JPS61225974A - Picture processing system - Google Patents

Abstract

PURPOSE:To obtain a picture with high quality regardless of the presence of nesting by deciding whether or not a constituting picture element such as a character is a constituting picture element such as a photograph at each picture element and applying a picture processing corresponding to each picture element depending on the result of decision so as to apply the picture processing of an original where a binary picture and a multi-value picture exist in mixture. CONSTITUTION:A picture processing unit 13 reads sequentially picture information of each picture element from a line buffer 12 under the control of a microcomputer 15 in parallel with the write of picture information to the line buffer 12 and applies a prescribed processing. When a deciding section 133 decides the picture element as a character constituting picture element, a multi-grading processing section 135 applies multi-grading processing and supplies an output. A picture output device 14 is such a device as a printer or a CRT display and prints out or displays a processed picture signal outputted sequentially from the picture processing unit 13 in parallel with the read of the picture reader 11.

Description

[Detailed Description of the Invention] [Technical Field] The present invention relates to an image processing method in a facsimile machine, an image processor, etc., and in particular to an image processing method for documents such as documents containing binary images such as characters and multi-gradation images such as photographs. Regarding.

[Prior art]

Conventionally, as a method of reading a document containing a binary image (black and white image) such as a single character and a multi-gradation image (shade image) such as a photograph and reproducing each image with high quality, for example, Japanese Patent Laid-Open No. 5
As shown in No. 8-153455, the scanned original image is divided into multiple blocks each consisting of multiple pixels, and each block is determined whether it is a text image area or a photo image area, and the image is applied to each area. There are known methods for performing image processing. Although this method can successfully reproduce an image containing a single character image and a photographic image, it not only requires block memory but also It is impossible to perform real-time processing in parallel with scanning, and furthermore, in the case of so-called nested images where text is included in a photo, it is impossible to distinguish between a photo area and a text area, making it impossible to expect a high-quality reproduced image.

〔the purpose〕

An object of the present invention is to perform image processing in real time on a document containing a mixture of binary images such as characters and multi-gradation images such as photographs, and to reproduce high-quality images regardless of the presence or absence of nesting. It is in.

[Configuration] The present invention focuses on each pixel of a document containing a mixture of binary images such as characters and multi-gradation images such as photographs, and determines for each pixel whether it is a constituent pixel such as a character or a constituent pixel such as a photograph. It is characterized in that it makes a determination and performs image processing adapted to each pixel based on the determination result. An embodiment of the present invention will be described below.

FIG. 1 shows the overall configuration of an embodiment of the present invention, in which a document 10
It consists of an image reading device 11 that reads the image, a line ballast 121, an image processing device 131, an image output device 14, and a microcomputer 15 that controls the operations of these parts. The image processing device 13 is the center of the present invention, and has one function: a difference detection section 131. It is divided into a difference accumulation section 1321, a determination section 133, a binarization processing section 134, and a multi-tone processing section 135.

The document 10 includes a mixture of binary images such as characters and multi-gradation images such as photographs. The image reading device 11 sequentially reads the document 10 in a one-dimensional scanning format, converts it into multi-value data by analog-to-digital conversion, etc., and outputs image information corresponding to the density of each pixel. Image reading device! The image information in pixel units output from 11 is sequentially stored in the line buffer 12.The line buffer 12 has a memory configuration for 3 lines, but the memory is not necessarily large enough to store all 3 lines worth of image information. It is not necessary to have a capacity.

The image processing bag [13, in parallel with the writing operation of image information to the line buffer 12 by the image reading device i1, under the control of the microcomputer 5, processes the target pixel P in the positional relationship shown in FIG. 8 pixels adjacent to it Q
Pixel information for a total of nine pixels of t + Qz t...Q8 is read out sequentially from the line buffer 12, and the following processing is performed.

First, the difference detection unit 131 calculates the density of the pixel of interest P by DP.
, pixels Q□, Q2, . . . Q, adjacent to the pixel of interest P,
The concentrations of DQ1, DQ2 . ...O Q As o, the density difference δ is determined by equation (1).

δ=IC, DP+C, DQ, +C2DQ2+-+C3o
Qe I ''(1) Knee, coefficients C8 and C1, C,
, . . . C8 has a different sign and 1. Assume that the condition Σ C3=0 is satisfied.

1=0 For example, using the pixel of interest P and the adjacent pixels 1, Q1, Q, S, Q on the upper, lower, left and right sides, and the pixel Q in the diagonal direction
1, Q1, Q1, Q, are ignored.

Co=2. C2=C,=C5=C,=-1/2,C,=
C:4=Cs=Ca=0. Further, the pixel of interest P and the adjacent upper and left pixels Q2. Using only Q, Co=2. C2=Ca” 1-C1
=C3=C15=Ca =C? =C8=0. Even with this, good image reproduction is possible.

Note that the density difference δ is the density DP of the pixel of interest P and the density DQ1 of the pixels Q11Q21...Q adjacent to the pixel of interest P.
, DQ2, .

Next, the difference accumulator 132 adds the density difference δ obtained by the difference detector 131 to the current cumulative difference value T based on equation (2) to obtain a new cumulative difference value T.

T==w1T+W2δ −(2
) where "r t W 2 is a weighting factor, especially W.

is the damping coefficient (w, <1). The first term on the right side of equation (2) reflects the trend of the density difference up to now, and the second term on the right side reflects the density difference between the pixel of interest and its adjacent pixels.

Next, the 1 determining unit 133 compares the difference cumulative value T obtained by the difference accumulating unit 132 with a predetermined threshold TH.
> TH, the pixel of interest P is determined to be a constituent pixel of a character, and in the case of 1T≦T, the pixel of interest P is determined to be a constituent pixel of a photograph. If the image information of the pixel of interest P is determined to be a constituent pixel of a photograph, the image information of the pixel of interest P is binarized and outputted by the binarization processing unit 134, and if the image information of the pixel of interest P is determined to be a constituent pixel of a photograph, the image information of the pixel of interest P is Multi-gradation processing of information is performed and output.

The image output device 14 is a printer, a CRT display, or the like, and prints out or displays processed image signals sequentially output from the image processing device 13 in parallel with the reading operation of the image reading device 11.

FIG. 3 shows a specific example of the configuration of the image processing device 13.

In FIG. 3, the functions of the binarization processing section 134 and the multi-gradation processing section 135 are configured in the same circuit block.

Every time one pixel of image information is written into the line buffer 12 of FIG. 1, image information having the positional relationship shown in FIG. 2 is extracted from the line buffer 12.

It is stored in a 3×3 register 301 via an input line 300. That is, the register 301 stores the pixel P of interest and its adjacent pixels Q1, Q2, . ...The image information of Q8 is stored. The image information in this register 301 is sent to an MTF circuit 302 and an image information calculation circuit 304. The 4° MTF circuit 302 is a circuit that performs MTF correction on the density of the pixel of interest P. Let the density of P be DP', and the pixel Q2 .
Focusing only on Q1, Q1, and Qwa, their concentrations are DQ2
．． Assuming DQ1, oQs, and DQ7, the following formula, DP' = 3XDP-(DQ2+DQ, +DQ!l+D
Q7)/2... (3) The 0-pixel information selection circuit 303 that obtains the image information of the target pixel P that has undergone MTF correction uses the MTF selection signal 310 instructed from the outside. , the image information of the pixel of interest P output from the register 301 is selected as it is, and if the signal 310 is “1”, the image information corrected by MTF in the MTF circuit 302 is selected and given to the digital comparison circuit 304. .

On the other hand, the image information calculation circuit 304 includes the difference detection section 131. It is equivalent to the difference accumulation unit 132 and the determination unit 133, and calculates equations (1) and (2) based on the 3×3 image information given from the register 301, and determines whether the pixel of interest is a constituent pixel of a character or a photograph. This is a circuit that determines whether a pixel is a constituent pixel or not. In this case, the image information calculation circuit 304 calculates the pixel of interest P.
If the pixel is determined to be a constituent pixel of a character, 0'' is output, and if the pixel is determined to be a constituent pixel of a photograph, 1'' is output.

The dither matrix 305 stores multi-level thresholds in a matrix form. Each threshold value of this dither matrix 305 is sequentially read out by a line/column address signal 311 and is applied to a dark value selection circuit 307 via a dither selection circuit 306.

The threshold selection circuit 307 is further provided with a predetermined fixed threshold by a line 312. The threshold value selection circuit 307 selects line 3 when the output of the image information calculation circuit 304 is II OIt, that is, when the pixel of interest P is determined to be a constituent pixel of a character.
If 12 fixed threshold values are selected and the output of the one-picture information calculation circuit 304 is '1'', that is, if the pixel of interest P is determined to be a constituent pixel of a photograph, the multi-level threshold from the dither selection circuit 306 is selected. The digital comparison circuit 304 calculates the image information (
or MTF-corrected image information) to the threshold selection circuit 30.
7, if the image information value is larger than the threshold value, the image information is set to 11111 (black), and if it is equal to or smaller than the threshold value, the image information is set to 110'1 (white), and the output line 313. In addition.

The fixed darkness value of line 312 can also be made variable depending on the condition of the document.

Although FIGS. 1 and 3 are embodiments in which the present invention is implemented by digital processing, the present invention can also be implemented by analog processing.

FIG. 4 shows an embodiment in which the present invention is implemented by analog processing, and particularly shows portions corresponding to the difference detection section 131, difference accumulation section 132, and determination section 133 in FIG. 1. In addition, in FIG. 4, the pixel of interest P in the positional relationship of FIG.
and the adjacent upper and left pixels Q2. It is said that only Q4 will be used.

In this case, equation (1)! i is simplified as follows.

δ=I (DP-DQ2)+(DP-DQ,)I
...(1)' In FIG. 4, the operational amplifiers 40f and 402 take the difference between the density signal DP of the pixel P of interest and the density signal DQ2 of the pixel Q2, and similarly the density signal DP of the pixel P of interest and the pixel The difference from the concentration signal oQa of Q4 is taken by operational amplifiers 403 and 404, and each operational amplifier 4゜l~
By rectifying the output of 404 with diodes 405 to 40B, a current value corresponding to the difference δ is obtained at point A. The amount of charge in the capacitor C increases in accordance with this current value, but the resistor R8 corresponds to the weighting coefficient w2 in equation (2),
Since the attenuation coefficient W in equation (2) is determined by the relationship between capacitor C and resistor R2, the voltage Va at point B reflects the cumulative difference value T in equation (2). The analog comparison circuit 409 is
The threshold voltage vT set by the variable resistor R3 and the voltage vT at points I and B are compared, and if vT>vT, 4, the binary processing signal 411 is turned on, and if VA≦vTH, the inverter 4 is turned on.
10, the multi-gradation processing signal 412 is turned on.

〔effect〕

As is clear from the above description, the following effects can be obtained according to the present invention.

(1) Since it is determined for each pixel whether the pixel of interest is a constituent pixel of text, etc. or a constituent pixel of a photograph, etc., optimal image processing of documents containing a mixture of binary images such as text and multi-gradation images such as photographs is possible in real time. This can be achieved with

(2) The memory used is for a maximum of three lines, and if simplified difference detection is applied, a shift register for at least one line is required, which significantly reduces the memory capacity compared to a method in which the original is divided into multiple blocks.

(3) Since processing is performed pixel by pixel, it is possible to distinguish between a single character and a photograph even in the case of nested text such as a photograph.

[Brief explanation of drawings]

FIG. 1 is an overall configuration diagram of an embodiment of the present invention, FIG. 2 is a diagram showing the positional relationship between a pixel of interest and adjacent pixels, and FIG. 3 is a diagram showing a specific example of the configuration of the main parts in FIG. 1. FIG. 4 is a diagram showing an embodiment of the main parts when the present invention is realized by analog processing. 10... Original document, 11... Image reading device. 12... Line buffer, 13... Image processing device. 131... Difference detection section, 132... Difference accumulation section, 133... Judgment section, 134... Binarization processing section, 135... Multi-tone processing section, 14... Image output device , 15... microprocessor.

Claims (5)

[Claims] (1) In a method for processing images in which binary images such as characters and multi-gradation images such as photographs coexist, the image to be processed is analyzed pixel by pixel to determine whether the pixel is a binary image constituent pixel or not. An image processing method characterized by determining whether a pixel is a gradation image constituent pixel, and performing separate image processing adapted to each of the determined binary image constituent pixel and multi-gradation image constituent pixel. (2) To determine whether a pixel is a binary image constituent pixel or a multi-gradation image constituent pixel, the density of the pixel of interest P is DP_1, the density of the eight pixels Q_1, Q_2,...Q8 adjacent to the pixel of interest P is DQ_1, respectively. When DQ_2,...DQ_8, the density difference δ is δ=|C_0DP+C_1DQ_1+C_2DQ_2+
...+C_8DQ_8| (then, C_0 and C_1
, C_2,...C_8 have different signs, ▲ mathematical formula, chemical formula,
If the cumulative value T of the concentration difference is defined as T = w_1T + w_2δ (where w_1 and w_2 are weighting coefficients), and the cumulative value T is larger than a predetermined value. The pixel of interest P is 2
2. The image processing method according to claim 1, wherein the pixel P of interest is otherwise used as a pixel constituting a multi-tone image as a pixel constituting a value image. (3) Determination as to whether the pixel is a binary image constituent pixel or a multi-gradation image constituent pixel is based on the density of the pixel P of interest and the pixels adjacent to the pixel P of interest.
The largest absolute value of the differences between the densities of pixels Q_1, Q_2, ...Q_8 is selected as the density difference δ, and the cumulative value T of the density differences is defined as T=w_1T+w_2δ (then, w_1 and w_2 are weighting coefficients), and when the cumulative value T is larger than a predetermined value, the pixel of interest P is set to 2.
2. The image processing method according to claim 1, wherein the pixel P of interest is a pixel constituting a value image, and in other cases, the pixel P of interest is a pixel constituting a multi-tone image. (4) 8 pixels Q_1, Q_2 adjacent to the pixel of interest P
, . . . Q_8. (5) The pixels forming the binary image perform binary processing, and the pixels forming the multi-tone image perform multi-tone processing. The image processing method described in Section 4.