JPH06326864A - Picture processing method - Google Patents

Abstract

PURPOSE:To provide the picture processing method in which production of a unique stripe pattern in a picture after picture processing is prevented to improve the picture quality. CONSTITUTION:Multi-value processing picture data are binarized and picture data of multi-value processing density are divided into plural meshes each comprising a matrix of min picture elements to discriminate whether or not an edge by a black and white picture elements is in existence in a noted mesh (23). When any edge is in existence, binarized picture data are provided as an output as they are and when no edge is in existence, the binarized picture data are rearranged in the matrix (24). When picture processing is applied to a picture without any contrast such as a photograph, production of unique pattern to a picture after picture processing is prevented to improve the picture quality.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method.

[0002]

2. Description of the Related Art Conventionally, in an image processing method applied to a digital facsimile, an error diffusion method has been used as a binarization method for representing halftone data consisting of a plurality of gradations in a pseudo manner. Floyd and Steinberg, "An Ada
ptive Algorithm for SpatialGreyscale "Proceedings
of the SID vol 17/2 Second Quarter 1976, pp75-7
7 (See SID: SOCIETY for IMFORMATION DISPLAY)).

FIG. 2 is a block diagram showing an error diffusion method binarization circuit using a conventional error diffusion method. In the figure,
Reference numeral 11 is an adder, 12 is a comparator, 13 is an error diffusion binarization circuit including the adder 11 and the comparator 12, and I (x,
y) is a multivalued density value as an input signal at the address (x, y), P (x, y) is a binary density value as an output signal at the address (x, y), and E (x, y) Is an error in the address (x, y), and TH is a fixed threshold value.

In the error diffusion binarization circuit 13,
The halftone data to be binarized is the address (x, y)
Is represented by a multivalued density value I (x, y). The density value I (x, y) is input to the error diffusion binarization circuit 13 as an input signal, image-processed and binarized to become the density value P (x, y). Then, the density value P (x, y)
Is output from the error diffusion binarization circuit 13 as an output signal.

A pixel having the density value I (x, y) is set to a density value P.
When expressed by the pixel of (x, y), the pixel of the density value P (x, y) is too bright or too dark as compared with the original pixel. Therefore, the difference between the density value I (x, y) and the density value P (x, y) is calculated as an error E (x, y). E (x, y) = I (x, y) -P (x, y) Then, the adder 11 weights and distributes the density of the error E (x, y), and distributes the address (x, y). ) Is added to the density value I (x, y) of the pixels around (). Next, in the peripheral pixels, the signal from the adder 11 is binarized by the fixed threshold TH.

The weighting of the density corresponding to the error E (x, y) can be expressed by the matrix shown in the equation a.

[0007]

[Equation 1]

In the above formula a, * indicates the pixel. In this case, among the pixels adjacent to the pixel, the density value I of all pixels printed after the pixel is printed.
The error E (x, y) is distributed to (x, y). That is, when the address of the pixel is (x, y) and the error at the address (x, y) is E (x, y), the address (x + 1, y) of the pixels adjacent to the pixel is
The density value I (x + 1, y) of the pixel of A = (7/16) · E (x, y) is added to obtain the density value I (x of the pixel of the address (x-1, y + 1). The density of B = (3/16) · E (x, y) is added to −1, y + 1), and C = is added to the density value I (x, y + 1) of the pixel at the address (x, y + 1). The densities of (5/16) · E (x, y) are added, and D = (1/16) · (x, y) for the density value I (x + 1, y + 1) of the pixel at the address (x + 1, y + 1). ) Is added.

Then, the above process is repeated every time the signal from the adder 11 is binarized by the fixed threshold value TH.

[0010]

However, in the conventional image processing method, the density value I (x, y) of all the pixels that are printed after the pixel is printed out of the pixels adjacent to the pixel. Since the error E (x, y) is distributed to each of the images and the distribution method is uniform, when image processing is performed on an image having no shading, such as a photograph, a unique image is obtained. A stripe pattern is generated and the image quality is degraded.

The present invention solves the problems of the conventional image processing method, prevents the occurrence of a unique stripe pattern in the image after the image processing, and improves the image quality. The purpose is to provide.

[0012]

Therefore, in the image processing method of the present invention, the image data of multi-valued density values is converted into two.
The image data having multivalued density values is binarized and divided into a plurality of meshes composed of a matrix of m × n pixels, and it is determined whether or not edges of black pixels and white pixels exist in the mesh of interest.

If the edge exists, the binarized image data is output as it is, and if the edge does not exist, the binarized image data is rearranged in the matrix.

[0014]

According to the present invention, as described above, in the image processing method of the present invention, the binarizing circuit is provided to binarize the image data of multi-valued density values and the edge determining means. Then, the multi-valued image data of the density value is divided into a plurality of meshes composed of a matrix of m × n pixels, and it is determined whether or not edges of black pixels and white pixels exist in the mesh of interest.

If an edge exists, the binarized image data is output as it is, and if an edge does not exist, the binarized image data is rearranged in the matrix by the rearrangement means. .

[0016]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 1 is a block diagram of an image processing apparatus to which an image processing method according to an embodiment of the present invention is applied. In the figure, 13 is an error diffusion binarization circuit, and 21 is an input means of the error diffusion binarization circuit 13. The input means 21 is provided corresponding to a scanner (not shown) for reading a printed image, an A / D converter (not shown) for converting analog data read by the scanner into digital data, and each pixel in the line direction. It comprises a correction circuit (not shown) for correcting variations in the image sensor (not shown). The input unit 21 outputs a multivalued density value I (x, y) corresponding to the address (x, y) in pixel units to the edge determination unit 23 and the error diffusion binarization circuit 13.

In the present embodiment, the density value I (x, y) has 64 gradations and is composed of values from 0 to 63. The density value I (x, y) is binarized in the error diffusion method binarization circuit 13 to be a binary density value P (x, y) which is output to the rearrangement means 24. The error diffusion binarization circuit 13 includes an adder 11 (see FIG. 2) and a comparator 12.

In the error diffusion binarization circuit 13, the pixel having the density value I (x, y) is converted into the density value P (x,
In terms of the pixel y), the pixel having the density value P (x, y) is too bright or too dark as compared with the original pixel. Therefore, the difference between the density value I (x, y) and the density value P (x, y) is calculated by the error E.
Calculate as (x, y). E (x, y) = I (x, y) -P (x, y) Then, the adder 11 weights and distributes the density of the error E (x, y), and distributes the address (x, y). ) Is added to the density value I (x, y) of the pixels around (). Next, in the pixels around the address (x, y), the signal from the adder 11 is binarized by the fixed threshold TH.

The processing is repeated every time the signal from the adder 11 is binarized by the fixed threshold value TH. In this embodiment, the error diffusion method by the error diffusion method binarization circuit 13 is used as a means for binarizing the multivalued density value I (x, y). Random dither methods such as the minimum mean error method can also be used.

By the way, the error diffusion binarization circuit 13
When image processing is performed on an image having no light and shade, such as a photograph, a unique stripe pattern occurs in the image after the image processing. Therefore, the edge determination means 23 divides the image data into a plurality of meshes each consisting of a matrix of m × n pixels, and determines whether or not edges of black pixels and white pixels exist in the mesh of interest. If an edge is present, the image is a shaded image, and a stripe pattern does not occur even if the image processing by the error diffusion method is performed.
Therefore, the rearrangement means 24 uses the error diffusion binarization circuit 1
The density value P (x, y) from 3 is output to the output means 25 as it is.

If there is no edge in the mesh of interest, a stripe pattern is generated when the image processing by the error diffusion method is performed. Therefore, the error diffusion binarization circuit 13
The density values P (x, y) from the above are rearranged by the rearrangement means 24, and the density values P (x, y) after rearrangement are output to the output means 25. In this embodiment, the edge determination means 23 divides the mesh into a plurality of meshes each having a matrix of 3 × 3 pixels. Then, within the mesh, the density value I
The difference between the maximum value M _MAX of (x, y) and the minimum value M _MIN of the density value I (x, y) is calculated, and when the difference is larger than the set edge determination threshold E _TH, an edge is present in the mesh. Is determined to exist.

M _MAX −M _MIN > E _TH Then, when there is no edge in the mesh of interest, a striped pattern occurs. Therefore, rearrangement means 24 is provided to rearrange the density value P (x, y) from the error diffusion binarization circuit 13. FIG. 3 is a diagram showing a rearrangement order in the embodiment of the present invention.

In this case, the rearrangement means 24 (FIG. 1) is 3 ×.
Image data is rearranged for each mesh formed of a matrix of 3 pixels. That is, the rearrangement unit 24 divides the image data having the density values P (x, y) binarized by the error diffusion method into a plurality of meshes each having a matrix of 3 × 3 pixels, and black in each mesh. The pixels are read out, and the read black pixels are written in the order shown by "1" to "9" in the figure.

Then, the rearranged image data is sequentially output to the output means 25. The output means 25 is
A printer, a display, an encoding device or the like (not shown) outputs image data as an image or a code. Next, the operation of the rearrangement means 24 will be described with reference to FIGS. FIG. 4 is a diagram showing an output example of image data from the error diffusion binarization circuit in the embodiment of the present invention, and FIG. 5 is a diagram showing an output example of image data from the rearrangement means in the embodiment of the present invention. is there. In the figure, the image data in which the density is uniform and the density value I (x, y) is fixed to “20 _{(H )} ” is binarized, and “1” is a black pixel,
"0" corresponds to a white pixel.

In the figure, numerals 31 and 32 are meshes composed of a matrix of 3 × 3 pixels. In this case, when the density value I (x, y) is input from the input means 21, the edge determination means 23 makes the image data of the density value I (x, y) a plurality of 3 × 3 pixel matrixes. Mesh 31,
Divide into 32. Then, within the meshes 31 and 32, the maximum value M _MAX of the density value I (x, y) and the density value I (x, y)
Minimum value M of the difference between _MIN calculated, the mesh if the difference is greater than the set edge determination threshold E _TH 31 and 32 that the
It is determined that there is an edge inside. In this case, it is assumed that it is determined that the edge does not exist in the meshes 31 and 32.

On the other hand, when the image data is input from the error diffusion binarization circuit 13, the rearrangement means 24 divides the image data into a plurality of meshes 31 and 32 formed of a matrix of 3 × 3 pixels. , The black pixels in each mesh 31, 32 are read and the number of black pixels is counted. In this case, there are three black pixels in the mesh 31 as shown in FIG.
Since there are individual pixels, black pixels are written in the order shown by "1" to "3" in FIG. By doing so, the image data as shown in FIG. 5 is obtained.

Since there are two black pixels in the mesh 32 as shown in FIG. 4, the black pixels are written in the order shown by "1" and "2" in FIG. By doing so, the image data as shown in FIG. 5 is obtained. It should be noted that the present invention is not limited to the above-described embodiments, and various modifications can be made based on the spirit of the present invention, and these are not excluded from the scope of the present invention.

[0028]

As described above in detail, according to the image processing method of the present invention, multivalued image data of density values are binarized and image data of multivalued density values is m × n. It is divided into a plurality of meshes each composed of a matrix of pixels, and it is determined whether or not edges of black pixels and white pixels are present in the mesh of interest.

If the edge exists, the binarized image data is output as it is, and if the edge does not exist, the binarized image data is rearranged in the matrix. Therefore, when image processing is performed on a non-shading image such as a photograph, it is possible to prevent a unique stripe pattern from occurring in the image after the image processing, and improve the image quality.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing an error diffusion method binarization circuit using a conventional error diffusion method.

FIG. 3 is a diagram showing a rearrangement order in the embodiment of the present invention.

FIG. 4 is a diagram showing an output example of image data from the error diffusion binarization circuit in the embodiment of the present invention.

FIG. 5 is a diagram showing an output example of image data from a rearrangement unit in the embodiment of the present invention.

[Explanation of symbols]

 31,32 mesh I (x, y), P (x, y) concentration value

Claims (1)

[Claims] 1. (a) Binary image data of multi-valued density values, (b) Divide image data of multi-valued density values into a plurality of meshes composed of a matrix of m × n pixels, (c) ) It is determined whether or not there are edges due to black pixels and white pixels in the mesh of interest. (D) If an edge exists, the binarized image data is output as it is, and (e) an edge exists. If not, the image processing method is characterized by rearranging the binarized image data in the matrix.