n pixels in the vertical direction. When the processing of step SP27 is performed on such an array of pixels Pv, an array of pixels Pva in FIG. 11B is generated, and each pixel value is 255 here.
[0054]
Then, when the low density region thickening process of step SP28 is performed on the array of pixels Pva in FIG. 11B, the array of pixels Pvai in FIG. 8C is generated. Here, the second pixel value from the left is generated. Remains 255 without being restored.
[0055]
FIG. 11D is a diagram illustrating a result of performing the process of step SP29 on the array of pixels Pvai in FIG. 11C and the array of pixels Pv in FIG. Here, the pixel value of the second pixel from the left has a large value, and it is shown that the degree to which a thin line exists in this portion is high.
[0056]
Here, FIG. 11 is a diagram for explaining the case where the processing of steps SP27 to SP29 is performed on the portion where the thin line to be detected exists, but there is a thin line to be detected with reference to FIG. An example will be given for the case of not. FIG. 12 shows a case where the same processing is performed when there is a black line (low density line) having a width of two pixels in the horizontal direction and a length of n pixels in the vertical direction. It is a figure for demonstrating.
[0057]
In this case, when the processing of step SP27 is performed on the array of pixels Pv, the array of pixels Pva in FIG. 12B is generated, but the pixel value of the second pixel from the left does not become 255. Has a value of zero. Then, when the high density area thickening process in step SP28 is performed on the array of the pixels Pva, the array of the pixels Pvai in FIG. 12C is generated. Here, the second pixel value and the third pixel from the left are generated. The value is restored to the original value of the reduced image G0, and the low density region returns to the width of 2 pixels. Accordingly, the pixels Mvb (i, j) of the difference image G3 are all zero, and it is determined that no thin line exists.
[0058]
In this manner, the thinning process and the thickening process are sequentially performed in the horizontal direction that is orthogonal to the reduction direction, and a difference image G3 that is a difference between the image G2 and the reduction image G0 obtained as a result is obtained. A black thin line expressed as a low density region having a width of one pixel can be extracted.
[0059]
FIGS. 13A and 13B are diagrams for explaining a case where such processing is performed in a plane. FIG. 13A shows a reduced image G0 before processing, and FIG. 13B shows a differential image G3 after processing. . As described above, in the lines L2, L3, L5, and L6, which are the horizontal pixel array of the reduced image G0 before processing, there is a low density region whose horizontal width is one pixel, and such a region exists. Is extracted as a region where the thin line is high in the difference image G3. On the other hand, in the lines L1, L4, and L7 of the reduced image G0 before processing, there is a low density region whose horizontal width is two pixels, but in such a region in the difference image G3, there is a thin line. Is considered to be low.
[0060]
<Combination of black and white thin line extraction results>
Next, in step SP30 (FIG. 5), the white thin line extraction result and the black thin line extraction result are synthesized. Specifically, by adding the pixel Mvw (i, j) and the pixel Mvb (i, j) in each pixel (i, j), fine line data on both the white fine line and the black fine line is obtained. Synthesize. That is, Mv (i, j) = Mvw (i, j) + Mvb (i, j).
[0061]
In step SP31, a process of expanding the high density area (high density area expansion process) is performed. This high-density area expansion processing can be realized by using, for example, a filter in which the maximum value among the pixel values of a total of nine pixels of the target pixel and its adjacent eight pixels is the pixel value of the target pixel. FIG. 14 is a diagram illustrating pixel values of images before (a) and after (b) processing when high density region expansion processing is performed using such a maximum value filter. As will be described later, this high-density region expansion processing is performed in order to add the influence of the filtering processing to the neighboring region where the thin line exists when performing the filtering processing using the thin line data as mask data. This is not necessary when the purpose is only to detect the position of the fine line.
[0062]
Furthermore, in step SP32, the image composed of each pixel Mv (i, j) is enlarged in the vertical direction. Specifically, each pixel Mv (i, j) is enlarged m times (= n times) in the vertical direction by using the nearest neighbor interpolation method (nearest neighbor method).
[0063]
This enlargement ratio m (= n) is the reciprocal of the enlargement ratio (reduction ratio) 1 / m (= 1 / n) in the reduced image generation unit 21, and as a result of this enlargement, it is the same as the number of pixels of the original image. Like that. Thereby, it becomes easy to specify the presence position of the fine line in association with each pixel position of the original image. Note that when it is not necessary to detect the presence position of a fine line, for example, when only the presence or absence of a fine line is detected, this enlargement process is unnecessary.
[0064]
As described above, a vertical thin line can be detected.
[0065]
<B3. Horizontal thin line detection>
Next, the detection of the horizontal thin line in step SP40 (FIG. 4) will be described. The fine line detection in the horizontal direction can be performed based on the same processing as the fine line detection in the vertical direction.
[0066]
FIG. 15 is a flowchart showing a procedure for detecting a thin line extending in the horizontal direction, and corresponds to FIG. In FIG. 15, a reference symbol having a subscript h representing “horizontal direction” corresponds to a reference symbol having “v” representing “vertical direction” as a subscript in FIG. 5. Further, the processes of steps SP41 to SP52 are the same as the processes of corresponding steps SP21 to SP32, respectively.
[0067]
Here, the difference from the thin line detection in the vertical direction is that the original image G is reduced in the “horizontal direction” (not in the vertical direction), and thickened in the “vertical direction” which is a direction orthogonal to the horizontal direction. The thin line is detected by performing the thinning process. That is, in the above-described thin line detection operation in the vertical direction, the vertical direction may be replaced with the horizontal direction, and the horizontal direction may be replaced with the vertical direction.
[0068]
Accordingly, in the processes corresponding to step SP22 and step SP26, the maximum and minimum calculations with the adjacent pixels in the “vertical direction” are performed. For example, in step SP23, when performing the high density region thinning process, the pixel value of each pixel of interest Ph (i, j) (FIG. 7) in the reduced image G0 and the adjacent pixel Ph (i, j + 1) in the vertical direction thereof. ) And the pixel value of each pixel Phi (i, j) in the first processed image G1. That is, Phi (i, j) = min {Ph (i, j), Ph (i, j + 1)}. The same applies to other processes.
[0069]
A thin line detection operation or the like can be performed by such processing in step SP40.
[0070]
<B4. Thin line data synthesis in both directions>
In the next step SP60 (FIG. 4), the fine line data in the vertical direction and the fine line data in the horizontal direction thus obtained are combined to generate fine line data (mask data) M in consideration of the fine line data in both directions. . Specifically, the absolute value of the difference between Mv (i, j) and Mh (i, j) is calculated as a new value M (i, j). That is, M (i, j) = abs {Mv (i, j) −Mh (i, j)}. The symbol abs {} represents an absolute value.
[0071]
This operation is based on exclusive OR. Here, the reason why exclusive OR (subtraction) is performed on Mv and Mh instead of OR (addition) is as follows. The values of both Mv and Mh are large, for example, in a portion where a vertical thin line exists and a horizontal thin line exists, that is, a portion where both thin lines intersect. However, such an intersecting portion is unlikely to be a portion where a line break occurs, and conversely, it is considered that a line break is particularly likely to occur when a thin line occurs only in one direction. For example, in FIG. 16, it is considered that line breaks are likely to occur not in the intersection portion A1, but in the thin line portion A2 extending in a direction shifted by a slight angle from the strict vertical direction. Therefore, by performing exclusive OR (subtraction) on Mv and Mh, the above intersection is excluded from the detection area as a thin line portion, and a case where a thin line exists only in one direction is characteristic. It is considered preferable to extract.
[0072]
<B5. Line break prevention treatment>
Next, in step SP70, line break prevention processing is performed based on the mask data M described above. This process is performed by the line break prevention processing unit 50.
[0073]
FIG. 17 is a flowchart showing the detailed operation in step SP70. By performing the processes of steps SP71 to SP75 in FIG. 17 for all the pixels in the original image G, it is possible to perform processing for preventing line breakage on the original image G.
[0074]
First, in step SP71, the corresponding weighting coefficient k is determined based on the value of the mask data M corresponding to the target pixel position of the original image G. The weighting coefficient k is a value represented by a function having the value of the mask data M as a variable, and is represented by a function as shown in FIG. 18, for example. FIG. 18 is a graph in which the horizontal axis represents the value of the mask data M, and the vertical axis represents the value of the weighting coefficient k (where k is a real number between zero and 1).
[0075]
Specifically, the value of k corresponding to the value of each mask data M can be stored in advance in the reference table LUT (FIG. 18). Thereby, k corresponding to the value of each mask data M can be referred to immediately. Even if it is difficult to express the relationship between M and k by a mathematical expression, the value of k can be easily determined.
[0076]
In addition, when the relationship between M and k can be expressed by a predetermined mathematical expression, k corresponding to M may be obtained from time to time based on the predetermined mathematical expression.
[0077]
Moreover, FIG. 19 is explanatory drawing which shows the outline | summary of a line break prevention process. Using the above weighting coefficient k, the result of filtering the original image G is weighted to the original image G to obtain a processed image as an output image.
[0078]
Therefore, in the next step SP73, filter processing is performed using the target pixel and its neighboring pixels. As a filter for this process, for example, a weighted average filter as shown in FIG. 20 can be used. This weighted average filter is a value obtained by multiplying the pixel value of the target pixel by 4/16, a value obtained by multiplying each pixel value of pixels adjacent to the target pixel in the left, right, upper, and lower directions by 2/16, and the target pixel. On the other hand, it is a filter that adds a value obtained by multiplying each pixel value of pixels adjacent in the oblique direction by 1/16 to obtain a new pixel value at the position of the target pixel. It can be performed.
[0079]
Then, in step SP75, as shown in FIG. 19, a value obtained by processing the original image G by the weighted average filter as shown in FIG. In setting the value, the above-described weighting coefficient k is used as the weighting coefficient. More specifically, a value obtained by multiplying the pixel value p2 obtained by performing weighted average filtering on the target pixel and its neighboring pixels by a weighting coefficient k, and a pixel value p1 of the target pixel and a coefficient (1-k ) Is added to a new pixel value at the target pixel position.
[0080]
By performing the operations in steps SP71 to SP75 for all the pixels in the original image G, the original image G can be subjected to processing for preventing line breakage.
[0081]
Here, the mask data M has a value as an index representing the degree of presence of a thin line that is a cause of the occurrence of line breaks for each pixel position of the original image G. Therefore, by determining how much the influence of the weighted average filter is applied according to the coefficient k determined based on the mask data M, the position where the thin line is present is specified, and the weighted average filter is appropriately applied. Can be performed in a specific area. In other words, the “blurring” effect by the weighted average filter can be selectively applied only to the portion where the line break is likely to occur.
[0082]
Here, as shown in FIG. 18, when the mask data M is equal to or less than the threshold value S1, the weighting coefficient k is set to zero. By setting the threshold value S1 as a predetermined value that is not zero (for example, about 20), two adjacent pixels in the reduced image G0 have slightly different pixel values (for example, the pixel values of both pixels are 245 and 255). Even if it is “falsely detected” as a thin line in step SP20 or the like (exactly when it is detected that the degree of the presence of the thin line is not zero), the influence is eliminated, and the above weighted average The result of the filtering process can be weighted selectively with respect to the position where the processing by the filter is necessary.
[0083]
The mask data M is obtained by performing the high density region expansion process in the above-described step SP31. Therefore, by further performing such weighted average filter processing, the influence of the filter processing can be applied not only to the thin line portion but also to the portion in the vicinity of the thin line, and thus processing to smoothly thicken the thin line can be performed. It becomes possible. Thereby, the effect of preventing line breakage can be further increased. At this time, by excluding (subtracting) the mask data M before the expansion process from the mask data after the expansion process, it is possible to “blur and thicken only the surroundings while maintaining the density of the thin line itself”. For this reason, the low-contrast thin line does not disappear.
[0084]
In the above example, the filtering process is performed in step SP73 even when k is not zero. However, when k is zero, the pixel value of the corresponding pixel of the original image is updated as it is without performing the filtering process. You may output as a pixel value.
[0085]
<C. Variations>
In the thin line detection operation in the vertical direction (step SP20) of the above-described embodiment, when creating the reduced image G0, the first processed image G1, the second processed image G2, and the difference image G3, all the pixels are all imaged. However, the present invention is not limited to this. For example, each image G0 to G3 may be created one line (row) at a time. The same applies to the horizontal thin line detection operation in step SP40, and the images G0 to G3 may be created for each column.
[0086]
In the above embodiment, the mask data M is obtained by synthesizing the thin line detection data for two directions in step SP60 or the like. However, the present invention is not limited to this, and the thin line detection data for one direction is used as the mask data. It may be used as M.
[0087]
In the above embodiment, a case where the inverse number m of the reduction ratio 1 / m when the original image is reduced in the vertical or horizontal direction is equal to the number n of pixels in the averaging process along the direction (m = n) will be described. However, it is not limited to this. For example, the reciprocal number m of the reduction ratio 1 / m when the original image is reduced in a predetermined direction may be smaller than the number n of pixels in the averaging process along the predetermined direction.
[0088]
FIG. 21 is a diagram for explaining the reduced image generation operation in step SP21 in such a case, and corresponds to FIG. 6 of the above embodiment.
[0089]
In the above embodiment, the average value of 8 pixels (n = 8) in the predetermined direction in the original image G is 1 pixel in the reduced image G0, and the same operation is shifted by 8 pixels (m = n = 8) in the predetermined direction. By repeating, the original image G was reduced to 1/8 in the predetermined direction to generate a reduced image. In this modification, the average value for every n pixels (n = 8) in the predetermined direction in the original image G Although the same operation is repeated while shifting by 4 pixels (m = 4 ≠ n), the original image G is reduced to ¼ in a predetermined direction to generate a reduced image. . According to this, when detecting a thin line extending in a specific direction by the averaging process of n pixels, it is possible to reduce a lack of information in a predetermined direction at the time of reduction, and thus “ Can be prevented.
[0090]
Here, the direction of the thin line to be detected can be adjusted by changing the value of the number of pixels n in the averaging process. That is, a thin line with a small deviation angle from the specific direction can be detected as the number n of pixels increases, and a thin line with a large deviation angle from the specific direction can be detected with a decrease in the number n of pixels. . In other words, as the number of pixels n is larger, the range of deviation of the fine line to be detected from the specific direction can be limited, and the thin line extending in the specific direction can be detected. In this way, by adjusting the number n of pixels in the averaging process, the directionality of the thin line to be detected can be adjusted, and fine line detection having directivity can be performed.
[0091]
Further, in the above-described embodiment, the detection of the thin line extending in the direction close to the vertical direction and the horizontal direction of the pixel array arranged in the vertical direction and the horizontal direction has been described. This can be used, for example, to avoid a line break phenomenon that occurs when the direction of the halftone dot pattern and the pixel arrangement direction have no inclination (in the case of being parallel). However, the present invention is not limited to this, and can also be applied to detection of a thin line extending in a specific direction other than the vertical direction and the horizontal direction. For example, in the case where the halftone dot pattern has a predetermined inclination in the pixel arrangement direction, the directionality of the thin line that becomes the line break generation source also changes. Depending on the directionality of the thin line, a thin line extending in that direction can be detected, and further, a line break prevention process can be performed based on the detection result. In this case, in accordance with the direction of the thin line that is the source of the line break, it is reduced in the direction of the thin line to be detected (diagonal direction), and the thinning process and the thickening process are sequentially performed in the orthogonal direction. Thus, a thin line can be detected. In the oblique reduction operation, an average of a plurality of corresponding pixels projected by coordinate transformation in the oblique direction may be obtained. In the thinning process and the thickening process, pixels adjacent in the vertical, horizontal, and diagonal directions of the pixel of interest in the pixel row arranged in the vertical and horizontal bidirectional directions are selected according to the direction of the thin line to be detected. The thinning and fattening process can be used properly as needed. Specifically, the adjacent pixel in the direction closest to the direction orthogonal to the direction of the thin line to be detected can be used.
[0092]
Furthermore, in the above-described embodiment, detection of thin lines extending in the vicinity of two directions perpendicular to each other (vertical direction and horizontal direction) has been described. However, the present invention is not limited to this, and any two arbitrary thin lines to be detected are different from each other. Even when there is directivity in the vicinity of directions, it is possible to detect a thin line extending in the vicinity of the two directions. According to this, for example, the same effect can be obtained also when halftone dot conversion is performed using a halftone dot pattern based on two axial directions that are not orthogonal.
[0093]
【The invention's effect】
As described above, according to the image processing apparatus of claim 1, the reduced image generated by reducing the original image in the predetermined direction is thinned in the direction orthogonal to the predetermined direction and then thickened. Since the second processed image is obtained and a difference image that is the difference between the second processed image and the reduced image is obtained, a thin line extending in a direction close to a predetermined direction is detected from the obtained difference image. Can do.
[0094]
According to the image processing apparatus of the second aspect, when the original image is reduced in the predetermined direction, an averaging process of n pixels along the predetermined direction is involved. Therefore, the fine line to be detected is adjusted by adjusting the value of n. The directivity of can be adjusted.
[0095]
According to the image processing apparatus of the third aspect, since the specific density area is a high density area, a high density thin line can be detected based on the difference image. According to this aspect, since the specific density area is a low density area, a low density thin line can be detected based on the difference image.
[0096]
According to the image processing apparatus of the fifth aspect, it is possible to detect a thin line in two directions in which line breakage is particularly likely to occur.
[0097]
According to the image processing device of claim 6, the difference image is subjected to an enlargement process that enlarges the difference image in a predetermined direction to a reciprocal multiple of the reduction ratio in the reduced image generation unit. Since the fine line index indicating the degree of presence or absence of the fine line is generated, the corresponding position of the detected fine line in the original image can be easily obtained.
[0098]
Further, according to the image processing apparatus of claim 7, the weighting coefficient determining means for determining a weighting coefficient when weighting the result of applying a predetermined filter process to the original image to the original image based on the fine line index. And a filter processing means for obtaining a processed image obtained by weighting the original image using the weighting coefficient as a result of applying a predetermined filter process to the original image. It is possible to perform a line break prevention process by specifying a thin line portion that is a high portion and selectively adding an influence of a filter process.
[0099]
According to the image processing apparatus of the eighth aspect, since the weighting coefficient determining means has a correspondence table between the thin line index and the weighting coefficient as a reference table in advance, the weighting coefficient determining means is based on the thin line index. The weighting factor can be easily determined.
[0100]
According to the image processing apparatus of the ninth aspect, when performing the enlargement process, the enlargement processing unit generates a thin line index by performing a process of expanding a high density region of the difference image, and based on the thin line index. Since the filter processing means performs the filter process using the weighting coefficient obtained in this way, the filter process can also be performed in the vicinity of the thin line.
[0101]
Furthermore, according to the recording medium of the tenth aspect, it is possible to obtain the same effect as that of the invention according to any one of the first to ninth aspects.
[Brief description of the drawings]
FIG. 1 is a diagram illustrating a hardware configuration of an image processing apparatus 1 according to an embodiment of the present invention.
FIG. 2 is a functional block diagram related to the image processing apparatus 1;
FIG. 3 is a detailed functional block diagram of the image processing apparatus 1;
FIG. 4 is a flowchart regarding processing operations in the image processing apparatus 1;
FIG. 5 is a flowchart relating to an operation of detecting a thin line extending in the vertical direction.
FIG. 6 is a conceptual diagram for explaining reduced image generation processing;
FIG. 7 is a conceptual diagram for explaining a target pixel Pv (i, j).
FIG. 8 is a diagram for explaining a white thin line detection operation;
FIG. 9 is a diagram for explaining a white thin line detection operation;
FIG. 10 is a diagram for explaining a white thin line detection operation;
FIG. 11 is a diagram for explaining a black thin line detection operation;
FIG. 12 is a diagram for explaining a black thin line detection operation;
FIG. 13 is a diagram for explaining a black thin line detection operation;
FIG. 14 is a diagram for explaining high-density area expansion processing (step SP31).
FIG. 15 is a flowchart relating to an operation of detecting a thin line extending in the horizontal direction.
16 is a diagram showing fine lines of an original image G. FIG.
FIG. 17 is a flowchart showing a detailed operation of a line break prevention process.
FIG. 18 is a diagram showing the relationship between mask data M and weighting coefficient k.
FIG. 19 is an explanatory diagram showing an outline of a line break prevention process.
FIG. 20 is a diagram illustrating a weighted average filter which is an example of a filter.
FIG. 21 is an explanatory diagram relating to a modified example of the reduced image generation operation.
FIG. 22 is a diagram for explaining a line break;
FIG. 23 is a diagram for explaining line breaks;
FIG. 24 is a diagram for explaining line breaks;
FIG. 25 is a diagram for explaining line breaks caused by USM processing;
[Explanation of symbols]
1 Image processing device
20 Fine line detector
20A Vertical direction thin line detector
20B Horizontal thin line detector
21 Reduced image generator
23bA, 23bB Black line extraction unit
23 wA, 23 wB white thin line extraction unit
50 Line break prevention processing section
G Original image
G0 reduced image
G3 difference image
LUT reference table
M mask data
k Weighting factor

Claims (10)

An image processing apparatus that performs processing on an original image,
Reduced image generating means for reducing the original image in a predetermined direction to generate a reduced image;
Thinning processing means for generating a first processed image by performing a thinning process for thinning a specific density region in a direction orthogonal to the predetermined direction with respect to the reduced image;
About the first processed image, a thickening processing unit that generates a second processed image by performing a thickening process of thickening a specific density region in a direction orthogonal to the predetermined direction;
Difference image generation means for generating a difference image that is a difference between the reduced image and the second processed image;
An image processing apparatus comprising: The image processing apparatus according to claim 1.
The reduced image generating means includes an averaging process of n pixels along the predetermined direction when the original image is reduced in the predetermined direction. The image processing apparatus according to claim 1.
The image processing apparatus, wherein the specific density area is a high density area. The image processing apparatus according to claim 1.
The image processing apparatus, wherein the specific density area is a low density area. The image processing apparatus according to claim 1.
The predetermined direction is two different directions,
The reduced image generating means, the thinning processing means, the thickening processing means, and the difference image generating means perform each processing relating to each of the two directions. The image processing apparatus according to any one of claims 1 to 5,
A thin line index representing the degree of presence or absence of a thin line at each pixel position of the original image by performing an enlargement process for enlarging the difference image in the predetermined direction to a reciprocal of the reduction ratio in the reduced image generation unit. Expansion processing means for generating
An image processing apparatus further comprising: The image processing apparatus according to claim 6.
A weighting coefficient determining means for determining a weighting coefficient when weighting the original image with a result of applying a predetermined filtering process to the original image based on the fine line index;
Filter processing means for obtaining a processed image obtained by weighting the original image using the weighting coefficient, as a result of applying a predetermined filter process to the original image;
An image processing apparatus further comprising: The image processing apparatus according to claim 7.
The image processing apparatus according to claim 1, wherein the weighting coefficient determining means has a correspondence relationship between the thin line index and the weighting coefficient in advance as a reference table. The image processing apparatus according to claim 7.
The enlargement processing unit generates the fine line index by performing a process of expanding a high density region of the difference image when performing the enlargement process. A computer-readable recording medium storing a program for causing a computer to function as the image processing apparatus according to any one of claims 1 to 9.

Images