JPH0354680A - Character area separating system for color image - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Summary] In a character area separation method for improving the image quality in digital image processing devices such as facsimiles and digital copying machines, halftone dots are created using a signal obtained by linearly combining color signals read from a color image. /Separated into non-halftone areas and separated into photo/non-photo areas based on the maximum density value signal of the color signal, and areas determined as non-halftone areas and non-photo areas are separated as text areas. .

[Industrial application field]

The present invention relates to a character region separation method for improving image quality in digital image processing devices such as facsimile machines and digital copying machines.

[Conventional technology]

Digital image processing is used to separate the image into text areas, other photographic areas, halftone dot areas, etc. Sharpening processes such as edge enhancement processing and binarization processing are applied to the text areas, and other areas are For example, in Japanese Patent Laid-Open No. 59-205, it is possible to prevent the occurrence of moiré and reproduce color images with high gradation by smoothing and applying dither processing.
This is known from Japanese Patent No. 872 and the like.

In the invention described in this publication, the character area is separated from other areas for each RGB primary color signal, that is, the density signal of each primary color, and by taking the logical sum of the separation results for each of these primary color signals, any one of the primary color signals is obtained. If the area is determined to be a text area, sharpening processing is applied to the image data of all colors.

[Problem to be solved by the invention]

However, in this conventional technology, processing must be performed to separate the character area and other areas for each primary color signal, so there was a problem that the time and amount of hardware required for the processing increased. .

[Means to solve the problem]

In the text area separation method, which reads a color image as a color signal and separates the text area, it separates the color image into halftone/non-halftone areas using a signal that is a linear combination of the color signals, and also separates the photo/non-halftone area using the maximum density value signal of the color signal. The area is separated into a non-photo area, and the area determined as a non-halftone area and a non-photo area is separated as a character area.

[For production]

When a text area is separated into a halftone/non-halftone area, it is separated as a non-halftone area, and when separated into a photo/non-photo area, it is separated as a non-photo area, so it is a non-halftone area and a non-photo. The area determined as an area becomes a character area.

〔Example〕

The figure shows the principle of the present invention, in which the image reading unit 1 obtains R, G, and B signals, which are density value data for each of the primary colors R, G, and B, and converts these signals into a linear combination signal. The generation unit 2 generates a linear combination signal J expressed by the following equation (1), which is obtained by linearly combining each color signal, and using this linear combination signal, the halftone/non-halftone area separation unit 3 generates a halftone/non-halftone area.
Separate into non-halftone areas.

J+ =a, R+az G+a3 B
(1) (Here, a, -a3j coefficients, R, G.B are amplitudes of each color signal.) As a separation method in this separation section 3, for example, Yorito Motoide Patent application filed on the date 1986
As described in No. 263884, a matrix consisting of M×M pixels is sequentially applied to each pixel point of a digital multi-tone human image obtained by linearly combining the above-mentioned color signals. It is detected from the density relationship with surrounding pixels in the matrix whether the central pixel is a peak or valley of density change, and NXN#ffi(
The number of pole pixels indicating a mountain and the number of pole pixels indicating a valley within a block are counted as a unit, and the pole pixel with the larger count value is set as the number of pole pixels of the block, and the block of interest is set as the block of interest. When the number of extreme pixels of the target block is smaller than the number of extreme pixels of surrounding blocks, it can be determined that a predetermined pixel in the block of interest belongs to a non-halftone area.

As mentioned above, since R, G, and B color signals are used in this embodiment, the maximum signal J2 of the density value generated in the output section 4 of No. 18 is given by the following (2
) as shown in the formula.

J2 = may (R, G, B) [2) The method of separating the photo/non-photo area into photo/non-photo areas in the photo/non-photo area separation section 5 using this maximum value signal is based on a patent application previously filed by the present applicant. A separation method such as that described in "Image processing method j" published as JP-A No. 61-225974 can be applied.

To briefly explain this separation method, for example, let the density of a pixel of interest P be DP, and eight pixels Q adjacent to this pixel of interest
, , Q2, Q, are respectively D cL ,
DQ2. """"" When D Qa is assumed, the density difference δ is calculated by the following equation (3).

δ" l Co DP +C+ DQ, +C2 DQ
2 10+Cs DQ8 1 −・−(3)
(Here, C., C+, C2, and 1"""Ca are coefficients, and Co and c,,c2...-...-.c
It has a different sign from s, and Co + C I -i-
C2 + """""+ Cs = 0.) Then, T = W, T + W2 δ (4) (Here, T is the cumulative value from the start of each scanning line, W1W2 is a weighting coefficient, w, is a decay factor of w, < 1), and when the value of T obtained by this equation (4) is larger than a predetermined darkness value, the above-mentioned pixel P is set as a non-photograph area. In the logical product means 6 which makes the determination, the logical product of the non-halftone output from the halftone/non-halftone dot q area fa section 3 and the non-photo output from the above photograph/non-photograph area separation section 5 is performed. By taking , the area that is a non-halftone area and a non-photo area is separated as a character area.

Note that the halftone/non-halftone area separation method and the photograph/non-photograph area separation method are not limited to those shown in the above embodiments, and may be those described in the above-mentioned applications and publications or other known linear methods. It goes without saying that a halftone/non-halftone region separation method and a photograph/non-photograph region separation method using combined signals can be applied.

〔Effect of the invention〕

According to the present invention, by taking the logical product of the non-dot output output when it is determined that it is not a halftone area and the non-photo output that is output when it is determined that it is not a photo area, Another effect can be achieved in that text areas can be separated from other areas in an image in which halftone dots, photographs and text are mixed.

[Brief explanation of drawings]

The figure is a bronc diagram showing an embodiment of the present invention.

Claims (1)

[Claims] In a character area separation method that reads a color image as a color signal and separates character areas, the color signal is separated into halftone/non-halftone areas by a linearly combined signal, and the density value of the color signal is A text area separation method for a color image, characterized in that the text area is separated into photo/non-photo areas based on a maximum value signal, and an area determined as a non-halftone area and a non-photo area is separated as a text area.