JP3893891B2 - Image processing device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to an image processing apparatus that performs image processing using an error diffusion method. More specifically, the present invention relates to an image processing apparatus that reduces an error-diffused image by thinning out pixels.
[0002]
[Prior art]
Conventionally, in an image processing apparatus, when reducing the number of gradations of input image data, for example, when binarizing image data of 256 gradations, an error diffusion method has been widely used. The error diffusion method calculates the density difference (error) between the density of the input image and the density of the output image for each pixel, performs a specific weighting operation on this error, and then distributes the density to the surrounding pixels to obtain the density. In this method, binarization is performed while the image is stored. As a technique for reducing an image, a thinning process for thinning out pixels is widely used because the process is simple.
[0003]
For this reason, conventionally, as the image reduction process, the error diffusion binarization process is performed after the error diffusion binarization process as shown in FIG. 9, or the error diffusion binarization is performed after the reduction process as shown in FIG. We were processing. However, in the process shown in FIG. 9, the density balance differs between the error diffusion binary image and the reduced binary image. In addition, thin lines and the like that existed in the error diffusion binary image are erased by the thinning process. On the other hand, in the process shown in FIG. 10, the density can be saved, but the thin lines and the like that existed in the case of the multi-valued image are erased by the thinning process. As described above, the reduction processing of the image has a problem that the density cannot be stored and the gradation reproducibility is deteriorated or the thin line is interrupted and reproduced.
[0004]
Thus, one technique for solving this problem is disclosed in, for example, Japanese Patent Laid-Open No. 5-63963. In this technique, binarization is performed by an error diffusion method using error data of pixels that are not thinned out, so that the binarization error is propagated between pixels that are not thinned out at the time of reduction. Thereby, a reduced image can be obtained without deteriorating the gradation reproducibility.
[0005]
[Problems to be solved by the invention]
However, in the technique disclosed in Japanese Patent Laid-Open No. 5-63963, binarization processing is performed using pixel data that is not thinned out as image data used for binarization processing, and then thinning processing is performed. ing. In other words, as a result, this technique does the same as the processing shown in FIG. Therefore, the image quality is the same as that obtained by the processing shown in FIG. For this reason, the problem that a thin line etc. are interrupted and reproduced is not solved. This is because information on the thinned pixels is discarded without being reflected in the image processing at all, and information on the thinned pixels is lost.
[0006]
Accordingly, the present invention has been made to solve the above-described problems, and is an image processing apparatus that can obtain a reduced image that has no loss of image information and is excellent in gradation reproducibility and that has been stored in density. It is an issue to provide.
[0007]
[Means for Solving the Problems]
An image processing apparatus according to the present invention, which has been made to solve the above problems, includes N-value conversion means for converting M-value image data into N (M> N) values for each pixel by an error diffusion method, and N-value conversion. A reduction means for reducing the N-valued image by means of pixel thinning, and a pixel that is not thinned out by the reduction means from the N-valued image by the N-value reduction means, and using the extracted pixels N-value average calculating means for calculating the average value of N-value image data around the target pixel, M-value average calculating means for calculating the average value of M-value image data around the target pixel, and N-value average calculating means A difference calculating unit that calculates a difference between the average value calculated and the average value calculated by the M value average calculating unit, and a correcting unit that corrects the image data of the M value based on the calculation result of the difference calculating unit, The reduction means is the M-value image data corrected by the correction means. N-value conversion means is characterized in reducing the image N-valued. Note that when calculating the weighted average error required by the N-value conversion means, only the N-value conversion error of pixels that are not thinned out is referred to. Here, the pixels that are not thinned out mean pixels that remain after the reduction process.
[0008]
In this image processing apparatus, when M-value image data is input, the average value of the M-value image data around the target pixel is calculated by the M-value average calculation means. Further, the M-value image data is converted to N-value by the error diffusion method for each pixel by the N-value conversion means. That is, processing for reducing the number of gradations of the input image is executed. Then, the N-value average calculating means extracts pixels that are not thinned out by the reducing means from the image that has been N-valued by the N-value converting means, and the extracted pixels are used to generate N-value image data around the target pixel. The average value of is calculated. Then, the difference calculation means calculates the difference between the average value calculated by the N value average calculation means and the average value calculated by the M value average calculation means. Then, the correction unit corrects the M-value image data based on the calculation result of the difference calculation unit. That is, the difference calculated by the difference calculating means is fed back to the M-value image data as error data. Then, the corrected M-value image data is converted to N-value by the N-value conversion means to obtain an N-value image. When this N-value image is obtained, it is reduced by the reduction means. Thus, a reduced image is obtained.
[0009]
As described above, according to the present image processing apparatus, the corrected M-value image data is converted into an N-value, and the N-value image is reduced. The correction of the M-value image data is performed based on the difference between the average value calculated by the N-value average calculation means and the average value calculated by the M-value average calculation means. For this reason, the information of the pixels to be thinned out is reflected in the corrected M value data. Therefore, even if the thinning process is performed, the pixel information to be thinned is not lost. As a result, it is possible to obtain a reduced image having no gradation of image information and excellent in gradation reproducibility in which density is preserved. That is, a reduced image in which fine lines are reproduced without interruption can be obtained.
[0010]
Here, in the image processing apparatus according to the present invention, it is desirable that the N-value average calculating means calculates the average value by the filter processing and switches the coefficient in the filter in accordance with the magnification of the reducing means. Thereby, the feedback amount to the M-value image data can be appropriately changed according to the magnification. Therefore, even when the magnification is changed, a reduced image with excellent gradation reproducibility in which density storage without image information loss is made can be obtained.
[0011]
DETAILED DESCRIPTION OF THE INVENTION
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, a most preferred embodiment embodying an image processing apparatus of the invention will be described in detail with reference to the drawings. In this embodiment, the image processing apparatus according to the present invention is applied to a digital copying machine.
[0012]
The digital copying machine according to the embodiment includes an image reader unit that reads an image of a document, and a print unit that prints out the read image on a recording sheet. Among these, the present invention is applied to the image reader unit. Therefore, the schematic configuration of the image reader unit is shown in FIG. 1, and the configuration of the control system is shown in FIG. As shown in FIG. 1, the image reader IR includes a document placing glass 2 on which the document D is placed, a document cover 1 for pressing the document D placed on the document placing glass 2, a main cover A scanner 3 having a light source L and a mirror 4a elongated in the scanning direction and moving in the sub-scanning direction to scan the entire document D, a slider driving unit 7 for controlling the movement of the scanner 3, and an analog image signal are obtained. Setting information from the CCD sensor 6, mirrors 4 b and 4 c for guiding the reflected light from the document D to the CCD sensor 6, the lens 5, the image processing unit 20, the operation panel 10 unit, and the operation panel unit 10 And a control unit 30 for overall control of the entire copying machine. The output from the image processing unit 20 is input to the image output unit 40.
[0013]
In the present embodiment, the print unit PR corresponds to the image output unit 40. However, the print unit IR is an example of the image output unit 40, and the image output unit 40 may be an external printer, an image memory, or the like.
[0014]
Here, the scanner 3, the mirrors 4b and 4c, the lens 5, and the CCD sensor 6 constitute an image input unit 7 shown in FIG. The image input unit 7 acquires image data. Specifically, reflected light obtained by scanning the document D by the scanner 3 is received by the CCD sensor 6 and is subjected to photoelectric conversion to obtain analog image data. The analog image data acquired by the CCD sensor 6 is transmitted to the image processing unit 20.
[0015]
The image processing unit 20 performs various processes on the image data transmitted from the image input unit 7. As shown in FIG. 3, the image processing unit 20 includes an A / D conversion unit 21, a shading correction unit 22, a reflectance / density conversion (hereinafter referred to as “LOG conversion”) processing unit 23, and an MTF process. A unit 24, a binarization unit 25, and a scaling processing unit 26 are provided.
[0016]
Here, the A / D converter 21 converts analog image data input from the CCD sensor 6 into digital image data. The shading correction unit 22 removes light amount unevenness in the main scanning direction on the image. Specifically, before the reading operation of the document D, the reflected light from the white plate for shading correction is received by the CCD sensor 6, the analog data obtained there is converted into digital data, and then the digital data is stored in the memory. Remember. When the document D is read, the read data of the document D is corrected using the digital data stored in the memory as a reference value.
[0017]
The LOG conversion processing unit 23 performs LOG conversion using a lookup table in order to obtain density scale image data. The MTF correction unit 24 corrects the sharpness of the image. The binarization unit 25 converts multi-value image data (256 gradations in the present embodiment) into binary image data using an error diffusion method. Details of the binarization unit 25 will be described later. The scaling processing unit 26 performs image enlargement / reduction processing by controlling memory writing and reading operations.
[0018]
The image data processed in each of the processing units 21 to 26 is transmitted from the image processing unit 20 to the image output unit 40. That is, as shown in FIG. 1, in this embodiment, the image data processed by the image processing unit 20 is output to the printing unit PR (an example of the image processing unit 40). Then, the image of the document D is reproduced on the recording paper by the printing unit PR.
[0019]
An operation panel unit 10 shown in FIG. 1 is used for setting conditions necessary for copying such as an image mode, a document size, and an exposure level by an operator, as well as occurrence of a jam and paper empty. It also informs the abnormal state of the copying machine. As shown in FIG. 4, the operation panel unit 10 includes a liquid crystal touch panel 11, a numeric keypad 12 for inputting the number of copies, a copy magnification, etc., a reset key 13 for returning various setting values in the copying machine to standard values, and a copy operation. A stop key 14 for stopping, a start key 15 for starting a copying operation, and a mode switching key 16 for switching a copy mode are provided. Here, the liquid crystal touch panel 11 displays not only the operation state of the copying machine such as the exposure level, the copying magnification, and the recording paper size, but also information related to the abnormal state of the copying machine such as the occurrence of a jam and the occurrence of paper empty. It is like that. The liquid crystal touch panel 11 can also specify an exposure level, a copy magnification, a recording paper size, and the like.
[0020]
Here, the binarization unit 25 shown in FIG. 3 includes an error diffusion binarization circuit shown in FIG. The error diffusion binarization circuit includes an average value calculation unit 60 for calculating an average value around the target pixel (multi-value average data Iav (n)) based on the input image data I (n), an adder 61, Comparator 62 that compares corrected image data Ir (n) output from adder 61 with threshold value TH and outputs binarized data B (n), and binarized data B output from comparator 62 The binarization error E (() is obtained from the selector 63 that selects the binarization reference value (HB, LB) according to (n), the output Ir (n) of the adder 61, and the output Br (n) of the selector 63. n), an error storage memory 65 for storing the binarization error E (n) output from the subtractor 64, and an error calculation unit 66 for calculating the weighted average error Eav (n). And a binary storage memory 71 for storing the binarized data B (n) The weighted binary average data Bav (n) is calculated by the binary average calculating unit 72, the multi-value average data Iav (n) calculated by the average value calculating unit 60, and the weighting calculated by the binary average calculating unit 72 And a subtractor 73 for calculating average difference data AVr (n) that is a difference from the binary average data Bav (n). Each of the error storage memory 65 and the binary storage memory 71 is provided with a plurality of FIFO memories, and can store data for several lines in the sub-scanning direction.
[0021]
The adder 61 receives input image data I (n), average difference data AVr (n), and weighted average error Eav (n). The adder 61 outputs the sum of these data as corrected image data Ir (n). That is, in the adder 61, the average difference data AVr (n) and the weighted average error Eav (n) are fed back to correct the input image data I (n).
[0022]
The comparator 62 outputs “1” as the binarized data B (n) when the corrected image data Ir (n) is larger than the threshold value TH, and “0” as the binarized data B (n) otherwise. "Is output. The selector 63 selects the binarized reference value HB when “1” is output from the comparator 62, and selects the binarized reference value LB when “0” is output. It is like that. The reference values HB and LB are predetermined fixed values, and in this embodiment, HB = 255 and LB = 0 are set.
[0023]
In addition, the average value calculation unit 60 calculates the multi-value average data Iav (n) based on the following equation using a 5 × 5 matrix size average value filter as shown in FIG. 6.
Iav = (ΣW_ij× I_ij) / ΣW_ij
Here, W is a filter coefficient, the suffix “i” indicates an address in the sub-scanning direction, and “j” indicates an address in the main scanning direction. Note that the filter shown in FIG. 6A is used when the magnification in the scaling processing unit 26 is “1/2”, and the filter shown in FIG. 6B is the scaling processing unit 26. This is used when the magnification at is 1/4.
[0024]
Returning to FIG. 5, the error calculation unit 66 performs weighting using a peripheral error weighting filter having a matrix size corresponding to the magnification of the scaling unit 26 based on the binarization error E (n), and the weighted average error Eav. (N) is calculated. The numerical value shown in the peripheral error weighting filter is a weighting coefficient, and the mark “*” in the filter indicates the pixel currently being processed (target pixel). The weighted average error Eav (n) is input to the terminal C of the adder 61. Note that the weighting coefficient is set to be larger as it is closer to the pixel of interest, and the sum is “1”.
[0025]
The binary average calculation unit 72 performs weighting using a weighting filter having a matrix size corresponding to the magnification of the scaling processing unit 26 based on the binarized data B (n), and weighted binary average data Bav (n ) Is calculated. The numerical value shown in the weighting filter is a weighting coefficient, and the mark “*” in the filter indicates the pixel currently being processed (target pixel). The weighted binary average data Bav (n) is input to the terminal B of the subtractor 73. Note that the weighting coefficient is set to be larger as it is closer to the pixel of interest, and the sum is “1”.
[0026]
The coefficients of the filters used in the error calculation unit 66 and the binary average calculation unit 72 are also switched in accordance with the magnification in the scaling processing unit 26. Switching control of these coefficients is performed by the control unit 30 according to the flowchart shown in FIG. That is, first, the magnification of the scaling unit 26 is confirmed (S1). Then, each of the error weighting coefficient, the binarized average weighting coefficient, and the average value filter coefficient is set according to the magnification (S2 to S4).
[0027]
By configuring the error diffusion binarization circuit as described above, the binarization error E (n) and the average difference data AVr (n) can be fed back to the input image data I (n). Therefore, the binarization unit 25 can execute a binarization process for diffusing the binarization error E (n) and the average difference data AVr (n) to surrounding pixels. The operations of the error diffusion memory 65, the error calculation unit 66, the binary storage memory 71, the binary average calculation unit 72, and the average value calculation unit 60 are controlled by the control unit 30.
[0028]
Next, the operation of the digital copying machine having the above configuration will be described. First, image information of the document is acquired by the image input unit 7. Then, the image data acquired by the image input unit 7 is transmitted to the image processing unit 20. Then, the image processing unit 20 sequentially performs A / D conversion, shading correction, LOG conversion processing, MTF correction, and binarization processing on the image data, and then executes scaling processing. Then, based on the image data that has been subjected to various image processes, a reproduction image of the original is copied on the recording medium by the printing unit PR, and the recording medium on which the image is formed is discharged out of the apparatus. This completes copying for one sheet.
[0029]
Next, the binarization process and the image reduction process during the series of copying operations described above will be described with reference to FIG. First, the binarization process in the binarization unit 25 will be described. When the image shown in FIG. 8A is input to the binarization unit 25, the adder 61, the image data I (n), the average difference data AVr (n) calculated by the subtractor 73, The weighted average error Eav (n) calculated by the error calculation unit 66 is added to calculate corrected image data Ir (n). The corrected image data Ir (n) is input to the terminal B of the comparator 62 and the terminal B of the subtractor 64, respectively. Then, the comparator 62 compares the corrected image data Ir (n) with the threshold value TH to calculate binarized data B (n).
[0030]
Then, the binarization reference value Br is selected by the selector 63 based on the binarization data B (n). The binarized reference value Br (n) selected here is input to the terminal A of the subtractor 64. The binarization reference value Br (n) is either HB or LB. Next, the subtractor 64 calculates a binarization error E (n) from the binarization reference value Br (n) and the corrected image data Ir (n). The binarization error E (n) is output to the error storage memory 65 and the error calculation unit 66 in order to calculate the weighted average error Eav (n). Here, binary error E (n) for several lines is stored in the sub-scanning direction by the error storage memory 65, and these are delayed by an arbitrary line and supplied to the error calculation unit 66. As a result, the error calculation unit 66 is supplied with data of an arbitrary matrix size. Then, the error calculation unit 66 calculates the weighted average error Eav (n) by multiplying the binarization error E (n) by the weighting coefficient (see FIG. 8B) according to the magnification of the scaling unit 26. Is done. The calculation result is input to the terminal C of the adder 61.
[0031]
The binarized data B (n) calculated by the comparator 62 is also output to the binary storage memory 71 and the binary average calculating unit 72 in order to calculate the weighted average binary data Bav (n). The Here, the binary storage memory 71 stores binarized data B (n) for several lines in the sub-scanning direction, which are delayed by an arbitrary line and supplied to the binary average calculation unit 72. . As a result, the binary average calculation unit 72 is supplied with data of an arbitrary matrix size. Then, the binary average calculation unit 72 multiplies the binarized data B (n) by a weighting coefficient (see FIG. 8C) corresponding to the magnification in the scaling processing unit 26 to give weighted binary average data Bav ( n) is calculated. This calculation result is input to the terminal B of the subtractor 73.
[0032]
As a method of extracting error data E (n) and binarized data B (n) of pixels that are not thinned out by the error calculation unit 66 and the binary average calculation unit 72, the memory write and read timings are thinned out. A known method such as a method of controlling by a signal or a method of selecting data read from a memory by a thinning signal may be used.
[0033]
Further, the terminal A of the subtractor 73 has the multi-value average data Iav (n) calculated by the average value calculation unit 60 by the average value filter (see FIG. 8D) according to the magnification of the scaling unit 26. Is entered. Then, the subtractor 73 calculates the difference between the multi-value average data Iav (n) and the weighted binary average data Bav (n), that is, average difference data AVr (n). The average difference data AVr (n) is input to the terminal B of the adder 61. As described above, the adder 61 adds the image data I (n), the average difference data AVr (n), and the weighted average error Eav (n) to calculate the corrected image data Ir (n). Thereafter, the same processing is performed for each pixel, and the binarization of the image ends. As a result, a binarized image shown in FIG. 8E is obtained.
[0034]
Then, the binarized image is input to the scaling processing unit 26. Then, the scaling processing unit 26 performs reduction processing by pixel thinning. That is, only the necessary binarization result is extracted in the scaling processing unit 26. Thereby, the reduced image shown in FIG. 8F is obtained.
[0035]
As described above, the binarization unit 25 corrects the input image data I (n) based on the weighted average error Eav (n) and the average difference data AVr (n). The average difference data AVr (n) is obtained from the weighted binary average data Bav (n) calculated by the binary average calculation unit 72 and the multi-value average data Iav (n) calculated by the average value calculation unit 60. It is calculated. Therefore, the corrected image data Ir (n) reflects the input image information in addition to the peripheral error. That is, the corrected image data Ir (n) reflects the thinned pixel information. For this reason, even if the thinning process is performed after the corrected image data Ir (n) is binarized, information on the thinned pixels is not lost. As a result, the finally obtained reduced image has no loss of image information and has excellent gradation reproducibility with density preservation.
[0036]
As described above in detail, according to the copying machine according to the embodiment, the binarization unit 25 causes the multi-value average data Iav (n) that is the average value in the local region of the input image data I (n). And a subtractor 73 for calculating average difference data AVr (n) that is a difference between the multi-value average data Iav (n) and the weighted binary average data Bav (n). ing. The calculation result of the subtracter 73 is input to the adder 61 together with the weighted average error Eav (n). Therefore, the adder 61 performs feedback correction on the input image data I (n) using the average difference data AVr (n) and the weighted average error Eav (n). The multi-value average data Iav (n) is calculated based on the image before the thinning process is performed, and the weighted binary average data Bav (n) is calculated based on the binarization result of the pixels that are not thinned. .
[0037]
Therefore, the average difference data AVr (n), which is the difference between the multi-value average data Iav (n) and the weighted binary average data Bav (n), is fed back to the input image data I (n), so that Missing information about the drawn pixels is prevented. Also, the weighted average error Eav (n) is calculated based on the binarization error of pixels that are not thinned out and fed back to the input image data I (n). For this reason, an error diffusion image having excellent gradation reproducibility can be obtained. That is, according to the copier according to the present embodiment, it is possible to perform a reduction process (thinning-out process) of an error diffusion image with no loss of image information and excellent gradation reproducibility.
[0038]
It should be noted that the above-described embodiment is merely an example and does not limit the present invention in any way, and various improvements and modifications can be made without departing from the scope of the invention. For example, in the above embodiment, the case where the invention is applied to a digital copying machine has been described. However, the present invention can also be applied to a printer, a facsimile, or the like in addition to a copying machine. Needless to say, the specific numerical values (for example, matrix size, etc.) exemplified in the above-described embodiment are merely examples.
[0039]
【The invention's effect】
As described above, according to the present invention, there is provided an image processing apparatus that is capable of obtaining a reduced image that is free of image information and is excellent in gradation reproducibility that has been stored in density.
[Brief description of the drawings]
FIG. 1 is a schematic configuration diagram of an image reader unit in a digital copying machine according to an embodiment.
FIG. 2 is a block diagram illustrating a control system of an image reader unit.
FIG. 3 is a block diagram illustrating a schematic configuration of an image processing unit.
FIG. 4 is a plan view showing a schematic configuration of an operation panel unit.
FIG. 5 is a block diagram illustrating a configuration of a binarization circuit.
FIG. 6 is a diagram illustrating an example of an average value filter used in an average value calculation unit.
FIG. 7 is a flowchart showing the processing content of coefficient switching control.
FIG. 8 is a conceptual diagram of processing operations in a binarization unit and a scaling processing unit.
FIG. 9 is a diagram for explaining an overview of conventional image reduction processing;
FIG. 10 is a diagram for explaining an outline of another conventional image reduction process.
[Explanation of symbols]
25 Binarization part
26 Scaling processing unit
60 Average value calculator
61 Adder
62 Comparator
63 selector
64 Subtractor
65 Error storage memory
66 Error calculator
71 Binary storage memory
72 Binary average calculator
73 Subtractor

Claims (2)

N-value conversion means for converting M-value image data into N (M> N) values for each pixel by an error diffusion method;
Reduction means for reducing the N-valued image by the N-value conversion means by pixel thinning;
N values for extracting pixels that are not thinned out by the reduction means from the N-valued image by the N-value conversion means and calculating an average value of N-value image data around the pixel of interest using the extracted pixels An average calculating means;
M value average calculating means for calculating an average value of M value image data around the target pixel;
Difference calculating means for calculating a difference between the average value calculated by the N value average calculating means and the average value calculated by the M value average calculating means;
Correction means for correcting image data of M values based on the calculation result of the difference calculation means,
The image processing apparatus according to claim 1, wherein the reduction unit reduces an image obtained by converting the M-value image data corrected by the correction unit into an N-value by the N-value conversion unit. The image processing apparatus according to claim 1,
The N-value average calculating means calculates an average value by filter processing and switches a coefficient in the filter in accordance with a magnification of the reduction means.

Images