JPH0583553A - Image processor - Google Patents

Abstract

PURPOSE:To improve the picture quality of a binary image. CONSTITUTION:A density judging part 1 judges density change tendency in a pixel matrix segmented by a matric register part 4, and the kind of an MTF correcting operation to be executed by a spatial frequency (MTF) correcting part 8 is selected corresponding to the density of images around the attention picture element and the resolution of the images. When MTF correction is turned to over correction and the noise of the images around the attention picture element is large, base dirt or the like is removed by applying an isolate point removing operation due to an isolate point removing part 11 concerning binary image signals BW.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention, when converting a multi-valued image signal into a binary image signal, determines the pixel of interest based on the level of the pixel of interest to be binarized and the level of the adjacent pixel adjacent to this pixel of interest. The present invention relates to an image processing device that performs a spatial frequency correction calculation for correcting a spatial frequency characteristic.

[0002]

2. Description of the Related Art For example, in an image processing apparatus such as a facsimile apparatus which reads and processes a document image, a read image is blurred due to a spatial frequency characteristic of an optical system of the document reading apparatus, and a reproduced image (recorded image) is generated. The inconvenience that the image quality deteriorates occurs.

Therefore, when a multi-valued image signal obtained by reading an original image is binarized into a binary image signal, based on the level of a target pixel to be binarized and an adjacent pixel adjacent to the target pixel, A spatial frequency correction calculation for correcting the spatial frequency characteristic of the pixel of interest is performed.

[0004]

However, such a conventional device has the following disadvantages.

That is, the reading state of the original image to be binarized, for example, the resolution in the sub-scanning direction and the S of the read image.
Since the same calculation is always applied regardless of the N-ratio and the like, depending on the reading state of the original image, the inconvenience may occur in that the effect of improving the image quality of the read image does not appear so much.

It is an object of the present invention to provide an image processing apparatus which solves the inconvenience of the conventional apparatus and is capable of performing an appropriate spatial frequency correction calculation according to the reading state of the original image.

[0007]

According to the present invention, when converting a multi-valued image signal into a binary image signal, attention is paid based on the level of a target pixel to be binarized and an adjacent pixel adjacent to the target pixel. In an image processing device for performing a spatial frequency correction operation for correcting the spatial frequency characteristic of a pixel, an arithmetic means capable of executing a plurality of spatial frequency correction operations having different weighting coefficients corresponding to a target pixel and an adjacent pixel and an input image characteristic. Accordingly, the control means is provided for selecting the spatial frequency correction calculation applied by the calculation means. Further, the control means selects the spatial frequency correction calculation in which the weighting coefficient of the adjacent pixel adjacent in the direction in which the pixel correlation of the input image is larger is selected. Further, the calculation means is
Each spatial frequency correction calculation is realized by a combination of a plurality of common calculation means, and the weighting coefficient of the adjacent pixel is expressed by a power of 2.

When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to the target pixel. In an image processing device that performs a spatial frequency correction calculation, a calculation means that can execute a plurality of spatial frequency correction calculations with different weighting coefficients corresponding to a pixel of interest and an adjacent pixel, and the calculation means are applied according to the characteristics of the input image. Control means for selecting the spatial frequency correction operation,
It is provided with an isolated point removing means for removing the isolated point image included in the output data of the arithmetic means.

When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to the target pixel. In an image processing apparatus for performing spatial frequency correction calculation, SN ratio calculation means for calculating the SN ratio of an input image, and when the value of the SN ratio calculated by this SN ratio calculation means is smaller than a predetermined value, isolation after spatial frequency correction is performed. The control means for removing the point image is provided.

When the multi-valued image signal is converted into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to the target pixel. In an image processing device for performing spatial frequency correction calculation, a calculation means capable of executing a plurality of spatial frequency correction calculations having different weighting coefficients corresponding to a target pixel and adjacent pixels, and density calculation for calculating an average density of the target pixel and adjacent pixels And a control means for selecting the spatial frequency correction calculation applied by the calculation means according to the calculation result of the density calculation means.

[0011]

Therefore, since the contents of the spatial frequency correction calculation are selected according to the characteristics of the input image such as the average density of the image, the spatial frequency correction can be performed more effectively and the read image The image quality can be greatly improved. Further, the calculation means for performing the spatial frequency correction calculation realizes the respective spatial frequency correction calculation by combining a plurality of common calculation means, and the weighting coefficient of the adjacent pixel is expressed by a power of 2.
The circuit amount of the calculation means can be reduced. Further, since the isolated points are removed from the image after the spatial frequency correction calculation based on the SN ratio of the image, the effect of overcorrection of the spatial frequency correction calculation can be removed.

[0012]

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

In this embodiment, as shown in FIG. 1A, the spatial frequency correction calculation is performed on the target pixel E to be binarized.
And adjacent pixels A, B, C, D, F, G,
Based on the H and I concentrations, one of the following four modes is applied.

1) Mode 1: Adjacent pixels in the main scanning direction and the sub scanning direction have the same weighting.

E '= 3E- (B + D + F + H) / 2

2) Mode 2: The weighting coefficient of adjacent pixels in the main scanning direction is large.

E '= 4E- (B + 2D + 2F + H) / 2

3) Mode 3: Large weighting coefficient of adjacent pixels in the sub-scanning direction

E '= 4E- (2B + D + F + 2H) / 2

4) Mode 4: Large weighting coefficient of adjacent pixels (large degree of spatial frequency correction calculation)

E '= 5E- (B + D + F + H)

For example, when the adjacent pixel D, F in the main scanning direction has a stronger correlation with the target pixel E, the calculation of the mode 2 in which the weighting coefficient of the adjacent pixel in the main scanning direction is large, the image quality is selected. The effect of improvement is great. When the correlation between the target pixel E and the adjacent pixels B and H in the sub-scanning direction is stronger, or when the transport speed in the sub-scanning direction is high, the weighting coefficient of the adjacent pixel in the sub-scanning direction is large. When the calculation of mode 3 is selected, the effect of improving the image quality is great. In addition, when the density of the entire pixel matrix shown in FIG.
Greatly improves the image quality of images such as characters.

Here, the correlation between the adjacent pixels D and F in the main scanning direction is stronger, or the adjacent pixels B and F in the sub scanning direction are
The determination of whether the correlation of H is stronger can be performed as follows.

For example, as shown in FIG.
The pixel matrix of (a) is divided in the sub-scanning direction, the differences in the densities of the respective divided blocks are compared, and if the density difference is small, it can be determined that the correlation in the sub-scanning direction is strong. ..

Further, as shown in FIG.
The pixel matrix of (a) is divided in the main scanning direction, the densities of the divided blocks are compared, and when the density difference is small, it can be determined that the correlation in the main scanning direction is strong. ..

Further, the average density of the entire pixel matrix is calculated, and when the average density is small, it can be determined that the area is a low density area.

On the other hand, when the operation of mode 4 is applied, the spatial frequency correction is overcorrected, and there is a possibility that the background stain will occur. Therefore, in the present embodiment, when applying the operation of mode 4, the background stain can be removed by applying the isolated point removal operation to the image after the spatial frequency correction operation. Further, this isolated point removal calculation is performed even when the SN ratio of the read image is poor so that the background stain of the read image can be removed.

In this isolated point removal calculation, for example, the target pixel E and the reference pixels A, B, C, D, F, G, H and I are
2A to 2E, the target pixel E is inverted to black, and the target pixel E and the reference pixels A, B, C, D, F, G are also inverted. , H,
When I matches any one of the patterns shown in FIGS. 2F to 2J, the target pixel E is inverted to white.

FIG. 3 shows an image processing apparatus according to an embodiment of the present invention.

In the figure, the line image sensor 1
Is for reading the original image in line units.
The image for a line is divided into pixels, and the analog image signal AV corresponding to each pixel is sequentially output. The analog image signal AV is amplified by the amplifier 2 with a predetermined gain and then analog / digital. It is added to the converter 3.

The analog / digital converter 3 converts the analog image signal AV into a digital image signal DV having a predetermined number of bits. The digital image signal DV is supplied to the matrix register section 4, the control section 5, and the line. Buffer 6
Has been added to.

The line buffer 6 is a FIFO buffer having a storage capacity capable of storing the digital image signal DV for one line, and its output signal is added to the matrix register section 4 as the digital image signal DVa one line before. At the same time, it is added to the line buffer 7.

The line buffer 7 is a FIFO buffer having a storage capacity capable of storing the digital image signal DV for one line, and its output signal is added to the matrix register unit 4 as the digital image signal DVb two lines before. ing.

The matrix register section 4 has a digital image signal DV applied from the analog / digital converter 3 and a digital image signal DVa applied from the line buffer 6.
Also, the pixel matrix data shown in FIG. 1A is cut out based on the digital image signal DVb added from the line buffer 7. Of the pixel matrix data cut out by the matrix register unit 4, Data DE of the target pixel E, reference pixels B, D, F, H
The data DB, DD, DF, and DH are added to the MTF (spatial frequency) calculator 7. Further, all the data of the pixel matrix is added to the density determination unit 9.

The MTF calculation unit 7 determines the data D of the target pixel E.
E and data DB, D of reference pixels B, D, F, H
Based on D, DF, DH, a spatial frequency correction operation of a mode designated by the mode signal MD added from the control unit 5 is performed, and the operation result is a binarization circuit as a corrected digital image signal DVc. Added to 10.

The density determination unit 9 is composed of the matrix register unit 4
Based on all the additional data of the pixel matrix added,
It is to determine that the density difference in the sub-scanning direction, the density difference in the main-scanning direction, and the average density of the pixel matrix are small, and when the density difference in the sub-scanning direction is small, a mode determination signal representing mode 3 Output the MS to the control unit 5,
When the density difference in the main scanning direction is small, the mode determination signal MS representing the mode 2 is output to the control unit 5, and when the average density is small, the mode determination signal MS representing the mode 4 is output to the control unit 5.

The binarization circuit 10 includes a corrected digital image signal D
Vc is binarized with a predetermined threshold value, and the processing result is the isolated point removing unit 11 as a binarized image signal BW.
And to the line buffer 12.

The line buffer 12 is a FIFO buffer having a storage capacity for storing the binarized image signal BW for one line, and its output signal is the isolated point removed as the binarized image signal BWa one line before. In addition to being added to the section 11, it is added to the line buffer 13.

The line buffer 13 is a FIFO buffer having a storage capacity for storing the binarized image signal BW for one line, and its output signal is an isolated point removed as the binarized image signal BWb two lines before. Added to part 11.

The isolated point removing unit 11 sequentially stores the signals representing the operation on / off state designated by the control unit 5, and the binarization circuit 10 for the pixel of interest designated by the control unit 5 for the action on. Input binarized image signal B
W is output as it is as a binary image signal BV to the external device, and for the target pixel for which operation ON is designated by the control unit 5, the binary image signal BW input from the binarization circuit 10 The isolated point removal calculation described above is performed based on the binarized image signals BWa and BWb input from the line buffers 12 and 13, and the calculation result is output as the binary image signal BV.

The control unit 5 drives the line image sensor 1 and controls the operation of a document conveying system (not shown) in order to perform a reading operation at a resolution designated and input by an upper device (not shown) or an operator. At the same time, the control unit 5 appropriately controls the operation of the analog / digital converter 3, the operation of the MTF correction unit 8, the operation of the binarization circuit 10, and the operation of the isolated point removal unit 11 as described later. Have control.

FIG. 4 shows an example of the matrix register section 4.

In the figure, the digital image signal DV output from the analog / digital converter 3 is registered in the register RA.
The output of the register RA is added to the density determination unit 9 as the data DA of the reference pixel A, and
It has been added to register RB. The output of the register RB is added to the density determination unit 9 as the data DB of the reference pixel B, is output to the MTF correction unit 8 and is also added to the register RC. The output of the register RC is added to the density determination unit 9 as the data DC of the reference pixel C.

The digital image signal DVa output from the line buffer 6 is added to the register RD, and the output of the register RD is added to the density determination unit 9 as the data DD of the reference pixel D and the MTF correction unit. 8 and is added to the register RE. The output of the register RE is added to the density determination unit 9 as the data DE of the target pixel E, is output to the MTF correction unit 8 and is also added to the register RF. Register RF
Is output as the data DF of the reference pixel F to the density determination unit 9
And is output to the MTF correction unit 8.

The digital image signal DVb output from the line buffer 7 is added to the register RG, and the output of the register RG is added to the density determination unit 9 as the data DG of the reference pixel G and to the register RH. Has been added. The output of the register RH is added to the density determination unit 9 as the data DH of the pixel of interest H, and M
It is output to the TF correction unit 8 and also added to the register RI. The output of the register RI is the data DI of the reference pixel I.
Is added to the density determination unit 9.

In this way, the matrix register unit 4
1A based on the digital image signal DV added from the analog / digital converter 3, the digital image signal DVa added from the line buffer 6, and the digital image signal DVb added from the line buffer 7. The pixel matrix data is cut out.

FIG. 5 shows an example of the MTF correction section 8.

In the figure, the data DE is the subtraction circuit 2
It is added to the negative side input ends of 1, 22, 23, and 24, and is also added to the input end E of the arithmetic unit 25. Further, data DB, DH, DD, and DF are added to the positive side input ends of the subtraction circuits 21, 22, 23, and 24, respectively.

The output of the subtraction circuit 21 is added to one input terminal of the addition circuit 26 as the calculation result of the calculation (BE). The output of the subtraction circuit 22 is added to the other input end of the addition circuit 26 as the calculation result of the calculation (HE). The output of the subtraction circuit 23 is added to one input terminal of the addition circuit 27 as the calculation result of the calculation (D−E).
The output of the subtraction circuit 24 is added to the other input terminal of the addition circuit 27 as the calculation result of the calculation (FE).

The output of the adder circuit 26 is calculated (B + H-2).
It is added to the input terminal A of the calculation unit 25 as the calculation result of E) and is also added to the shift register 28. The output of the adder circuit 27 is added to the input terminal C of the calculation unit 25 as a calculation result of the calculation (D + F-2E) and is also added to the shift register 29.

The shift register 28 shifts down the input data by 1 bit to form a 1/2 value, and its output is input to the input terminal B of the arithmetic unit 25 as an operation (B + H-2E) / 2. Has been added. Shift register 2
Reference numeral 9 is for shifting down the input data by 1 bit to form a 1/2 value, and its output is the operation (D + F-
2E) / 2 is added to the input terminal D of the calculation unit 25.

When the operation of mode 1 is designated by the mode signal MD applied from the control section 5, the operation section 25 changes from the data (E) of the input terminal E to the input terminal B.
Data ((B + H-2E) / 2) and the data at the input terminal D ((D + F-2E) / 2) are subtracted, and the operation of mode 1 is realized, and the operation result is used as the corrected digital image signal DVc. Output.

When the mode signal MD applied from the control unit 5 designates the operation of the mode 2, the operation unit 25 changes the data (E) from the input terminal E to the data (E) at the input terminal B. (B + H-2E) / 2) and the data (D + F-2E) at the input terminal C are subtracted, so that the mode 2
Is calculated and the calculation result is corrected digital image signal D
Output as Vc.

When the operation of mode 3 is designated by the mode signal MD applied from the control section 5, the operation section 25 changes from the data (E) at the input terminal E to the data (E) at the input terminal A. B + H-2E) and the data at the input terminal D ((D + F-2E) / 2) are subtracted, whereby the operation of mode 3 is realized and the operation result is output as the corrected digital image signal DVc.

When the operation of mode 4 is designated by the mode signal MD applied from the control section 5, the operation section 25 changes from the data (E) at the input end E to the data (E) at the input end A. B + H-2E) and the data (D + F-2E) at the input terminal C are subtracted, and thereby the operation of mode 4 is realized, and the operation result is output as the corrected digital image signal DVc.

In this way, since the MTF correction unit 8 is realized by a combination of operations common to the operations of modes 1, 2, 3, and 4, the device configuration can be reduced and the operation speed can be improved. You can

FIG. 6 shows an example of processing of the density determination section 9.

First, the input data DA, DB, DC, D
Based on D, DE, DF, DG, DH, and DI, the density difference BMS in the main scanning direction, the density difference BSS in the sub-scanning direction, and the average density ASS are calculated (process 101), and the density difference B
A threshold Tm for determining the MS, a threshold Ts for determining the concentration difference BSS, and a threshold Ta for determining the average concentration ASS are set (process 102). The set values of these threshold values Tm, Ts, Ta can be set to predetermined fixed values or can be formed from the fixed values corresponding to the setting of the read density.

Next, it is checked whether or not the density difference BMS is larger than the threshold value Tm (decision 103). When the result of the judgment 103 is YES, it means that the density difference in the main scanning direction is large. Is determined to be larger than the threshold value Ts (decision 104). When the result of the determination 104 is YES, the density difference in the main scanning direction and the sub scanning direction are both large, and the correlation between the target pixel E and the adjacent pixels A, B, C, D, F, G, H, and I. Since there is no bias in the relationship, it is further checked whether the average density ASS is larger than the threshold value Ta (decision 105).

If the result of judgment 105 is YES, it means that the density of the pixel matrix is sufficiently large, and therefore mode 1 is set in order to set the normal MTF correction calculation (process 106). When the result of determination 105 is NO, it means that the density of the pixel matrix is small, and therefore mode 4 is set in order to set the MTF correction calculation in which the weighting coefficient of the adjacent pixel is large (process 107).

When the result of the determination 104 is NO, the density difference in the main scanning direction is large and the density difference in the sub-scanning direction is small, and further, the average density ASS.
Is larger than the threshold value Ta (decision 10
8).

If the result of decision 108 is YES, it means that the density of the pixel matrix is sufficiently high and the correlation in the sub-scanning direction is strong, so mode 3 is set in which the weighting coefficient of the adjacent pixel in the sub-scanning direction is large ( Processing 10
9). When the result of determination 108 is NO,
The process proceeds to step 107 and mode 4 is set.

When the result of judgment 103 is NO, it means that the density difference in the main scanning direction is small, and therefore it is further checked whether or not the average density ASS is larger than the threshold value Ta (decision 110). If the result of the determination 110 is YES, it means that the density of the pixel matrix is sufficiently high and the correlation in the main scanning direction is strong, so that mode 2 in which the weighting coefficient of the adjacent pixels in the main scanning direction is large is set (process 111). .. When the result of determination 110 is NO, the process proceeds to step 107 and mode 4 is set.

When the MTF correction mode is set for each target pixel E in this way, the density determination section 9 outputs a mode determination signal MS representing the content thereof to the control section 5.

7 and 8 show an example of processing of the control unit 5.

The control unit 5 sets a threshold value Tsn for determining the SN ratio of the image (process 201), and based on the digital image signal DV output from the analog / digital converter 3, the periphery of the pixel of interest E. The SN ratio of the image is calculated (process 202). Then, it is checked whether or not the calculated SN ratio is smaller than the threshold value Tsn (decision 203), and when the result of the judgment 203 is YES, it means that the image noise is large, and therefore the mode signal MD corresponds to the value corresponding to mode 1. Is set (process 204), and the isolated point removing unit 11 is set.
Is turned on (process 205).

When the result of the judgment 203 is NO, it is checked at that time whether the resolution in the main scanning direction notified from the host device is rough (judgment 206).
When the result of determination 206 is YES, it is checked whether the resolution in the sub-scanning direction is rough (determination 207). When the result of determination 207 is NO, the mode determination signal MS of the concentration determination unit 9 is input (process 208), and the value corresponding to the mode determination signal MS is set in the mode signal MD,
The operation mode of the MTF correction unit 8 is set (process 20).
9). Then, at this time, it is checked whether or not the mode 4 is set (decision 210), and the result of the decision 210 is YES.
When, the operation of the isolated point removing unit 11 is turned on (process 211).

When the result of judgment 207 is YES, it means that the resolution in the main scanning direction and the resolution in the sub-scanning direction are both coarse, and therefore mode 4 is set in mode signal MD.
Is set (step 212), the process proceeds to step 211, and the isolated point removing unit 11 is turned on.

When the result of judgment 206 is NO, it is checked whether or not the resolution in the sub-scanning direction is rough (judgment 213). When the result of judgment 213 is NO, the conveying speed at the time of reading the original is high. It is checked whether or not it is set to (determination 214). When the result of the determination 214 is NO, it means that the resolution in the main scanning direction and the resolution in the sub scanning direction are not coarse, so a value representing mode 3 is set in the mode signal MD (process 215).

When the result of judgment 213 is YES and when the result of judgment 214 is YES, the routine proceeds to step 208, and the subsequent steps are executed.

In this way, the control unit 5 controls the target pixel E
When the noise around the image is large, M in mode 1
The TF correction calculation is selected, and the isolated point removing unit 11 is set to the operation-on state to remove noise included in the binarized image.

When the noise in the image around the target pixel E is small, the MTF correction calculation mode is determined based on the mode determination signal MS output from the density determination section 9. Further, in this case, when the mode 4 is set, the isolated point removing unit 11 is turned on to prevent image deterioration due to MTF overcorrection. However, at this time, when the resolutions in the main scanning direction and the sub scanning direction are both coarse, mode 4 is set, the isolated point removing unit 11 is turned on, and the resolutions in the main scanning direction and the sub scanning direction are set. If neither is coarse, mode 3 is set.

With the above configuration, when the reading resolution of the original is specified by the upper device (not shown) and the reading start is notified, the control section 5 drives the original conveying system (not shown) in a predetermined manner and The line image sensor 1 is driven at the line reading timing.

As a result, the analog image signal AV corresponding to the image on the read line is output from the line image sensor 1.
Are sequentially output in the order of reading, and the analog image signal A
V is a digital image signal D from the analog / digital converter 3.
Converted into V, added to the matrix register unit 4 and added to the line buffer 6, and further to the control unit 5
Can be added

Further, the digital image signal DVa is output from the line buffer 6 in synchronization with the input of the digital image signal DV and added to the matrix register section 4 and the line buffer 7. Also, from the line buffer 7, the digital image signal DVa is synchronized with the input of the digital image signal DVa.
b is output and added to the matrix register unit 4.

Therefore, when the analog image signal AV for one pixel is output from the line image sensor 1, the digital image signals DV, DVa, DVb added to the matrix register section 4 are sequentially updated in synchronization with it. As a result, the data output from the matrix register unit 4 to the MTF correction unit 8 and the density determination unit 9 becomes the target pixel E.
Changes with the movement of.

The density determination section 9 is composed of the matrix register section 4
Based on the data input from the above, the above-mentioned processing is performed to form the mode determination signal MS, which is output to the control unit 5. The control unit 5 calculates the noise of the image around the target pixel E based on the sequentially input digital image signal DV, and based on the calculated value and the mode determination signal MS input from the density determination unit 9, By performing the above-mentioned processing, the mode signal M
While setting the contents of D, the operation on / off of the isolated point removing unit 11 is set.

As a result, the MTF correction unit 8 has the control unit 5
The MTF correction calculation processing corresponding to the value of the mode signal MD added from is executed, and the correction digital signal DVc obtained by the calculation result is sequentially output to the binarization circuit 10.

As a result, the binarization circuit 10 outputs the binarized image signal BW to the isolated point removing section 11 and the line buffer 12. Further, the binarized image signal BWa is output from the line buffer 12 at the timing when the binarized image signal BW is input and added to the isolated point removing unit 11 and the line buffer 13. Well, from the line buffer 13, the binarized image signal B
The binarized image signal BWb is output at the timing when Wa is input and is input to the isolated point removing unit 11.

On the other hand, the isolated point removing section 11 sequentially stores the signals of the operation on / off state designated by the control section 5, and outputs the signals of the operation on / off state corresponding to the target pixel to be the processing object of the isolated point removal. When it is taken out and represents the operation-on state, the isolated point removal processing described above is executed, and the processing result is output as the binary image signal BV. When the signal in the operation on / off state corresponding to the pixel of interest represents the operation off state, the above-described isolated point removal processing is not performed, and the binarized image signal BW (BWa) of the pixel of interest is directly converted into the binary value. The image signal BV is output.

In this way, in this embodiment, each time the target pixel E is switched, the image reading resolution, the tendency of density change in the pixel matrix, and the average density in the pixel matrix are applied. Since the MTF correction calculation is selected, the image quality of the binary image can be significantly improved.

If the SN ratio of the image is bad, or
When the side effect of MTF overcorrection appears, the isolated points appearing in the binary image are removed, so that background stains can be removed and the image quality of the binary image can be further improved. it can.

In the above embodiment, four MTFs are used.
Although the correction calculation is set and any one of them is selected, the type of the MTF correction calculation is not limited to this. Further, in the above-described embodiment, the MTF
When selecting the correction calculation mode, the tendency of the density change in the pixel matrix is emphasized, but another condition can be applied for the condition setting for this selection.

Further, in the above-described embodiment, the size of the pixel matrix to which the MTF correction calculation is applied is set to 3 × 3, but the size of this pixel matrix is
It is not limited to this.

[0085]

As described above, according to the present invention,
Since the contents of the spatial frequency correction calculation are selected according to the characteristics of the input image such as the average density of the image, the spatial frequency correction can be performed more effectively and the image quality of the read image is greatly improved. can do. Further, the calculation means for performing the spatial frequency correction calculation realizes the respective spatial frequency correction calculation by combining a plurality of common calculation means, and the weighting coefficient of the adjacent pixel is expressed by a power of 2. The circuit amount of the calculation means can be reduced. Moreover, since the isolated points are removed from the image after the spatial frequency correction calculation based on the SN ratio of the image, the effect of overcorrection of the spatial frequency correction calculation can be removed.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing a pixel matrix for MTF correction according to an embodiment of the present invention, and a mode of block division in the pixel matrix for determining a tendency of density change.

FIG. 2 is a schematic diagram for explaining isolated point removal processing.

FIG. 3 is a block diagram showing an image processing apparatus according to an embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration example of a matrix register section.

FIG. 5 is a block diagram showing a configuration example of an MTF correction unit.

FIG. 6 is a flowchart showing a processing example of a density determination unit.

FIG. 7 is a flowchart showing a part of the processing of the control unit.

FIG. 8 is a flowchart showing another part of the processing of the control unit.

[Explanation of symbols]

 4 matrix register section 5 control section 8 MTF correction section 10 binarization circuit 11 isolated point removal section 21, 22, 23, 24 subtraction circuit 25 arithmetic section 26, 27 addition circuit 28, 29 shift register

Claims (6)

[Claims] 1. When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to the target pixel. In an image processing device that performs a spatial frequency correction calculation, a calculation means that can execute a plurality of spatial frequency correction calculations with different weighting coefficients corresponding to a pixel of interest and an adjacent pixel, and the calculation means are applied according to the characteristics of the input image An image processing apparatus comprising a control means for selecting a spatial frequency correction calculation. 2. The image processing apparatus according to claim 1, wherein the control means selects a spatial frequency correction calculation in which a weighting coefficient of an adjacent pixel adjacent in a direction in which a pixel correlation of the input image is larger is selected. .. 3. The calculation means realizes each spatial frequency correction calculation by a combination of a plurality of common calculation means, and the weighting coefficient of adjacent pixels is expressed by a power of two. The image processing apparatus according to item 1. 4. When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to this target pixel. In an image processing device that performs a spatial frequency correction calculation, a calculation means that can execute a plurality of spatial frequency correction calculations with different weighting coefficients corresponding to a pixel of interest and an adjacent pixel, and the calculation means are applied according to the characteristics of the input image. An image processing apparatus comprising: a control unit for selecting a spatial frequency correction calculation and an isolated point removal unit for removing an isolated point image included in output data of the calculation unit. 5. When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to this target pixel. In an image processing apparatus for performing spatial frequency correction calculation, SN ratio calculation means for calculating the SN ratio of an input image, and when the value of the SN ratio calculated by this SN ratio calculation means is smaller than a predetermined value, isolation after spatial frequency correction is performed. An image processing apparatus comprising a control means for removing a point image. 6. When converting a multi-valued image signal into a binary image signal, the spatial frequency characteristic of the target pixel is corrected based on the level of the target pixel to be binarized and the adjacent pixel adjacent to this target pixel. In an image processing device that performs spatial frequency correction calculation, a calculation means capable of executing a plurality of spatial frequency correction calculations having different weighting coefficients corresponding to the pixel of interest and adjacent pixel, and density calculation for calculating the average density of the pixel of interest and adjacent pixel An image processing apparatus comprising: a means and a control means for selecting a spatial frequency correction calculation applied by the calculation means according to the calculation result of the density calculation means.