JPH0423566A - Picture processing unit - Google Patents

Abstract

PURPOSE:To decrease notch of a picture with simple processing and hardware constitution by replacing a picture element density of an edge part into a density larger in the white level or black level direction and binarizing the picture with error dispersion processing. CONSTITUTION:The processing unit consists of a picture read section 1, an A/D converter section 2, an edge picture element separation section 3, an edge picture element density correction section 4, an error dispersion processing section 5, and a picture output section 6. While picture elements not for the edge among multi-value picture information are converted into binary picture information by the error dispersion method as they are, the density of white level picture elements of the edge is corrected into a higher white level and the density of black level picture elements of the edge is corrected into a higher black level respectively and the result is converted into binary picture information by the error dispersion method. Thus, white/black level inversion of picture elements at the edge is reduced by simple processing and hardware constitution and the notch of the picture is decreased.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an image processing device that converts multilevel image information obtained by reading an image into binary image information using an error diffusion method.

[Prior Art] As a method of converting halftone image information into binary image information,
The error diffusion method, which is a random dither method, is well known.

Now, for example, when one image is read at a density of O to 15 (white: o, black: 15), image information J (X, Y) of each coordinate X, Y as shown in FIG. 3(a) is obtained. Suppose that we obtain

In the error diffusion method, multilevel image information J(X.

■) When converting the image information I (X, Y) into binary image information I (X, Y), the feedback error e (X, Y) that is calculated when converting the surrounding pixels of one pixel (X, Y) and the threshold value TH, when J (X, V) + e (X, Y)≧TH, I (X, Y) is black, J (x, ■) + e (x + v)
<When TH, I(X, Y) is white.

Now, it is assumed that the image information 1(X, Y) in FIG. In this case, if the feedback error e (X, Y) = 0, the densities 0 to 7 of each pixel of the image information J (X, '/) are converted to white, the densities 8 to 15 are converted to black, and the same The black-and-white edges of the drawing image never fade.

However, for example, if the feedback error e(X, Y)=-], the pixel with density 8 is converted to white, so the image information I(3, 2) becomes white, as shown in FIG. Become.

As described above, conventionally, when the pixel density is close to the threshold value at the black and white edge portions of an image, the black and white are inverted, causing a notch in the binarized image.

In order to improve this inconvenience, for example, Japanese Patent Application Laid-Open No. 62139
As seen in Publication No. 472, the run length of black and white pixels of one page of images is measured, and the feedback error e is determined by the run length.
A method has been proposed in which the feedback rate for calculating (X, Y) is changed.

This proposal is characterized in that the run length of white pixels or black pixels becomes long at the edge portion of an image, so the feedback error e(X, Y) is adjusted by the run length.

[Problems to be Solved by the Invention] However, there are problems in that the process of adjusting the feedback rate by the run length of black and white pixels as described above is difficult and the hardware configuration becomes complicated.

SUMMARY OF THE INVENTION An object of the present invention is to provide an image processing apparatus that can solve the above problems and reduce image notches with simple processing and hardware configuration.

[Means for Solving the Problems] To this end, in the first invention, pixels in the edge portion that is the boundary between the white area and the black area in one page of the image are determined, and the pixels in the edge portion of the multilevel image information are identified. Pixels that are not at the edges are converted directly into binary image information using the error diffusion method, while white pixels at the edges are converted into constant image information with a density that is whiter than other white pixels, regardless of the pixel density. , the black pixels in the edge area are each replaced with constant image information that has a blacker density than other black pixels, and then the image information in the edge area is converted into binary image information using the error diffusion method. I have to.

In addition, in the second invention, instead of replacing the image information in the "-2 process, the white pixel density of the edge part is further moved to the white side by a certain value, and the black pixel density is further moved by a certain value to the black side. After correction, it is binarized.

[Operation] Since the difference between the density levels of white pixels and black pixels in the edge portion and the threshold value at the time of binarization becomes larger, black and white inversion is reduced, and notches in the image are reduced.

Further, the processing for determining edges within an image is well known as a simple technique, and the processing for replacing pixel data at edge portions and correcting density is also simple. This results in
Notches can be reduced with simple processing and hardware configuration.

[Embodiments] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 shows a block diagram of an image processing apparatus according to an embodiment of the present invention. In the figure, an image reading section 1 reads a document image using a line image sensor to obtain an analog image signal, and an A/D converting section 2
The method converts the image signal into a digital signal to obtain multilevel image information.

The edge pixel separation unit 3 separates pixels at an edge portion, which is a boundary between a white area and a black area, from other pixels in a read image of one page using a known technique. The edge pixel density correction section 4 corrects the density of pixels in the edge portion.

The error diffusion processing section 5 processes the image information other than the edge portion output from the edge pixel separation section 3 and the edge pixel density correction section 4.
The image information of the edge portions whose density has been corrected in the above are each sensitized by the error diffusion method. The image output unit 6 outputs the binarized image information as an image by displaying it on a screen or recording it on recording paper.

Assume now that the image reading section 1 reads one page of the original image using the above configuration. Then, both signals outputted from the image reading section 1 are converted into multivalued image information by the A/D converting section 2 and inputted to the edge pixel separating section 3.

Now, for example,
Assume that image information of each coordinate x, Y (X, Y) is input to the edge pixel separation unit 3 as shown in FIG. 2(a).

The edge pixel separation unit 3 examines the density difference between two adjacent pixels in the image information J (X, Y), and determines two pixels whose density difference is larger than a preset value as an edge portion.

As a result, in the case of FIG. 3A, each pixel at the coordinates of X2 and X=3 is determined to be a pixel in the edge portion.

The edge pixel separation unit 3 collects image information J (
X, Y) is output to the edge pixel density correction section 4, and image information J (X, Y) other than the edge portion is output to the error diffusion processing section 5.

The edge pixel density correction unit 4 distinguishes the input image information J(X, Y) of the edge portion into black and white. Then, the black image information J(X, Y) is replaced with a density of 15+α□ which is higher than the maximum density value of 15 by a certain value.

Further, the white image information J (X, Y) is replaced with a density O-α2 that is lower than the lowest density value 0 by a certain value. The image information replaced in this way, +(X, Y), is output to the error diffusion processing section 5.

As a result, image information J (X, Y) as shown in FIG. 2(b) is input to the error diffusion processing section 5 from the edge pixel separation section 3 and the error diffusion processing section 5.

The error diffusion processing unit 5 converts this multilevel image information J(X, Y) into binary image information I(X,'/) using a predetermined error diffusion method.

That is, as explained in FIG. 3, the feedback error e(X, Y) that occurs when binarizing surrounding pixels and the threshold value
Using H, J(X, Y)+e(X, Y)≧TI(+71, ■(
X, V) are black, and when J(X, Y)+e(X, Y)<T+, I(x, Y) is white.

In this case, the threshold value TH is set to a value of TH:8, for example, but the difference between this threshold value TH and the pixel density of the edge portion is large. As a result, the feedback error e(X, Y
) increases to some extent in the positive or negative direction, binary image information I(X, Y) is obtained in which the pixels at the edge are not inverted in black and white, as shown in FIG. 2(c).

By the way, in this case, multilevel image information J (X
, Y), the resulting binarization error is E (X, Y), then this error (X, Y) becomes the feedback error e(X, Y) of the surrounding pixels. Affect.

In this embodiment, since one black pixel is further subjected to density correction in the black direction, the generated error ε(X, Y) takes a positive value. In this case, the feedback error e(X, Y) when binarizing the image information J(X, Y) of the surrounding pixels increases in the positive direction.

When the feedback error e(X, Y) increases, the image information I(X, Y
) has a higher probability of becoming black.

On the other hand, since the density of white pixels is further corrected in the white direction, the error t (X, Y) becomes negative, the feedback error e (X, Y) for the surrounding pixels decreases, and the density of the surrounding pixels decreases. Information I (X, '/) has a high probability of becoming white.

This makes it easier for the same white or black pixels to be continuous, increasing the notch reduction effect.

The image information I(X, Y) thus binarized is displayed or recorded by the image output section 6.

As described above, in this embodiment, the pixel density at the edge is replaced with a larger density value in the white direction or black direction, and then binarized by error diffusion processing, which reduces black and white inversion of pixels. , the notches in the output image are reduced.

Further, there is a known technology that allows the edge pixel separation section 3 to be easily constructed, and the processing of the edge pixel density correction section 4 is also simple. Therefore, the notches can be reduced with simple processing and hardware configuration.

By the way, in the above embodiment, the image information J(X,
Y) is replaced with data of a constant density of white or black regardless of the original density, but as shown in Figure 2(d), in the case of a black pixel, the original density is replaced. On the other hand, it is further corrected towards black by a constant value β□, and in the case of a white pixel, a constant value β2
However, the correction may be further made in the white direction.

In this case, the density of the original image can be reflected in the output image.

[Effects of the Invention] As described above, according to the present invention, pixels that are not in the edge part of the multilevel image information are directly converted into z-value image information by the error diffusion method, while white pixels in the edge part are converted into z-value image information. After correcting the density further toward white and black pixels further toward black, the image information is converted to binary image information using the error diffusion method. Therefore, with simple processing and hardware configuration, black and white pixels at edges can be corrected. Inversion can be reduced, and image notches can be reduced.

[Brief explanation of the drawing]

FIG. 1 is a block configuration diagram of an image processing apparatus according to an embodiment of the present invention, FIG. 2 is an explanatory diagram showing a procedure for processing image information, and FIG. 3 is an explanatory diagram showing a state in which a notch occurs in conventional processing. It is a diagram. 1... Image reading section, 2... A/D conversion section, 3...
Edge pixel separation section, 4... Edge pixel density correction section, 5
. . . error diffusion processing unit, 6 . . . image output unit. Agent Patent attorney Crest 1) Makoto 2nd figure 3rd figure χ −→− ↓

Claims (2)

[Claims] (1) In an image processing device that converts multilevel image information obtained by reading an image into binary image information using an error diffusion method, pixels at the edge, which is the boundary between the white area and the black area, in one page of the image are edge pixel discriminating means for discriminating; non-edge portion binarization means for converting pixels that are not at the inner edge of the multi-level image information into binary image information by an error diffusion method; A white pixel density correction means replaces a white pixel with constant image information having a whiter density than other white pixels, and a white pixel density correction means replaces a white pixel with constant image information having a density that is even whiter than other white pixels, and a white pixel density correction means that replaces a black pixel at the inner edge part of the multivalued image information with a whiter density than other black pixels. A black pixel density correction means replaces the image information with constant image information having a side density; An image processing device comprising: edge portion binarization means for converting into value image information. (2) In an image processing device that converts multivalued image information obtained by reading an image into binary image information using an error diffusion method, pixels at the edge, which is the boundary between the white area and the black area, in one page of the image are edge pixel discriminating means for discriminating; non-edge portion binarization means for converting pixels that are not at the inner edge of the multi-level image information into binary image information by an error diffusion method; A white pixel density correction means for further correcting the density of the white pixel toward the white side by a certain value; and a black pixel density correction means for further correcting the density of the black pixel at the inner edge portion of the multivalued image information toward the black side by a certain value. , edge portion binarization means for converting the image information corrected by the black pixel density correction means and the white pixel density correction means into binary image information by an error diffusion method as edge portion image information. An image processing device characterized by: