JPH10210298A - Image processing method and device therefor - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and device capable of performing the excellent processing of an image. SOLUTION: A main signal S which is a pixel level of a pixel under consideration is guided to a contrast extracting part 10, a multiplication part 45, a filter operation part 41. Then a contrast C near the pixel under consideration is extracted in the contrast extracting part 10. Power P and power (1-P) are guided in a table reference part 20 according to the contrast C transmitted from the contrast extracting part 10. Then the power P and the power (1-P) are guided to the multiplication parts 45, 46 respectively. On the other hand, a filtered signal S' is generated by filtering the image by using a given image filter at the filter operation part 41. The main signal S and the filtered signal S' are P-folded or (1-P)-folded at the multiplication parts 45, 46 respectively and transmitted to an adder 47. Then an output signal S" is generated.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method and apparatus for processing an image obtained from a scanner, a digital camera, or another apparatus.

[0002]

2. Description of the Related Art FIG. 10 is a diagram showing a configuration of a conventional image processing apparatus. The main signal S of the pixel of interest is calculated by the filter
1 and enters the main signal adjustment unit 42. In the filter operation section 41, the main signal S is operated by a two-dimensional image filter to generate a filtered signal S '. Then, the main signal adjustment unit 42 generates a signal P · S obtained by multiplying the main signal S by P based on a preset magnification P. Here, P is a value satisfying 0 ≦ P ≦ 1. On the other hand, the filtered signal S ′ generated by the filter operation unit 41
Is a signal (1P) multiplied by (1-P) by a preset magnification (1-P) in the filtered signal adjustment unit 43.
−P) · S ′. Then, the signal P · S from the main signal adjustment unit 42 and the signal (1−1) from the filtered signal adjustment unit 43
The output signal S ″ is generated by combining the P) · S ′ with the adder 44.

For example, in such a conventional image processing apparatus, when performing a smoothing process for removing noise from an image, an image filter having a size of 5 × 5 as shown in FIG. 11 is used. Then, the smaller the value of the preset magnification P (closer to 0), the greater the ratio of the output signal S ″ to be smoothed.

Therefore, the effect of filtering in image processing using an image filter is determined by the filter operation unit 41.
And the magnification P set in advance. Note that the magnification P and the magnification (1-P) indicate the mixing ratio of the main signal S and the filtered signal S ′ in the output signal S ″.

[0005]

However, when a smoothing process for removing noise is performed by the image processing apparatus as described above, the edge portion of the image is also smoothed and the edge of the original image is deteriorated. Problem.
Such edge deterioration significantly reduces the quality of an image.

Further, when the size of an image filter to be applied is increased to increase the noise suppressing effect, there is a problem that the deterioration of the edge is correspondingly increased.

[0007] Such a problem occurs not only in image smoothing for removing noise, but also in image processing using an image filter such as image sharpening.

[0008] Also, in the median filter, edge deterioration occurs. This will be described using a median filter having a size of 5 × 1 as an example.

In the case of an ideal edge having no noise as shown in FIG. 12 (see FIG. 12A), even if the median filter processing is performed on the target pixel, FIG.
As shown in (1), since the original edge level L2 is preserved, the edge is not deteriorated.

However, noise is mixed in an actual image, and an edge where noise N is mixed as shown in FIG. In this case, if a median filter is applied to the target pixel shown in FIG. 13A to perform a median filter process, the noise N may have a median value as shown in FIG. Receiving
It will deteriorate.

Such a problem arises from the fact that an image filter having the same characteristics is applied to all pixels constituting an image, and the magnification P is the same for all pixels.

[0012] The present invention has been made in view of the above problems, and has as its object to provide an image processing method and apparatus capable of performing favorable processing on an image.

[0013]

According to one aspect of the present invention, there is provided a method for performing a predetermined process on an original image for each pixel, wherein a target pixel to be processed is processed. A filter operation step of performing an arithmetic operation on the main signal expressing the pixel level of the pixel level using an image filter of a predetermined size to generate a filtered signal, and extracting a contrast from a pixel group including the target pixel and its neighboring pixels. A desired contrast deriving step, a mixing ratio deriving step of deriving a mixing ratio between the main signal and the filtered signal according to the contrast obtained in the contrast deriving step, and mixing the main signal and the filtered signal based on the mixing ratio And an output signal generating step of generating an output signal.

According to a second aspect of the present invention, there is provided the method according to the first aspect, wherein the noise component contained in the original image is removed using an image filter.
A step of setting a mixing ratio of the main signal higher than the filtered signal when the contrast is high, and setting a higher mixing ratio of the filtered signal than the main signal when the contrast is low; The calculation step includes a step of selecting a pixel level having a median value from among pixel levels of a plurality of pixels existing within a predetermined size to obtain a filtered signal.

According to a third aspect of the present invention, there is provided an apparatus for performing a predetermined process on an original image for each pixel, wherein the main signal representing a pixel level of a pixel of interest to be processed has a predetermined size. Filter operation means for performing an operation process using the image filter of the above to generate a filtered signal, contrast derivation means for obtaining a contrast from a pixel group including the pixel of interest and its neighboring pixels, and contrast obtained by the contrast derivation means. A mixing ratio deriving unit that derives a mixing ratio between the main signal and the filtered signal, and an output signal generating unit that generates an output signal by mixing the main signal and the filtered signal based on the mixing ratio. ing.

According to a fourth aspect of the present invention, in the apparatus according to the third aspect, the processing for removing a noise component contained in the original image using an image filter is performed.
Means for setting a higher mixing ratio of the main signal than the filtered signal when the contrast is high, and setting a higher mixing ratio of the filtered signal than the main signal when the contrast is low; The arithmetic means includes means for selecting a pixel level having a median value from among pixel levels of a plurality of pixels existing within a predetermined size to obtain a filtered signal.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION

<Overall Configuration of Apparatus> An example of the overall configuration of an apparatus to which an embodiment of the present invention is applied will be described. FIG. 1 is a diagram showing an overall configuration of an apparatus to which an embodiment of the present invention is applied. The image input unit 100 optically reads a document, such as an input scanner, and generates digital image data. The generated image data is stored in the image processing unit 20.
0 is transferred. The image data is stored in the image processing unit 200.
Is subjected to a predetermined process, and is output to an image output unit 300 such as an output scanner. In the output scanner 300, the image is recorded on a recording medium such as a film. Note that an operation input unit 201 and an information display unit 202 are provided to cause the image input unit 100, the image processing unit 200, and the image output unit 300 to perform desired operations of the operator. The image processing unit 2
00 is provided with an image processing device 250 to which the embodiment of the present invention is applied.

<Overview of Image Filter Operation> m × n
Filter operation processing for smoothing or sharpening an image by scanning a matrix image filter (m and n are integers of 2 or more) in an image plane is referred to as image filtering.

FIG. 2 is an explanatory diagram showing an outline of image filtering. With respect to the image I shown in FIG. 2A, the X direction (horizontal direction) is set as a main scanning direction, and the Y direction (vertical direction) is set as a sub-scanning direction, and a pixel (target pixel) OP to be processed is scanned pixel by pixel. While filtering sequentially. At the time of filtering, the target pixel OP is positioned at the center of the image filter F as shown in FIG. FIG.
In the example of (b), a 5 × 5 matrix image filter is used.

As the matrix image filter, an image filter generally called a median filter, an image filter that performs filtering based on a weighting coefficient assigned to each matrix component, and the like are applied.

First, the filtering by the median filter will be described. The median filter is used for an image smoothing process. At the time of filtering, pixels within the range of the filter (window) are taken as samples, their pixel levels are rearranged in the order of magnitude, and the center value is taken. Value.

For example, the target pixel OP is located at the center of the median filter F as shown in FIG. FIG.
In the example of (b), a 5 × 5 matrix median filter is used. If a median filter having a size of 5 × 5 is used as described above, the filtered signal value of the target pixel OP is generated by changing the pixel level of the target pixel OP to the median value of the 25 pixel levels. I do.

Next, the filtering will be described based on the weighting factors assigned to the respective components in the matrix. A value obtained by performing a weighted average of the pixel level of the target pixel OP and the pixel levels of the peripheral pixels based on the weighting coefficient assigned to each component of the image filter is set as a filtered signal value of the target pixel OP.

That is, assuming that arbitrary natural numbers are m and n, the image filter F becomes (2m + 1) × as shown in FIG.
It is _assumed that the matrix has a size of (2n + 1) and each component is assigned a weighting coefficient of k- _mn to kmn. A filtered signal S ′ _xy obtained by performing filtering on the target pixel located at the coordinates (x, y) using the image filter F is

[0025]

(Equation 1)

## EQU1 ## Here, S _xy is the pixel level of the pixel located at the coordinates (x, y) of the original image.

In the image filter shown in FIG.
If the weighting coefficients corresponding to the target pixel and its surrounding pixels are all positive numbers, the filter is a smoothing filter for performing image smoothing. In contrast, only the weighting coefficients k ₀₀ corresponding to the pixel of interest is a positive number, if all the weighting coefficients of the other peripheral pixels is negative, a sharpening filter for performing image sharpening. That is, the frequency characteristic of the weighting coefficient becomes the characteristic of the image filter.

<Image Processing Apparatus> Next, an image processing apparatus according to an embodiment of the present invention will be described. FIG. 4 is a configuration diagram illustrating the image processing apparatus according to the present embodiment. The input of this image processing device is a main signal S which is the pixel level (density value) of the pixel of interest. The main signal S is supplied to a contrast extraction unit 10, a multiplication unit 45, and a filter operation unit 4.
Led by one.

The contrast extracting section 10 extracts the contrast of the image. The upper neighboring pixel level of the target pixel is YU, the lower neighboring pixel level is YL, and the left neighboring pixel level is X.
Let L be the pixel level on the right and XR be the pixel level on the right.

[0030]

(Equation 2)

As a result, the contrast C can be obtained. Here, the upper neighboring pixel level YU may be the pixel level of the pixel P1 adjacent above the target pixel as shown in FIG. 5A, or may be the upper pixel level YU of the target pixel as shown in FIG. The pixel level of the pixel P5 located a few pixels away may be used. In addition, if the pixel level of one pixel is set to the neighboring pixel level, a correct contrast cannot be obtained if the pixel contains a lot of noise.
As shown in (c), a plurality of pixel groups P above the pixel of interest
The average value or median of the pixel levels of G1 may be used as the upper neighboring pixel level YU. Similarly, the lower neighboring pixel level YL, the left neighboring pixel level XL, and the right neighboring pixel level XR are shown in FIG.
The adjacent pixels P3, P4 and P2 shown in FIG. 5A may be used, or the pixels P7 and P which are several pixels away from the pixel of interest shown in FIG.
8 and P6, or the pixel groups PG3 and PG3 shown in FIG.
The average value or the median value of PG4 and PG2 may be used.

As another method for obtaining the contrast, the maximum pixel level among a plurality of pixels including the target pixel OP shown in FIGS. 5A, 5B, and 5C is Dmax, and the minimum pixel level is Dmax. Let Dmin be the pixel level of

[0033]

(Equation 3)

Thus, the contrast C can be obtained. Further, as shown in FIG. 6, a pixel level indicating the maximum value is selected from the pixel group PG including the target pixel OP and its surrounding pixels, and is set as the maximum pixel level Dmax, and the minimum value is indicated from the pixel group PG. By selecting a pixel level and setting it as the minimum pixel level Dmin, the contrast C can be derived from Expression 3.

In general, a high contrast indicates an edge portion of an image, and a low contrast indicates a flat portion with a small change in density of the image. Therefore, by determining the contrast C as described above, it is possible to determine what portion of the image the target pixel OP is.

The contrast C obtained by the contrast extracting unit 10 is sent to the table reference unit 20. The table reference section 20 includes a storage medium such as a memory therein, and obtains a magnification P corresponding to the input contrast C by referring to a table stored in advance. The magnification P is a value within a range from “0” to “1”. FIG. 7 is a diagram illustrating an example of the relationship between the input (contrast C) in the table reference unit 20 and the obtained magnification P. As shown in FIG. 7, when the contrast C is smaller than the predetermined value C1, the value of the magnification P is "0". When the value is larger than the predetermined value C2, the value of the magnification P becomes "1". Then, the contrast C is changed from C1 to C
As the number increases to 2, the magnification P increases at a constant rate. In the case of the relationship as shown in FIG. 7, when the contrast C has a value larger than C2, the target pixel corresponds to an edge portion of the image. When the contrast C has a value smaller than C1, the target pixel corresponds to the image. It has been determined that it corresponds to a flat part. Note that the relationship between the contrast C and the magnification P may be a relationship other than that shown in FIG.

Then, the table reference unit 20 sends the magnification P obtained as described above to the multiplication unit 45, calculates the magnification (1-P), and sends the obtained value to the multiplication unit 46. Note that the magnification P and the magnification (1-P) are equivalent to the output signal S ″
5 shows the mixing ratio of the main signal S and the filtered signal S ′.

On the other hand, the filter operation unit 41 performs filtering on the target pixel using a preset image filter, and generates a filtered signal S ′.

When the image filter is a median filter, a preset median filter is applied in the filter operation section 41. For example, when a median filter having a size of 5 × 5 is set, a filtered signal S ′ is generated by selecting a median value among pixel levels of 25 pixels located in the filter area. When a median filter having a size of 3 × 3 is set, 9
The filtered signal S 'is generated by selecting the median of the pixel levels of the pixels. In the filtering of an image using such a median filter, by selecting a median value from a group of pixels in a filter region including a target pixel, noise having a value greatly different from an average pixel level of the group of pixels is removed. Is done.

When the image filter used in the filter operation unit 41 is an image filter that performs filtering based on a weighting coefficient assigned to each component in a matrix, each of the image filters shown in FIG. The calculation shown in Expression 1 is performed based on the weighting coefficient assigned to the component, and a filtered signal S ′ of the pixel of interest is generated. Note that the image filter shown in FIG. 11 functions as a smoothing filter for removing a noise component of an image.

The filtered signal S ′ thus generated by the filter operation unit 41 is sent to the multiplication unit 46.

The product of the main signal S and the magnification P is derived in the multiplier 45 to generate a signal PS. The multiplier 46 calculates the product of the filtered signal S 'and the magnification (1-P). Thus, a signal (1-P) .S 'is generated. And adder 4
7, finally the signal PS and the signal (1-P) .S '
And an output signal S ″ is generated, and transmitted to the subsequent device. Therefore, when the contrast C detected by the contrast extracting unit 10 is large, the signal P ·
The value of S increases and the value of the signal (1-P) · S ′ decreases. Conversely, when the contrast C is small, the value of the signal P · S becomes small, and the value of the signal (1-P) · S ′ becomes large. Such a relationship has the effect of transmitting the main signal S to the subsequent stage for the edge portion of the image and transmitting the filtered signal S 'to the subsequent stage for the flat portion of the image.

Then, as described above, the filter operation unit 41
If the processing performed in step (1) is smoothing processing for the purpose of removing noise components, etc., the edge information of the original image should be accurately transmitted to the subsequent stage without smoothing the edge part of the image. As for the flat part of the image, the information of the image from which noise has been removed by the filter operation unit 41 can be transmitted to the subsequent stage.

When the image processing apparatus performs an image sharpening process, for example, the table reference unit 20
The relationship between the contrast C and the magnification P used in (1) is as shown in FIG. 8, and the image filter used in the filter operation unit 41 is the image filter shown in FIG.

In such a case, the filtered signal S 'obtained from the filter operation unit 41 is a signal in which the high frequency component is emphasized. Therefore, the contrast C of the image is C1
For smaller portions, the magnification P is increased in order to transmit the main signal S to the subsequent stage in order not to emphasize the noise component. On the other hand, since the portion where the contrast C is larger than C2 is the edge portion of the image, the filtered signal S 'is transmitted to the subsequent stage to further enhance the edge.

When the processing for the target pixel is completed in this way, the same processing is performed with the next pixel to be processed in the next scanning order as the target pixel. By repeating this, predetermined processing ends for the entire image.

By adopting such an apparatus configuration, in the image smoothing processing, the edge of the image can be satisfactorily removed without deteriorating the edge of the image and the flat portion of the image. Then, in the image sharpening process, it is possible to enhance only the edge portion of the image without enhancing the noise component in the flat portion of the image.

<Modification> The image processing apparatus described in this embodiment can easily determine whether or not an image is an edge portion. No processing is performed on a flat portion having a small density change, and the present invention can also be applied to image processing which produces a special effect of an image by blurring only the edge portion of the image. In this case, the relationship between the contrast C and the magnification P referred to by the table reference unit 20 is as shown in FIG. 8, and the image filter applied by the filter calculation unit 41 is a median filter or FIG. The image filter is assigned a weighting coefficient for smoothing as shown in FIG.

Further, the two-dimensional contrast is obtained in deriving the contrast. However, the present invention can be applied to a one-dimensional image obtained by a CCD line sensor or the like. The extraction may be performed only in one direction (the direction in which the pixels of the CCD line sensor are arranged).

[0050]

As described above, according to the first and third aspects of the present invention, an image filter of a predetermined size is used for a main signal expressing the pixel level of a target pixel to be processed. A filtered signal is generated by performing arithmetic processing, a contrast is obtained from a pixel group including the pixel of interest and its neighboring pixels, and a mixing ratio between the main signal and the filtered signal is derived according to the obtained contrast. To generate an output signal by mixing the main signal and the filtered signal. Therefore, in the image smoothing process, the edge of the image can be satisfactorily removed without deteriorating the edge of the image and the flat portion of the image. In the image sharpening process, it is possible to emphasize only the edge portion of the image without enhancing the noise component in the flat portion of the image. That is, the image quality is not reduced.

According to the second and fourth aspects of the present invention, when the contrast is high, the mixing ratio of the main signal is set higher than that of the filtered signal, and when the contrast is low, the filtering ratio of the main signal is higher than that of the main signal. The ratio of the mixed signals is set high. In addition, a pixel level having a median value among the pixel levels of a plurality of pixels existing within a predetermined size is selected as a filtered signal. Therefore, while the edge information of the original image can be left for the edge portion of the image, the noise component of the flat portion of the image can be satisfactorily removed. Even if the size of the image filter is increased to further enhance the noise suppression effect, the edge portion of the image does not deteriorate.

[Brief description of the drawings]

FIG. 1 is a diagram showing an overall configuration of an apparatus to which an embodiment of the present invention is applied.

FIG. 2 is an explanatory diagram illustrating an outline of image filtering.

FIG. 3 is a diagram illustrating weighting coefficients of respective components of an image filter.

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

FIG. 5 is an explanatory diagram for explaining an image area for extracting a contrast.

FIG. 6 is an explanatory diagram for explaining an image area for extracting a contrast.

FIG. 7 is a diagram illustrating a relationship between an input and a magnification in a table reference unit.

FIG. 8 is a diagram illustrating a relationship between an input and a magnification in a table reference unit.

FIG. 9 is a diagram illustrating an example of an image filter for image sharpening processing.

FIG. 10 is a diagram showing a configuration of a conventional image filtering device.

FIG. 11 is a diagram illustrating an example of an image filter for an image smoothing process.

FIG. 12 is an explanatory diagram for describing processing by a median filter of an ideal edge having no noise.

FIG. 13 is an explanatory diagram for describing processing by a median filter of an edge mixed with noise.

[Explanation of symbols]

 Reference Signs List 10 Contrast extraction unit 20 Table reference unit 45, 46 Multiplication unit 47 Adder 41 Filter operation unit 100 Image input unit 200 Image processing unit 201 Operation input unit 202 Information display unit 250 Image filtering unit 300 Image output unit

Claims (4)

[Claims] 1. A method for performing a predetermined process for each pixel on an original image, wherein a main signal representing a pixel level of a pixel of interest to be processed is calculated using an image filter of a predetermined size. A filter operation step of performing processing to generate a filtered signal; a contrast deriving step of obtaining a contrast from a pixel group including the target pixel and its neighboring pixels; and a main signal according to the contrast obtained in the contrast deriving step. A mixing ratio deriving step of deriving a mixing ratio between the filtered signal and the filtered signal; and an output signal generating step of generating an output signal by mixing the main signal and the filtered signal based on the mixing ratio. An image processing method comprising: 2. The method according to claim 1, wherein a process of removing a noise component included in the original image using an image filter is performed. Setting the mixing ratio of the main signal higher than the filtered signal, and setting the mixing ratio of the filtered signal higher than the main signal when the contrast is low. The image processing method according to claim 1, further comprising: selecting a pixel level having a median value from among pixel levels of a plurality of pixels existing within the predetermined size as the filtered signal. 3. An apparatus for performing a predetermined process for each pixel on an original image, wherein a main signal representing a pixel level of a target pixel to be processed is calculated using an image filter of a predetermined size. Filter operation means for performing a process to generate a filtered signal; contrast derivation means for obtaining a contrast from a pixel group including the pixel of interest and its neighboring pixels; and the main signal according to the contrast obtained by the contrast derivation means A mixing ratio deriving unit that derives a mixing ratio between the filtered signal and the filtered signal; and an output signal generating unit that generates an output signal by mixing the main signal and the filtered signal based on the mixing ratio. An image processing apparatus characterized by the above-mentioned. 4. The apparatus according to claim 3, wherein a process for removing a noise component contained in the original image using an image filter is performed. Means for setting the mixing ratio of the main signal higher than the filtered signal, and setting the mixing ratio of the filtered signal higher than the main signal when the contrast is low. An image processing apparatus comprising: means for selecting a pixel level having a median value from among pixel levels of a plurality of pixels existing within the predetermined size and using the selected pixel level as the filtered signal.