JPH08116443A - Image processing unit - Google Patents

Abstract

PURPOSE: To prevent appearance of a boundary in an area of the same input density in the image processing unit binarizing a halftone image to generate image data of pseudo halftone. CONSTITUTION: An image processing circuit 1 is made up of an original image storage device 3 storing part of halftone original image information, a RAM 5 storing each calculation result, plural weight coefficient matrices M1, M2,... a ROM 7 storing a threshold level T or the like, a CPU 9 to conduct various image processing, an output image storage device 11 storing data after error diffusion processing, and a random number generating circuit 12. The halftone image data are binarized while being subjected to binarized error distribution through the use of a weight coefficient matrix determined by random numbers generated by the random number generating circuit 12 for respective picture elements.

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, and more particularly, to a digital electrophotographic copying machine having a function of binarizing a halftone image to generate pseudo halftone image data. The present invention relates to an image processing apparatus in a machine, a thermal transfer type or an inkjet type printer, or the like.

[0002]

2. Description of the Related Art Conventionally, as a means for binarizing a halftone image to create a pseudo halftone image, a systematic dither method for binarizing the image according to a threshold matrix (dither matrix) table, etc. The various dither methods are widely used. However, these conventional methods have a contradiction that the matrix table needs to be large in order to improve the gradation reproducibility, and the matrix table needs to be small in order to obtain high resolution. It was difficult to achieve both reproducibility and high resolution.

In addition to the above, there is an error diffusion method as a method for achieving both gradation reproducibility and high resolution, which is given a relatively good evaluation among various conventional methods. The error diffusion method distributes the error between the actual density and the binarized value, which occurs when binarizing a halftone pixel, to the peripheral pixels, corrects the density, and This is a process of making a determination based on a predetermined threshold value and determining one of two values.

In the following description, the input density of each pixel takes an integer value from 0 to 255, and the output density is 0 or 2.
Let's say it is 55. In the uniform density area A where the input density is the same throughout, the distribution of the output density is the same throughout the binarization process. For example, consider a uniform density region B having an input density of 128. In area B, the output density is 0
The ratio of pixels with output density of 255 is almost 5 over the entire area.
It will be 0% each.

[0005]

However, in this way,
Even in an area where the output values of various types are evenly distributed over the entire area, the binarization processing by the error diffusion method changes the arrangement pattern of the two output values due to the effect of the binarization error distributed from the peripheral pixels. It may happen.

For example, in the uniform density region B,
A pattern B01 in which pixels having an output density of 0 (represented by white pixels) and pixels having an output density of 255 (represented by black pixels) appear alternately as shown in FIG. 4A, and as shown in FIG. There is a pattern B02 in which pixels having an output density of 0 and pixels having an output density of 255 are randomly arranged. Therefore, in the conventional error diffusion method, even though the area has a uniform input density,
There is a problem that a pseudo boundary is generated where the pattern described above changes.

The present invention has been made in order to solve the above-mentioned problems, and it is possible to create a more preferable pseudo-halftone image in which a boundary does not occur regardless of the distributed binarization error. An object is to provide an image processing device.

[0008]

According to the first aspect of the present invention,
In an image processing device for binarizing a halftone image to create pseudo halftone image data, a binarization processing means for binarizing the density of each pixel of the halftone image by an error diffusion method,
An image including a coefficient matrix setting unit that randomly selects a plurality of weighting coefficient matrices provided for each pixel and sets the weighting coefficient matrix for error distribution in the error diffusion processing. It is a processing device.

According to a second aspect of the present invention, the binarization processing means corrects the density of each pixel of the halftone image by the sum of binarization errors distributed from peripheral pixels to obtain a corrected input density. Input density calculation means, output density determination means for comparing the corrected input density obtained by the corrected input density calculation means with a threshold value to determine the output density of each pixel, and the output determined by the output density determination means. Based on the weighting coefficient matrix, the binarization error computing means for computing the binarization error generated in each pixel from the density and the correction input density calculated by the correction input density computing means, and the binarization error based on the weighting coefficient matrix. The image processing apparatus according to claim 1, further comprising: a binarization error distribution unit that weights and distributes to peripheral pixels.

According to a third aspect of the present invention, the coefficient matrix setting means selects a random number generating means and a plurality of weighting coefficient matrices provided for each pixel on the basis of the random number generated by the random number generating means. The image processing apparatus according to claim 1 or 2, further comprising: a weighting coefficient matrix selecting unit which is set as a weighting coefficient matrix for error distribution in the error diffusion processing.

[0011]

In the image processing apparatus according to the first aspect, the binarization processing means binarizes the density of each pixel of the halftone image by the error diffusion method. On the other hand, the coefficient matrix setting means randomly selects a plurality of weighting coefficient matrices provided for each pixel, and sets the weighting coefficient matrix for error distribution in the error diffusion processing.

By randomly setting different weighting coefficient matrices, pixels having an output density of 0 and output density of 25 in the output image binarized by the error diffusion method.
5 pixels will be randomly arranged. Therefore, regardless of the influence of the distributed binarization error, it is possible to prevent a boundary from occurring in a region having a uniform input density.

The binarization processing means corrects the density of each pixel of the halftone image by the sum of the binarization errors distributed from the peripheral pixels by the corrected input density calculation means to obtain the corrected input density, The output density determining means compares the corrected input density calculated by the corrected input density calculating means with a threshold value to determine the output density of each pixel, and the binarization error calculating means is determined by the output density determining means. The binarization error generated in each pixel is calculated from the output density and the corrected input density calculated by the corrected input density calculation means, and the binarized error distribution means calculates the binarized error based on the weighting coefficient matrix. It is realized by weighting and distributing to peripheral images.

The coefficient matrix setting means is, for example,
Weighting coefficient matrix selection means, for each pixel,
It is realized by selecting from a plurality of weighting coefficient matrices provided based on the random number generated by the random number generating means and setting it as a weighting coefficient matrix for error distribution of the error diffusion processing.

[0015]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the image processing circuit 1. The image processing circuit 1 that performs error diffusion processing and the like in this embodiment is configured as a computer,
The input data is subjected to required arithmetic processing. The image processing circuit 1 includes an original image storage device (image memory) 3 for storing a part of halftone original image information input by a digital electric signal, a RAM 5 for storing each calculation result, an image processing program and an image processing program. A plurality of weighting coefficient matrices M1, M2, ... For the distribution of the binarization error of the peripheral pixels, and the threshold value T
And the like, a CPU 9 for performing various kinds of image processing using the RAM 5, an output image storage device (image memory) 11 for storing data after error diffusion processing, and a random number generation circuit 12. There is.

Halftone image data of an original image input from an image inputting means (not shown) such as an image pickup device or an external storage device for storing image data already taken, is stored in the original image storage device 3 via the bus 13. Memorized in. The original image data may be input pixel by pixel, line by line, or sent for one screen. The CPU 9 performs error diffusion processing on the stored halftone image information using the image processing program in the ROM 7 and the data such as the matrix.

A procedure for determining a corrected input density and a matrix in the error diffusion process to obtain an output value will be described below with reference to FIGS. The number of gradations of the original image data of halftone is 256 gradations of 0 to 255. Figure 2
3 is a flowchart showing a processing procedure of binarization processing in the present embodiment.

First, the input density Ix, y of the first pixel of interest is retrieved from the original image storage device 3 in which the original image data input by the image input means is stored (S10). Further, the threshold value T is retrieved from the ROM 7 (S15). Next, Equation 1
As described above, the corrected input density I′x, y is calculated by adding the weighting error sum E of the binarization errors in the peripheral pixels distributed from the peripheral pixel to the target pixel to the input density Ix, y (S2).
0).

[0019]

[Equation 1]

This error sum E is obtained in step S11, which will be described later, in the pixel processed before the current pixel of interest is processed.
It is a value already accumulated and accumulated in the binarization error storage buffer of each pixel set in the RAM 5 due to the binarization error distribution of 0.

Next, the threshold value T and the corrected input density I'x, y
(S50) and the binary output density Ie is determined (S60, S70). When I′x, y ≧ T, the output density Ie becomes “255” which is one of the two values (S60), and when I′x, y <T, the output density Ie is the other of the two values. Is set to "0" (S70). The output density Ie thus determined is stored in the output image storage device 11 (S80).

Next, the binarization error e generated in the target pixel is obtained (S90). The binarization error e is obtained by subtracting the output density Ie from the corrected input density I'x, y as shown in the equation 2.

[0023]

[Equation 2]

Next, the random number generating circuit 12 is driven to generate a random number R within a predetermined range (S95). The random number R output from the random number generation circuit 12 in this embodiment is a ROM
Multiple weighting coefficient matrices M1 and M existing in 7
It is obtained in order to randomly select one weighting coefficient matrix from 2, ... Various methods are known for randomly selecting, but for example, if the weighting coefficient matrices M1, M2, ... Are two, the random number R is a value that can take 0 or 1 at random,
When R = 0, the weighting coefficient matrix M1 may be used for distribution of the binarization error, and when R = 1, the weighting coefficient matrix M2 may be used for distribution of the binarization error.
Alternatively, the random number R is a value that can take an integer randomly, and when the random number R is an odd number, the weighting coefficient matrix M1 is used for distribution of the binarization error, and when the random number R is even, the weighting coefficient matrix M2 is used for distribution of the binarization error. You may use it.

Of course, even when the weighting coefficient matrix is 3 or more, the output value of the random number R and the calculation based on the output value are used to randomly generate the numerical values corresponding to the respective weighting coefficient matrices, and based on the numerical values. It is possible to randomly select one weighting coefficient matrix from the three or more weighting coefficient matrices.

Next, the weighting coefficient matrix is determined according to the generated random number R, and the corresponding weighting coefficient matrix is retrieved from the ROM 7 (S100). Using this weighting coefficient matrix, the binarization error generated in the target pixel is distributed to the peripheral pixels (S110).

For example, it is assumed that two weighting coefficient matrices M1 and M2 are stored in the ROM 7 as shown in equations 3 and 4.

[0028]

(Equation 3)

When it is determined that the weighting coefficient matrix M1 is selected for the pixel of interest this time by the value of the random number R and is used for binarization error distribution, that is,
The coefficient value having the positional relationship as shown in FIG. 3A with respect to the target pixel represented by “*” is obtained from the ratio of each component of the weighting coefficient matrix M1 and the sum of all components, and the target pixel is obtained. The binarization error e of 1 is multiplied and weighted, and added to the buffer for storing the binarization error of the pixel at the corresponding position.

For example, for the pixel to the right of the pixel of interest, the weighting factor is "5/21 (the component to the right of" * "in the matrix M1 of equation 3 is 5: the sum of all components: 2"
1) ”is obtained, and as a result,“ e × 5/21 ”is added to the buffer of the pixel on the right.

When it is determined that the weighting coefficient matrix M2 of the equation 4 is selected for the pixel of interest this time by the value of the random number R and is used for binarization error distribution,
That is, the coefficient value having the positional relationship as shown in FIG. 3B with respect to the pixel of interest represented by “*” is obtained from the ratio of each component of the weighting coefficient matrix M2 and the sum of all components,
The binarization error e of the pixel of interest is multiplied and weighted, and added to the buffer for storing the binarization error of the pixel at the corresponding position.

In this way, the processing of the pixel at the position of the coordinates (x, y) is completed, and it is determined whether or not the next unprocessed pixel exists (S120). The pixel at the coordinate is set as the pixel of interest (S130), and the above-described processing is repeated. In the processing order of the pixels in this embodiment, the pixels on the left of the uppermost line are processed horizontally to the right, and the same processing is sequentially moved downward by one line to finally process the pixels on the lower right. Is the order.

When the above-mentioned processing is executed for all the pixels, a negative determination is made in step S120, and the binarization processing ends. In this embodiment, as described above,
In the error diffusion method, a weighting coefficient matrix for binarizing error distribution is randomly determined from a plurality of prepared weighting coefficient matrices by random number generation. As described above, since the same weighting coefficient matrix is not always used for each pixel and a plurality of weighting coefficient matrices are not used even in the same repetition, the output image binarized by the error diffusion method is used. Pixels having an output density of 0 and pixels having an output density of 255 are randomly arranged. Therefore, regardless of the influence of the distributed binarization error, it is possible to prevent a boundary from occurring in a region having a uniform input density.

In the above embodiment, step S20 corresponds to the process as the corrected input density calculating means,
50, S60, and S70 correspond to the process as the output density determining unit, step S90 corresponds to the process as the binarizing error calculating unit, step S110 corresponds to the process as the binarizing error distributing unit, and the random number The generation circuit 12 corresponds to the random number generation means, and step S100 corresponds to the processing as the weighting coefficient matrix selection means.

[Others] In the above embodiment, the weighting coefficient matrix has two types having different matrix shapes,
That is, although 2 rows and 5 columns and 3 rows and 3 columns are used, shapes other than these may be combined. Further, even if the shape is the same, two or more matrices having different values of each component may be provided, and the matrix may be randomly determined from among them.

Further, although the random number is generated by the circuit, C
You may obtain a random number by the program process by PU9. In addition to the configuration in which the weighting coefficient matrix of the above-described embodiment is randomly selected for each pixel, or instead of the configuration, a threshold value is randomly calculated and set for each pixel, or a plurality of threshold values are randomly determined for each pixel. The pattern may be randomized by selection.

[Brief description of drawings]

FIG. 1 is a block diagram illustrating an image processing circuit according to an embodiment.

FIG. 2 is a flowchart of the binarization process.

FIG. 3 is an explanatory diagram showing a relationship between a target pixel and peripheral pixels in the error distribution process.

FIG. 4 is an explanatory diagram showing a difference in a pixel distribution pattern after binarization processing by an error diffusion method in a uniform uniform density region.

[Explanation of symbols]

1 ... Image processing circuit 3 ... Original image storage device 5 ...
RAM 7 ... ROM 9 ... CPU 11 ... Output image storage device 12 ... Random number generation circuit 13 ... Bus

Claims (3)

[Claims] 1. An image processing apparatus for binarizing a halftone image to create pseudo halftone image data, wherein a binarization processing means binarizes the density of each pixel of the halftone image by an error diffusion method. And a coefficient matrix setting means for randomly selecting from a plurality of weighting coefficient matrices provided for each of the pixels and setting it as a weighting coefficient matrix for error distribution in the error diffusion processing. Image processing device. 2. The binarization processing means corrects the input density by correcting the density of each pixel of the halftone image based on the sum of the binarization errors distributed from the surrounding pixels to obtain a corrected input density. Output density determination means for comparing the corrected input density calculated by the corrected input density calculation means with a threshold value to determine the output density of each pixel, and the output density determined by the output density determination means and the corrected input density Binarization error calculation means for calculating a binarization error generated in each pixel from the corrected input density obtained by the calculation means, and the binarization error is weighted based on the weighting coefficient matrix to the peripheral pixels. The image processing apparatus according to claim 1, further comprising: a binarization error distribution unit for distributing. 3. The coefficient matrix setting means selects a random number generating means and a plurality of weighting coefficient matrices provided for each pixel based on the random number generated by the random number generating means, and the error diffusion is performed. The image processing apparatus according to claim 1 or 2, further comprising: a weighting coefficient matrix selecting unit which is set as a weighting coefficient matrix for processing error distribution.