JP2003046781A - Method and device for image processing - Google Patents

Abstract

PROBLEM TO BE SOLVED: To perform processing according to the strength of removing, the type of document and resolution, etc., in detecting and removing dust noise on a document image. SOLUTION: Upon detection of a defective pixel by a defective pixel detection section 105, if the difference between a gradation value of a median value in pixels in a filter range preset by a filter range setting section 107 and a gradation value of the defective value is no less than a threshold value preset by threshold value setting section 108, a median filter 106 replaces the defective pixel with the median value. In this case, the filter range or the threshold value may be changed according to the setting by a filter strength setting section 109.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and an image processing apparatus, and more particularly to an image processing method and an image processing apparatus for correcting dust noise due to dust contained in image data read by an image reading apparatus or the like. It is a thing.

[0002]

2. Description of the Related Art Conventionally, as one of methods of generating electronic image data, there is a method of reading an image drawn on a document with an optical image scanner and obtaining electronic image data representing the image of the document. With this method, you can easily generate image data, but
Dirt or the like attached to the reading surface of the image scanner is also read as a part of the image. Therefore, the generated image data includes an image such as dirt as noise (dust noise) that is not originally a component of the image. In order to correct the dust noise due to such dust and the like, conventionally, the read image data is image-reproduced by a display device or the like, and the operator often observes the image and acknowledges that it is dust noise. Above, there was no effective method other than retouching correction using image editing software.

Therefore, an invention relating to a method for automatically detecting and removing such dust noise in a short time is as follows.
Japanese Patent Application No. 20 dated April 2, 2001 by the applicant of the present application
The patent application has been filed as 01-103642.

[0004]

However, Japanese Patent Application No. 2
In the technique disclosed in No. 001-103642, a defective pixel detected as dust noise is always corrected in a fixed procedure based on the values of its neighboring pixels. This fixed procedure does not always give good results for all images, so it was necessary to make the degree of correction variable.

The present invention has been made in view of the above-mentioned conventional example, in which the strength for removing dust noise can be set variably,
Alternatively, another object is to provide an image processing method and apparatus capable of optimizing the dust noise removal processing according to the type of the original image.

[0006]

In order to solve the above problems, the image processing method of the present invention has the following configuration.

Setting step capable of changing the range in the vicinity of the pixel of interest, the first threshold value, or both, the gradation level of the pixel of interest in the image data, and the gradation level of pixels in the range in the vicinity of the pixel of interest. And a difference between the gradation level of the replacement pixel and the gradation level of the replacement pixel is equal to or larger than the first threshold, the replacement step of replacing the gradation level of the target pixel with the gradation level of the replacement pixel. .

Preferably, in the setting step, the range in the vicinity of the pixel of interest is changed to be widened or narrowed.

In the setting step, the first threshold value is changed so as to be lowered or raised.

In the setting step, the range in the vicinity of the pixel of interest is widened and the first threshold is lowered, or the range in the vicinity of the pixel of interest is narrowed and the first threshold is raised.

With this configuration, it is possible to set the threshold value for replacing the pixel of interest and the value of the replacement pixel.

Preferably, the setting step changes the range in the vicinity of the pixel of interest, the first threshold value, or both in accordance with the type of image containing the pixel of interest.

Preferably, the method further includes a prohibition step in which the pixel of interest is not replaced in the replacement step according to the setting of the document type.

Preferably, in the setting step, the range in the vicinity of the pixel of interest or the first threshold value or both of them are changed according to the resolution when the image including the pixel of interest is digitized.

Preferably, in the replacing step, as the replacing pixel, among the pixels in the range near the target pixel,
A pixel having an intermediate value of the gradation level as the gradation level is used.

Preferably, a target pixel detecting step of detecting a target pixel group whose center pixel is a pixel having a gradation level difference from a neighboring pixel of not less than a second threshold value from the image data; The reference value is obtained based on the gradation level value of the pixel in the first range, and the pixel whose gradation level difference between the reference value and each pixel of the target pixel group is equal to or larger than the third threshold is determined. A defective pixel detecting step of detecting as a defective pixel is further provided, and the defective pixel detected by the defective pixel detecting step is set as the central pixel.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of an image processing method and an image processing apparatus according to an embodiment of the present invention will be described in detail below with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing the functional arrangement of an image processing apparatus having an image processing unit 102 to which the present invention is applied. In FIG. 1, 101 is an image reading device that reads image data, 103 is an interface (I / F) for inputting image data, 104 is an image memory that stores image data, 105 is a defective pixel extraction unit, and 106 is Is a median filter, 107 is a filter range setting unit, 108 is a gradation level threshold setting unit, 109
Is a filter strength setting unit.

With the above arrangement, the image processing method according to the first embodiment will be described in detail with reference to the flowchart of FIG.

First, in step 201, image data is input from the image reading apparatus 101 to the image memory 104 through the interface 103. Next step 20
In 2, the defective pixel extracting unit 105 extracts pixels having the feature of dust noise from the image data read from the image memory 104. Here, the procedure for extracting defective pixels will be described.

<Extraction of Defective Pixels> FIG. 6 is a detailed block diagram of the defective pixel extraction unit 105 for extracting defective pixels. In FIG. 6, 601 is a gradation level difference ΔL1 calculation unit, 602 is a gradation level difference ΔL2 setting unit, 603 is a target pixel extraction unit, 604 is a range setting unit, 605 is an average value calculation unit, and 606 is a gradation level difference. A ΔL3 calculation unit, 607 is a gradation level difference ΔL4 setting unit, and 608 is a pixel extraction unit. The defective pixel extraction unit extracts pixels having the characteristics of dust noise as follows.

First, the ΔL1 calculator 601 reads out image data from the image memory 104, pays attention to arbitrary pixel data A among them, and as shown in FIG.
Difference ΔL1x between the left and right adjacent pixels E and F
Then, the gradation level difference ΔL1y between the target pixel A and the upper and lower adjacent pixels C and H is calculated as follows. It should be noted that the symbols A, B, etc. indicating the pixels are also used as they are as the symbols indicating the density gradation of the pixel. Further, as a method of selecting the target pixel A, a method of sequentially selecting in the raster scanning order is desirable in order to scan the entire image without omission.

ΔL1x = | A × 2- (E + F) | (1) ΔL1y = | A × 2- (C + H) | And This calculation process is based on the gradation level difference Δ
This is performed by the L1 calculation unit 601. In the case of color image data, for example, the following is performed. The difference Δ in gradation level between the pixel of interest and the left and right pixels for each color component of RGB
L1xr, ΔL1xg, and ΔL1xb are obtained, and the largest value is used as ΔL1x in Expression (1). ΔL
Similarly, y is obtained. This method is the same as in the following description and other embodiments, and can be obtained in this manner when obtaining the difference in gradation level.

The method of calculating ΔL1 is not limited to the above example, and either ΔL1x or ΔL1y may be used. In addition, when calculating ΔL1x or ΔL1y,
The calculation method may be such that only the left side and the upper side are used without using both the upper and lower directions. In that case, the gradation level difference between the pixel of interest and the adjacent pixel is used as ΔL1x or ΔL1y. Furthermore, when determining ΔL1, only adjacent pixels are used in the present embodiment, but pixels that are two or more pixels away from the pixel of interest may be used.

ΔL1 is a value calculated in order to utilize the local correlation that the image data locally has a high degree of correlation of the pixel values forming the image data. for that reason,
The pixel to be compared with the target pixel can be selected as long as it can be expected to be correlated with the target pixel. That is, although it is possible to use a pixel that is more than two pixels away from the pixel of interest, it does not matter which pixel it is at all and it must be a pixel in the vicinity of the pixel of interest. Adjacent pixels are preferably selected as in the embodiment. This range also varies depending on the resolution when reading the image data. For example, when attention is paid to two pixels separated by a certain number of pixels, even if the high resolution has a certain degree of correlation, the low resolution decreases the degree of correlation. Note that here, the range in the image data where it can be expected to be correlated with the target pixel is referred to as the vicinity of the target pixel.

Next, the target pixel extraction unit 603 makes the first
The first threshold Δ set by the threshold ΔL2 setting unit 602 of
L2 is compared with the gradation level difference ΔL1, and ΔL1 ≧ ΔL2
It is determined whether or not If ΔL1 ≧ ΔL2 is not satisfied, the pixel of interest is moved to the next unprocessed pixel.

The first threshold value ΔL2 is a threshold value for judging the possibility that the pixel of interest is noise, and the number of pixels to be subsequently judged whether or not it is noise is narrowed down. Is a value for. Since it is a value for narrowing down the target, if the value of the first threshold value ΔL2 is larger than 0, for example, if it is 1, the object can be achieved (the gradation level is represented by an integer such as 0 to 155). To). In addition, as a method of determining the first threshold value ΔL2, for example, a relatively uniform local image that does not include the boundary of an object is experimentally acquired from a natural image or the like, and the gradation level ΔL1 (ΔL1R) at that is acquired. A method of calculating the above equation and determining the minimum numerical value of ΔL1R <ΔL2 as ΔL2 is conceivable. ΔL2 setting unit 602
Can be configured with a memory that stores a value that is input in advance.

When ΔL1 ≧ ΔL2,
The target pixel extraction unit 603 extracts the target pixel as the target pixel, and then also extracts the neighboring pixels around it as the target pixel. In the following, the target pixel that is a candidate for the correction process is simply referred to as the target pixel. In FIG. 3, when the target pixel A is the target pixel, for example, the adjacent pixels C, E,
F and H are also extracted as target pixels. This is because dust noise normally has some spread.
This extraction processing is performed by the target pixel extraction unit 603.

The method of extracting the target pixel is not limited to the above example, and the target pixel and at least one of the upper, lower, left and right adjacent pixels may be extracted as the target pixel. Further, it is also possible to extract pixels that are more than two pixels away from the pixel of interest. However, this range is also limited to the range in the vicinity described above. It should be noted that the extraction referred to here is not limited to literal extraction, but it is possible to perform substantially the same processing as extraction, such as storing the address of a pixel in the image memory and making it available for reference. The operation of setting the state is also included.

Next, paying attention to the extracted target pixel data, the average value calculation unit 605 calculates the average value La of the pixel data in the range set by the range setting unit 604, and the gradation level difference ΔL3 is calculated. The unit 606 calculates the gradation level difference ΔL3 between the target pixel data and the average value La. Range setting section 6
When the range set in 05 is 3 × 3 pixels as shown in FIG. 3, the gradation level difference ΔL3 is calculated as follows.

ΔL3 = | A−La | = | A− (A + B + C + D + E + F + G + H + I) / 9
│ The above formula is a formula for calculating ΔL3 for the target pixel A, but regarding each target pixel extracted together with one target pixel, they are regarded as the target pixel A, and the same process as the above formula is performed for all target pixels. ΔL3 is calculated for.

Next, the second threshold value ΔL4 set by the gradation level difference setting unit 607 is compared with the gradation level difference ΔL3 to determine whether ΔL3 ≧ ΔL4.

This determination is a process for determining a defective pixel to be actually corrected from the pixels narrowed down as the target pixel of the correction process. The pixels narrowed down as the target of the correction process may include, for example, pixels forming the boundary of the object and pixels forming the gradation. Therefore, by determining whether the difference in gradation level between the pixel of interest and the average value of pixels in a wide range in the vicinity thereof is equal to or larger than the second threshold value ΔL4, the pixels forming a part of these normal images are corrected. Exclude from the target pixel.

As a method of determining the second threshold value ΔL4, for example, a local image forming a boundary of an object or a gradation portion is experimentally acquired from a natural image or the like, and a gradation level ΔL3 (ΔL3R) there is calculated. , Δ
A method of determining the minimum numerical value of L3R <ΔL4 as ΔL4 can be considered.

If ΔL3 ≧ ΔL4 as a result of the above judgment, the pixel of interest is finally judged as a defective pixel and extracted.
This extraction processing is performed by the pixel extraction unit 608. ΔL3
<ΔL4, the target pixel is not determined to be a defective pixel,
The pixel of interest is moved to the next, and the process from the extraction of the target pixel by the target pixel extraction unit 603 is sequentially repeated.

By this procedure, dust noise existing in image data can be detected automatically and efficiently.
Dust noise is detected while gradually reducing the number of candidate pixels. Since the determination that requires time for processing is performed at a stage where the number of pixels is narrowed down to some extent, it is not necessary to perform detailed determination processing for all pixels, and the time required for noise detection can be shortened. That is, in each stage, a candidate for noise, which is a defective pixel, is detected based on the gradation level difference between the pixel of interest and its neighboring pixels. Then, while changing the neighborhood range stepwise (while expanding the neighborhood range), the gradation level difference with the pixels in the neighborhood range is obtained, and the defective pixel is determined based on the difference. At this time, in the first stage of noise detection, the neighborhood range is narrowed to improve the processing efficiency, and in the advanced stage, the neighborhood range is expanded to improve the accuracy of the determination.

<Correction of Defective Pixel> When the defective pixel is extracted by the defective pixel extraction unit 105 as described above, the process returns to step S203 of FIG. 2 and the defective pixel correction process is performed. In step S203, the so-called median filter is processed by the median filter 106.
Will be done at. The process (median filter process) performed by the median filter 106 will be described in detail below.

The median filter 106 first pays attention to the extracted defective pixel, and obtains the intermediate value of the pixel data in the range set by the filter range setting unit 107 (hereinafter referred to as the filter range). The filter range is 3 as shown in Fig. 3.
In the case of × 3 pixels, the target pixel is A, and the surrounding 8 pixels are B to
I. Here, it is assumed that D <C <B <E <G <F <H <I <A when these pixels are arranged in the order of gradation levels.

Next, the median filter 106 finds the pixel having the intermediate value of the gradation level (in this case, the fifth gradation level G) from the surrounding 8 pixels including the target pixel A, The difference between the gradation level of the intermediate value pixel G and the gradation level of the target pixel A is the gradation level threshold setting unit 10
When it is equal to or higher than the threshold of the gradation level set in 8, the gradation level of the target pixel A is replaced with the gradation level of the intermediate value pixel G.

By such median filter processing of replacing the pixel of interest with the intermediate value of its neighboring pixels,
The defective pixel A has been repaired. In this case, if the image data is composed of RGB image data, RG
B If the intermediate values of each gradation level are different pixels, different pixel data may be replaced.

The filter strength setting unit 109 is a setting unit for the user to set the strength with which dust noise should be removed. For example, "standard""strong"
Setting contents such as "weak" are conceivable. Depending on the contents set by the filter strength setting unit 109,
The filter range set by the range setting unit 107 and the gradation level threshold set by the gradation level threshold setting unit 108 are variably controlled.

For example, the "strong" setting makes the filter range wider than the "standard" setting. By doing so, the defective pixel is easily replaced by the intermediate pixel. That is, as the filter range is expanded, pixels having a weak correlation with the defective pixel are included in the filter range.
For this reason, the range of the gradation level of the pixels in the filter range is widened, and the pixel having a larger gradation level difference from the defective pixel is easily selected as the intermediate value pixel. Therefore, even if the threshold value has the same gradation level as the “standard”, the defective pixel is likely to be replaced with the intermediate value pixel.

In this case, for example, the filter range is set to 7 × 7 pixel range in “standard”, the filter range is set to 3 × 3 pixel range in “weak”, and the filter range is set to 9 × 9 in “strong”.
A filter range is set such as a pixel range.

Further, by setting the threshold value of the gradation level to be smaller in the "strong" setting than in the "standard" setting, the defective pixel can be easily replaced with the intermediate value pixel. Of course, both the filter range and the threshold of the gradation level may be variably controlled.

On the other hand, when the filter strength setting is "weak", the filter range is narrowed contrary to "strong", the gradation level threshold is increased, or both are performed. Thus, it is possible to make it difficult for the defective pixel to be replaced with the intermediate value pixel. Further, by setting these combinations, not only “standard”, “strong”, and “weak” but also more detailed settings are possible.

As described above, the dust noise can be removed with the strength desired by the user.

The filter strength setting unit 109 may have a structure capable of independently setting the filter range and the threshold of the gradation level.

By connecting an output device such as a display device or a printing device to the image processing device, an image from which dust noise is removed can be output.

<Noise Removal Processing by Computer> The configurations of FIGS. 1 and 6 described above can also be realized by a general-purpose digital computer such as a personal computer. To this end, the computer program describing the procedure of FIG. 2 is loaded into the memory 1312 of the digital computer having the configuration of FIG. 7, and the program is executed by the CPU 1315, thereby realizing the function of each block of FIG. .

In FIG. 7, the computer main body 1310
Are connected to the image reading apparatus 101 and the printing apparatus 1302. The image reading apparatus 101 has an interface 13
11 and the printing apparatus 101 is connected via the interface 1317. In the memory 1312, a general-purpose memory portion for storing programs and a work area and an image memory portion corresponding to the image memory 104 in FIG. 1 are secured. Secondary storage device 1
314 is a fixed medium device such as a hard disk,
An optical disk, a removable magnetic disk, a magneto-optical disk, or a replaceable non-volatile memory unit, which can supply or store image data or a computer program shown in FIG. The CPU 1315 manages the hardware and software resources of the computer and executes the computer program including the procedure shown in FIG. 2 to realize that the computer functions as the configuration of FIG. The operation unit 1316 is a keyboard or the like for the operator to input as necessary. Threshold ΔL2,
ΔL4, the strength of the filter, and the like can be set by the operation unit 1316.
You can also enter from. Display 1313 is CR
It is a display device using T, liquid crystal, etc., and is used for image display and the like.

When the procedure of FIG. 2 is executed by the computer having such a configuration, the printing device 1302 can print out the noise-removed image. In this case, it can be applied to a copying machine or the like in order to remove noise without the intervention of an operator.

(Second Embodiment) An image processing method and an image processing apparatus according to a second embodiment of the present invention will be described with reference to FIG. FIG. 4 is a block diagram showing a functional configuration of the image processing apparatus. The blocks having the same functions as those in the first embodiment are designated by the same reference numerals and the description thereof will be omitted. In FIG. 4, 401 is an image processing unit to which the present invention is applied, and 402 is a document type setting unit.

With respect to the image processing method according to the second embodiment having the above-mentioned structure, parts different from those of the first embodiment will be described. The difference from the first embodiment is that the filter range and the threshold of the gradation level are changed according to the type of the original in the present embodiment.

The document type setting unit 402 is a setting unit for setting the type of document that the user intends to read with the image reading apparatus 101. For example, setting contents such as “photograph” and “printed matter” can be considered. Typically,
Since many people and natural landscapes are often photographed in a “photograph” document, there is a strong correlation between the gradation level of a pixel and the gradation level of a pixel that is close to it. Is often smooth. On the other hand, in the case of a “printed matter” document, since it is often composed of characters and charts, there is little correlation between the gradation level of a certain pixel and the gradation level of a neighboring pixel, In many cases, the gradation level changes rapidly.

Therefore, the filter range set by the range setting unit 107 and the gradation level threshold set by the gradation level threshold setting unit 108 are variably controlled according to the setting contents. . For example, the type of document is "printed"
In the “photograph” setting, the filter range is made wider than in the setting, so that the defective pixel is easily replaced by the intermediate value pixel. That is, by widening the filter range, a pixel having a weak correlation with the defective pixel, that is, a pixel having a larger gradation level difference from the defective pixel is selected as the intermediate value pixel, and the same gradation as that of the “printed matter” is selected. Even at the level threshold, defective pixels are likely to be replaced by intermediate-value pixels.
Also, by setting the gradation level threshold to be smaller in the “photograph” setting than in the “printed matter” setting, the defective pixel can be easily replaced with the intermediate value pixel. Of course, both the filter range and the threshold of the gradation level may be variably controlled.

By setting these combinations,
It is possible to set not only "photograph" and "printed matter" but also an intermediate document.

Further, in the case of a “printed matter” document, control may be performed to prohibit the processing of the median filter 106.

As described above, the dust noise can be removed with the optimum setting according to the type of the original of the image data.

(Third Embodiment) An image processing method and an image processing apparatus according to a third embodiment of the present invention will be described with reference to FIG. FIG. 5 is a block diagram showing a functional configuration of the image processing apparatus, blocks having the same functions as those in the first embodiment are designated by the same reference numerals, and description thereof will be omitted. In FIG. 5, 501 is an image processing unit to which the present invention is applied, and 502 is a resolution setting unit.

In the above-mentioned configuration, the image processing method according to the third embodiment will be described by focusing on the differences from the first embodiment.

The resolution setting unit 502 is a setting unit for setting the resolution when the user reads with the image reading apparatus 101. For example, "300dpi""600dpi"
Such setting contents are possible. The physical distance between pixels is 1/2 in "600dpi" compared to "300dpi"
Therefore, the change in gradation level between adjacent pixels is smooth.

The filter range set by the range setting unit 107 and the gradation level threshold set by the gradation level threshold setting unit 108 are variably controlled according to the setting contents of the resolution. . For example, in the case of "600dpi" setting compared to the case of "300dpi" setting, by making the filter range twice as wide, the size of the image becomes the same, and the defective pixel is easily replaced by the intermediate value pixel. In other words, by widening the filter range, pixels with weak correlation with defective pixels, that is, pixels with a larger gradation level difference with defective pixels, are selected as intermediate value pixels, and the same gradation as "300 dpi" is selected. Even with the level threshold, the median filtering effect similar to that at "300 dpi" can be obtained. Also,
Even if the threshold value of the gradation level is set smaller in the “600 dpi” setting than in the “300 dpi” setting, the defective pixel can be easily replaced with the intermediate value pixel. Of course, both the filter range and the threshold of the gradation level may be variably controlled.

By setting these combinations, it is possible to set for any resolution in this way.

As described above, it is possible to remove the dust noise with the optimum setting according to the resolution of the image data.

Even when the present invention is applied to a system composed of a plurality of devices (eg, host computer, interface device, reader, printer, etc.), a device composed of one device (eg, copying machine, facsimile). Device).

Further, an object of the present invention is to supply a storage medium (or a recording medium) recording a program code of software for realizing the functions of the above-described embodiments to a system or an apparatus, and a computer of the system or the apparatus ( Alternatively, it is achieved by the CPU or MPU) reading and executing the program code stored in the storage medium.

In this case, the program code itself read from the storage medium realizes the function of the above-described embodiment, and the storage medium storing the program code constitutes the present invention.

Further, by executing the program code read by the computer, not only the functions of the above-described embodiment are realized, but also the operating system (OS) running on the computer is executed based on the instruction of the program code. ) And the like perform a part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

Furthermore, after the program code read from the storage medium is written in the memory provided in the function expansion card inserted into the computer or the function expansion unit connected to the computer, based on the instruction of the program code. It also includes a case where a CPU or the like included in the function expansion card or the function expansion unit performs a part or all of the actual processing, and the processing realizes the functions of the above-described embodiments.

[0070]

As described above, according to the present invention, the dust noise due to dust or the like is appropriately removed according to the strength of removing the dust noise, the type of the original image, or the resolution of the image data. It becomes possible.

[Brief description of drawings]

FIG. 1 is a block diagram of an image processing apparatus according to a first embodiment of the present invention.

FIG. 2 is a flowchart illustrating the operation of the first embodiment of the present invention.

FIG. 3 is a diagram showing a pixel data arrangement according to the first embodiment of the present invention.

FIG. 4 is a block diagram of an image processing apparatus according to a second embodiment of the present invention.

FIG. 5 is a block diagram of an image processing apparatus according to a third embodiment of the present invention.

FIG. 6 is a block diagram of a defective pixel extraction unit according to the first embodiment of the present invention.

FIG. 7 is a block diagram of a computer that implements the present invention.

Claims (20)

[Claims] 1. A setting step capable of changing a range in the vicinity of a pixel of interest, a first threshold value, or both, a gradation level of a pixel of interest in image data, and a floor of a pixel in a range in the vicinity of the pixel of interest. A replacement step of replacing the gradation level of the target pixel with the gradation level of the replacement pixel when the difference from the gradation level of the replacement pixel obtained based on the gradation level is equal to or more than the first threshold value; An image processing method comprising: 2. The image processing method according to claim 1, wherein in the setting step, the range in the vicinity of the target pixel is changed to be widened or narrowed. 3. The image processing method according to claim 1, wherein in the setting step, the first threshold value is changed to be lowered or raised. 4. The setting step increases the range in the vicinity of the pixel of interest and lowers the first threshold value, or narrows the range in the vicinity of the pixel of interest and raises the first threshold value. The image processing method according to claim 1. 5. The setting step changes the range in the vicinity of the pixel of interest, the first threshold value, or both in accordance with the type of image containing the pixel of interest. The image processing method according to item 1. 6. The image processing method according to claim 5, further comprising a prohibition step of not performing the replacement of the target pixel in the replacement step according to the setting of the document type. 7. The setting step changes the range in the vicinity of the pixel of interest, the first threshold value, or both in accordance with the resolution when the image including the pixel of interest is digitized. Item 5. The image processing method according to any one of Items 1 to 4. 8. The replacement step comprises, as the replacement pixel,
The image processing method according to claim 1, wherein among the pixels within the range near the target pixel, a pixel having an intermediate value of the gradation level as a gradation level is used. 9. A target pixel detection step of detecting, from image data, a target pixel group whose center pixel is a pixel whose difference in gradation level from a neighboring pixel is equal to or greater than a second threshold value; A reference value is obtained based on the gradation level value of the pixel within the first range, and the pixel whose difference in gradation level between the reference value and each pixel of the target pixel group is the third threshold value or more is defective. 9. The image processing according to claim 1, further comprising a defective pixel detecting step of detecting as a pixel, wherein the defective pixel detected by the defective pixel detecting step is set as the central pixel. Method. 10. A setting means capable of changing a range in the vicinity of the pixel of interest, a first threshold value, or both, a gradation level of the pixel of interest in the image data, and a floor of a pixel in the range in the vicinity of the pixel of interest. Replacement means for replacing the gradation level of the target pixel with the gradation level of the replacement pixel when the difference from the gradation level of the replacement pixel obtained based on the gradation level is equal to or more than the first threshold value; An image processing apparatus comprising: 11. The image processing apparatus according to claim 10, wherein the setting unit widens or narrows a range in the vicinity of the target pixel. 12. The image processing apparatus according to claim 10, wherein the setting unit changes the first threshold value so as to decrease or increase it. 13. The setting means widens a range in the vicinity of the pixel of interest and lowers the first threshold value, or narrows a range in the vicinity of the pixel of interest and raises the first threshold value. The image processing apparatus according to claim 10. 14. The setting unit changes the range in the vicinity of the pixel of interest, the first threshold value, or both in accordance with the type of image including the pixel of interest. The image processing device according to item 1. 15. The image processing apparatus according to claim 14, further comprising a prohibition unit that controls the replacement unit so that the pixel of interest is not replaced according to the setting of the document type. 16. The setting means changes the range in the vicinity of the pixel of interest, the first threshold value, or both depending on the resolution when the image including the pixel of interest is digitized. Item 14. The image processing device according to any one of items 10 to 13. 17. The replacement unit uses, as the replacement pixel, a pixel having an intermediate value of a gradation level as a gradation level among pixels in a range in the vicinity of the pixel of interest. The image processing device according to item 10. 18. A target pixel detection means for detecting, from image data, a target pixel group whose center pixel is a pixel having a gradation level difference from a neighboring pixel of not less than a second threshold value; A reference value is obtained based on the gradation level value of the pixel within the first range, and the pixel whose difference in gradation level between the reference value and each pixel of the target pixel group is the third threshold value or more is defective. 18. The image processing according to claim 10, further comprising: defective pixel detection means for detecting as a pixel, wherein the defective pixel detected by the defective pixel detection means is used as the central pixel. apparatus. 19. By executing by a computer,
A computer program that realizes the image processing device according to any one of claims 10 to 18. 20. A computer-readable storage medium storing the computer program according to claim 19.