JPS6143078A - Processing system of original picture - Google Patents

Abstract

PURPOSE:To discriminate properly an intermediate and binary parts with the simple circuit constitution by executing the binary picture processing when a density difference exceeds a range of setting values and executing the half tone picture processing when the set value is within the range. CONSTITUTION:Density data Dn+1 from an A/D converter 1 and density data Dn from an FF2 are inputted as address signals to an arithmetic ROM5. In said ROM5 contents obtained by executing the algorithm arithmetic beforehand are stored. ''1'' at a logical level and zero at the logical level are outputted when said arithmetic is decided to be the binary processing and when it is decided to be the half tone procesing, respectively. After ''1'' or ''0'' deciding signal Gn sequentially outputted from the ROM5 is provisionally latched by a latch 6, it is inputted to a threshold ROM7. The threshold outputted from the ROM7 is inputted to a comparator 8. Density data Dn-1 from an FF3 is inputted to the other terminal of the comparator 8, and it compares these data and outputs a black/white binary signal.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a document image processing method employed in facsimile machines, digital copying machines, etc., and in particular, a method for clearly distinguishing text parts and parts such as photographs and pictures using a simple structure. This invention relates to a method for making a determination.

[Conventional technology]

When reading an image of a document that contains a mixture of parts such as photographs and pictures that are better recorded in halftones and parts such as characters and graphs that are better recorded simply in black and white, it is necessary to If processing is applied, the resolution of the text part will deteriorate, making it difficult to read the text.For this reason, it is necessary to clearly distinguish between the text part and the halftone part of the manuscript, and to distinguish between the text part and the halftone part. Although it is necessary to perform different image processing for each image, conventionally, the following method was used to distinguish the original image: (1) The image data of the read original is passed through a bandpass filter in the form of an analog signal, The amount of high frequency components in the output is checked, and a portion where a large amount of high frequency components are present is determined to be a character portion.

(2) Convert the image data of the read and captured original into multivalued digital data, and use the multivalued digital data to convert the pixel of interest and the 8 pixels in the same 3x3 area centered on the pixel of interest. Each density variation with respect to the target pixel is checked, and based on the variation, it is determined whether or not it is a text portion.

However, the method (1) above generally has the disadvantage of poor judgment accuracy, and the method (2) above is superior in terms of judgment accuracy, but requires complex calculation processing. There was a problem that the circuit scale became large and the cost increased.

[Problem that the invention seeks to solve]

This invention has been made to solve the above problem, and is intended to accurately distinguish between the halftone part and the binary part with a simple circuit configuration. and operation] In the present invention, the density level Dn of the pixel of interest an and the density level DK of one pixel an+t adjacent to the pixel of interest an in the main scanning direction.
When the concentration difference 1 pn with l+1 is once larger than the first set value ca (1Dn-Dn+tl >aa) or the concentration difference is greater than the first set value Cd (IDn-Dn+t
l≦Cd) and the average density is the second set value CII and the third set value QwDn+Dn+t
It is outside the range of Dn+Dn++ (・20w
or □≦CB), it is determined that it is a character image, and binarization is performed using a normal fixed slice level.
n-1-tl≦ca) and the average concentration is within the range of the second set value am and the third set value OW D D + D n
+1 (CB<<OW), it is determined that the image is a photograph or painting, and halftone image processing is performed using a dither method using a dither matrix, for example.〇 [Example] In Fig. 2 , Attention pixel 1! The density level of L11 is n, and the density levels of pixels an-1 and an'+1 adjacent to each other in the main scanning direction are Dn ``It'' + 1. In the present invention, the pixel of interest a!I and the adjacent pixel a n-1
By examining only the density state with -1, the target pixel &
fl[Determine whether to perform halftone processing or binarization processing◎ The control flow for this determination is shown in FIG.

First, the density difference 1 between the pixel of interest 6n and the adjacent pixel atl+1
Dn-Dn+11 (=D) is calculated, and if the density difference is greater than the set value 0 (L), it is determined that the image is a character image, and normal binarization processing with a fixed slice level is performed. If the value is less than the set value 0 (L), the pixel of interest an and the adjacent pixel an+ are determined, and nx within a minute area is determined.
Halftone processing is performed using a dither method in which the black and white determination slice level of n pixels is changed for each pixel. If the average density 0"+9°1" is a value outside the range of the set values OB and C., it is determined that the image is a character image, and normal binarization processing is performed.

The above set values Od, C+a and Qv can be set to various values depending on the type of document, and an example thereof is shown in FIG.

In Figure 4, image data is A/D converted to digital data with 16 gradations from "1" to "16", and Cde=9.
, 0B=3, 0v=10. Halftone processing is performed in the shaded areas in the diagram, and normal binarization processing is performed in other areas. The density level pn of pixel an is "8" and the adjacent pixel all++1.

If the density level Dn+t is "8", halftone processing is performed, and the density level Dn of the pixel of interest am is "13".
Then, the density level Dfi+1 of the adjacent pixel an+1 is “15
”, binarization and processing are performed.

Next, FIG. 1 shows a specific configuration example of the present invention.

In FIG. 1, a document image read by a solid-state image pickup device (not shown) such as a COD is sequentially input to a MU/D converter 1 pixel by pixel.

A/D converter 1 converts the input analog data into 16
It performs analog-to-digital conversion into digital data, and its 4-pit parallel output is input to the 7-lip flop 2 at the first stage of the shift register 4. The shift register 4 constitutes a two-stage shift register by flip-flops 2 and 3, and when the output of the flip-flop 3 is the density data Dn-* of the pixel an-1, the output of the flip-flop 2 is the pixel of interest. am concentration data Dn, and the output of the A/D converter 1 is the pixel a n −
) The output of the A/DK converter 1 and the output of the flip-flop 2, which are density data Dn+t of 1, are input to the arithmetic ROM 5 as an address signal. The calculation ROM 5 stores the contents of all the calculations of the algorithm shown in FIG.
When it is determined that it is value processing, a logic level of "1" is output, and when it is determined that it is halftone processing, a logic level of "0" is output.
" is output. That is, the set value Od of the algorithm shown in FIG.

OB and Qv are each "9" as shown in Figure 4.
, "3" and "10", each unit in the shaded area shown in FIG.
In order to calculate the ratio of the signal, @arithmetic i(0M5's memory content is determined. ◎If the memory content is like this, for example, D n-1-1 [signal of gradation level "6" is output. When a signal of ``8'' is output at the gradation level as K D n, a signal of ``O'' at the logic level is output from the X ROM 5, and D n -1-
When a signal of gradation level "2" is output as t and a signal of gradation level "2" as Dn, a signal of logic level "1" is output from calculation R6M5.

The determination signal ()n of "0" or "l" sequentially output from the calculation ROM 5 is once latched by the latch 6 and then input to the width value ROM 7. Therefore, when the flip-flop 3 outputs the density data Dn-s of the pixel an-t, the latch 6 outputs the determination signal Gn-1 of the pixel ILn-1.Threshold RQM7 is A fixed one-step value for binarization processing and a dither matrix for halftone processing are stored in advance, respectively, and a judgment signal G! is sent from the latch 6. When a signal with logic level "l" is output as l-1, the fixed threshold value used for binarization processing is output, and when a signal with logic level "0" is output, it is output in the dither matrix used for halftone processing. The signal output from the declared value ROM 7 is input to the comparator 8. The other input terminal of the comparator 8 receives the pixel an- from the flip-flop 3.
1 concentration data DXI-s is input, and the comparator 8
compares these and outputs a black and white binary code.

Snawachi, density data D of pixels corresponding to the character image from flip-flop 3! When 1-1 is output, the comparator 8 performs binarization processing using a normal fixed slice level, and the flip-flop 3 outputs density data Da of pixels corresponding to a halftone image such as a photograph. When -1 is output, the comparator 8 performs binarization processing using slice levels based on a dither matrix.

By the way, when we conducted experiments by appropriately changing the set values oa, on, and Ow to various values, we found that the accuracy of judgment was comparable to that of the conventional method that refers to all 3×3 surrounding pixels. I understand.

〔Effect of the invention〕

As explained above, (1) According to the present invention, it is possible to determine whether an image is a character image or a picture by using a simple calculation method based on the density information of only the pixel of interest and one pixel adjacent in the main scanning direction. In order to be able to accurately determine
This provides an excellent effect in that the circuit scale of the determination circuit can be reduced without reducing the reliability of the determination, and as a result, the cost of the device can be reduced.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the configuration of an embodiment of this invention, Fig. 2 is an explanatory diagram showing three adjacent pixels, Fig. 3 is a flowchart showing the judgment algorithm of this invention, and Fig. 4 is judgment based on the same flowchart. 0 which is a figure showing an example of an aspect

Claims (1)

[Claims] Find the density difference and average density between the pixel of interest and one pixel adjacent to the pixel of interest in the main scanning direction, and when the density difference is greater than a first set value or the density difference is less than or equal to the first set value. and binary image processing is performed when the average density is outside the range of the second and third set values, and the density difference is less than or equal to the first set value and the average density is outside the range of the second and third settings. A document image processing method characterized in that halftone image processing is performed when the value is within a range of values.