JPS59205872A - Method and apparatus for processing color picture - Google Patents

Abstract

PURPOSE:To eliminate the deterioration in picture quality by binary-coding other chrominance component signal by a prescribed threshold signal and recording it when a chrominance signal of one color separation picture signal is decided as high resolution. CONSTITUTION:The color separation signal resulted from reading of an original from an input device 101 is converted for the complementary color and then A/D-converted and inputted to binary-coding circuits 104Y, 104M and 104C and area identifying circuits 108Y, 108M and 108C, respectively. As the result of identification of the recognizing circuit 108, when an OR circuit 109 judge that the inside of any inside of a noted picture element block of an original is a high resolution region, the picture element data are compared by a single picture element matrix table 105 as to all chrominance components. When all three color signals, Y, M and C are judged to be in a contrast picture region, the binary-coding is processed by a dither matrix table memory 106. Thus, the deterioration in the picture quality due to the overlap of the single threshold processing and the dither processing at each chrominance component is prevented.

Description

[Detailed description of the invention] The present invention relates to a color image processing method and apparatus.

In particular, the present invention relates to color image processing that is effective for color recording devices that perform binary recording for high-resolution input devices.

In a color recording device using a conventional binary recording device, the input mtc data is converted into a dither matrix or its modified matrix.
I) 0, which is two-dimensionally expressed in gradation by Tx, has the disadvantage that in high-resolution areas such as characters and symbols, thin lines are divided by dithering, deteriorating image quality. An example of a solution for monochromatic, e.g. black and white images is the SID 77 Di
GEST ”'CIBM KwanY, Wong announced “Adaptive Switching of D
ispersed and C1uBtered Ha
lftone Patterns for Bi-Lev
el Image Rendition J is here. Laplacian gradient and mini
Two methods have been proposed: -max and of corner sums. As a result, for a monochromatic image, it is determined whether the state of the original image is a grayscale image area or a high-resolution image area such as text, and it is determined whether to perform dither processing or single threshold processing when performing binarization processing.

However, in the case of color images, the input device usually separates the original data into six color bands of R (red), G (green), and B (blue) and converts the signals, so each of the three color signals is When the method is applied, a certain color signal (for example, R) within the input pixel area of interest is determined to be a high-resolution area, and other color signals (
For example, G) may be determined as a grayscale image area.

Therefore, from the printed results, for example, the R signal is subjected to single threshold processing, and the G signal is subjected to dither processing.When overlapping colors, not only is sufficient color reproduction achieved, but even thin line areas are separated by the dithering of the G signal. It has the disadvantage of deteriorating image quality.

The present invention eliminates the above drawbacks by determining whether a color-separated image signal is a high-resolution image or a gray-scale image, and when it is determined that one color component signal has high resolution, other color component signals are divided into two by a predetermined threshold signal. The present invention relates to a color image processing method and apparatus characterized by performing value conversion processing.

Before explaining the present invention, a conventional high-resolution, gray-scale image region identification method for monochrome images will be explained with reference to FIG. 1st
The figure shows an example shown in Japanese Patent Laid-Open No. 58-3374, in which image input data 200 is alternately stored in both lineback memories 202 and 206. In the state shown in FIG. 1, 202 acts as an input line backer memory, and 203 acts as an output line buffer memory. Both 202 and 206 are buffer memories for regular lines. A block memory 204 stores image data for pixels of a document. This is to identify whether the square pixel area of the document is high-resolution image data or grayscale image data by secondarily checking the nxn pixel data, and 205 is a discriminator for this purpose. The identification method is to find the maximum value Pmax and minimum value Pm1n of pixel data in the block memory of 204, and if (Pmax - Pmin)≧A, the image in the block is identified as a high-resolution area, and the threshold value changeover switch is 209 to the B side and select the single threshold matrix table 207.

If (Pmax - Pm1n) < A, it is identified as a gray area and the dither matrix table A of 206 is selected based on the output signal of the discriminator 205. A comparator 208 compares the threshold data selected at 209 with each pixel data from the block data memory. Therefore, the output of 208 is converted into a binary signal and transmitted to a binary on-off printer 210, where a binary image is recorded.

FIG. 2 shows one embodiment of the present invention.

This is a document reading input device with 101 six-color color separation filters. -The color separation signal from the input device is separated into R (red), G (green), and B (blue) signals and sent to the well-known 102 complementary color conversion circuit. enter. 102, the RGB filter signal is C (cyan)
2M (magenta).

The complementary color is converted to Y (yellow) and converted into digital data through the A/D converter 103. It is assumed that the input device 101 has highlight and dark setting knobs. 104
are the various memories 202 to 204 in FIG. 1, and the comparator 208.
It mainly consists of a binary comparator circuit. 105 and 106 are threshold value matrix tables, for example, 105 is a single threshold value matrix table, and 106 is a dither matrix memory. 107
is a threshold value changeover switch and corresponds to 209 in FIG.

108 is a circuit block including a buffer memory, a block memory, and an area identification circuit, which corresponds to 201 to 205 in FIG.

104C to 108C are area identification circuits and binarization processing units for cyan components, and 104M to 108M
and 104Y to 108Y constitute an area identification circuit and a binarization processing section for magenta and yellow components, respectively. 10
9 is the identification output CI r M of the area identification circuit for each color.
011, which is a circuit that takes the logical sum (OR) of l + Yl
0 is a printer that performs printing processing by superimposing each color toner according to data from each binarization processing circuit.

That is, when analog R, G, and B signals are input in parallel by scanning a document, the complementary color conversion circuit 102 converts them into analog Y, M,
It is converted to a C signal and output to Para. This Y, M, C
The signals are stored in the A/D conversion circuit memory and block memory, respectively. At the same time, it is also stored in the buffer 7 memory (+7 out of 108) and the block memory. The data in the block memory in block memory 108 is identified by a discriminator as shown in Fig. 1, and when it is determined that it is a high-resolution area such as a character, an identification signal of 1 is output. If it is determined to be a region, it outputs 1, switches all of the switchers 107Y, 107M, and 1070 to the fixed threshold value signal source 105 side, and binarizes the pixel data of each block memory in 104 with a constant threshold signal. .

Therefore, if even one of the three color component identification circuits determines that the target pixel block of the document is a high-resolution area by the OR circuit 109 of the identification results C1+Ml+Yl from 108G, 108M, and 108Y, other color components are determined. Even if the region is determined to be a light and dark region, pixel data comparison is performed for all color components using a single-threshold matrix (IJ rack table). As a result, the dither matrix and the single-threshold matrix are mixed and reproduced for each color signal component without being superimposed. Therefore, the sharpness of the image quality is improved, and the overall image quality is improved. If 3 electric signal components Y.

Binarization processing is performed by the dither matrix table memory 106 only when both M and C are recognized as grayscale image regions.

Note that the above recognition can also be performed for one page of the reproduced portion instead of the block area of 1 x 1.

By providing a single threshold matrix table and a dither memory circuit for each color, it is possible to change the single threshold level for each color or change the array of dither tables, thereby achieving appropriate quality color reproduction. can be made possible.

In addition, as a single threshold value and a dither pattern table in the above explanation, for example, when the input image data range is D to level 15 and the area is identified in a block pixel area of 4 x 4 pixels, if you want to slice everything at level 5, use U (A ) as shown in the figure, and this is called a single threshold matrix table.
A table in which the slice level within the array is changed as shown in (B) is called a dither matrix table.

It is also possible to form a black component signal BK from the B, G, and R signals, output Y, M, C, and BK, perform area discrimination and binarization processing separately from color signals, and output the signal to the logic circuit 109. .

[Brief explanation of the drawing]

FIG. 1 is a conventional image processing circuit diagram, FIG. 2 is a color image processing circuit diagram according to the present invention, and FIG. 3 is an explanatory diagram of threshold values in FIG. 2, 104Y, 104M. i Port 4C is a binarization circuit, 108y, 108M, 10'
8c is an area recognition circuit ○ Applicant: Canon Co., Ltd.

Claims (1)

[Claims] (1) It is characterized by determining whether a color-separated image signal is a high-resolution image or a grayscale image, and when it is determined that one color component signal has high resolution, the other color component signals are binarized using a predetermined threshold signal. color image processing method. (Two-color separation image signal generation means, means for determining whether the image signal is a high-resolution image or a grayscale image, and when it is determined that one color component signal is a high-resolution image area, the image signals of other color components are set to a predetermined threshold value. A color image processing device characterized by having means for performing binarization processing based on a signal.