JPH04154371A - Device and method for processing image - Google Patents

Abstract

PURPOSE:To obtain the image of a satisfactory picture quality in both of a character/ line drawing part and a photograph/half tone part by replacing original picture density data of a picture element extracted by an edge extracting part with a value for showing the maximum density, using its value as the original picture density data and executing a binarization processing of the picture element by an average error minimum method. CONSTITUTION:A binarization processing part 4 sends original picture density data Dm,n of notice picture elements Pm,n and original picture density data of a picture element of its periphery to an edge extracting part 5, and the edge extracting part 5 derives the difference sum of density of the notice picture element and the peripheral picture element, reads out a prescribed value from a prescribed value store part 49, and compares this value and the difference sum. Based on a result of comparison, in the case the notice picture element is an edge part, a density replacing part 6 sets its original picture density data Dm,n to 255 being the maximum gradation, and sets the above data to a value as it is, in the case said picture element is not the edge part. Thereafter, binary data subjected to binarization processing by an average error minimum method by using this binarized threshold is obtained, and by displaying and printing it, a pseudo half tone image is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Summary] The present invention relates to an image processing device and an image processing method for obtaining a false halftone image using the minimum average error method in digital image processing.

The aim is to obtain images with good quality in both text/line drawings and photographs/halftones.

an edge extraction section that extracts pixels forming edge portions of an image from original image density data; a density replacement section that replaces the original image density data of the pixels extracted by the edge extraction section with a value indicating a maximum density; The image forming apparatus is configured to include a binarization processing section that performs binarization processing of one pixel by the minimum average error method using the value indicating the maximum density after the replacement as the original image density data for the pixel.

[Industrial Field of Application] The present invention relates to a three-image processing device and an image processing method, and more particularly to an image processing device and an image processing method for obtaining a pseudo halftone image by a minimum average error method in digital image processing.

In digital image processing, pixels with a density between the maximum density and the minimum density (intermediate density) are processed (
halftone processing) and binarize.

As this method, the systematic dither method and the minimum average error method are known.

[Conventional technology]

The systematic dither method, which is widely used as a halftone processing method, requires a simple hardware configuration, so
Virtual halftone processing (expression) is possible at low cost.

However, although it has these advantages, it also has the following disadvantages.

(2) When the original is a halftone image, such as a printed image, a periodic striped pattern (so-called moiré) that does not exist in the original appears in the image obtained as a result of processing.

■ If the original contains text or line drawings, the image resulting from the processing of those parts will be cut off, resulting in extremely poor reproducibility.

■ If the matrix size is increased relative to the resolution of the reading system in order to obtain multiple gradations, the resolution will deteriorate, making it impossible to achieve multiple gradations and high resolution at the same time.

On the other hand, the minimum average error method is a method that can simultaneously achieve multiple gradations and high resolution. This method is described by Manfred
, R,5chroeder, rlmage
sfro+*Computers J, IEE
E Spectrum+Vo1. 6. pp
Published in 66-781969 and subsequently J, Fja
rvis+ C,N.

Judice+ and l+1. An improved method was developed by H.N1nke.
echniques for the Display
of Continuous Tone Pic
tures on flilevelDispla
y”, Computer Graphics
and ++*ageProcessing, Vol.
, 5+1) G113-40.1976.

This method calculates the difference when a pixel with an intermediate density is binarized in the same way as a pixel with a maximum density and a pixel with a minimum density.
This is compensated for by correcting the density of pixels around the pixel of interest, thereby making the original image density and output image density equal.

[Problem to be solved by the invention]

According to the minimum average error method, the number of expressed gradations is not uniquely determined by the matrix size, as is the case with the systematic dither method, and moiré does not occur.
This method is attracting attention as it is superior in terms of gradation and resolution.

However, the minimum average error method, which has these advantages, also has
There are problems as follows.

■ Compared to simple binarization processing, the reproducibility of characters and line drawings is still inferior.

■ Spatial lag occurs in the appearance of dots (missing edges, thinning, etc.) in areas where the density is low.

For this reason, it is not possible to obtain images of good quality in the 9 character/line drawing areas, photographs, and halftone areas.

According to the inventor's study, one of the causes of the above-mentioned problems (1) and (2) is that the edge portion of the image is extracted from the quantized original image density data used for binarization processing in the minimum average error method. The problem is that it is difficult to extract.

An object of the present invention is to provide an image processing apparatus and an image processing method that can obtain images of good quality in both the 5-character/line drawing part and the photograph/halftone part.

[Means for Solving the Problems] FIG. 1 is a diagram showing the basic configuration of the present invention, and 2 shows an image processing apparatus according to the present invention.

In FIG. 1, ■ is an input section, 2 is a data processing section, 3 is a quantization processing section, 4 is a binarization processing section, 5 is an edge extraction section,
6 is a density replacement section, and 7 is an output section.

The edge extracting unit 5 extracts pixels forming an edge portion of an image from quantized original image density data regarding a plurality of pixels obtained by reading an image.

The density replacement unit 6 sets the original image 4 degree data of the extracted pixel to a value indicating the maximum density. This value indicating the maximum density corresponds to the maximum value of the quantized original image density data, that is, one value after the binarization process (usually "1").

The binarization processing unit 4 performs binarization processing for each of the five pixels using the minimum average error method. At this time, the pixels extracted as edge parts are used as the original image density data.
The value indicating the maximum concentration after substitution is used.

[For production]

The binarization processing section 4 sends the quantized original image density data to the edge extraction section 5 for binarization processing of a certain image.

When the edge extraction section 5 receives the original image density data, it uses this data to extract pixels forming an edge portion.

This extraction is performed by comparing all, part, or the sum of the differences between the original image density data of the pixel to be processed (the pixel of interest) and the original image density data of predetermined surrounding pixels with a predetermined value. . Here, 2. For example, even if the original image density data of a contiguous pixel is relatively small, if the original image density data of one peripheral pixel is extremely small, the difference will be relatively large.

That is, it becomes larger than a predetermined value and can be extracted as an edge portion.

The density replacement unit 6 is processed in the edge extraction unit 5.

The original image density data of the pixels determined to be the ji portion is set as the maximum value.

The binarization processing unit 4 uses this binarized dark value as a dark value to perform binarization processing on three pixels by the minimum average error method. Now, let's look at the example above.

Since the relatively small original 4 degree data is replaced by the maximum 4 degrees, the original 4 degree data is always higher than the binarized darkness value. Therefore, as a result of binarization, the pixel is set to be a black pixel, that is, the maximum density (the larger value of the maximum density and the minimum density). This makes it possible to eliminate chipping and thinning of the edges of two characters and line drawings.

S example] The image processing apparatus shown in FIG. 1 will be further explained.

The input unit 1 is used to obtain original image data by reading a document including, for example, a single character/line drawing part, a photograph/halftone part, and is composed of, for example, a scanner.

Original image data is obtained for each pixel of the scanner.

The scanner repeats scanning in the main scanning direction and the sub-scanning direction. Therefore, for each pixel arranged in a two-dimensional plane,
You will get the original image data.

The output unit 7 outputs data after binarization processing (binary data).
It is used to generate images based on 1. For example,
Consists of printer, display, etc. Binary data is
1j or rQ, and the former is output as a black pixel (pixel with maximum density).

The latter is output as a white pixel (pixel with the lowest density).

The data processing unit 2 includes a CPU (central processing unit) and a memory, and performs a process of converting original image data into binary data.

The quantization processing unit 3 quantizes the original image data to generate original image density data. Quantization is 1. For example, the density of each pixel is
It is expressed using 256 gradations from 0" to 255 degrees. Therefore, the binarized darkness value and the predetermined value for edge extraction are also expressed using these gradations.

FIG. 2 is a block diagram of an embodiment, showing details of the binarization process according to the present invention.

In FIG. 2, 41 is an original image density data line buffer, 42 is a coefficient matrix for weighting, 43 is a binarization error data line buffer, 44 is an average error calculation unit, 45 and 46 are adders, and 47 is a binarization , 48 is a binary data output section, 49 is a predetermined value storage section, and 9 or more are the binarization processing section 4
Configure.

First, the binarization process using the minimum average error method will be briefly explained.

Now, it is assumed that pixels on the n-th dot in the main scanning direction and the m-th line in the sub-scanning direction in the a-image space are focused and binarized. Pixel of interest P1. The original image density data D1.7 etc. of □ are
Quantized to any value from minimum density to maximum density.

Binarization errors generated in peripheral pixels that have already been binarized and stored in advance are read out, and a weighted average is determined based on predetermined weighting coefficients. Add this value to the original density data D1, 1 of the pixel of interest P1, 7 to obtain correction data Dm+R1,
By comparing with the binarized dark value Tf (a value uniquely equivalent to either the maximum density or the minimum density), binary data is obtained. The difference.

It is stored as the binarization error E,,7.

By repeating the above steps over the entire image space, binary data that has been binarized using the minimum average error method is obtained, and by displaying and printing this, a pseudo halftone image is obtained.

Next, the binarization processing of the second embodiment will be specifically explained.

The input unit (image reading unit) 1 reads one image in the horizontal scanning direction N(
Integer) is read by a line-type image sensor lined with dots to obtain original image data. The quantization processing unit 3 quantizes this original image data according to its density, and the original image density data D
, 2, etc. are obtained for each pixel.

The binarization processing unit 4 converts this original image density data D, , +R
etc. are stored in the corresponding position (bit) of each pixel in the original image density data line buffer (hereinafter referred to as buffer) 41. Buffer 41 is comprised of random access memory. m and n are positive integers, and 1≦n≦N.

After this storage, the binarization processing unit 4 immediately stores the pixel of interest P.
1.7 original image density data D11 and surrounding pixels (
For example, two predetermined four pixels Pll-11 shown in FIG.
M+ pH+11-1+ P@+ll'l+ PI@+
1. +1) original image density data is sent (input) to the edge extraction section 5.

The edge extraction unit 5 first calculates the sum of differences in density between the pixel of interest and surrounding pixels (sum of differences), that is, (D, . . . ).

D, -1,,,)+ (D,,,,-D,,,-1)+
(D...

Find D, Il+1 )±(D, , -D, , .).

Here, these four surrounding pixels are the pixels to the top, bottom, left and right of the pixel of interest in the two-image space6, that is, the pixels that have the greatest relationship with the pixel of interest, and the weighting for the pixel of interest is the one with the highest value of the coefficient. It is a large pixel. By selecting these four, it is possible to relatively accurately extract points (that is, edge portions) where the original image density data has a large change with two relatively simple calculations.

In addition, around 1 ii! A wider range of pixels may be used as the ii element. For example, the eight pixels shown in the figure can be used for the buffer 41 in FIG. The range to be used as peripheral pixels can also be determined by referring to a coefficient matrix 42, for example, weighting, which will be described later. That is,
For weighting of a predetermined value or more, a pixel having a coefficient K is set as a peripheral pixel, and its original image density data D is used.

Further, instead of the sum of differences, a difference with one predetermined pixel may be used. For example, +D @, 11 D m, and I+ -1 may be obtained and used using the original image density data of the pixel on the left.

Further, instead of the sum of differences, differences between all of a plurality of predetermined pixels may be used. For example, using the four pixels mentioned above, D m, 11 D *-1+ D ? ,,
.

D5. . , l D 11 ' l 9,
You may use this.

Next, the extraction unit 5 reads a predetermined value from the predetermined value storage unit 49 and compares it with the determined sum of differences. As the four predetermined value storage sections, predetermined areas of the memory are used. The predetermined value is 6, for example "64", that is, about 1 of the maximum gradation.
/4.

Then, if the sum of differences is larger than a predetermined value, it is determined that the pixel of interest is in the area. If the sum of differences is smaller than a predetermined value, it is determined that the pixel of interest is not an edge portion. The determination result is sent to the density replacement section 6.

The density replacement unit 6 performs correction on the edge portion based on the 1 determination result to obtain a correction value D@+It-. That is, when the pixel of interest is an edge portion (extracted),
The original image density data D, 7° is the maximum gradation of "25".
5″°. If it is not an edge portion, the original image density data is left unchanged.

On the other hand, the average error calculation unit 44 performs a weighted average calculation. For this purpose, the average error calculation unit 44 calculates the weighting corresponding to the pixel of interest P1.1 from the coefficient matrix 42, and calculates the binary values generated in each surrounding pixel. The weighting for each error is a coefficient K
Also read M -1+ II -1 etc. From the binarization error data line buffer (hereinafter referred to as buffer) 43, the binarization error E generated in each pixel that has already been binarized is
Read II-1+ +1-1, etc. Here, the binarization process is the n-1st dot of the 5th m-1th line, the nth dot, the n+1th dot, the n-1st dot of the mth line, the dot, etc.
shall be carried out in this order. In this case, the pixels that have already been binarized are P 11-1+ 11-1+ P m-1,
R+ Pll-1+ II. l + P @ + II
There are four -1. Therefore, 1 weighting is a coefficient matrix of 4
2 and the data read from the buffer 43 are
The average error calculation unit 44 calculates a weighted average calculation of the zero-order data corresponding to one pixel, X (K, -, .

Ell-1+ 11-1 + Ka-1,・Ell-
11. -+.

E,, □l1. ．． . ,+to,,,-・E,, ,,-,
)/(K,-,,.

Do -+ on 10, 1+ on +, -+ on -, -+).

Next, the binarization processing unit 4 inputs the correction value of the original image density data of the pixel of interest P II) R, , , and the result of the weighted average calculation to the adder 45 and adds them to obtain the addition result D. m+
11'' and sends it to the binarization section 47.

The 24a conversion unit 47 compares the addition result D@+n'' with a predetermined binary value TH to obtain binary data ○m1N.The binary value TH is set to 2, for example. “12
8'', that is, 1/2 of the maximum gradation.

That is, 1),,, when //>TH ○,,,,=2ss(!
4 pixels) When D m+R”≦TH ○,,,,=O
(white pixel). That is, it is determined whether the pixel is a black pixel or a white pixel.

The binary data 01 and 6 are set to either the highest value "255" or the lowest value "0°' of the three gradations, and then sent to the binary data output section 4.

The binary data output section 48 has +Om+R of "°255".
”, the signal “B” (black pixel) is sent to the output unit 7, and when O@+R is “0”, the signal “0” (white pixel) is sent to the output unit 7.

On the other hand, the 2II processing unit 4 calculates the difference between the addition result 1.1'' and the binary data 011 + TI using the adder 46, and converts this into the binary data generated by the binarization of the target pixel P@, TI. conversion error E, + n (-D m+ 11
”-〇@+n) and stored in the corresponding position of the buffer 43.

Repeat the above steps in the main scanning direction and sub-scanning direction 24:t
The output “1” or “0” of the L data output unit 48 is sent to the output unit 7.
By transferring the image to 1, a pseudo halftone image which has been binarized according to the present invention can be obtained.

FIG. 3 is a diagram showing the processing results. In particular, the same figure (A)
2 shows the processing results according to the present invention, and FIG. 3B shows the processing results according to the conventional minimum average error method.

The processing data of the example illustrated in FIG. 3(A) is as follows.

[1-layer original image data: 240 (dpi) 64 gradations (0
: White ~ 3F = Black) [2] Processing details: /Umbrella Image Processi
ng 5sourceFile: Tsuyoshiin9.
c*/[Triple printing 240 (dpi), F
The processing data of the example shown in FIG. 3(B) of MLBP 111 is as follows.

[Single-layer original image data: 240 (dpI).

64 gradations (0: white to 3Fl) C2') Processing details: /Umbrella Image Process
ing 5sourceFile: min, c city/ ・Weighting is coefficient matrix (K,,,) fA 5urv
Eye or Techniques for the
Display of ContinuousT
one Pictures on B11eve
Based on the example in lDisplaysJ.

Average miscalculation calculation Average error = Σ(K,,,・E,1,)/Σ, 7 [3] Print output 240 (dpi), FMLB
P 111 Comparing Figure 3 (A) with Figure 3 (B), it is clear that the reproducibility (resolution) of a single character/line drawing has been improved, and the delay in don'ts has also been improved.

In this embodiment, the average error calculation section 44 refers to four binarization errors for its calculation, but it may use a wider range of binarization errors.

In addition, the quantization gradation is 25 from "0" to "'255".
Needless to say, it is not limited to six gradations.

Note that when changing the number of quantization gradations, it is necessary to change the predetermined value for edge extraction as appropriate.

Although the present invention has been described above with reference to embodiments, the present invention can be modified in various ways according to its gist.

The present invention can be widely applied to devices 2 that perform digital image processing, such as digital copying machines, image scanners, facsimile machines, etc., when obtaining pseudo halftone images through binarization processing.

〔Effect of the invention〕

As explained above, according to the present invention, pixels corresponding to the edge portion of a 9M image are extracted and the pixels corresponding to the edges of the 9M image are extracted and 9 by correcting the concentration data.
Not only can the reproducibility of characters and line drawings be improved, but also chipping and thinning of the edges and edges can be eliminated, so good image quality can be obtained for both single characters and line drawings and photographs and halftone parts.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the principle configuration of the present invention. FIG. 2 is a block diagram of an embodiment. FIG. 3 is a processing result. FIG. 1 is an input section, 2 is a data processing section, 3 is a quantization processing section, and 4
is a binarization processing unit, 5 is a processing and extraction unit, 6 is a density replacement unit,
7 is an output section.

Claims (3)

[Claims] (1) An edge extraction unit (5) that extracts pixels forming the edge portion of the image from the original image density data;
) for replacing the original image density data of the pixel extracted with a value indicating the maximum density; and for the extracted pixel, using the value indicating the replaced maximum density as the original image density data. , a binarization processing unit (4) that performs pixel binarization processing using the minimum average error method. (2) The image processing apparatus according to claim 1, wherein the extraction is performed by comparing a sum of differences between original image density data of the pixel of interest and its surrounding pixels with a predetermined value. (3) Using the difference between the quantized original density data of the pixel of interest and its surrounding pixels, extract the pixel that will become the edge of the image, and set the original density data of the extracted pixel to a value indicating the maximum density. An image processing method, comprising: replacing the extracted pixel, and performing binarization processing on the extracted pixel by a normal error minimum method using a value indicating the maximum density after the replacement.