JPH08163360A - Method and device for picture binarization - Google Patents

Abstract

PURPOSE: To provide the picture binarizing device which converts a multilevel picture to a binary picture of high quality where the occurrence of a fixed false image is reduced. CONSTITUTION: A binarizing means 1 of this device binarizes a correction picture signal Sc and obtains the error at this time to output it as an error signal Se . This error signal Se is stored in an error storage means 2. A correction means 3 reads out the error signal Se of picture elements around a noticed picture element from the error storage means 2 and adds this signal to a multilevel picture signal S0 of the noticed picture element to generate the correction picture signal Sc . The correction means 3 periodically charges the picture element, whose error signal Se should be read out, in accordance with the position of the noticed picture element to add this signal. This operation is successively repeated for the other picture elements to obtain a binary picture. Since the position of the picture element whose error signal Se should be taken out is periodically changed, a fixed pattern hardly occurs in the binary picture.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image binarization method and an image binarization apparatus suitable for a document image filing apparatus, an electronic copying machine, a facsimile apparatus, etc., which binarizes a multivalued image by an error diffusion method. .

[0002]

2. Description of the Related Art Binary recording, which is easy to record and excellent in stability, has been conventionally used for recording multivalued images having gradation.

In this binary recording, a multi-valued image obtained by scanning an original is converted into a binary value of white or black according to the pixel value and recorded. In this binary recording, an error diffusion method, a dither method, a density pattern method or the like which is known as a pseudo halftone expression method is used so that an image part belonging to an intermediate gradation looks gray when recorded. . Among these pseudo-halftone expression methods, the error diffusion method is generally used because it can express gradation more smoothly and has better resolution than other pseudo-halftone expression methods.

This error diffusion method will be described by taking as an example the case where the pixel value has 256 gradations of 0 to 255 (black = 0, white = 255). For example, if a multi-valued image signal having a pixel value of "239" is binarized with a threshold value of "128",
It becomes "1", but the true pixel value of white for "1" is "25".
The error of 239−255 = −16 occurs for “5”.
In the error diffusion method, this error is distributed and added to the multi-valued image signals of pixels to be binarized thereafter, so that the average gradation of the binary image can be compared with that of the multi-valued image in a wide range. It is intended to approach the gradation.

There are various kinds of error distribution methods. For example, as shown in FIG. 3, when a certain pixel D ₀ is binarized, each of the main scanning direction A and the sub-scanning direction B is determined. A method is often used in which the errors of the previous pixels D ₁ and D _{2 in} the direction are added by ½.

[0006]

However, in the above-described conventional error distribution method, the errors of the pixels having the same positional relationship (the left pixel D ₁ and the upper pixel D _{2 in} the example of FIG. 3) are always added. Therefore, there is a problem that a certain pattern is likely to occur in the binary image, and the image quality of appearance is not so good.

Therefore, the present invention has been made in view of the above circumstances, and an image binarization method and an image 2 capable of converting a multi-valued image into a high-quality binary image with less occurrence of constant false images. It is an object to provide a quantizer.

[0008]

An image binarization method according to claim 1, wherein a pixel value of a pixel of interest next to be binarized in a multi-valued image is binarized for a pixel near the pixel of interest. In the image binarization method for obtaining a binary image by an error diffusion method in which the error at the time is added at a predetermined ratio and the pixel value after the addition is binarized, the pixel of the error addition target is set to the position of the target pixel. It is characterized in that the addition is carried out by periodically changing it accordingly.

An image binarization apparatus according to a second aspect of the present invention generates a corrected image signal based on the input multi-valued image signal and obtains a binary image by the error diffusion method of binarizing the corrected image signal. In the image binarization device, the correction image signal is binarized by a predetermined threshold value, a binarization unit that obtains an error when binarizing the corrected image signal and outputs the error signal, and an error that stores the error signal. A correction image is obtained by reading the error signal when binarizing the pixels in the vicinity of the target pixel from the storage unit and the multi-valued image signal of the target pixel to be binarized next, and adding the error signal at a predetermined ratio. A means for generating a signal, comprising a correction means for periodically changing the pixel from which the error signal is read according to the position of the pixel of interest to perform the addition.

According to another aspect of the image binarizing apparatus of the present invention, the correcting means outputs a main scanning direction periodic signal which periodically changes every time binarization of a predetermined number of main scanning lines is completed. A direction counting means, a main scanning direction error selecting means for selecting an error signal corresponding to the main scanning direction cycle signal from the error storing means, and a cycle every time binarization of a predetermined number of pixels in the sub-scanning direction is completed. Sub-scanning direction counting means for outputting a sub-scanning direction periodical signal that changes with time, a sub-scanning direction error selecting means for selecting an error signal corresponding to the sub-scanning direction periodical signal from the error storage means, and both selection means. And an addition unit that adds the selected error to the multi-valued image signal of the pixel of interest to be binarized at a predetermined ratio to generate the corrected image signal.

[0011]

According to the image binarization method of the first aspect, since the pixels to be subjected to error addition periodically change, it is difficult for a constant pattern to occur in the binary image.

According to the image binarizing apparatus of the second aspect,
The binarizing means binarizes the corrected image signal, calculates the error at that time, and outputs it as an error signal. The error signal is stored in the error storage means. The correction unit reads the error signal of the pixel in the vicinity of the multi-valued image signal of the pixel of interest from the error storage unit and adds the error signal to generate a corrected image signal. The addition is performed by periodically changing the pixel from which the error signal is read according to the position of the pixel of interest. By repeating this operation for other pixels in sequence, a binary image can be obtained.
Since the position of the pixel from which the error signal is taken out periodically changes, it is difficult for a fixed pattern to occur in the binary image.

According to the image binarization apparatus of the third aspect,
The binarizing means binarizes the corrected image signal, calculates the error at that time, and outputs it as an error signal. The error signal is stored in the error storage means. The main scanning direction counting means and the sub scanning direction counting means respectively output a main scanning direction periodic signal and a sub scanning direction periodic signal which change periodically. The main scanning direction error selecting means and the sub scanning direction error selecting means select an error signal corresponding to the main scanning direction periodic signal and the sub scanning direction periodic signal from each error storage means. The adding means adds each selected error signal to the multi-valued image signal of the pixel of interest to generate a corrected image signal. By repeating this operation for other pixels in sequence, a binary image can be obtained. Since the position of the pixel from which the error signal is extracted changes periodically,
A fixed pattern is less likely to occur in the value image.

[0014]

Embodiments of the present invention will now be described in detail with reference to the drawings.

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

This image binarization apparatus generates a corrected image signal S _c based on the input multi-valued image signal S ₀ and binarizes the corrected image signal S _c by the error diffusion method. is intended to obtain, with binarizes the corrected image signal S _c with a predetermined threshold, the binarizing means 1 for outputting an error when the binarization thereof calculated as an error signal S _e, the error signal S _e
Error storage means 2 for storing the error signal and a multi-valued image signal S ₀ of a pixel to be binarized next (hereinafter referred to as “target pixel”), and an error signal S _e when the pixel near the target pixel is binarized.
Is read from the error storage means 2 and added at a predetermined ratio to generate a corrected image signal S _c .

Next, the details of each of the above parts will be described.

When the correction image signal S _c from the correction unit 3 has 256 gradations, the gradation value N of the correction image signal S _c has a predetermined threshold value, for example, " If it is larger than 128 "," 1 "is output to the outside, and if it is smaller," 0 "is output to the outside as the binarized signal S ₁ . Further, the binarizing means 1 is configured such that, when the binarized signal S ₁ obtained by the threshold value processing is “1”, the true pixel value “255” of white corresponding to the “1” and the corrected image signal S _{When the} binarized signal S ₁ is “0”, the error (N-255) from the gradation value N of _c is corrected to the true pixel value “0” of black corresponding to the “0” and the corrected image signal S. The error from the tone value N of _c (N-
0) is output to the error storage means 2 as the error signal S _e .

The error storage means 2 is adapted to store at least one main scanning line above one main scanning line to be binarized and one pixel to the left of the target pixel, the error signal S _e. ing.

The correction means 3 stores the error signal S _{e obtained} by binarizing the pixels in the vicinity of the target pixel as the error storage means 2.
The corrected image signal S _c is generated by adding the read signal from the pixel at a predetermined ratio, and the pixel from which the error signal S _e is read is periodically changed according to the position of the pixel of interest to perform the addition. As a specific configuration example, the correction unit 3 includes a main scanning direction counting unit 30 that outputs a main scanning direction periodic signal S _m.
And the main scanning direction error selecting means 3 for selecting the error signal S _e corresponding to the main scanning direction periodic signal S _m from the error storing means 2.
1, the sub-scanning direction periodic signal S _n and the sub-scanning direction counting means 32 for outputting the sub-scanning direction error selection means 33 for selecting the error signal S _e from the error memory unit 2 corresponds to the sub scanning direction period signal S _n And an adding means 34 for adding each selected error signal S _e to the multi-valued image signal S ₀ of the target pixel at a predetermined ratio to generate a corrected image signal S _c .

The main-scanning-direction counting means 30 supplies to the main-scanning-direction error selecting means 31 a main-scanning-direction periodic signal S _m which changes periodically each time the binarization of a predetermined number of, for example, one main-scanning line is completed. It is what is output. This main scanning direction counting means 3
0 is the main scanning direction periodic signal S _m , for example, “0” when the main scanning line to be binarized is counted from above and is an even number.
In the case of an odd number, "1" is output. The main scanning direction periodic signal S that periodically changes each time the binarization of two or more continuous main scanning lines is completed.
You may make it output _m .

In the main scanning direction error selecting means 31, the main scanning direction periodic signal S _m from the main scanning direction counting means 30 is "0".
In the case of, the error signal S _{e obtained} by binarizing the pixel on the upper left of the pixel of interest (the pixel for error addition) is the error signal S _e on the left of the pixel of interest when the main scanning direction periodic signal S _m is “1”. The error signal S _{e obtained} by binarizing a pixel (pixel for error addition) is selected and output to the adding means 34 as an error in the main scanning direction.

The sub-scanning direction counting means 32 outputs a sub-scanning direction periodic signal S _n which changes periodically every time the binarization of a predetermined number in the sub-scanning direction, for example, one pixel, is completed. Is output to. For example, when the sub-scanning line of the pixel of interest is counted from the left, “0” is output as the sub-scanning line selection signal S _n when the pixel is an even-numbered pixel and “1” when it is an odd-numbered pixel. The sub-scanning direction periodic signal S _n that periodically changes may be output every time the binarization of two or more consecutive pixels in the sub-scanning direction is completed.

In the sub-scanning direction error selecting means 33, the sub-scanning direction periodic signal S _n from the sub-scanning direction counting means 32 is "0".
In case of, the pixel above the pixel of interest (pixel for error addition)
Is binarized, the error signal S _e when the sub-scanning direction periodic signal S _n is “1” is the binarized error signal when the pixel at the upper right of the pixel of interest (pixel for error addition) is binarized. S _e is selected and output as an error in the sub-scanning direction.

The adding means 34 includes a main scanning direction error (error signal S _e ) selected by the main scanning direction error selecting means 31 and a sub scanning direction error selecting means 33 for the input multi-valued image signal S ₀ of the target pixel. The selected sub-scanning direction error (error signal S _e ) is added by a predetermined ratio, for example, 1/2, and the corrected image signal S _c (= S ₀ + (S _e + S _e ) / 2) is set to 2
It is output to the digitizing means 1.

Next, the operation of this embodiment will be described with reference to FIG. FIG. 2 is a diagram for explaining the addition by the adding means 34. In the figure, A indicates the main scanning direction and B indicates the sub scanning direction.

The multi-valued image signal S ₀ of each pixel on the uppermost main scanning line of the multi-valued image is sequentially input to the adding means 34 from left to right, and when the signal input on the main scanning line is finished. ,
Signals are sequentially input on the main scanning line therebelow. Here, the pixel (n) to the left of the pixel of interest (n, m) to be binarized next is
The binarization is completed up to -1, m), and the error storage means 2 stores
The binarized pixel, that is, the pixel on the main scanning line one above the main scanning line of the pixel of interest (n, m) and the pixel of interest (n,
It is assumed that the error of the pixel (n-1, m) on the left side of (m) is stored.

Next, as a target pixel (n, m), for example, a pixel (12, 12) having a main scanning line and a twelfth sub-scanning line
When the multi-valued image signal S ₀ of 12) is input to the adding means 34, the main scanning direction counting means 30 causes the main scanning direction of the target pixel (n, m) at the time of inputting the multi-valued image signal S _0. A main scanning direction periodic signal S _m corresponding to whether the line is an even numbered line or an odd numbered line is output to the main scanning direction error selection means 31. In the case of the present embodiment, since the main scanning line of the target pixel (12, 12) is an even number, “0” is output as the main scanning direction periodic signal S _m . Further, the sub-scanning direction counting means 32 uses the target pixel (n,
The sub-scanning direction periodic signal S _n according to whether the sub-scanning line m) is an even-numbered line or an odd-numbered line is output to the sub-scanning direction error selecting means 33. In the case of the present embodiment, since the sub-scanning line of the target pixel (12, 12) is an even-numbered sub-scanning line, the sub-scanning direction periodic signal S _n
"0" is output as.

Then, the main scanning direction error selecting means 31 selects the corresponding error signal S _e from the error storing means 2 as the main scanning direction error based on the main scanning direction selecting signal S _m from the main scanning direction counting means 30. And outputs it to the adding means 34.
In the case of this embodiment, since the main scanning direction selection signal S _m is “0”, the pixel (1) at the upper left of the pixel of interest (12, 12) is selected.
The error signal S _e when binarizing (1, 11) is selected. Further, the sub-scanning direction selecting means 33 selects the error signal S _e from the error storing means 2 as the sub-scanning direction error based on the sub-scanning direction periodic signal S _n from the sub-scanning direction counting means 32 and adds it to the adding means 34. Output to. In the case of this embodiment, since the sub-scanning direction periodic signal S _n is “0”, the error signal S _e obtained when the pixel (12, 11) above the target pixel (12, 12) is binarized is select.

Next, the adding means 34 selects each of the selecting means 3 for the input multi-valued image signal S ₀ of the target pixel (12, 12).
And an error signal S _e of the error signal S _e and the sub-scanning direction error as the main scanning direction error from 1,33 to incremented by 1/2 to generate a corrected image signal S _c and the binarizing means 1 Output.

The binarizing means 1 binarizes the corrected image signal S _e from the adding means 34 with a predetermined threshold value and outputs it as a binary image signal S ₁ of "0" or "1", and also an error signal. S _e is generated and output to the error storage means 2. The error signal S _e is stored in the error storage means 2. The above operation is performed for other pixels (13, 12), (14, 12), (15, 1).
A binary image is obtained by sequentially repeating the steps 2), ....

Thus, for the pixel of interest (12, 12), the upper left pixel (11, 11) and the upper pixel (12, 12)
11) each error signal S _e is selected, and the target pixel (13,
For 12), the error signals S _e of the upper pixel (13, 11) and the left pixel (12, 12) are selected. Also,
Regarding the pixel of interest (12, 13), the upper left pixel (11,
12) and each error signal S _{e of the} upper right pixel (13, 12)
Is selected. Regarding the pixel of interest (13, 13), the left pixel (12, 13) and the upper right pixel (14, 1)
Each error signal S _{e in} 2) is selected. That is, the target pixel (12, 12), (13, 12), (12, 13),
As shown in (13, 13), the positions of the pixels from which the error signal S _e is extracted are different from each other within the size of 2 × 2, which is 2 × 2.
The position of the pixel from which the error signal S _e is taken out periodically changes for each size of.

The present invention is not limited to the above embodiment,
Modifications can be made without changing the gist of the invention. For example, in the present embodiment, the position of the pixel from which the error signal S _e is taken changes periodically for each 2 × 2 size, but the size is not limited and may be 3 × 3 or more. Further, the combination of pixels from which the error signal S _e is taken out is not limited and can be freely combined. For example, the pixel of interest (1
2, 12) the upper left pixel (11, 11) and the upper pixel (12, 11), and the target pixel (13, 12) the upper pixel (13, 11) and the left pixel (12, 1).
Although the error signal S _e is taken out from 2), the upper pixel (13, 11) and the left pixel (12, 12) and the target pixel (13, 12) are reversed with respect to the target pixel (12, 12). For 12), the positions of the pixels from which the error signal S _e is taken out may be changed like the upper left pixel (11, 11) and the upper pixel (12, 11).

[0034]

According to the present invention described in detail above, the following effects can be obtained.

According to the first aspect of the present invention, since the pixels to be error-added periodically change, the multi-valued image can be converted into a high-quality binary image with few false images.
A valuation method can be provided.

According to the second aspect of the present invention, since the position of the pixel from which the error signal is taken out is periodically changed, the multi-valued image can be converted into a high-quality binary image with less generation of constant false images. An image binarization device can be provided.

According to the third aspect of the present invention, the position of the pixel from which the error signal is extracted is periodically changed based on the periodically changing main scanning direction periodic signal and sub-scanning direction periodic signal. It is possible to provide an image binarization device capable of converting a high-quality binary image in which a certain false image is less likely to occur.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of the present invention.

FIG. 2 is a diagram for explaining the operation of this embodiment.

FIG. 3 is a diagram for explaining a conventional image binarization method.

[Explanation of symbols]

 1 Binarizing Means 2 Error Storage Means 3 Correcting Means 30 Main Scanning Direction Counting Means 31 Main Scanning Direction Error Selecting Means 32 Sub-scanning Direction Counting Means 33 Sub-scanning Direction Error Selecting Means 34 Adding Means

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G06F 15/68 320 A

Claims (3)

[Claims] 1. A multi-valued image, in which a pixel value of a pixel of interest to be binarized next is added with a predetermined ratio of an error when binarizing a pixel near the pixel of interest, and the pixel after the addition is added. Image 2 that obtains a binary image by the error diffusion method that binarizes the values
In the binarization method, the image binarization method is characterized in that the pixel for error addition is periodically changed according to the position of the pixel of interest to perform the addition. 2. An image binarization apparatus for obtaining a binary image by an error diffusion method of generating a corrected image signal based on an input multi-valued image signal and binarizing the corrected image signal, The signal is binarized with a predetermined threshold value, and the error when binarized is calculated and output as an error signal 2
Value conversion means, error storage means for storing the error signal,
Next, the multi-valued image signal of the pixel of interest to be binarized is read from the error storage means with an error signal when binarizing the pixels in the vicinity of the pixel of interest and added at a predetermined ratio to generate a corrected image signal. An image binarization device, further comprising: a correction unit that periodically changes a pixel from which an error signal is read according to a position of a pixel of interest to perform the addition. 3. The correction means includes a main scanning direction counting means for outputting a main scanning direction periodic signal which periodically changes each time binarization of a predetermined number of main scanning lines is completed, and the error storage means. A main-scanning-direction error selecting unit that selects an error signal corresponding to the main-scanning-direction periodic signal, and a sub-scanning-direction periodic signal that periodically changes each time binarization of a predetermined number of pixels in the sub-scanning direction is completed. A sub-scanning direction counting means for outputting,
A sub-scanning direction error selecting means for selecting an error signal corresponding to the sub-scanning direction periodic signal from the error storing means, and a multi-valued image signal of the pixel of interest for binarizing the error selected by the both selecting means next. The image binarization apparatus according to claim 2, further comprising: an addition unit configured to add at a predetermined ratio to generate the corrected image signal.