JPH0670164A - Image processor - Google Patents

Abstract

PURPOSE:To prevent the break and the omission of the thin lines by applying an operational expression suited to the vertical, horizontal and oblique thin lines. CONSTITUTION:The data corrected by an arithmetic circuit 20 are compared with the prescribed threshold value through a binarizing circuit 21 and then binarized. The binarized data are outputted as the correction data. Meanwhile the circuit 20 selects an MTF (space frequency transmission characteristic) correcting operational expression based on a selection signal S1 serving as the correction data on a picture element B of a line Ln-1 set before a noted picture element E and a selection signal S2 serving as the correction data on a picture element D preceding the element E.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus for correcting deterioration of MTF (spatial frequency transfer characteristic) of an image.

[0002]

2. Description of the Related Art Generally, since the photoelectric conversion output of an image reading device is deteriorated due to document contrast and MTF of an optical system, an image sensor, etc., for example, a block of 3 × 3 pixels or 5 × 5 pixels ( In the matrix), the target pixel is subjected to MTF correction based on the surrounding pixels, so that fine lines and contours are clearly reproduced.

For example, as shown in FIG. 2, 3 × 3 pixels A to
When the pixel of interest E is corrected in I, the correction calculation of E ′ = 3E− (1/2) (B + D + F + H) (1) is performed in the method of using the pixels B, D, F, and H on the upper, lower, left, and right sides. Further, in the method including all the pixels A to I, the correction calculation of E ′ = 3E− (1/4) (A + B + C + D + F + G + H + I) (2) is performed. Then, this E ′ is binarized by a predetermined threshold value, and MTF correction is performed.

[0004]

However, as shown in FIG.
When the horizontal thin line is read as shown in (a) and multi-valued data for three lines is obtained as shown in FIG. 5 (b), when the calculation is performed by the equation (1), it is shown in FIG. 5 (b). The target pixel data “5” in the matrix 1 is E ′ = 3.5− (1/2) (3 + 3 + 5 + 3) = 8, and the target pixel data “5” in the matrix 2 is
Becomes E ′ = 3.5− (1/2) (3 + 5 + 5 + 3) = 7. Therefore, 8 or more is 1 (black), and 7 or less is 0.
When binarized as (white), the pixel data of interest “5” in the matrix 2 is binarized as white, which causes a problem of line breakage.

Further, when a relatively thick horizontal thin line is read as shown in FIG. 6 (a) to obtain multi-valued data for five lines as shown in FIG. 6 (b), according to the equation (1), When calculated, each pixel data “6” of interest in the matrices 1, 2 and 3 shown in FIG. 6B is: E ′ = 3.6− (1/2) (2 + 2 + 6 + 6) = 10 E ′ = 3.6− (1/2) (2 + 6 + 6 + 6) = 8 E '= 3.6- (1/2) (6 + 6 + 6 + 6) = 6. Therefore, if binarization is performed with the same threshold value, the pixel of interest in the matrix 3 becomes white, so that it looks like a hollow character and becomes an image that is very difficult to see.

Further, as shown in FIG. 10, when a thin horizontal line and a vertical line intersect and the value of the pixel of interest is equal to that of each of the upper, lower, left and right pixels, the pixel of interest is determined to be white and a line break occurs. That is, when the value of the pixel of interest and the values of the upper, lower, left, and right pixels are “6”, E ′ = 3 · 6− (1/2) (6 + 6 + 6 + 6) = 6 in Equation (1), and in Equation (2), Since '= 3.6- (1/4) (6 + 6 + 6 + 6 + 6 + 6) = 6, the pixel of interest is determined to be white when binarized by setting 8 or more to 1 (black) and 7 or less to 0 (white).

In view of the above-mentioned conventional problems, the present invention is an MTF.
An object of the present invention is to provide an image processing apparatus capable of appropriately correcting an image that cannot be corrected by correction, and particularly capable of preventing thin line breaks and hollow lines.

[0008]

In order to achieve the above-mentioned object, the first means uses one of a plurality of predetermined arithmetic expressions to determine the value of the pixel of interest based on each value of the pixel of interest and its surrounding pixels. Calculation means for selectively performing MTF correction, binarization means for binarizing the value of the pixel of interest calculated by the calculation means into white or black with a predetermined threshold value, and binarization by the binarization means. It is characterized by further comprising: data of a line before the target pixel and selection means for selecting an arithmetic expression of the arithmetic means based on the previous data.

The second means is a calculating means for performing MTF correction of the value of the pixel of interest with one predetermined arithmetic expression on the basis of each value of the pixel of interest and its surrounding pixels, and the pixel of interest calculated by the calculating means. Binarizing means for selectively binarizing the value of 1 into white or black with one of a plurality of threshold values, and an error in the arithmetic expression calculated by the arithmetic means, and the binary value is calculated based on this error. And a selecting means for selecting a threshold value of the converting means.

A third means is to calculate the value of the pixel of interest by a predetermined arithmetic expression based on the respective values of the pixel of interest and its diagonal pixels.
The calculation means for TF correction and the value of the pixel of interest calculated by the calculation means are selectively changed to white or black by a predetermined threshold value.
And a binarizing means for digitizing.

According to a fourth means, the value of the pixel of interest is calculated based on the respective values of the pixel of interest and its surrounding pixels by the MTF.
The value of the pixel of interest is calculated by the second arithmetic expression based on each value of the first arithmetic means for correcting and the pixel around the pixel of interest.
And a second arithmetic means for correcting and a binarizing means for selectively binarizing one of the values of the pixel of interest calculated by the first or second arithmetic means into white or black with a predetermined threshold value. It is characterized by having.

The fifth means calculates the value of the pixel of interest based on the respective values of the pixel of interest and the pixels around the pixel of interest using the first arithmetic expression MTF.
The value of the pixel of interest is corrected by the second arithmetic expression based on the respective values of the first arithmetic means for correcting and the pixel of interest and the surrounding pixels.
The second calculation means for correction and the selection means for comparing the respective values of the target pixel calculated by the first and second calculation means and selecting the larger value, and the value selected by the selection means. And a binarizing means for selectively binarizing the above with a predetermined threshold value into white or black.

In the sixth means, the first computing means in the fourth or fifth means computes the value of the pixel of interest based on the respective values of the pixel of interest and the pixels only in the upper, lower, left and right thereof, and then in the second means. The calculation means calculates the value of the pixel of interest on the basis of the respective values of the pixel of interest and the pixels that are only oblique thereto.

[0014]

With the above arrangement, the first means can selectively MTF-correct the value of the pixel of interest with one of a plurality of predetermined arithmetic expressions, and the data of the line before the binarized pixel of interest can be obtained. Since the arithmetic expression is selected on the basis of the previous data, it is possible to prevent the thin line from being broken or hollow even if it is binarized with a predetermined threshold value.

In the second means, the value of the pixel of interest calculated by the calculating means can be selectively binarized into white or black by one of a plurality of threshold values, and the value of the pixel of interest in front of the binarized pixel of interest. Since the threshold value is selected based on the data of the line and the previous data, even if the pixel of interest is corrected by a predetermined arithmetic expression, it is possible to prevent the thin line from being broken or the void.

In the third means, the value of the target pixel is corrected based on the respective values of the target pixel and its diagonal pixels only.
It is possible to prevent vertical and horizontal fine lines from being broken or hollow.

In the fourth means, the pixel of interest is corrected by the first and second arithmetic expressions. Therefore, by using an arithmetic expression suitable for vertical and horizontal thin lines and an arithmetic expression suitable for diagonal thin lines,
It is possible to prevent line breaks and hollows between vertical and horizontal thin lines and diagonal thin lines.

In the fifth means, the calculation result corrected by the first and second calculation formulas is automatically selected.
By using an arithmetic expression suitable for a horizontal thin line and an arithmetic expression suitable for an oblique thin line, it is possible to automatically prevent line breaks or voids in vertical and horizontal thin lines and oblique thin lines.

In the sixth means, the first computing means computes the value of the pixel of interest based on the respective values of the pixel of interest and the pixels only in the upper, lower, left, and right thereof, and the second computing means calculates the value of the pixel of interest. Since the value of the pixel of interest is calculated based on each value of the diagonal pixels only, it is possible to prevent line breaks and hollow lines between the vertical and horizontal thin lines and the diagonal thin lines with a simple circuit configuration.

[0020]

Embodiments of the present invention will be described below with reference to the drawings. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention, and FIG. 2 is an explanatory diagram showing processing blocks of the image processing apparatus of FIG.

In FIG. 1, the data read by the image reading device 11 is the flip-flops 12 and 13.
Is delayed by one pixel each, and is delayed by the memory 14 by one line.
6, each pixel is delayed by one pixel, and the memory 17 is delayed by one line.
19 delays each by one pixel.

Therefore, as shown in FIG. 2, 3 × 3 = 9
For the pixels A to I, the data of the pixels A to C on the line L _n-1 before the pixel of interest E are flip-flops 19,
18, output from the memory 17, line L of the target pixel E
Pixels D to F of n are flip-flops 16 and 1 respectively.
5, the data of the pixels G to I on the line L _{n + 1} next to the target pixel E, which is output from the memory 14, is output from the flip-flops 13 and 12 and the image reading device 11, respectively.

The arithmetic circuit 20 outputs a 2-bit selection signal S1,
The value of the target pixel E can be corrected by the following four types of correction arithmetic expressions according to S2.

E ′ = 3E− (½) (B + D + F + H) (1) E ′ = 5E− (B + D + F + H) (3) E ′ = 3E− (B + H) (4) E ′ = 3E− (D + F) ( 5) That is, in the formula (1), correction is performed by the pixels on the upper, lower, left, and right sides of the target pixel E, and in the formula (3), the weighting coefficients of the target pixel E and the pixels on the upper, lower, left, and right are different from the formula (1), In (), correction is performed only on pixels above and below the target pixel E, and in equation (5), correction is performed only on the left and right pixels of the target pixel E.

The data corrected by the arithmetic circuit 20 is
The binarizing circuit 21 compares the binarized data with a predetermined threshold value, outputs the binarized data as correction data, and the memory 22 delays the data by one line, and then the flip-flop 23 stores one pixel. It is delayed by one minute, and is also delayed by one pixel by the flip-flop 24. Then, the flip-flops 23 and 24
, The selection signals S1 and S2 are applied to the arithmetic circuit 20, respectively.

That is, the selection signal S1 is the correction data for the pixel B on the line L _n-1 before the target pixel E, and the selection signal S2 is the correction data for the pixel D before the target pixel E. Then, the arithmetic circuit 20 selects an equation as follows according to the selection signals S1 and S2.

When S1 = S2 = 0 or S1 = S2 = 1 ... Equation (1) When S1 = 1 and S2 = 0 ... Equation (5) When S1 = 0 and S2 = 1 ... Equation (4) Therefore, the pixel B on the line L _n−1 before the pixel of interest E and the pixel D on the previous line are both white (S1 = S2 = 0) or black (S1 = S2 = 0).
= S2 = 1), the pixel B on the line Ln−1 before the pixel of interest E is black (S
1 = 1) and the previous pixel D is white (S2 = 0), it is corrected by the equation (5) using only the pixels D and F on the left and right, that is, the main scanning direction, and the pixel before the target pixel E is corrected. The pixel B of the line L _n-1 is white (S1 = 0) and the previous pixel D is black (S2
= 1), the pixels B, H in the vertical direction, that is, in the sub-scanning direction
It is corrected by the equation (4) using only

Then, when the horizontal thin line is read as shown in FIG. 5A to obtain multi-valued data for three lines as shown in FIG. 5B, the pixel data of interest "5" in the matrix 2 is obtained. “E ′ = 3 · 5− (3 + 3) = 9, so if 8 or more is 1 (black) and 7 or less is 0 (white), the pixel data of interest“ 5 ”in the matrix 2 is It is binarized as black, and therefore line breaks can be prevented from occurring.

The equations in the above embodiment are examples, and the weighting coefficient and the pixel of interest are arbitrary, and 3 ×
Instead of 3 pixels, for example, a block of 5 × 5 pixels may be used as a processing unit.

Next, a second embodiment will be described with reference to FIG. In FIG. 3, the flip-flops 12, 13, 15, 16, 18 and the memory 1 for obtaining 3 × 3 pixels are shown.
Since 4 and 17 have the same configuration, description thereof will be omitted. Then, the arithmetic circuit 20a corrects, for example, only by E '= 3E- (1/2) (B + D + F + H) (1), and the binarization circuit 21a changes this correction data to a variable threshold value according to the selection signals S1 and S2. Binarize with.

Further, the correction amount calculation circuit 25 calculates, for example, ΔE = 2E− (1/2) (B + D + F + H) (6), and when the calculation result ΔE is 2 or more, it is “1”.
When it is smaller than 2, "0" is output. And
This data is delayed by one line by the memory 22, then delayed by one pixel by the flip-flop 23, and delayed by one pixel by the flip-flop 24. The selection signals S1 and S2 are output from the flip-flops 23 and 24, respectively, to the binarization circuit 21a.
Applied to.

That is, the selection signal S1 is based on the correction data of the pixel B of the line L _n-1 before the target pixel E, and the selection signal S2 is based on the correction data of the pixel D before the target pixel E. Then, the binarizing circuit 21a sets the threshold value to "8" when S1 = S2 = 0, sets the threshold value to "6" when S1 + S2≥1, and outputs the data E'calculated by the arithmetic circuit 20a.
Is binarized.

Therefore, when the horizontal thin line is read as shown in FIG. 5A and multivalued data for three lines is obtained as shown in FIG.
The pixel data of interest "5" in the matrix 2 is E '= 3.5- (1/2) (3 + 5 + 5 + 3) = 7, but since it is binarized with the threshold value "6", line breakage may occur. Can be prevented.

Further, when a relatively thick horizontal thin line is read as shown in FIG. 6 (a) to obtain multi-valued data for 5 lines as shown in FIG. 6 (b), according to equation (1), When calculated, each target pixel data “6” in the matrix 3 shown in FIG. 6B becomes E ′ = 3 · 6− (1/2) (6 + 6 + 6 + 6) = 6, but it is binary with the threshold value “6”. Therefore, it is possible to prevent the characters from becoming hollow characters.

Next, a modification of the second embodiment will be described with reference to FIG. In the second embodiment described above, the correction amount calculation circuit 25 is configured to change the threshold value of the binarization circuit 21a. However, as shown in FIG. 4, the pixel B of the line Ln-1 before the target pixel E is changed. Correction data (selection signal S1),
Correction data for the pixel D before the pixel of interest E (selection signal S2)
The threshold value of the binarization circuit 21a may be changed by.

That is, for example, when S1 = S2, the threshold value is set to "8", and when S1 ≠ S2, that is, the correction data of the pixel B and the correction data of the previous pixel D on the previous line L _n-1. If is different, the threshold is set to "6". Therefore, the reproducibility of black can be improved by lowering the threshold value at the edge portion where the white pixel and the black pixel are adjacent to each other.

Next, a third embodiment will be described with reference to FIG. In this embodiment, the arithmetic circuit 21b is configured so that E '= 3E- (1/2) (A + C + G + I) (7) That is, the target pixel E is corrected using only the diagonal pixels A, C, G, and I. Has been done.

Therefore, as shown in FIG. 10, when a horizontal line and a vertical line cross each other and the values of the pixel of interest and the upper, lower, left, and right pixels are the same, for example, when the values of each pixel are both "6", E ' Since = 3.6- (1/2) (0 + 0 + 0 + 0) = 18, the pixel E of interest is determined to be black when binarized by setting 8 or more to 1 (black) and 7 or less to 0 (white).

Here, if only the equation (7) is calculated,
As shown in FIG. 8, when diagonal lines intersect, as in the conventional example, E '= 3.6- (1/2) (6 + 6 + 6 + 6) = 6 in the formula (1), and in the formula (2). , E ′ = 3.6− (1/4) (6 + 6 + 6 + 6 + 6 + 6) = 6, so if 8 or more is 1 (black) and 7 or less is 0 (white), the pixel of interest is determined to be white. .

Therefore, as shown in FIG. 9, a circuit 20a operated by the equation (1) and a circuit 20b operated by the equation (7) are provided, and the operation results of both circuits 20a, 20b are compared by the selection circuit 26 and are large. By selecting one, vertical, horizontal,
It is possible to prevent diagonal line breaks. That is, FIG.
In the case of the diagonal line shown in, the equation (1) becomes E ′ = 3.6− (1/2) (6 + 6 + 6 + 6) = 6, and in the equation (7), E ′ = 3.6− (1/2) Since (0 + 0 + 0 + 0) = 18, the pixel of interest E can be selected by selecting equation (7).
Can be determined to be black.

Further, for example, in the case of the data as shown in FIG. 8 (c), in the formula (1), E '= 3.6- (1/2) (6 + 4 + 4 + 6) = 8, and in the formula (7), Since E ′ = 3.6− (1/2) (4 + 4 + 4 + 4) = 10, the pixel of interest E can be selected by selecting the equation (7).
Can be determined to be black. In the example shown in FIG. 9, the selection circuit 26 automatically compares the calculation results of the two circuits 20a and 20b and selects the larger one. 20
It may be configured to select b.

[0042]

As described above, according to the first aspect of the present invention, the value of the target pixel is selectively selected by one of a plurality of predetermined arithmetic expressions based on each value of the target pixel and its surrounding pixels. MTF
Computation means for correcting, binarizing means for binarizing the value of the pixel of interest calculated by the computing means into white or black with a predetermined threshold, and the pixel of interest binarized by the binarizing means. Since the data of the previous line and the selection means for selecting the arithmetic expression of the arithmetic means on the basis of the previous data are provided, it is possible to prevent the thin lines from being broken and the hollow lines even when binarized with a predetermined threshold value. it can.

According to a second aspect of the present invention, an arithmetic means for performing MTF correction of the value of the pixel of interest with one predetermined arithmetic expression on the basis of each value of the pixel of interest and the surrounding pixels, and the arithmetic means. Binarizing means for selectively binarizing the value of the target pixel into white or black with one of a plurality of threshold values, and an error in the arithmetic expression calculated by the arithmetic means, and based on this error, the above-mentioned Since the selection means for selecting the threshold value of the binarization means is provided, even if the pixel of interest is corrected by a predetermined arithmetic expression, it is possible to prevent the thin line from being broken or the hollow portion.

According to a third aspect of the present invention, an arithmetic means for performing MTF correction of the value of the pixel of interest with a predetermined arithmetic expression based on the respective values of the pixel of interest and its diagonal pixels only, and the attention calculated by the arithmetic means. Since a binarizing means for selectively binarizing the pixel value into white or black with a predetermined threshold value is provided, it is possible to prevent line breaks or hollow lines in vertical and horizontal fine lines.

According to a fourth aspect of the present invention, there is provided a first arithmetic means for performing MTF correction of the value of the pixel of interest by the first arithmetic expression based on the respective values of the pixel of interest and the pixels around it, and the periphery of the pixel of interest. A second calculation means for MTF correcting the value of the pixel of interest based on each value of the pixel of the second calculation formula and one of the values of the pixel of interest calculated by the first or second calculation means. Since a binarizing means for selectively binarizing white or black with a threshold value of is used, by using an arithmetic expression suitable for vertical and horizontal thin lines and an arithmetic expression suitable for diagonal thin lines, It is possible to prevent the horizontal thin line and the diagonal thin line from being broken or hollow.

According to a fifth aspect of the present invention, the first calculation means for MTF-correcting the value of the target pixel with the first calculation formula based on the respective values of the target pixel and its surrounding pixels, the target pixel and its surroundings. The second calculation means for MTF-correcting the value of the pixel of interest based on the respective values of the pixel of No. 2 and the respective values of the pixel of interest calculated by the first and second calculation means are compared. Since the selection means for selecting the larger value and the binarization means for selectively binarizing the value selected by the selection means into white or black with a predetermined threshold value are provided, By using an arithmetic expression suitable for a thin line and an arithmetic expression suitable for a slanting thin line, it is possible to automatically prevent line breaks and hollows in vertical and horizontal thin lines and slanting thin lines.

The invention according to claim 6 is the invention according to claim 4 or 5.
In the invention described above, the first computing means computes the value of the pixel of interest based on the respective values of the pixel of interest and the pixels only in the upper, lower, left, and right thereof, and in the second computing means, the pixel of interest and the pixels only in the diagonal thereof. Since the value of the pixel of interest is calculated based on each value of
With a simple circuit configuration, it is possible to prevent vertical or horizontal thin lines and diagonal thin lines from being broken or hollow.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of an image processing apparatus according to the present invention.

FIG. 2 is an explanatory diagram showing processing blocks of the image processing apparatus of FIG.

FIG. 3 is a block diagram showing an image processing apparatus of a second embodiment.

FIG. 4 is a block diagram showing an image processing device of a modified example of the second embodiment.

FIG. 5 is an explanatory diagram showing thin lines and their multivalued data.

FIG. 6 is an explanatory diagram showing another thin line and its multivalued data.

FIG. 7 is a block diagram showing an image processing apparatus of a third embodiment.

FIG. 8 is an explanatory diagram showing a case where vertical and horizontal fine lines are subjected to MTF correction with diagonal pixels.

FIG. 9 is a block diagram showing an image processing apparatus of a fourth embodiment.

FIG. 10 is an explanatory diagram showing a case in which vertical and horizontal fine lines are subjected to MTF correction with pixels on the top, bottom, left, and right.

[Explanation of symbols]

 20, 20a, 20b Arithmetic circuit 21, 21a, 21b Binarization circuit 25 Correction amount calculation circuit 26 Selection circuit

Claims (6)

[Claims] 1. A calculation means for selectively MTF-correcting the value of a target pixel with one of a plurality of predetermined calculation expressions based on the respective values of the target pixel and its surrounding pixels, and the calculation means. Binarizing means for binarizing the value of the pixel of interest into white or black with a predetermined threshold value, and the data on the previous line and the previous data of the pixel of interest binarized by the binarizing means. An image processing apparatus comprising: a selection unit that selects an arithmetic expression of the calculation unit. 2. A calculation means for performing MTF correction on the value of the target pixel with a predetermined one calculation formula based on the respective values of the target pixel and its surrounding pixels, and a plurality of values of the target pixel calculated by the calculation means. Binarizing means for selectively binarizing into white or black with one of the thresholds, and an error in an arithmetic expression calculated by the arithmetic means, and the threshold of the binarizing means based on the error. An image processing apparatus comprising: a selection unit for selecting. 3. A calculation unit for performing MTF correction on the value of the pixel of interest by a predetermined calculation formula based on each value of the pixel of interest and its diagonal pixels only, and a value of the pixel of interest calculated by the calculation unit is set to a predetermined value. An image processing apparatus comprising: a binarizing unit that selectively binarizes white or black with a threshold value. 4. A first MTF-correcting value of a pixel of interest based on respective values of the pixel of interest and pixels around it
And a second arithmetic means for MTF correcting the value of the pixel of interest with a second arithmetic expression based on each value of pixels around the pixel of interest, and arithmetic operation by the first or second arithmetic means An image processing apparatus comprising: a binarizing unit that selectively binarizes one of the values of the noted pixel of interest with a predetermined threshold value into white or black. 5. A first arithmetic expression for MTF-correcting a value of a target pixel based on respective values of the target pixel and its surrounding pixels.
And a second calculating means for performing MTF correction of the value of the pixel of interest with a second arithmetic expression based on each value of the pixel of interest and its surrounding pixels, and calculation by the first and second calculating means. Selecting means for comparing the respective values of the selected target pixels and selecting the larger value; and binarizing means for selectively binarizing the value selected by the selecting means into white or black with a predetermined threshold value. And an image processing apparatus including. 6. The first computing means computes the value of the pixel of interest based on the respective values of the pixel of interest and the pixels only on the top, bottom, left, and right thereof, and the second computing means calculates only the pixel of interest and its diagonal. 6. The image processing apparatus according to claim 4, wherein the value of the pixel of interest is calculated based on each value of the pixel.