JPH01255379A - Color image processing method - Google Patents

Abstract

PURPOSE:To obtain a quantized image with a high quality by weighting at least two of respective color resolved images by means of different weighting matrixes. CONSTITUTION:A processor 51 reads a first weighting matrix from a read-only memory 52, a memory 48 is accessed, a red color resolved image is read, and it is successively binarized (quantized) by using the first weighting matrix with a mean error minimizing method. A second weighting matrix is read, a memory 49 is accessed, a green color resolved image is read, the binarizing processing is executed in the same way, further, a third weighting matrix is read, a memory 50 is accessed, a blue color resolved image is read, and the binarizing processing is executed. Since respective mask original plate use different weighting matrixes in such a way, a fringe pattern generated to the image is made into a non-correlation. Thus, the quantized image with the high quality can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application 1] The present invention relates to a color image processing method, and more particularly to a method for quantizing a color image.

[Prior art 1] When processing a color image, a commonly used method is to create color-separated images by dividing the original image into three colors: red, green, and blue, or yellow, magenta, and cyan, and then process each color-separated image. has been done. The minimum average N% difference method is often used as a method for quantizing these color-separated images. In this method, each pixel of each color-separated image is sequentially extracted by main scanning and sub-scanning in a predetermined direction, and the extracted pixel of interest is quantized in the following manner.

First, evaluation pixels that are placed in the vicinity of the pixel of interest and have already been quantized are extracted, each evaluation pixel is weighted in the sixth order, and the total sum is calculated. As this weighting value, a weighting matrix as shown in (a) of ttS3 is often used. Here, the "symbol type" indicates the position of the pixel of interest. Next, the elements t>to are quantized so that the sum of the weighted quantization errors of the evaluation pixels and the sum of the 11-way weighted quantization errors of the pixel of interest are minimized.

For example, when quantizing and binarizing a pixel at the position of the back (xt y), a threshold T is determined as follows, and binarization is performed according to the threshold.

T=0.5+(1/Σ0kl) kl ×Σ(I kl E (x + + y + 1)k
Also, the stock E(XIF) is the quantization error in the warehouse (x-y), and αkl is the corresponding element of the weight matrix as shown in FIG. 3(a).

The quantized image obtained in this way has superior properties such as higher resolution, no false contours, and less image roughness compared to other methods such as Kuman's dither method. There is However, in the quantized image obtained by this method, an fI pattern or fringe pattern appears. As a way to alleviate this problem, the 1983 National Conference of the Television Network Society 15-5
, P2O3 [Study of binary display method with constraint conditions for halftones]
A method using conditional random numbers in the minimum average error method and a pixel distribution method within a mesh as disclosed in Japanese Patent Application Laid-open Nos. 150382-1982 and 278274-1980 have been proposed.

Problem to be Solved by the Invention 1 However, these methods cause problems such as the generation of new granular noise, and can only be used in fields where such noise is not a problem.

The present invention has been made to solve the above problems,
The purpose is to provide a color image processing method to obtain high quality quantized images.

[Means for Solving 1Wjl] In order to achieve the above object, the present invention is characterized in that at least two of each color-separated image are weighted using different weight matrices.

Alternatively, scanning and quantizing at least two color-separated images in different directions is considered a *sign.

[Effect 1] Therefore, the striped patterns occurring in at least two of each color separation image are uncorrelated.

[Embodiment 1] Hereinafter, a color copying apparatus using the color image processing method of the present invention for binarization will be described as an embodiment of PIS1 with reference to the drawings.

First, the mechanical configuration of this color copying apparatus will be explained. 12
The figure is a sectional view of a monochrome printer and a color image recording device in a combined arrangement. In FIG. 2, a monochrome laser printer 1 is arranged as a monochrome printer.

The polygon scanner 2 removes the electrostatic latent image by irradiating a charged photoreceptor 3 with laser light.
The toner image is transferred onto a recording paper 22, which is either a punched paper or a plastic sheet, and is fixed by a fixing device 6. This monochrome laser printer 1 records data transmitted from an image conversion section 40, which will be described later. Color image recording device r! 120 is a device that records an image using a mask original created from a color image to image information.The configuration of this device is to record an image in black and white, which is output from a monochrome laser printer 1, when it is necessary to output a normal black and white image. The recording paper 22 on which the image has been fixed is sent to the paper discharge tray 11 or 12.

On the other hand, this color image recording device 20 uses the recording paper 22 outputted from the monochrome laser printer 1 as a mask original plate 22.
R, 22G, 22[3 (hereinafter also referred to as intermediate sheet 22)
QI? an exposure table 25 that brings the intermediate sheet 22 into close contact with the photosensitive pressure sensitive paper 24; three primary color light sources 21R, 21G, and 21B for exposure; and microcapsules coated on the surface of the exposed pressure sensitive paper 24. A storage section 27 stores a color developer sheet 26 coated with a color developer that develops color by reacting with a dye precursor contained therein; A pressure developing means 28 that crushes the hardened microcapsules with pressure and develops an image on the color developer sheet 26, and a heat fixing device W that promotes color development.
129, a color image output tray 30, a manual feed tray 31 for feeding the intermediate sheet from the outside, an intermediate sheet output tray 32 for ejecting the exposure agent intermediate sheet 22, and a color developer sheet manual feed tray 33. .

Next, the effect of color mode will be explained. In the case of color mode, the mask original 22R created by the monochrome laser printer 1 passes through the intermediate sheet path switching section 10, and the intermediate sheet leading edge position is determined by the intermediate sheet leading edge positioning roller 34. The mask original plate 22R is electrostatically attached to an insulating intermediate sheet conveyance means 35 (for example, polyethylene terephthalate) that is routed around the intermediate sheet for conveying the intermediate sheet, using an electrostatic generator 38 such as a corotron, and exposed. It is transported to unit 36. The mask original plate 22R is positioned by moving the intermediate sheet conveyance means 35 so as not to cause color shift in the color latent image formed on the photosensitive pressure sensitive paper 24. In the exposure unit 36, the mask original 22R and the photosensitive pressure sensitive paper 24 are brought into close contact with each other by the exposure table 25 and exposed to light @21H, thereby forming a latent image corresponding to the mask original 22H.

After exposure, the mask original 22R is transferred to the intermediate sheet discharge tray 32.
is discharged to. The same effect is applied to the mask originals 22G and 22B, thereby forming a color latent image on the photosensitive pressure sensitive paper 24. The exposed photosensitive pressure sensitive paper 24 is superimposed on a color developer sheet 26 and developed by a pressure developing means 28. The color developer sheet is then discharged to the color paper tray 30 via a heat fixing device fi29. The used photosensitive and pressure sensitive paper 24 after pressure development is wound up by the winding means 37.

The monochrome laser printer 1 is not limited to laser printers, but can be used to create three-color separation documents of red, *color, and blue, and can be used with any type of desktop, such as an LED printer, liquid crystal printer, or other thermal printing printers. The printer is loaded on top of the color image recording device and docked, and in the case of a larger printer, the created three-color separation document is loaded from the manual feed tray 31. 1, a color image can be easily obtained.

Next, a method for creating each of the mask originals 22R, 22G, and 22+3 output by this monochrome laser printer 1 will be explained. First, with reference to FIG. 1, the image conversion device rPi 40 for processing an original image to create images to be formed on each of the mask originals 22R, 22B, and 22G will be described.

The image scanner 41 reads an original image and outputs signals for each of the three primary colors of the original image from three terminals. A correction device 42 is provided at each terminal of the image scanner 41.
．． 43 and 44 are installed to correct the signals of each color to suit the processing described below.

Are these correction devices r1142, 43.44 analog-to-digital conversion? It is connected to memories 48, 49, 50 via A/D converters 45, 46, 47, and the corrected signals of each color are quantized into multi-values. It is assumed that the number of quantization bits at this point is sufficiently large and the quantization error can be ignored. A processor 51 is connected to these memories 48, 49, and 50. A one-way only memory 52 is connected to this processor 51 . This read-only memory 52 stores a program for operating the processor 51 and weight matrices for the minimum average error method as shown in FIGS. 3(a) to 3(c). Each of these mm matrices is
This will be referred to as the third matrix. Note that the symbol "book" at t5 in these weight matrices indicates that the position corresponds to the pixel of interest. A memory 53 and the monochrome laser printer 1 are further connected to the processor 51, and the memory 53 is connected to the processor 5.
1, and output it to the monochrome laser printer 1.
stores the mask original based on the signal. Furthermore, memory 4
8 to 50 and 53 are configured as the same integrated circuit.

The image read by the INNONO scanner 41 is divided into three primary colors of red, green, and blue, each corrected and converted into analog-to-digital, and stored in the memory 48, 49, 50 according to the synchronization signal from the processor 51. Ru. The images stored in these memories 48, 49, 50 correspond to color separated images.

The processor 51 performs the following processing according to chart 70 shown in FIG. 4. First, in step S61,
Processor 51 reads the first weight matrix from read-only memory 52 .

Next, in step S62, the processor 51 accesses the memory 48 to read out the red color-separated image, and scans each pixel with respect to the original image, with the right direction being the main scan and the downward direction being the sub scan. The average error minimum method 1 is then sequentially binarized (quantized) using the first weight matrix, and the binarized value of each pixel is stored in the memory 53. Next, in step 863, the values of each pixel stored in the memory 53 are printed onto the recording paper 22 by the monochrome laser printer 1, and output as a mask original 2211.

Next, the process proceeds to step S64, where the processor 51 reads out the second weight matrix from the read-only memory 52.
Next, the process advances to step S65, where the processor 51
9 to read out the green color separated image and use the weight matrix of 1jS2 to process step S
0th order step 866 where the binarization process is performed in the same way as in step 62.
In step S63, a mask original 22G is obtained.

Next, the process proceeds to step S67, in which the processor 51 reads out the PIS3 weight matrix from the read-only memory 52.Next, the process proceeds to step 868, in which the processor 51 accesses the memory 50, reads out the blue color separation image, and reads out the blue color separation image. 3 weight ma) The above step S using IJ hooks
In step S69, the same process as in steps S63 and S66 is performed to obtain a mask original 22B.

Each mask original plate 22R, 22 formed as described above
Since G and 22B use different weight matrices, the respective mask originals 22R and 22G.

The striped pattern occurring in the 22B image becomes uncorrelated. In particular, this embodiment uses a weight matrix that tends to appear like vertical stripes, as shown in (b) of FIG. 3, and a weight matrix that tends to appear like horizontal stripes, as shown in (c) of FIG. , there is almost no correlation between the striped patterns of each color separation image using these matrices.

Next, an example of tjrJ2 of the present invention will be described below.

This second embodiment has the same configuration as the first embodiment. However, the read-only memory 52 stores only 17 sets of weight matrices (the first weight matrix), and the programs are also different. will be changed as follows.

First, in step S71, the processor 51 reads the first weight matrix stored in the read-only memory 52.Then, the processor 51 proceeds to step 872 and reads the first weight matrix stored in the read-only memory 52.
8 to read out the red color-separated image, scan each pixel of the original image with the right direction as the main scan and the downward direction as the sub scan, and use the minimum average error method described above to read out the red color separated image.
The binarized pixels and elements are sequentially binarized (quantized) using the weight matrix of t51 and stored in the memory 53.First, in step S73, the binarized pixels and elements are stored in the memory 53. The value of each pixel is printed on a recording paper 22 by a monochrome laser printer 1, and then printed on a mask original plate 2.
It will be 2R.

Next, the process proceeds to step 874, where the processor 51
9 and rewrites the green color separated image stored in the memory 49 so that only the top and bottom are upside down.0 Next, the process proceeds to step S75, and the processor 51 accesses the memory 49 and rewrites the green color separated image stored in the memory 49 so that only the top and bottom are upside down. The resulting green color separated image is read out, and the upside down green color separated image is scanned with the right direction as the main scan and the downward direction as the sub scan, and as in step S72 described above, fI%1 is scanned. The process proceeds to step 37G in which the processor 51 binarizes the binarized data using the weight matrix and stores it in the memory 53.
is accessed to rewrite the binarized image of the upside down green color separation image stored in the memory 53 so that it is further upside down. Therefore, this memory 53
The image stored in the 0
Next, the process advances to step S77, where the monochrome laser printer 1
The image is printed onto the recording paper 22 and becomes a mask original 22G.

Next, the process proceeds to step 87B, where the processor 51
0 and rewrites the blue color separated image stored in the memory 50 so that the top and bottom and left and right sides are reversed. Next, the process proceeds to step 879, where the processor 51 accesses the memory 50 and rewrites the blue color separated image stored in the memory 50 in step S75. Process in the same way as . Next, the process proceeds to step 880, where the memory 53 is accessed and the binarized image of the blue color separation image, which is stored in the memory 53 and whose vertical and horizontal directions are reversed, is further reversed vertically and horizontally. Rewrite it as Therefore.

The image stored in this memory 53 is equivalent to the one obtained by scanning the blue component of the original image in the left direction in the main scan direction and in the upward direction in the sub scan direction and converting it into a binarized image. The image is printed onto the recording paper 22 and becomes a mask original 22B.

Each mask original plate 22R922G formed as described above,
22B was scanned in different directions and obtained by the minimum mean error method. In the minimum average error method, even if the same image is processed, if the scanning direction is different, the quantized images will not be the same, and the generated striped patterns will also be different. Therefore, each mask original plate 22R obtained by quantizing each color separated image
, 22G, and 22I3 are naturally uncorrelated.

Therefore, each of the mask master plates 22R, 22G, and 22B obtained in the first and ft52 embodiments has uncorrelated striped patterns, so these mask master plates are processed as described above to form a color developing sheet. In the image obtained on 2G, each stripe pattern is inconspicuous.

Although this embodiment is written on the premise of binarization, these ideas can also be applied to multivalues such as ternary and quaternary values.

However, the present invention may be applied without departing from the spirit of the present invention.

E Effect of the Invention 1 As detailed above, in the present invention, the striped pattern generated by the minimum average error method is set to be uncorrelated with respect to at least two colors. When images are combined, striped patterns are not noticeable, and the same resolution as before can be obtained.

[Brief explanation of the drawing]

1 to 5 are detailed explanations of the present invention.
IS1 is a block diagram showing the configuration of the image conversion unit 40, FIG. 2 is a sectional view showing the mechanical configuration of the color copying device, FIG. 3 is a diagram showing the first to third weight matrices, fj
Fig. S4 is a 70-chart showing the action in the first embodiment, and Fig. 5 is a 70-chart showing the action in the second embodiment. In the figure, 52 is a read-only memory that stores the first to third weight matrices.

Claims (1)

[Claims] 1. A method of creating two or more color-separated images from an original image, sequentially scanning each pixel of each color-separated image, and quantizing the scanned pixel of interest, comprising: Quantization errors in pixels that are near the pixel of interest being scanned and that have already been quantized are weighted using a set weight matrix, and the sum of the weighted quantization errors and the sum of the quantization errors in the pixel of interest are In the color image processing method, the quantization value of the pixel of interest is determined such that the quantization value of the pixel of interest is reduced, wherein the quantization error in at least two color-separated images separated from the original image is weighted using different weight matrices. A color image processing method characterized by: 2. A method in which two or more color-separated images are created from the original image, each pixel of each color-separated image is sequentially scanned, and the pixel of interest being scanned is quantized, and the pixel of interest being scanned is quantized. Quantization errors in pixels that are close to the pixel and have already been quantized are weighted, and the quantization value of the pixel of interest is determined such that the sum of the weighted quantization errors and the sum of the quantization error of the pixel of interest is small. 1. A color image processing method for determining a color image processing method, characterized in that the scanning of at least two color-separated images separated from the original image is performed in different directions.