CN100550981C - Image processing device and method - Google Patents

Abstract

The invention discloses an image processing device and method, which are used for inhibiting noise of an acquired or recorded image in a color filter array domain. The method according to the present invention first divides the image into a green plane, a red plane and a blue plane. Then, each color plane is covered by a working window in sequence, so that each pixel on each color plane is positioned in the center of the working window in sequence and is designated as a central pixel. Then, in the working window, other pixels except the central pixel are classified into edge pixels and non-edge pixels. Finally, a filtered center pixel is calculated according to a formula and the filtered center pixel is substituted for the center pixel, thereby suppressing the noise of the image in the color filter array domain.

Description

Image processor and method

Technical field

The present invention relates to a kind of image processor and method, particularly relate to a kind of image processor and method that is used for noise suppressed (Noise suppression).

Background technology

In digital image systems such as digital camera, digital camera, the raw video of self-induction or one main body/background that acquisition component obtained (Raw image) is subjected to multiple noise easily and (is not present in this main body or the background, but come across the element in this image) interference, thereby influence the quality of this image.Described noise may be due to the characteristic of because the assembly of image system own (as, inductive component), perhaps since image system in the image processing process, noise is added in the image in order to reach some purpose.In an image, may be oriented to the pixel that contains noise or the characteristic or the character of pixel region, and may be oriented to the pixel of an edge or fine detail or the characteristic or the character of pixel region is very difficult to distinguish.Therefore, existing noise suppressing method usually removes the pixel or the pixel region of edge or fine detail, but not remove pixel or the pixel region that really contains noise, and therefore cause this image to produce blur effect (Blurring effect), therefore reduce the quality of image in the pixel of described edge that is removed or fine detail or the position at pixel region place.In addition, in chromatic image, aforesaid blur effect will cause color bleeding to surmount the edge of another pixel.

In prior art, when this main body/background by a device for image, as a digital camera, when carrying out imaging, this image captured to the color filter array with particular color channel pattern (Color filterarray, CFA).A pattern that is usually used in pick-up image is the Bayer pattern, and it has following color channels:

G?R?G?R?G?R...

B?G?B?G?B?G...

G?R?G?R?G?R...

Therefore, in the CFA of Bayer pattern each location of pixels (Pixel location) only all have with can be combined into full-color three kinds of Color plane (green, red, blueness) in the relevant intensity level (Intensity value) of a kind of Color plane.Two other assessment that lacks color is undertaken by the color interpolation method (Color interpolation) that is well known in the art on each location of pixels.Because most of traditional noise suppressed or the technology that removes are to carry out on the image with full-color Pixel Information, thus the color interpolation method through be everlasting noise suppressed or remove before carry out.Because color interpolation method itself can produce noise, cause the noise in the original picked image to mix, thereby may lose script and become characteristics of noise, and obtain becoming the feature of image characteristics with described noise by the generation of color interpolation method.

Traditionally, required internal memory and the handling procedure of noise removal of carrying out the full-color pixel that is obtained by the processing of color interpolation method will increase to three times (all having three times resolution because of each pixel), therefore will be very difficult and expensive by the hardware improvement if want.Carry out color space conversion (Color space conversion) after some noise reduction techniques is attempted in colored interpolation to yuv space, and only consider that (Y, noise suppressed Chrominance) is to reduce hardware burden for the look news.Yet, can propagate like this even than the more additional noise of color interpolation, therefore can't improve by hardware easily equally.

Therefore, noise suppressed needs to distinguish edge pixel and non-edge pixel, and need prior to finishing in the CFA territory, to promote treatment effeciency, reduce hardware cost simultaneously before any color interpolation.

Summary of the invention

Therefore, main purpose of the present invention is to provide a kind of image processor and method, and especially, image processor of the present invention and method can promote the efficient of noise suppressed, also can reduce hardware cost.

Be used to suppress in the color filter array territory noise of an image that captures or write down according to a kind of image treatment method of a preferred embodiment of the present invention.Further, this method comprises the following step:

At first, the image with this acquisition or record is divided into a green color plane, a red plane and a blue color planes.

Subsequently,, hide each Color plane with an operation window in regular turn, cause each pixel on each Color plane to be positioned at the central authorities of this operation window in regular turn, and be designated as a center pixel at this green color plane, this redness plane and this blue color planes.

Then, in this operation window, other pixel outside this center pixel of classifying is N edge pixel (C _{Edge, i}) and M non-edge pixel (C _{Non, j}).N and M are all the integer greater than 1, and i is an integer index of 1 to N, and j is an integer index of 1 to M.

At last, according to following formula, calculating one center pixel (C after filtration _Fc), and replace this center pixel (C with this filtered center pixel _c):

$C_{\mathrm{fc}}=\frac{C_c+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{edge},i}\·\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}{C}_{\mathrm{non},j}\·\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}{1+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}.$

In above formula, WEI1 is one first distribution function, and WEI2 is one second distribution function.

By this image treatment method, the noise that is present in the image of this acquisition in this CFA territory or record can be suppressed.

Be used to suppress in the color filter array territory noise of an image that captures or write down equally according to a kind of image processor of another preferred embodiment of the present invention.This image processor comprises a storage element and an image process unit.

This storage element is in order to a green color plane, a red plane and a blue color planes of the image of accepting this acquisition or record respectively.

This image process unit then is connected to this storage element, and at this green color plane, this redness plane and this blue color planes, hide each Color plane with an operation window in regular turn, cause each pixel on each Color plane to be positioned at the central authorities of this operation window in regular turn, and be designated as a center pixel (C _c).

This image process unit and in this operation window, other pixel outside this center pixel of classifying is N edge pixel (C _{Edge, i}) and M non-edge pixel (C _{Non, j}).N and M are all the integer greater than 1, and i is an integer index of 1 to N, and j is an integer index of 1 to M.

This image process unit and according to following formula calculates after filtration a center pixel (C _Fc), and replace this center pixel (C with this filtered center pixel _c):

$C_{\mathrm{fc}}=\frac{C_c+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{edge},i}\·\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}{C}_{\mathrm{non},j}\·\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}{1+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}.$

In above formula, WEI1 is one first distribution function, and WEI2 is one second distribution function.

By this image processor, the noise that is present in the image of this acquisition in this CFA territory or record can be suppressed.

Can be further understood by the following detailed description and accompanying drawings about the advantages and spirit of the present invention.

Description of drawings

Fig. 1 shows the flow chart according to the image treatment method of a specific embodiment of the present invention.

Fig. 2 A shows the green color plane schematic diagram with the operation window covering according to an instantiation of the present invention.

Fig. 2 B shows the red floor map with the operation window covering according to an instantiation of the present invention.

Fig. 2 C shows the blue color planes schematic diagram with the operation window covering according to an instantiation of the present invention.

Fig. 3 A shows the schematic diagram according to a Rayleigh distribution function of the present invention.

Fig. 3 B shows the schematic diagram according to the 2nd Rayleigh distribution function of the present invention.

First step function after the Rayleigh distribution function that Fig. 4 A shows Fig. 3 A is simplified.

Second step function after the 2nd Rayleigh distribution function that Fig. 4 B shows Fig. 3 B is simplified.

Fig. 5 shows the functional block diagram according to the image processor of a specific embodiment of the present invention.

The reference numeral explanation

1: image processor 12: image capture/record cell

14: separative element 16: storage element

18: image process unit 20: operation window

S50～S60: process step

Embodiment

The invention provides a kind of image processor and method that can effectively suppress the noise in the image.Specific embodiments of the invention and practical application embodiment below will be described in detail in detail, use proving absolutely feature of the present invention, spirit and advantage.

See also Fig. 1, Fig. 1 shows the flow chart according to the image treatment method of a specific embodiment of the present invention.This image treatment method can be used to a color filter array (Color filter array, CFA) noise of the image of inhibition one acquisition or record in the territory.Especially, this CFA's is arranged as a Bayer pattern.

As shown in Figure 1, this method comprises the following step:

Step S51 captures or writes down an image.

Step S53 is divided into a green color plane (Color plane), a red plane and a blue color planes with the image of this acquisition or record.

Step S55, at this green color plane, this redness plane and this blue color planes, hide each Color plane with an operation window (Working window) in regular turn, cause each pixel (Pixel) on each Color plane to be positioned at the central authorities of this operation window in regular turn, and be designated as a center pixel (C _c).

Step S57, in this operation window, other pixel outside this center pixel of classifying is N edge pixel (C _{Edge, i}) and M non-edge pixel (C _{Non, j}).N and M are all the integer greater than 1, and i is an integer index of 1 to N, and j is an integer index of 1 to M.

Step S59, according to following formula, calculating one center pixel (C after filtration _Fc), and replace this center pixel (C with this filtered center pixel _c):

$C_{\mathrm{fc}}=\frac{C_c+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{C}_{\mathrm{edge},i}\·\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}{C}_{\mathrm{non},j}\·\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}{1+\underset{i=1}{\overset{N}{\mathrm{\Σ}}}\mathrm{WEI}2\left(\right|C_c-C_{\mathrm{edge},i}\left|\right)+\underset{j=1}{\overset{M}{\mathrm{\Σ}}}\mathrm{WEI}1\left(\right|C_c-C_{\mathrm{non},j}\left|\right)}.$ [formula 1]

Wherein WEI1 is one first distribution function (Distribution function), and WEI2 is one second distribution function.

See also Fig. 2, Fig. 2 shows the Color plane schematic diagram with operation window 20 coverings according to an instantiation of the present invention.As shown in Figure 2, the image of aforesaid acquisition or record is divided into green (Fig. 2 A), red (Fig. 2 B) and blue three Color plane such as (Fig. 2 C), and hides by operation window 20.

Further, comprise center pixel G in the scope that is hidden by operation window 20 on the green color plane of Fig. 2 A _cAnd G ₁To G ₁₂Deng totally 13 pixels.In these 13 pixels, G ₂, G ₆, G ₇And G ₁₁Be classified as edge pixel etc. pixel, and G ₁, G ₃, G ₄, G ₅, G ₈, G ₉, G ₁₀And G ₁₂Be classified as non-edge pixel etc. pixel.Therefore, the method according to this invention, this center pixel G of this green color plane _cRecomputating via [formula 1] becomes filtered center pixel G _Fc:

G _fc＝{G _c+[G ₁·WEI1(|G _c-G ₁|)]+[G ₂·WEI2(|G _c-G ₂|)]+[G ₃·WEI1(|G _c-G ₃|)]+[G ₄·WEI1(|G _c-G ₄|)]+[G ₅·WEI1(|G _c-G ₅|)]+[G ₆·WEI2(|G _c-G ₆|)]+[G ₇·WEI2(|G _c-G ₇|)]+[G ₈·WEI1(|G _c-G ₈|)]+[G ₉·WEI1(|G _c-G ₉|)]+[G ₁₀·WEI1(|G _c-G ₁₀|)]+[G ₁₁·WEI2(|G _c-G ₁₁|)]+[G ₁₂·WEI1(|G _c-G ₁₂|)]}/{1+WEI1(|G _c-G ₁|)+WEI2(|G _c-G ₂|)+WEI1(|G _c-G ₃|)+WEI1(|G _c-G ₄|)+WEI1(|G _c-G ₅|)+WEI2(|G _c-G ₆|)+WEI2(|G _c-G ₇|)+WEI1(|G _c-G ₈|)+WEI1(|G _c-G ₉|)+WEI1(|G _c-G ₁₀|)+WEI2(|G _c-G ₁₁|)+WEI1(|G _c-G ₁₂|)}

In addition, other pixel on this green color plane is also calculated its filtered pixel value by this aforesaid mode in regular turn.

Further, comprise center pixel R in the scope that is hidden by operation window 20 on the red plane of Fig. 2 B _cAnd R ₁To R ₈Deng totally 9 pixels.In these 9 pixels, R ₂, R ₄, R ₅And R ₇Be classified as edge pixel etc. pixel, and G ₁, G ₃, G ₆And G ₈Be classified as non-edge pixel etc. pixel.Therefore, the method according to this invention, this center pixel R on this redness plane _cRecomputating via [formula 1] becomes filtered center pixel R _Fc:

R _fc＝{R _c+[R ₁·WEI1(|R _c-R ₁|)]+[R ₂·WEI2(R _c-R ₂|)]+[R ₃·WEI1(|R _c-R ₃|)]+[R ₄·WEI2(R _c-R ₄|)]+[R ₅·WEI2(|R _c-R ₅|)]+[R ₆·WEI1(|R _c-R ₆|)]+[R ₇·WEI2(|R _c-R ₇|)]+[R ₈·WEI1(R _c-R ₈|)]}/{1+WEI1(|R _c-R ₁|)+WEI2(|R _c-R ₂|)+WEI1(R _c-R ₃|)+WEI2(|R _c-R ₄|)+WEI2(R _c-R ₅|)+WEI1(|R _c-R ₆|)+WEI2(|R _c-R ₇|)+WEI1(R _c-R ₈|)}

Similarly, other pixel on this redness plane is also calculated its filtered pixel value by this aforesaid mode in regular turn.

Further, comprise center pixel B in the scope that is hidden by operation window 20 on the blue color planes of Fig. 2 C _cAnd B ₁To B ₈Deng totally 9 pixels.In these 9 pixels, B ₂, B ₄, B ₅And B ₇Be classified as edge pixel etc. pixel, and B ₁, B ₃, B ₆And B ₈Be classified as non-edge pixel etc. pixel.Therefore, the method according to this invention, this center pixel B on this redness plane _cRecomputating via [formula 1] becomes filtered center pixel B _Fc:

B _fc＝{B _c+[B ₁·WEI1(|B _c-B ₁|)]+[B ₂·WEI2(|B _c-B ₂|)]+[B ₃·WEI1(|B _c-B ₃|)]+[B ₄·WEI2(|B _c-B ₄|)]+[B ₅·WEI2(|B _c-B ₅|)]+[B ₆·WEI1(|B _c-B ₆|)]+[B ₇·WEI2(|B _c-B ₇|)]+[B ₈·WEI1(|B _c-B ₈|)]}/{1+WEI1(|B _c-B ₁|)+WEI2(|B _c-B ₂|)+WEI1(|B _c-B ₃|)+WEI2(|B _c-B ₄|)+WEI2(|B _c-B ₅|)+WEI1(|B _c-B ₆|)+WEI2(|B _c-B ₇|)+WEI1(|B _c-B ₈|)}

Similarly, other pixel on this blue color planes is also calculated its filtered pixel value by this aforesaid mode in regular turn.

In a specific embodiment, all be the Rayleigh distribution function according to first distribution function of the present invention and second distribution function.See also Fig. 3, Fig. 3 shows the schematic diagram according to Rayleigh distribution function of the present invention.In Fig. 3, transverse axis is represented the absolute value of center pixel and other pixel value difference; And the longitudinal axis represents this absolute value is brought into the value of Rayleigh distribution function gained.In this specific embodiment, when this other pixel is non-edge pixel, Rayleigh distribution function as shown in Figure 3A (first distribution function) is employed, and when this other pixel was edge pixel, the Rayleigh distribution function shown in Fig. 3 B (second distribution function) was employed.

In addition, in practical application, after can simplifying by the Rayleigh distribution function according to first distribution function of the present invention and second distribution function, and be performed as one first step function (Step function) and one second step function.See also Fig. 4 A and Fig. 4 B, first step function after the Rayleigh distribution function that Fig. 4 A shows Fig. 3 A is simplified; Second step function after the Rayleigh distribution function that Fig. 4 B shows Fig. 3 B is simplified.In practical application, this first step function is applicable to the calculating of described non-edge pixel; This second step function then is applicable to the calculating of described edge pixel.

Please note, the visual demand of image treatment method of the present invention and comprise different first distribution function and second distribution functions of many groups, for example, many groups first distribution functions of respectively corresponding different ISO values and second distribution function are to promote the effect of noise suppressed by this.

By said method, the present invention is according to the judgement of edge pixel and non-edge pixel, and utilize this first distribution function and second distribution function to recomputate the pixel value of each pixel in each Color plane respectively, and replace original pixel value with the pixel value that recomputates, use the noise of the image of this acquisition that suppresses to be present in this CFA territory or record.

Further see also Fig. 5, Fig. 5 shows the functional block diagram according to the image processor of a specific embodiment of the present invention, and this image processor is used to suppress in a color filter array (CFA) territory noise of an image that captures or write down.And especially, this CFA's is arranged as a Bayer pattern.

As shown in Figure 5, this image processor 1 comprises one image capture/record cell (Imagecapturing/recording unit), 12, one separative element (Splitting unit), 14, one storage element (Storage unit), 16 and one image process unit (Image processing unit) 18.

This image capture/record cell 12 is as CCD or CMOS sensing component, in order to this image of acquisition/record.This separative element 14 is connected to this image capture/record cell 12, and in order to this image is separated into a green color plane, a red plane and a blue color planes.In addition, this storage element 16 is linked to this separative element 14, and in order to accept this green color plane, this redness plane and this blue color planes respectively.

Further, this image process unit 18 is connected to this storage element 16, and at this green color plane, this redness plane and this blue color planes, hide each Color plane with an operation window in regular turn, cause each pixel on each Color plane to be positioned at the central authorities of this operation window in regular turn, and be designated as a center pixel (C _c).

This image process unit 18 and in this operation window, other pixel outside this center pixel of classifying is N edge pixel (C _{Edge, i}) and M non-edge pixel (C _{Non, j}).N and M are all the integer greater than 1, and i is an integer index of 1 to N, and j is an integer index of 1 to M.

In addition, this image process unit 18 and calculate after filtration a center pixel (C according to aforesaid [formula 1] _Fc), and replace this center pixel (C with this filtered center pixel _c).By above-mentioned each unit, image processor of the present invention can suppress to be present in the noise of the image of this acquisition in this CFA territory or record.

Note that this image processing unit is in order to calculate this filtered center pixel (C _Fc) formula identical with aforesaid [formula 1], and first distribution function wherein and second distribution function can be Rayleigh distribution function or step function equally, therefore here do not give unnecessary details.

In sum, image processor of the present invention and method at pixel of different nature (edge pixel and non-edge pixel), are calculated with corresponding distribution function, to reach the effect of noise suppressed in the color filter array territory.In addition, image processor of the present invention and method can promote usefulness, the reduction hardware cost that overall image is handled.The more important thing is that image processor of the present invention and method can suppress the noise in the image effectively, promote image quality, also can use different distribution functions in response to varying environment to reach optimum efficiency.

By the above detailed description of preferred embodiments, be to wish to know more to describe feature of the present invention and spirit, and be not to come category of the present invention is limited with above-mentioned disclosed preferred embodiment.On the contrary, its objective is that hope can contain in the category of claim of being arranged in of various changes and tool equality institute of the present invention desire application.Therefore, the category of claim of the present invention should be done the broadest explanation according to above-mentioned explanation, contains the arrangement of all possible change and tool equality to cause it.

Claims (6)

1. An image processing method for suppressing noise in a captured or recorded image in a color filter array domain, the method comprising the following steps: segmenting the captured or recorded image into a green plane, a red plane, and a blue plane; and For the green plane, the red plane, and the blue plane, sequentially cover each color plane with a working window, so that each pixel on each color plane is sequentially located at the center of the working window, and is designated as a central pixel C _c , and perform the following steps: In the working window, classify other pixels other than the central pixel into N edge pixels C _{edge, i} and M non-edge pixels C _{non, j} , N and M are integers greater than 1, and i ranges from 1 to N an integer index, j is an integer index from 1 to M; and A filtered central pixel C _fc is calculated and replaced by the filtered central pixel C _c according to the following formula: ${\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; fc</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span>$ where WEI1 is a first distribution function, and WEI2 is a second distribution function, wherein the first distribution function is a Rayleigh distribution function, and the second distribution function is another Rayleigh distribution function. 2. The image processing method as claimed in claim 1, wherein the first distribution function is implemented as a first step function, and the second distribution function is implemented as a second step function. 3. The image processing method as claimed in claim 1, wherein the arrangement of the color filter array is a Bayer pattern. 4. An image processing device for suppressing noise in a captured or recorded image in a color filter array field, the image processing device comprising: a storage unit for respectively receiving a green plane, a red plane and a blue plane of the captured or recorded image; and An image processing unit, connected to the storage unit, for the green plane, the red plane and the blue plane, sequentially cover each color plane with a working window, so that each pixel on each color plane is sequentially located in the working the center of the window, and is designated as a central pixel C _c , And the image processing unit classifies other pixels other than the central pixel into N edge pixels Ce _dge,i and M non-edge pixels C _non,j in the working window, where N and M are integers greater than 1, i is an integer index from 1 to N, j is an integer index from 1 to M, And the image processing unit calculates a filtered central pixel C _fc according to the following formula, and replaces the central pixel C _c with the filtered central pixel: ${\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; fc</span>}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; edge</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; WEI</span>}\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; (</span><span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; c</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; non</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; |</span><span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; ;</span>$ where WEI1 is a first distribution function and WEI2 is a second distribution function, wherein the first distribution function is a Rayleigh distribution function and the second distribution function is another Rayleigh distribution function. 5. The image processing device as claimed in claim 4, wherein the first distribution function is implemented as a first step function, and the second distribution function is implemented as a second step function. 6. The image processing device as claimed in claim 4, wherein the color filter array is arranged in a Bayer pattern.

Images