JPH05143728A - Image processor - Google Patents

Abstract

PURPOSE:To provide image processor which reduce granular noise generated in the case of reading a recycled paper original without affecting the resolution or reproducibility of an original image. CONSTITUTION:In respect to the original image read by an image reading part 11, the correcting amount of an attention picture element is calculated by a correcting amount calculation part 13, further, a prescribed coefficient is multiplied by a coefficient multiplication part 14, and the code of the correcting amount is decided by a code deciding part 15. A granular noise detection processing part 12 detects the granular noise generated in the case of reading the recycled paper original and based on the result and the judged code, a selecting part 16 selects either the calculated correcting amount itself or the correcting amount multiplied the coefficient. Then, the selected correcting amount is added to the attention picture element by an adder 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus such as a facsimile or a digital copying machine which handles recycled paper as an original.

[0002]

2. Description of the Related Art When reading an image having a background color such as a newspaper or a blueprinted original, a conventional image processing apparatus creates a density histogram by, for example, prescanning, and binarizes the read image from the result. A method of determining a threshold value when performing, or a process of detecting an average value of image densities in a block unit composed of a plurality of pixels and dynamically changing the binarization threshold value (so-called ABC: Automa
The background noise is removed by tic background control processing, etc.).

[0003]

Recently, recycled paper has been attracting attention for the purpose of effective use of resources. However, since recycled paper contains a lot of granular noise peculiar to high-quality paper, when an image on recycled paper is read by a conventional facsimile or the like and binarized, granular noise occurs in the background, There have been problems such as deterioration of image quality and reduction of compression rate at the time of encoding.

For such recycled paper originals, the above-mentioned ABC
Even if the (automatic background control) processing and the like were performed, it was not sufficient to remove the high-density granular noise. Also,
When resolution correction processing (so-called edge enhancement processing) is not performed in order to eliminate the influence of granular noise, the resolution of the read image is reduced, and when the binarization threshold value is changed. There is a problem that a document with a low density level cannot be reproduced.

The present invention has been made to solve the above-mentioned problems, and an image for reducing the granular noise generated when a recycled paper original is read without affecting the resolution and reproducibility of the original image. The purpose is to provide a processing device.

[0006]

[Means for Solving the Problems] and

In order to achieve the above object, the image processing apparatus of the present invention has the following configuration. That is, in an image processing apparatus that performs edge enhancement processing on each pixel of a read original image, a detection unit that detects granular noise generated when a recycled paper original is read, and edge enhancement processing is performed on the read pixel. Calculation means for obtaining a correction amount for the calculation, determination means for determining a sign of the correction amount obtained by the calculation means, multiplication means for multiplying the correction amount by a predetermined coefficient, and a determination result of the detection means and the determination means. It has a selection unit for selecting either the correction amount or a correction amount multiplied by a coefficient, and an addition unit for adding the correction amount selected by the selection unit to the target pixel.

[0007]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described in detail below with reference to the accompanying drawings. FIG. 1 is a block diagram showing the basic arrangement of the image processing apparatus according to this embodiment. In the figure, reference numeral 11 denotes an image reading unit having a function of A / D converting the read analog data into digital data after shading correction and the like. In the following embodiments, the image data obtained here will be described as density data. That is, the case where the data value is large is treated as the black side. A granular noise detection processing unit 12 determines whether or not the pixel of interest from the read image data is the granular noise generated when the recycled paper document is read. A correction amount calculation unit 13 for edge enhancement processing calculates a correction amount for correcting the target pixel value from the target pixel and its peripheral pixels. 14
Is a coefficient multiplication unit, which multiplies the correction value obtained by the correction amount calculation unit 13 by a predetermined coefficient of "1" or less.

Reference numeral 15 is a code determination unit, which determines the code of the above-mentioned correction value. Reference numeral 16 is a selection unit, which is a code determination unit 1
The correction amount to be added to the read image data is selected from the determination result of No. 5 and the determination result of the granular noise detection processing unit 12.
Specifically, when the sign determination result is negative, the result obtained by the correction amount calculation unit 13 is selected regardless of the granular noise detection result. On the other hand, when the sign determination result is positive, a desired correction amount is selected according to the granular noise detection result. That is, when it is determined that the pixel of interest is noise, the correction amount in the coefficient multiplication unit 14 is selected, and when it is determined that the pixel of interest is an effective pixel (that is, other than granular noise), correction is performed. The correction amount in the amount calculation unit 13 is selected. Then, the selected correction result is read by the adder 17 and added to the image data, so that the edge enhancement processing for reducing the influence of the granular noise is performed.

When a recycled paper original is targeted, the granular noise only needs to consider isolated black pixels, and does not need to consider isolated white pixels. Therefore, in the present embodiment, as described above, the edge enhancement amount is controlled using the grain information noise determination result only when the correction value to be added to the read image is positive (only the occurrence of isolated black pixels is reduced. Let me). Next, the processing of each unit described above will be described in detail. Figure 2
FIG. 6 is a diagram showing a timing example of an image signal output from the image reading unit 11. As shown in the figure, the read image data includes one pixel data in synchronization with an image clock, one line of image data in synchronization with a line synchronization signal, and one page of image data in synchronization with a page synchronization signal. Image data shall be input respectively.

FIG. 3 is a diagram showing the structure of the granular noise detection processing unit 12 in this embodiment. In the figure, reference numeral 31 is a pixel block extraction unit, which extracts pixels to be referred to in order to detect granular noise. Specifically, the reference pixel is 3 × shown in FIG.
If it has three pixels, it is realized by the configuration shown in FIG.
51a-i are 1-pixel delay elements, 52a-b are 1-line delay elements. FIG. 6 shows a timing chart of the pixel block extraction unit 31. Each signal corresponds to A to I shown in FIG. 4, and here, the case where the number of main scanning pixels is “8” is shown. As shown in FIG. 6, by passing the input image data through each delay element, it is possible to refer to a plurality of necessary pixels at the same timing. An average value calculator 32 calculates an average density in a block (a rectangular area including a reference pixel) using a plurality of extracted reference pixel values. When referring to the pixel shown in FIG. 4, specifically, the following calculation is performed.

I _AV = (I _A + I _B + I _C + I + _D + I <sub>E
+ I F + I G + I</sub> H + I I) / 9 where, I _AV: a block average density, I _N: is the concentration of Gasomoto N. A threshold value calculation unit 33 calculates a threshold value for binarizing the image data in the pixel block based on the average density calculated by the average value calculation unit 32. Specifically, as an example, the threshold value TH is calculated by the following calculation.
Ask for.

TH = I _AV + α Here, α is a correction value (α << I _AV ) and the detection level can be controlled by the value of the correction value α. Three
A binarization processing unit 4 binarizes the pixel of interest in the block by the threshold value calculated by the threshold value calculation unit 33. In this way, by performing the dynamic binarization process using the average value in the block, it is possible to obtain a binary image in which high frequency components are emphasized.

That is, the target granular noise can be detected. For example, as an example, the reference pixel in the block is shown in FIG.
In the case of the value shown in (the pixel value is assumed to be a value of 0 to 64), when the binarization processing is performed with the normal threshold value (TH = 32), the pixel of interest is determined to be white, and the granular noise is detected. Not done. However, when the binarization processing is performed using the average density in the block, the threshold value becomes TH = 4 (when α = 0), the target pixel is determined to be black, and the pattern matching processing unit described later is used. At 35, granular noise can be detected. 35
Is a pattern matching processing unit, and is a binarization processing unit 34.
A pattern of granular noise is detected from the image pattern in the binarized block by template matching.
As an example, the pattern shown in FIG. 8 is determined to be granular noise. As shown in FIG. 9, for example, the pattern matching processing unit 35 can be easily realized by a pixel block extraction unit 91 and a ROM (read only memory) 92 that have the same configuration as that described in the average value calculation unit 32. is there. This R
The reference pixel is addressed to the OM 92 (9 bits in this case)
Then, the determination result is programmed in advance as the memory content for the address. Then, the pattern matching result is obtained by sampling the ROM data output with the above-described image clock.

Next, the above-mentioned correction amount calculation unit 13 operates as shown in FIG.
As shown in 0, a pixel block extraction unit 101 that extracts a reference pixel (the configuration is the same as that shown in FIG. 5), and correction for edge enhancement of the pixel of interest from the pixel value of interest and its surrounding reference pixel value Correction amount calculation processing unit 102 for calculating the amount
It consists of and. Specifically, in the calculation by the Laplacian filter of the coefficient as shown in FIG. 11, the correction amount is calculated by the following equation with E shown in FIG. 4 as the pixel of interest.

[0015] _{E E = {4 × I E} - (I B + I D + I F + I H)} / 4 where, E _E: is a correction amount of the pixel E, I _N: is the concentration of pixels N. The arithmetic expression can be easily realized by the wired logic. In addition, edge enhancement is performed by adding the correction amount obtained here to the target pixel. As a specific example, when the above calculation is performed on the pixel shown in FIG. 7, the correction value for the pixel of interest is: E _E = {4 × 20− (2 + 2 + 2 + 1)} / 4 = 18
When the above value is added to the target pixel as it is (target pixel value = 38), it is reproduced as a black pixel by the binarization process.

In the coefficient multiplying unit 14, a multiplication circuit multiplies the calculated correction amount by a coefficient smaller than "1" which is determined in advance. That is, the edge enhancement processing with the reduced edge enhancement amount is performed. Specifically, when the coefficient is 1/2, the pixel value of interest in the image shown in FIG. 7 is (20 + 9 = 29).
Therefore, it will not be reproduced by the binarization process. next,
The sign determination unit 15 determines whether the correction amount calculation result E _E is positive or negative. Specifically, the sign bit of the operation data is output.

The selection unit 16 is a data selector, and selects and outputs the above-described two correction amounts according to the determination results of the code determination unit 15 and the granular noise detection processing unit 12. Table 1 shows the selection conditions.

[0018]

[Table 1]

That is, when the sign of the correction value is positive and the granular noise is detected, the correction value multiplied by the coefficient is selected (the edge emphasis amount is reduced). The edge enhancement process is performed by adding the selection result here to the original image by the adder 17. Through the above processing, the edge emphasis processing for reducing the granular noise seen in the recycled paper document is performed. In the case of a facsimile, for example, the processed output image data is coded after the binarization process and sent to the communication line.

Next, the processing of this embodiment will be described with reference to the flowchart shown in FIG. First, image data is read by the image reading unit 11 (step S121), and a correction amount for edge enhancement is calculated (step S122). Then, the sign determination is performed on the obtained correction value (step S123), and when the sign is negative, the correction value is added as it is to the read pixel (step S124). But,
If it is positive, it is determined whether or not the pixel of interest is granular noise according to the result of the granular noise detection process (step S125) (step S126). As a result, if it is determined to be granular noise, a coefficient is added to the calculated correction value. It is multiplied by (<1) (step S127), and the correction value multiplied by the coefficient is added to the target pixel (step S128). On the other hand, when it is not determined that the noise is granular noise, the calculated correction value is directly added to the target pixel (step S124). The above processing is performed for all pixels (step S129), and the processing ends.

As described above, according to the present embodiment, the influence on the image resolution and reproducibility is minimized by the simple pattern matching processing, and the grain information noise when the recycled paper original is read is reduced. It becomes possible to do.

[0022]

Other Embodiments In this embodiment, the reference block for detecting the granular noise has been described by taking 3 × 3 pixels as an example, but the present invention is not limited to this, and any size window may be used. Good. The determination pattern is not limited to that shown in the embodiment, and various patterns are conceivable, and the present invention is not limited to the above-mentioned embodiment.

Further, the above-mentioned pattern matching is not performed by matching with the granular noise pattern, but conversely, the edge pattern can be detected, and the judgment can be made by considering other than the edge portion as noise. In this case, for example, a pattern as shown in FIG. 13 is detected, and other patterns are processed as a noise detection block. Further, the granular noise detection processing can be performed not only by making a determination based on a binary pattern as shown in the embodiment, but also by directly detecting from the density gradient of multi-valued image data in a block. For example,
In the case of the reference pixel shown in FIG. 4, detection may be performed by the following determination formula.

[0024] _{D = SIG (I E -I A} ) · SIG (I E
_{-I B) · SIG (I E} -I C) · SIG (I E -I
_{D) · SIG (I E -I} F) · SIG (I E -I G) ·
Here _{SIG (I E -I H) ·} SIG (I E -I I), SIG (X): 1 ... (X ≧ 0), 0 ... (X <
0), and _IN : the density of the pixel N.

When D = 1 in the above determination formula, it is determined to be a noise image. That is, when the pixel of interest has an extreme value, it is determined to be granular noise. Further, in the above-described embodiment, the image data is described as the density data, but the same applies only when the brightness data is used, but the sign determination result is reversed.
The present invention may be applied to a system including a plurality of devices or an apparatus including a single device.
It goes without saying that the present invention can also be applied to the case where it is achieved by supplying a program to a system or an apparatus.

[0026]

As described above, according to the present invention, it is possible to suppress the influence on the resolution and reproducibility of the read image and reduce the influence of the granular noise generated when the recycled paper original is read. Become.

[Brief description of drawings]

FIG. 1 is a block diagram showing a basic configuration of an image processing apparatus in this embodiment.

FIG. 2 is a diagram showing a timing of an input pixel image signal.

FIG. 3 is a diagram showing a configuration of a granular noise detection processing unit.

FIG. 4 is a diagram showing positions of reference pixels on an image.

FIG. 5 is a diagram showing a configuration of a pixel extraction unit.

FIG. 6 is a time chart of a pixel extracting unit.

FIG. 7 is a diagram showing a specific example of a granular noise image.

FIG. 8 is a diagram showing a granular noise pattern.

FIG. 9 is a diagram showing a configuration of a pattern matching processing unit.

FIG. 10 is a diagram showing a configuration of a correction amount calculation processing unit.

FIG. 11 is a diagram showing coefficients of an edge enhancement filter.

FIG. 12 is a flowchart showing a process in this embodiment.

FIG. 13 is a diagram showing an edge pattern in another example.

[Explanation of symbols]

 11 Image Reading Section 12 Granular Noise Detection Processing Section 13 Correction Amount Calculation Section 14 Coefficient Multiplying Section 15 Code Determining Section 16 Selection Section 17 Adder

Claims (1)

[Claims] 1. An image processing apparatus for performing edge enhancement processing on each pixel of a read original image, a detection unit for detecting granular noise generated when a recycled paper original is read, and for the read pixel. The calculating means for obtaining the correction amount for edge enhancement, the judging means for judging the sign of the correction amount obtained by the calculating means, the multiplying means for multiplying the correction amount by a predetermined coefficient, the detecting means and the judging means. An image comprising: a selection unit that selects either the correction amount or a correction amount multiplied by a coefficient according to the determination result; and an addition unit that adds the correction amount selected by the selection unit to the pixel of interest. Processing equipment.