JPH11205599A - Image processor and image processing method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processor which performs gradation representation satisfactorily and an image processing method. SOLUTION: A binarization error HG is calculated with HG=D+RG-GSLVB (S5) when an attentional pixel is binarized into a black pixel (S3). Also, a binarization error is calculated with HG=D+RG-GSLVW (S6) when the attentioal pixel is binarized into a white pixel. In such cases, GSLVB =255+(255-D)×HENB, GSLVW=D/HENW, HENB and HENW are constants. Consequently, it is possible to perform γ correction at the same time when error diffusion processing is performed because a binarization error is corrected in accordance with value of density data D.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is applied to a digital copying machine, a facsimile machine, an image scanner and the like, and is used for reproducing halftones applied to multi-value density data of an optically read image. The present invention relates to an image processing device and an image processing method.

[0002]

2. Description of the Related Art A digital copying machine includes a scanner for optically reading a document image, an image processing unit for processing data of the image read by the scanner, and an image processing unit. And an image forming unit that records an image on recording paper based on the image. The image forming unit is configured to form an image by, for example, an electrophotographic process, and includes a photosensitive member, a laser scanning unit for selectively exposing the photosensitive member to write an electrostatic latent image, and a photosensitive member. And a developing unit for developing the above electrostatic latent image into a toner image.

The relationship between the density on the original and the output data of the scanner is not always linear, and the relationship between the data input to the image forming unit and the density of the recorded image formed on the recording paper is not necessarily linear. Absent. Therefore, in order to correct such reading characteristics and developing characteristics and make the relationship between the density of the original image and the density of the recorded image recorded on the recording paper linear, a gamma correction circuit is provided in the image processing unit. Provided. The gamma correction circuit may be used for the purpose of intentionally processing an image in addition to the correction of the reading characteristics and the development characteristics, but such data processing is performed in the original sense of the gamma correction. is not.

[0004] The γ correction circuit is realized by, for example, a memory storing a γ correction table. Specifically, as shown in FIG. 6, the gamma correction table is configured to output the data GOUT after the gamma correction by giving the density value D of the pixel to be processed as an address. The relationship between the density value D and the data GOUT after the γ correction is not linear, the gradient of the data GOUT with respect to the density value D is gentle in a certain density region, and the density value D is Is steeper. In the concentration area where the gradient is gentle,
Γ correction data G having the same value for a plurality of different density values D
OUT is assigned, the gradation of the original data is lost to some extent, and the number of gradations that can be expressed is reduced. This is because of the γ by digital processing using the γ correction table.
Correction is an unavoidable problem.

On the other hand, the image processing section may be provided with an error diffusion circuit for performing an error diffusion process for favorably reproducing a halftone image such as a photographic image. In the error diffusion process, when a certain pixel is binarized, a binarization error which is an error between the multi-value density data before binarization and the density data after binarization of this pixel is calculated. That is, in the binarization processing, the density data of the target pixel is compared with a certain threshold value, and the pixel is determined as a white pixel or a black pixel. For this reason, for a pixel having intermediate density data, an error necessarily occurs between a pixel before binarization and a pixel after binarization.

[0006] This binarization error is distributed to peripheral pixels with appropriate weighting. Then, in the binarization processing for a certain pixel of interest, the density data of the pixel of interest and the binarization error distributed from the peripheral pixels are added, and the addition result is compared with a predetermined binarization threshold. Is done. For example,
As shown in FIG. 2, the above-described binarization error occurs when binarization is performed on a pixel P0 of a certain line along the main scanning direction during reading by a scanner provided in the digital copying machine. I do. This binarization error is caused by adding an error diffusion coefficient of 1/4 or 1/8 to pixels P1 to P6 around the pixel P0.
Multiplied by

Conversely, focusing on the pixel on the side where the binarization error is distributed, as shown in FIG. 3, the pixel of interest A has pixels B, C, D, E, and F around it. , G, the binarization error is distributed. Therefore, attention pixel A
Is performed based on the value obtained by adding the error distributed from the surrounding pixels B, C, D, E, F, and G to the density data of the pixel A.

[0008] In this way, by distributing the binarization error generated in each pixel to peripheral pixels, the expression of halftone is achieved. The binarization process is performed based on a certain threshold value. For example, when binarizing multi-valued data of 256 gradations (white: “0”, black: “255”), the binarization threshold is, for example, an intermediate value “12”.
8 ". Then, the peripheral pixels B, C, D, E,
Based on the cumulative error value, which is a value obtained by accumulating the binarization errors distributed from F and G, and the value of the target pixel, the target pixel becomes a white pixel according to the following equations (1) and (2). Or, it is determined as a black pixel.

Black judgment: (cumulative error value) + (pixel value of interest) ≧ 128 (1) White judgment: (cumulative error value) + (pixel value of interest) <128 (2) After this binarization, the binarization error of the pixel of interest is
It is calculated according to equation (3) or equation (4). That is,
At the time of black determination, the binarization error has a negative value, and at the time of white determination, the binarization error has a positive value.

For black determination: (binary error) = (cumulative error value) + (pixel value of interest) −255 (3) For white determination: (binary error) = ( (Cumulative error value) + (target pixel value) −0 (4)

[0011]

In the above-mentioned prior art, the data GOUT in which the number of gradations is reduced by the gamma correction is used.
Is subjected to the error diffusion processing, but the gradation lost by the γ correction cannot be restored by the error diffusion processing. Therefore, since the expression of the halftone image is performed with the reduced number of gradations, the density change necessarily becomes stepwise, and it is difficult to express the smooth halftone image.

Further, since a memory for storing the γ correction table and an input / output circuit related thereto are required, there is a problem that the circuit scale of the image processing unit is large and the production cost is high. Therefore, an object of the present invention is to provide an image processing apparatus and an image processing method capable of solving the above-described technical problem and performing good gradation expression.

It is another object of the present invention to provide an image processing apparatus and an image processing method capable of favorably expressing gradation with a small-scale circuit.

[0014]

According to the first aspect of the present invention, there is provided an image processing apparatus for binarizing a multi-value density of each pixel constituting an image by an error diffusion process. An apparatus, in which each pixel constituting an image is sequentially regarded as a target pixel, and an added value of a density of the target pixel and a binarization error which is an error generated at the time of binarization in surrounding pixels having a predetermined positional relationship is calculated. An error diffusion processing unit (steps S1, S2, and S3 in FIG. 4) for binarizing the target pixel into a high-density pixel or a low-density pixel based on the binarization reference value. An image processing apparatus comprising: a binarization error calculating unit (steps S5 and S6 in FIG. 4) having a binarization error correction unit that corrects the binarization error according to the density value of the pixel of interest. is there.

According to the above arrangement, the binarization error when the target pixel is binarized is corrected in accordance with the density value of the target pixel. It can be carried out. For example, if a correction corresponding to γ correction for correcting the reading characteristics of the scanner and the recording characteristics of the image forming means is performed on the binarization error, it is not necessary to provide another circuit for γ correction. The configuration of the image processing apparatus can be simplified. That is, a memory for storing the γ correction table and a circuit related thereto are not required. Further, while the number of gradations is reduced in the γ correction using the γ correction table, according to the present invention, since the number of gradations does not decrease, the halftone image is particularly well represented. be able to.

The above-mentioned binarization error correcting means is characterized in that the ratio of the increase in the number of appearances of the high-density pixels to the increase in the density value of the target pixel is smaller in the low-density region than in the medium-density region. Preferably corrects the binarization error so as to be larger than the medium density region. Thereby, substantial gamma correction can be performed.
More specifically, the binarization error calculation means calculates a high value in accordance with the density value D of the target pixel, the cumulative error value RG obtained by adding the binarization errors of peripheral pixels, and the density value D of the target pixel. For a pixel binarized into a density pixel, HG = D + RG−GSLVB, where GSLVB = UL + (UL−D) × HENB,
UL is the upper limit of the density value D, and HENB is a constant.
HG = D + RG−GSLVW where GSLVW = (D−LL) / HENW, LL
Is the lower limit of the density value D, and HENW is a constant. Thus, a binarization error may be obtained.

According to a second aspect of the present invention, there is provided an image processing method for binarizing a multi-value density of each pixel constituting an image by an error diffusion process, wherein each pixel constituting the image is sequentially set as a target pixel. The sum of the density of the target pixel and a binarization error, which is an error generated when binarizing pixels in a predetermined positional relationship around the target pixel, is illuminated with a binarization reference value, and the target pixel is set to a high density pixel or a low density An image processing method comprising: a step of binarizing a pixel; and a step of calculating a binarization error at the time of binarization of the target pixel by adding a correction corresponding to a density value of the target pixel. is there.

Thus, the same effect as the first aspect of the invention can be obtained.

[0019]

Embodiments of the present invention will be described below in detail with reference to the accompanying drawings. FIG. 1 is a block diagram showing an electrical configuration of a portion related to image processing of a digital copying machine to which an embodiment of the present invention is applied. The document set on the digital copying machine is read by the scanner 1. The scanner 1 includes an image sensor for reading an image, such as a CCD image sensor or a CIS image sensor. The image sensor may be an area sensor that reads two-dimensional data, or a niria image sensor that reads line data. Usually, a line image sensor is used to make the device inexpensive.

The image data of the document read by the scanner 1 is supplied to an input interface unit 2, where the signal sample and hold processing is performed. In this embodiment, the input interface unit 2 is configured by an analog circuit, and the above processing is performed in an analog manner. The image data processed by the input interface unit 2 is applied to a digital AGC circuit 3, where a gain control for keeping the level of a signal (image data) within a desired range is performed, and an analog signal is converted to a digital signal. Is performed.

Next, the gain-controlled image data is applied to a shading correction circuit 4 to reduce or eliminate shading distortion. The shading distortion refers to unevenness in illumination of a reading light source when reading an original by the scanner 1 or unevenness in density between pixels due to other causes. The image data from which the shading distortion has been reduced or removed is next provided to the area separating circuit 5. In the region separating circuit 5, the input image data is image data (character data) obtained by reading characters, image data (photo data) obtained by reading a photograph, or a printed photograph, such as newspaper or magazine. It is determined whether the image data is image data (halftone data) of halftone dots obtained by reading the photographic image.

If character data, photograph data and halftone data are mixed in the input image data, area separation of each data is performed. On the output side of the region separation circuit 5, a filter unit 6 that performs different processing for each type of region-separated data is connected. The filter unit 6 includes a differentiation filter 6A, an integration filter 6B, an integration differentiation filter 6C, and a data through filter 6D. Then, the filter unit 6 receives, from the region separation circuit 5, a region separation signal indicating which region each pixel belongs to, and performs an appropriate filter process on each pixel. That is, the character data is stored in the differential filter 6.
A, and a process for sharpening the outline is performed. The halftone data is supplied to the integration filter 6B, and the data is smoothed. Also, the boundary pixels of each region are given to the integral / differential filter 6C, and are subjected to a process for alleviating a sudden change in density of different pixel information (in this case,
Data through filter 6D may be selected. ). Further, data other than the character data and the halftone dot data, that is, the photographic data, is given to the data through filter 6D and sent to the next circuit as it is. The determination of the filter corresponding to the region separation signal can be selected by a register.

The data from the filter section 6 is applied to a zoom / smoothing circuit 9. In the case of enlarging or reducing an image, the zoom / smoothing circuit 9 performs an enlarging or reducing process and a process of correcting image distortion accompanying the enlarging or reducing process. If the image is not enlarged or reduced, the zoom / smoothing circuit 9 does not perform any processing on the image data.

The image data that has undergone the processing up to the zoom / smoothing circuit 9 is then subjected to processing in one of the following circuits depending on the type, and multi-valued data (for example,
Conversion from 8 bits (0 to 255)) to binary data (0 or 1) is performed. In other words, if the image data is photographic data or halftone data and should be subjected to halftone processing, it is provided to the error diffusion circuit 11 and subjected to processing for good halftone expression. so,
It is binarized. In this embodiment, the error diffusion circuit 11
Is characterized by its configuration and processing.

On the other hand, if the image data is character data and data to be subjected to simple binarization processing, it is supplied to the binarization circuit 12. The binarization circuit 12 adjusts a slice level for binarization to distinguish a background from characters, line drawings, and the like. At this time, an automatic density adjustment process is also performed so that the density becomes appropriate. The output of the binarizing circuit 12 is supplied to an isolated point removing circuit 13 to remove isolated black points and white points that appear due to noise or the like.

The binarized data generated by the above processing is applied to an output circuit 14 and output to a printing circuit (not shown). In the apparatus of this embodiment, the gamma correction for correcting the reading characteristics of the scanner and the development characteristics in the image forming unit that forms an image by an electrophotographic process in response to a signal from the printing circuit is performed by the error diffusion circuit 11. This is performed together with the diffusion process.

FIG. 2 is a diagram for explaining the processing in the error diffusion circuit 11. In the error diffusion processing, when a certain pixel P0 is binarized, a binarization error is obtained by a method described later based on the multi-value density data of the pixel P0 and the binarization result (white or black). This is a process of distributing the binarization error to pixels around the pixel P0 at a predetermined distribution ratio. In this embodiment, the binarization error generated at the pixel P0 is caused by the pixels P1 and P5 adjacent to the pixel P0 on the downstream side in the main scanning direction RM and the sub-scanning direction RS when the image is read by the scanner 1, respectively. On the other hand, it is distributed by multiplying the error diffusion coefficient by 1/4. Further, a pixel P2 adjacent to the pixel P1 on the downstream side in the main scanning direction RM, a pixel P5, two pixels P3 and P4 on the upstream side in the main scanning direction RM, and a pixel P5
Is distributed to the pixel P6 adjacent on the downstream side in the main scanning direction RM by multiplying the error diffusion coefficient by 1/8.

The error diffusion circuit 11 sequentially processes each pixel constituting the image as a target pixel. In the error diffusion process for the pixel of interest, an error distributed from surrounding pixels is added to the density data of the pixel of interest, and the result of this addition is subjected to a binarization process. This processing will be described with reference to FIG. When the pixel of interest A is considered as the center, the binarization errors HG (G), HG (E), and HG (D) at the pixels G, E, and D of the previous line L1 are distributed to the pixel of interest A. The distribution is multiplied by a factor of 1/8. Alternatively, the binarization error HG (F) at the pixel F is distributed by multiplying the error diffusion coefficient by 1/4. With respect to the current line L2 including the pixel of interest A, the binarization error HG (B) at the pixel B adjacent on the upstream side in the main scanning direction RM is distributed by multiplying the error diffusion coefficient by 1/4. , The binarization error HG (C) at the pixel C located on the upstream side in the main scanning direction RM is distributed by multiplying the error diffusion coefficient by 1/8.

In the binarization processing for the target pixel A, first, a cumulative error value RG (A) obtained by adding the errors distributed to the target pixel A is calculated according to the following equation (5).

[0030]

(Equation 1) Then, the binarization processing target value T shown in the following equation (6), which is obtained by adding the multi-value density data D (A) given from the zoom / smoothing circuit 9 and the cumulative error value RG (A) for the target pixel A, By comparing (A) with a predetermined binarization threshold TH, binarization for the target pixel A is performed.

T (A) = D (A) + RG (A) (6) FIG. 4 is a flowchart for explaining the processing performed by the error diffusion circuit 11. First, the density data D input from the zoom / smoothing circuit 9 and the accumulated error value RG
Are added, and the binarization processing target value T is calculated (step S1). The binarization target value T and the binarization threshold “1”
28 (step S2), and when the binarization target value T is greater than or equal to the binarization threshold "128", the pixel of interest is binarized to a black pixel (step S3), and the binarization process is performed. When the target value T is less than the binarization threshold “128”, the target pixel is binarized to a white pixel (Step S)
4).

When the target pixel is binarized to a black pixel, the binarization error HG of the target pixel is further obtained by the following equation (7). When the pixel of interest is binarized to a white pixel, the binarization error HG of the pixel of interest is determined by the following equation (9). HG = D + RG-GSLVB (7) GSLVB = 255 + (255-D) × HENB (8) HENB is a constant, for example, “1”. .

HG = D + RG−GSLVW (9) where GSLVW = D / HENW (10) HENW is a constant, for example, “2”. GSLVB and GSLV of the above equations (8) and (10)
W is a reference value when calculating the binarization error, and is a variable that changes according to the density data D. Therefore, the binarization error HG is corrected according to the value of the density data D.

More specifically, the binarization error HG when the target pixel is binarized to a black pixel takes a negative value. On the other hand, the reference value GSLVB when the target pixel is binarized into a black pixel takes the maximum value when the density data D is “128” (minimum density data determined as a black pixel), and the density data D is The larger the value, the smaller the value. Therefore,
As the density data D increases, the absolute value of the binarization error HG decreases more quickly than when the reference value GSLVB is a constant value (for example, “255”). Therefore, as the density data D of the target pixel increases, the number of black pixels appearing in pixels around the target pixel sharply increases.

The binarization error HG when the target pixel is binarized to a white pixel takes a positive value. The reference value GSLVW when the target pixel is binarized to a white pixel takes the maximum value when the density data D is 127, and takes a smaller value as the density data D becomes smaller. But the (10)
Reference value GS determined according to density data D according to the equation
Due to the LVW, the binarization error HG decreases more slowly as the density data D becomes smaller than when the reference value GSLVW is set to a constant value (for example, “0”). Therefore, as the density data D of the target pixel becomes smaller, the number of appearances of black pixels in pixels around the target pixel gradually decreases.

FIG. 5 illustrates the relationship between the original image density and the number of black pixels per unit area (corresponding to the density of the output image) when the above-described error diffusion processing is applied to an original image in a region of uniform density. It is a graph shown typically. The straight line C0 fixes the reference value GSLVB to “255” and sets the reference value GSLVW
Is fixed to “0”. Also, the curves C1,
C2 and C3 are constants HENB and HENW, respectively.
Is set as follows.

C1: HENB = 1, HENW = 2 C2: HENB = 2, HENW = 3 C3: HENB = 3, HENW = 4 The ratio of the change in the number of black pixels to the original image density is large in the high density region and low in the high density region. It is small in the density region and takes a value in the middle density region. By comparing the straight line C0 representing the nilia characteristic with the other curves C1 to C3, it is understood that the appearance of black pixels is suppressed particularly in the intermediate density region. Thus,
Curve input / output characteristics can be realized, and correction of reading characteristics in the scanner 1 and development characteristics in the image forming unit;
That is, γ correction is realized. At this time, unlike the related art using the γ correction table, the number of gradations does not decrease, so that the density change of the halftone image can be expressed smoothly.

As described above, according to this embodiment, γ correction is substantially realized by improving the error diffusion processing without using a γ correction table. Therefore, since the number of gradations does not decrease, it is possible to express a halftone image particularly well. In addition, since a memory for storing the γ correction table and an input / output circuit related to the memory are not required, a circuit configuration for image processing can be simplified, thereby reducing costs. it can.

Although one embodiment of the present invention has been described, the present invention can be implemented in other embodiments. For example, in the above embodiment, a digital copier is taken as an example. However, the present invention relates to an apparatus for processing image data obtained by optically reading an image, such as an image scanner or a facsimile apparatus. Can be widely applied.

In addition, various design changes can be made within the scope of the technical matters described in the claims.

[Brief description of the drawings]

FIG. 1 is a block diagram showing an electrical configuration related to image processing of a digital copying machine to which an embodiment of the present invention is applied.

FIG. 2 is a diagram for explaining an error diffusion process.

FIG. 3 is a diagram for explaining an error diffusion process.

FIG. 4 is a flowchart illustrating a process performed by an error diffusion circuit.

FIG. 5 is an illustrative graph showing a relationship between a document image density and the number of black pixels.

FIG. 6 is a diagram for explaining a conventional γ correction process.

[Explanation of symbols]

 1 scanner 11 error diffusion circuit

Claims (2)

[Claims] An image processing apparatus for binarizing a multi-value density of each pixel constituting an image by an error diffusion process, wherein each pixel constituting an image is sequentially set as a pixel of interest, and the density of the pixel of interest and its surroundings are determined. The target pixel is binarized to a high-density pixel or a low-density pixel by illuminating an addition value with a binarization error, which is an error generated at the time of binarization, in a pixel having a predetermined positional relationship with a binarization reference value. Error diffusion processing means; and a binarization error calculation means having a binarization error correction means for correcting a binarization error at the time of binarization of the pixel of interest in accordance with the density value of the pixel of interest. Image processing apparatus. 2. An image processing method for binarizing a multi-value density of each pixel constituting an image by an error diffusion process, wherein each pixel constituting an image is sequentially set as a pixel of interest, and the density of the pixel of interest and its surroundings are determined. The target pixel is binarized to a high-density pixel or a low-density pixel by illuminating an addition value with a binarization error, which is an error generated at the time of binarization, in a pixel having a predetermined positional relationship with a binarization reference value. An image processing method, comprising: calculating a binarization error when binarizing a target pixel by adding a correction according to a density value of the target pixel.