CN103491282A - Image denoising method and device - Google Patents

Abstract

The embodiment of the invention relates to an image denoising method and device. The image denoising method comprises the steps: (1), obtaining a first pixel mean of a first pixel block in a current image, and obtaining a first pixel variance of the first pixel block by the utilization of the first pixel mean, (2), determining a pixel difference value block of the first pixel block in the current image and a second pixel block in a reference image, (3), obtaining a second pixel mean of the pixel difference value block according to the pixel difference value block, and obtaining a second pixel variance of the pixel difference value block by the utilization of the second pixel mean, (4), determining the block distance between the first pixel block and the second pixel block, (5), selecting the largest pixel variance in the first pixel variance and the second pixel variance as a noise value of a first pixel in the first pixel block when the block distance is smaller than a preset first distance threshold value and the first pixel variance is greater than a preset variance threshold value, and (6), determining a first denoising weight when time domain denoising is carried out on the first pixel by the utilization of the block distance and the noise value.

Description

Denoising method of images and device

Technical field

The present invention relates to communication technique field, relate in particular to a kind of Denoising method of images and device.

Background technology

At present, in the video image gathered in terminal, mixing a large amount of noises, this noise is except dark current noise, and most noise is random noise.

The interference produced in order to reduce as much as possible random noise, usually adopt motion detection and motion compensation technique, the image of arbitrary frame in video image is carried out to the differentiation of dynamic area and static region, for zones of different, adopt different noise adaptive filter methods, thereby improve de-noising effect.For example, the violent zone of moving, the weight of airspace filter is large; On the contrary, static zone, the weight of time-domain filtering is large.So both random noise can be effectively removed, the appearance of motion blur can well be avoided again.

In the de-noising scheme of prior art, the current frame image equal proportion is divided into to a plurality of block of pixels, calculate the first block of pixels in current image frame and and the reference image frame after frame accumulation de-noising in piece distance (position of the position of the first block of pixels and the second block of pixels is correspondence position) between the second block of pixels, according to the piece distance, determine time domain de-noising weight.In the prior art scheme, the piece distance is compared with default distance threshold, the if block distance surpasses distance threshold, and this piece is apart from corresponding block of pixels, and its time domain de-noising weight reduces, and de-noising weight in spatial domain increases; The if block distance is no more than distance threshold, and this piece is apart from corresponding block of pixels, and its time domain de-noising weight increases, and de-noising weight in spatial domain reduces.

But, also expose following defect in the de-noising scheme of prior art: 1) owing to there being random noise, calculate two pieces between block of pixels apart from the time need to consider its impact.But in the prior art, the value of random noise has probabilistic characteristics, therefore, the piece of calculating is apart from there being certain error, and when contrast of video images is low, the weight allocation that forms time domain, spatial domain is inaccurate, causes de-noising effect undesirable; 2), in dynamic area, due to the piece calculated, apart from the error existed, make the time domain weight is arranged to the excessive result that causes image streaking; Equally, at stagnant zone, due to the piece calculated, apart from the error existed, make the time domain weight is arranged to the too small noise that causes to eliminate sordid result.

Summary of the invention

The embodiment of the present invention provides a kind of Denoising method of images and device, realized adjusting according to the pixel information self-adapting of the block of pixels of processing in present image the de-noising weight in time domain, spatial domain, effectively removing under the prerequisite of stagnant zone random noise, avoiding the image streaking of dynamic area.

In first aspect, the embodiment of the present invention provides a kind of Denoising method of images, described method comprises: the first pixel average of obtaining the first block of pixels in present image, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, described the first block of pixels comprises a plurality of pixels;

Determine in described present image the pixel value difference piece of the second block of pixels in the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

According to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Determine the piece distance between described the first block of pixels and described the second block of pixels;

When described distance is less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

In the first in possible implementation, after the described piece distance of determining between described the first block of pixels and described the second block of pixels and described described distance and the described noise level of utilizing, determine that the first de-noising weight that described the first pixel is carried out to the time domain de-noising also comprises before:

When described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

At the second in possible implementation, described described distance and the described noise level of utilizing, determine that the first de-noising weight that described the first pixel is carried out to the time domain de-noising specifically comprises:

Judge whether described distance is greater than default second distance threshold value;

If described distance is greater than described second distance threshold value, described the first de-noising weight is 0;

If described distance is less than described second distance threshold value, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

If described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, and described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

If described distance is less than described second distance threshold value, and described distance be less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

At the third in possible implementation, described described the second pixel average and the described pixel noise value utilized, determine described the first pixel is carried out also comprising after the first de-noising weight of time domain de-noising:

According to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

In conjunction with the third possible implementation of first aspect or first aspect, in the 4th kind of possible implementation, described the first pixel average of obtaining the first block of pixels in present image also comprises before:

Obtain first filter value of described the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Described according to described the first de-noising weight, determine described the first pixel is carried out also comprising after the second de-noising weight of spatial domain de-noising:

According to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

In conjunction with the 4th kind of possible implementation of first aspect or first aspect, in the 5th kind of possible implementation, described first filter value of described the first block of pixels after time-domain filtering that obtain also comprises before:

Described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels;

Describedly obtain the first pixel average of the first block of pixels in present image and specifically comprise: the first pixel average of obtaining the first block of pixels of preliminary treatment described in described present image;

Describedly determine in described present image in the first block of pixels and reference picture that the pixel value difference piece of the second block of pixels specifically comprises:

Determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

In second aspect, the embodiment of the present invention provides a kind of Image Denoising by Use device, and described device comprises:

Acquiring unit, for obtaining the first pixel average of present image the first block of pixels, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, and described the first block of pixels comprises a plurality of pixels;

Determining unit, for determining the pixel value difference piece of the second block of pixels in described present image the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

Described acquiring unit also for, according to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Described determining unit also for, determine the piece distance between described the first block of pixels and described the second block of pixels;

Choose unit, for when described distance, being less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Described determining unit also for, utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

In the first in possible implementation, described choose unit also for, when described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

At the second, in possible implementation, described determining unit also comprises:

Whether judgment sub-unit, be greater than default second distance threshold value for judging described distance;

First processes subelement, if be greater than described second distance threshold value for described distance, described the first de-noising weight is 0;

Second processes subelement, if be less than described second distance threshold value for described distance, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

Described second process subelement also for, if described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

Described second process subelement also for, if described distance is less than described second distance threshold value, and described distance is less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

At the third in possible implementation, described determining unit also for, according to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

In conjunction with the third possible implementation of second aspect or second aspect, in the 4th kind of possible implementation, described acquiring unit also for, obtain first filter value of tell the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Described determining unit also for, according to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

In conjunction with the 4th kind of possible implementation of second aspect or second aspect, in the 5th kind of possible implementation, described device also comprises:

Pretreatment unit, for described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels;

Described acquiring unit specifically for, obtain the first pixel average of the first block of pixels of preliminary treatment described in described present image;

Described determining unit specifically for, determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

Therefore, the Denoising method of images and the device that by the application embodiment of the present invention, provide, at first terminal utilizes the first pixel average of the first block of pixels, obtains the first pixel variance yields, utilize the second pixel average of pixel value difference piece, obtain the second pixel variance yields, then, piece between the first block of pixels and the second block of pixels distance is compared with the first distance threshold, when the piece distance is less than the first default distance threshold, and when the first pixel variance yields is greater than default variance threshold values, choose in the first pixel variance yields and the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in the first block of pixels, finally, utilize piece distance and noise level, determine the first de-noising weight that the first pixel is carried out to the time domain de-noising, determine the second de-noising weight of carrying out the spatial domain de-noising according to the first de-noising weight, avoided in prior art, because the value of random noise has uncertainty, cause computing block apart from there being certain error, form time domain, the weight allocation in spatial domain is inaccurate, the image streaking that causes the undesirable and dynamic area of de-noising effect, the stagnant zone noise is eliminated sordid problem.And then realized adjusting according to the pixel information self-adapting of the block of pixels of processing in present image the de-noising weight in time domain, spatial domain, and effectively removing under the prerequisite of stagnant zone random noise, avoid the image streaking of dynamic area.

The accompanying drawing explanation

The Denoising method of images flow chart that Fig. 1 provides for the embodiment of the present invention one;

The block of pixels schematic diagram that Fig. 2 provides for the embodiment of the present invention;

Piece distance and the first de-noising weight relationship figure that Fig. 3 provides for the embodiment of the present invention;

The Image Denoising by Use structure drawing of device that Fig. 4 provides for the embodiment of the present invention;

The Image Denoising by Use device hardware structure diagram that Fig. 5 provides for the embodiment of the present invention.

Embodiment

For the purpose, technical scheme and the advantage that make the embodiment of the present invention clearer, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is the present invention's part embodiment, rather than whole embodiment.Embodiment based in the present invention, those of ordinary skills, not making under the creative work prerequisite the every other embodiment obtained, belong to the scope of protection of the invention.

Understanding for ease of real-time to the present invention, be further explained explanation below in conjunction with accompanying drawing with specific embodiment, and embodiment does not form the restriction to the embodiment of the present invention.

Embodiment mono-

The Denoising method of images that the Fig. 1 of below take provides as the example detailed description embodiment of the present invention one, the Denoising method of images flow chart that Fig. 1 provides for the embodiment of the present invention one, in embodiments of the present invention, the subject of implementation of carrying out following step can be for having the terminal that gathers image function, for example: digital camera, smart mobile phone, Ipad, also can realize by the software module be carried in terminal.As shown in Figure 1, this embodiment specifically comprises the following steps:

Step 110, obtain the first pixel average of the first block of pixels in present image, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, described the first block of pixels comprises a plurality of pixels.

Particularly, terminal is extracted the first two field picture as present image from the video sequence gathered, and present image is carried out to the division of equal proportion, obtains a plurality of block of pixels, and each block of pixels comprises a plurality of pixels, as shown in Figure 2.Be understandable that, when a two field picture is carried out to Image Denoising by Use, the embodiment of the present invention is by each pixel is carried out to denoising Processing, and then the purpose of denoising Processing is carried out in realization to image.

For example, the equal proportion block of pixels that each block of pixels is 5*5, certainly, also can be divided into the block of pixels of other equal proportions, as 3*3,4*4 etc.In embodiments of the present invention, take and the first pixel in the first block of pixels is carried out to denoising Processing describe as example.

Terminal is obtained the first pixel average ave1 of the first block of pixels in present image, and described ave1 is specially the mean value of the pixel value of each pixel in the first block of pixels, but through type (1) is determined:

$\mathrm{<figure-callout\; id="1"\; label="ave"\; filenames="HDA0000385733580000022.png"\; state="\{\{state\}\}">ave</figure-callout>}1=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}\mathrm{pixel}(x+i,y+j)$ Formula (1)

Wherein, pixel (x+i, y+j) is the pixel value that in the first block of pixels, position (x, y) locates.

After getting the first pixel average ave1 of the first block of pixels, terminal utilizes the first pixel average ave1 to obtain the first pixel variance yields current_sigma2 of the first block of pixels, but described current_sigma2 through type (2) is determined:

$\mathrm{current}\_\mathrm{sigma}2=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}{[\mathrm{pixel}(x+i,y+j)-\mathrm{ave}1]}^2$ Formula (2)

Wherein, pixel (x+i, y+j) is the pixel value that in the first block of pixels, position (x, y) locates.

Step 120, determine in described present image that the pixel value difference piece of the second block of pixels in the first block of pixels and reference picture, described the second block of pixels are the block of pixels with described the first block of pixels correspondence position.

Particularly, terminal is determined the pixel value difference piece of the second block of pixels in the first block of pixels and reference picture, and described the second block of pixels is the block of pixels with the first block of pixels correspondence position.

Further, when terminal is carried out denoising Processing to present image again, carry out the reference picture of denoising Processing as present image from the image obtained memory after the accumulation de-noising, and this reference picture is also carried out to the equal proportion identical with present image and divide, obtain a plurality of block of pixels, each block of pixels comprises a plurality of pixels, and according to the first block of pixels the position in present image, determine the second block of pixels of correspondence position from reference picture.

Terminal is subtracted each other the pixel value of corresponding pixel points in the pixel value of each pixel in the first block of pixels and the second block of pixels, obtains the pixel value difference of each pixel, and this pixel value difference is formed to the pixel value difference piece.

For example, the first block of pixels is 5*5, and the second block of pixels is also 5*5, and the second block of pixels is the block of pixels with the first block of pixels correspondence position.Each block of pixels comprises 25 pixels, and the pixel value correspondence of each pixel is subtracted each other, and obtains the pixel value difference of 25 pixels, and 25 pixel value differences are formed to the pixel value difference piece.

Step 130, according to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece.

Particularly, the definite pixel value difference piece according to step 120, terminal is obtained the second pixel average ave2 of this pixel value difference piece, and described ave2 is specially the mean value of the pixel value of each pixel in pixel value difference, but through type (3) is determined:

$\mathrm{ave}2=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}\mathrm{residue}(x+i,y+j)$ Formula (3)

Wherein, residue (x+i, y+j) is the pixel value that in the pixel value difference piece, position (x, y) locates.

After getting the second pixel average ave2 of pixel value difference piece, terminal utilizes the second pixel average ave2 to obtain the second pixel variance yields residue_sigma2 of pixel value difference piece, but described residue_sigma2 through type (4) is determined:

$\mathrm{residue}\_\mathrm{sigma}2=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}{[\mathrm{residue}(x+i,y+j)-\mathrm{ave}2]}^2$ Formula (4)

Wherein, residue (x+i, y+j) is the pixel value that in the pixel value difference piece, position (x, y) locates.Be understandable that, the formula that solves described the second pixel variance yields residue_sigma2 is of a great variety, in the embodiment of the present invention, the form to give an example only, through type (4) obtains, in actual applications, also can obtain by the formula of other kinds, be not restricted to this.

Step 140, determine the piece distance between described the first block of pixels and described the second block of pixels.

Particularly, terminal determines that piece between the first block of pixels and the second block of pixels is apart from tMSE.

Be understandable that, solve the described formula apart from tMSE of a great variety, for example, by getting the form of absolute difference after subtracting each other between each pixel in the first block of pixels and the second block of pixels, or, form again square after getting absolute difference, perhaps, to get the form of squared difference after subtracting each other between each pixel in the first block of pixels and the second block of pixels, or, will get the form of difference cube after subtracting each other between each pixel in the first block of pixels and the second block of pixels.The above-mentioned form to give an example, illustrated block, apart from definite method of tMSE, in actual applications, also can obtain by the formula of other kinds, is not restricted to this.

In the embodiment of the present invention, take the form of getting squared difference after subtracting each other between each pixel in the first block of pixels and the second block of pixels is example, and the method for piece apart from tMSE of determining is described.

Described tMSE is specially the distance between the first block of pixels and the second block of pixels, but through type (5) is determined:

$\mathrm{tMSE}=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}{[\mathrm{Pixel}\_\mathrm{curr}(x+i,y+j)-\mathrm{Pixel}\_\mathrm{ref}(x+i,y+j)]}^2$ Formula (5)

Wherein, Pixel_curr (x+i, y+j) is the pixel value that in the first block of pixels, position (x, y) locates; Pixel_ref (x+i, y+j) is the pixel value that in the second block of pixels, position (x, y) locates.

Step 150, be less than the first default distance threshold when described distance, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels.

Particularly, after definite piece is apart from tMSE, when piece is less than the first default distance threshold apart from tMSE, and, when the first pixel variance yields current_sigma2 is greater than default variance threshold values, terminal is chosen in the first pixel variance yields current_sigma2 and described the second pixel variance yields residue_sigma2 the maximum pixel variance yields as the noise level noise_sigma2 of the first pixel in the first block of pixels.

Further, but the noise level noise_sigma2 through type (6) of the first pixel determine:

Noise_sigma2=MAX (residue_sigma2, current_sigma2) formula (6)

For example, if the first pixel variance yields current_sigma2 is greater than the second pixel variance yields residue_sigma2, terminal is chosen the noise level noise_sigma2 of the first pixel variance yields current_sigma2 as the first pixel in the first block of pixels; Otherwise, choose the noise level noise_sigma2 of the second pixel variance yields residue_sigma2 as the first pixel in the first block of pixels.

Step 160, utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

Particularly, after the noise level noise_sigma2 that determines the first pixel, terminal utilizes piece apart from tMSE and noise level noise_sigma2, determines the first de-noising weight AlphaIIR that the first pixel is carried out to the time domain de-noising.

Further, described described distance and the described noise level of utilizing, determine that the first de-noising weight that described the first pixel is carried out to the time domain de-noising specifically comprises:

Whether the terminal judges piece is greater than default second distance threshold value thresh apart from tMSE; If block is greater than second distance threshold value thresh apart from tMSE, and the first de-noising weight is arranged to AlphaIIR is 0 to terminal, that is to say and does not need the first pixel is carried out to the time domain de-noising;

If block is less than second distance threshold value thresh apart from tMSE, and piece is greater than the 3rd default distance threshold Pa apart from tMSE, and terminal the first de-noising weight AlphaIIR is set to α _b, wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

If block is less than second distance threshold value thresh apart from tMSE, and piece is less than the 4th default distance threshold Pb apart from tMSE, and terminal the first de-noising weight AlphaIIR is set to α _a, wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

If block is less than second distance threshold value thresh apart from tMSE, and piece is less than the 3rd default distance threshold Pa apart from tMSE, be greater than the 4th default distance threshold Pb, the first de-noising weight AlphaIIR calls distance function by terminal and determines, wherein, the first de-noising weight AlphaIIR is directly proportional apart from tMSE to the piece in distance function, and with noise level, noise_sigma2 is inversely proportional to.

As example and non-limiting, but described distance function through type (7) is determined:

AlphaIIR=distance[Index] formula (7)

Wherein, formula (7) described [Index] but concrete through type (8) determine:

Index=tMSE/noise_sigma2 formula (8)

Wherein, described tMSE is described distance, and described noise_sigma2 is described noise level.

In embodiments of the present invention, described second distance threshold value thresh is the arbitrary value that is greater than the 3rd distance threshold Pa.

Further, as shown in Figure 3, piece distance and the first de-noising weight relationship figure that Fig. 3 provides for the embodiment of the present invention.In Fig. 3, the transverse axis X-axis be piece apart from tMSE, longitudinal axis Y-axis is the first de-noising weight AlphaIIR.Second distance threshold value, the 3rd distance threshold and the 4th distance threshold are positioned on X-axis, and the setting of distance threshold can set up on their own according to actual conditions; The interval point Max α of the first de-noising weight _a, Max α _bbe positioned on Y-axis.

From Fig. 3, can be clear that, if block is greater than second distance threshold value thresh apart from tMSE, and terminal the first de-noising weight AlphaIIR is set to 0, that is to say and does not need the first pixel is carried out to the time domain de-noising;

If block is less than second distance threshold value thresh apart from tMSE, and piece is greater than the 3rd default distance threshold Pa apart from tMSE, and terminal the first de-noising weight AlphaIIR is set to α _b, wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

If block is less than second distance threshold value thresh apart from tMSE, and piece is less than the 4th default distance threshold Pb apart from tMSE, and terminal the first de-noising weight AlphaIIR is set to α _a, wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

If block is less than second distance threshold value thresh apart from tMSE, and piece is less than the 3rd default distance threshold Pb apart from tMSE, be greater than the 4th default distance threshold Pb, the first de-noising weight AlphaIIR calls distance function by terminal and determines, that is to say in Fig. 3 definite with the distance between some B by some A.

Alternatively, after embodiment of the present invention step 140 and before step 160, also comprise the steps, specific as follows:

When described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

Particularly, after definite piece is apart from tMSE, when piece is greater than the first default distance threshold apart from tMSE, and, when the first pixel variance yields current_sigma2 is less than default variance threshold values, terminating machine is chosen in the first pixel variance yields current_sigma2 and described the second pixel variance yields residue_sigma2 the minimum pixel variance yields as the noise level noise_sigma2 of the first pixel in the first block of pixels.

Further, but the noise level noise_sigma2 through type (9) of the first pixel determine:

Noise_sigma2=MIN (residue_sigma2, current_sigma2) formula (9)

For example, if the first pixel variance yields current_sigma2 is less than the second pixel variance yields residue_sigma2, terminal is chosen the noise level noise_sigma2 of the first pixel variance yields current_sigma2 as the first pixel in the first block of pixels; Otherwise, choose the noise level noise_sigma2 of the second pixel variance yields residue_sigma2 as the first pixel in the first block of pixels.

Alternatively, before embodiment of the present invention step 110, also comprise that terminal obtains the first filter value Pout_time (x of the first block of pixels after time-domain filtering, y) and the second filter value Pout_spa (x of the first block of pixels after airspace filter, y) step, above-mentioned the first filter value Pout_time (x obtained, y) and the second filter value Pout_spa (x, y) can be used for determining in subsequent step and determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising.Concrete steps are as follows:

Obtain first filter value of tell the first block of pixels after time-domain filtering;

Particularly, terminal to before step 160, is at first carried out time-domain filtering to the first block of pixels in execution step 110, obtain the first filter value Pout_time (x of the first block of pixels after time-domain filtering, y), the described time-domain filtering to block of pixels is prior art, at this, no longer repeats.

Further, the first filter value Pout_time (x, y) but through type (10) determine:

Pout_time (x, y)=strenth*pixel_ref (x, y)+(1-strenth) * pixel_curr (x, y) formula (10)

Wherein, but the time-domain filtering intensity level that strenth is manual adjustment, and described pixel_curr (x, y) works as the pixel value of the first pixel of pre-treatment in the first block of pixels; Described pixel_ref (x, y) be in the second block of pixels with the pixel value of the second pixel of the first block of pixels point correspondence position.Be understandable that, the formula that solves described the first filter value Pout_time (x, y) is of a great variety, in the embodiment of the present invention, the form to give an example only, through type (10) obtains, in actual applications, also can obtain by the formula of other kinds, be not restricted to this.

Obtain second filter value of described the first block of pixels after airspace filter;

Particularly, terminal is getting the first filter value Pout_time (x, y) afterwards, also the first block of pixels is carried out to airspace filter, obtain the second filter value Pout_spa (x of the first block of pixels after airspace filter, y), the described airspace filter to block of pixels is prior art, at this, no longer repeats.

Further, the second filter value Pout_spa (x, y) but through type (11) determine:

$\mathrm{Pout}\_\mathrm{spa}(x,y)=\frac{1}{M\&times;N}\underset{i=0}{\overset{M}{\mathrm{\&Sigma;}}}\underset{j=0}{\overset{N}{\mathrm{\&Sigma;}}}\mathrm{pixel}\_\mathrm{curr}(x+i,y+j)$ Formula (11)

Wherein, described pixel_curr (x, y) works as the pixel value of the first pixel of pre-treatment in the first block of pixels.Be understandable that, solve the formula method except above-mentioned formula (11) of a great variety of described the second filter value Pout_spa (x, y), after also can carrying out airspace filter to the first pixel by two-sided filter (bilateral filter), obtain the second filter value Pout_spa (x, y), in the embodiment of the present invention, the form to give an example only, through type (11) obtains, in actual applications, also can obtain by the formula of other kinds, be not restricted to this.

Alternatively, tell the first filter value step of the first block of pixels after time-domain filtering obtaining before, also comprise that terminal carries out to the first block of pixels and the second block of pixels the step that filtering is carried out in filtering, by this filter step, can carry out filtering to the random noise of block of pixels to step 160 in execution step 110, make terminal during to step 160, can more effectively determine the first weight that pixel is carried out to the time domain de-noising in execution step 110.Concrete steps are as follows:

Described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels.

Particularly, terminal is carried out respectively the filtering processing to the first block of pixels and the second block of pixels, described filtering processing that is to say carries out preliminary treatment to the first block of pixels and the second block of pixels, by the first block of pixels and the second block of pixels are carried out to filtering, can make terminal during to step 160, can more effectively determine the first weight that pixel is carried out to the time domain de-noising in execution step 110.

Wherein, describedly obtain the first pixel average of the first block of pixels in present image and specifically comprise: the first pixel average of obtaining the first block of pixels of preliminary treatment described in described present image;

Describedly determine in described present image in the first block of pixels and reference picture that the pixel value difference piece of the second block of pixels specifically comprises:

Determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

Be understandable that, after described the first block of pixels and described the second block of pixels are carried out to the filtering processing, the first pixel average of the first pixel average of obtaining and first block of pixels of not carrying out the filtering processing equates; In like manner, definite pixel value difference piece is also identical.

Alternatively, after embodiment of the present invention step 160, also comprise that terminal is according to the definite step of the first pixel being carried out to the second de-noising weight of spatial domain de-noising of the first de-noising weight.In embodiments of the present invention, after terminal is determined the first de-noising weight, utilize 1 to deduct the first de-noising weight, and then determine the second de-noising weight.By this step, can make terminal carry out the time-space domain de-noising to the first pixel.

Alternatively, execute above-mentioned according to described the first de-noising weight, after determining described the first pixel being carried out to the second de-noising weight of spatial domain de-noising, terminal is also according to the first filter value Pout_time (x obtained, y), the second filter value Pout_spa (x obtained, y), the first de-noising weight AlphaIIR and the second de-noising weight, determine the step of the first pixel being carried out to the output valve P_out of time-space domain de-noising, by this step, terminal is carried out the denoising Processing of time-space domain to the first pixel according to definite output valve P_out.

In embodiments of the present invention, but output valve P_out through type (12) determine:

P_out=AlphaIIR*Pout_time+ (1-AlphaIIR) * Pout_spa (formula 12)

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

Be understandable that, in Denoising method of images as described above, only take and the first pixel in the first block of pixels is carried out to denoising Processing describe as example, in actual applications, need carry out denoising Processing to each pixel in each block of pixels, and then reach the purpose of image being carried out to denoising Processing, because the process of each pixel being carried out to denoising Processing is identical, at this, no longer repeat.

Therefore, the Denoising method of images provided by the application embodiment of the present invention, at first terminal utilizes the first pixel average of the first block of pixels, obtains the first pixel variance yields, utilize the second pixel average of pixel value difference piece, obtain the second pixel variance yields, then, piece between the first block of pixels and the second block of pixels distance is compared with the first distance threshold, when the piece distance is less than the first default distance threshold, and when the first pixel variance yields is greater than default variance threshold values, choose in the first pixel variance yields and the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in the first block of pixels, finally, utilize piece distance and noise level, determine the first de-noising weight that the first pixel is carried out to the time domain de-noising, determine the second de-noising weight of carrying out the spatial domain de-noising according to the first de-noising weight, avoided in prior art, because the value of random noise has uncertainty, cause computing block apart from there being certain error, form time domain, the weight allocation in spatial domain is inaccurate, the image streaking that causes the undesirable and dynamic area of de-noising effect, the stagnant zone noise is eliminated sordid problem.And then realized adjusting according to the pixel information self-adapting of the block of pixels of processing in present image the de-noising weight in time domain, spatial domain, and effectively removing under the prerequisite of stagnant zone random noise, avoid the image streaking of dynamic area.Simultaneously, the Denoising method of images that the embodiment of the present invention provides, also can realize adjusting the de-noising weight in the situation that can't obtain noise level, do not need to demarcate noise grade, also reduced workload.

Embodiment bis-

Correspondingly, the embodiment of the present invention also provides a kind of Image Denoising by Use device, and in order to realize the Denoising method of images in previous embodiment one, as shown in Figure 4, described Image Denoising by Use device comprises: acquiring unit 410, determining unit 420 and choose unit 430.

In described device, acquiring unit 410, for obtaining the first pixel average of present image the first block of pixels, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, and described the first block of pixels comprises a plurality of pixels;

Determining unit 420, for determining the pixel value difference piece of the second block of pixels in described present image the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

Described acquiring unit 410 also for, according to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Described determining unit 420 also for, determine the piece distance between described the first block of pixels and described the second block of pixels;

Choose unit 430, for when described distance, being less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Described determining unit 420 also for, utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

Described choose unit also for, when described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

Described determining unit 420 also comprises: whether judgment sub-unit 421 is greater than default second distance threshold value for judging described distance;

First processes subelement 422, if be greater than described second distance threshold value for described distance, described the first de-noising weight is 0;

Second processes subelement 423, if be less than described second distance threshold value for described distance, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

Described second process subelement 423 also for, if described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

Described second process subelement 423 also for, if described distance is less than described second distance threshold value, and described distance is less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

Described determining unit 420 also for, according to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

Described acquiring unit 410 also for, obtain first filter value of tell the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Described determining unit 420 also for, according to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

Described device also comprises: pretreatment unit 440, for described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtains preliminary treatment the first block of pixels and preliminary treatment the second block of pixels.

Described acquiring unit concrete 410 for, obtain the first pixel average of the first block of pixels of preliminary treatment described in described present image;

Described determining unit 420 specifically for, determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

Therefore, the Image Denoising by Use device provided by the application embodiment of the present invention, at first terminal utilizes the first pixel average of the first block of pixels, obtains the first pixel variance yields, utilize the second pixel average of pixel value difference piece, obtain the second pixel variance yields, then, piece between the first block of pixels and the second block of pixels distance is compared with the first distance threshold, when the piece distance is less than the first default distance threshold, and when the first pixel variance yields is greater than default variance threshold values, choose in the first pixel variance yields and the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in the first block of pixels, finally, utilize piece distance and noise level, determine the first de-noising weight that the first pixel is carried out to the time domain de-noising, determine the second de-noising weight of carrying out the spatial domain de-noising according to the first de-noising weight, avoided in prior art, because the value of random noise has uncertainty, cause computing block apart from there being certain error, form time domain, the weight allocation in spatial domain is inaccurate, the image streaking that causes the undesirable and dynamic area of de-noising effect, the stagnant zone noise is eliminated sordid problem.And then realized adjusting according to the pixel information self-adapting of the block of pixels of processing in present image the de-noising weight in time domain, spatial domain, and effectively removing under the prerequisite of stagnant zone random noise, avoid the image streaking of dynamic area.

Embodiment tri-

In addition, the Denoising method of images that the embodiment of the present invention provides also can realize by following form, in order to realize the Denoising method of images in previous embodiment one of the present invention, as shown in Figure 5, described Image Denoising by Use device comprises: network interface 510, processor 520 and memory 530.System bus 540 is for interconnection network interface 510, processor 520 and memory 530.

Network interface 510 is for externally carrying out interactive communication.

Memory 530 can be permanent memory, and for example hard disk drive and flash memory, have software module and device driver in memory 530.Software module can be carried out the various functional modules of said method of the present invention; Device driver can be network and interface drive program.

When starting, these software modules are loaded in memory 530, then by processor 520, are accessed and carry out as given an order:

Obtain the first pixel average of the first block of pixels in present image, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, described the first block of pixels comprises a plurality of pixels;

Determine in described present image the pixel value difference piece of the second block of pixels in the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

According to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Determine the piece distance between described the first block of pixels and described the second block of pixels;

When described distance is less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

Further, after the component software of described processor 520 references to storage 530, carry out the instruction of following process:

When described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

Further, after the software module of described processor 520 references to storage 530, execution utilizes described distance and described noise level, determines that the instruction that described the first pixel is carried out to the first de-noising weight process of time domain de-noising is specially the instruction of carrying out following process:

Judge whether described distance is greater than default second distance threshold value;

If described distance is greater than described second distance threshold value, described the first de-noising weight is 0;

If described distance is less than described second distance threshold value, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

If described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, and described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

If described distance is less than described second distance threshold value, and described distance be less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

Further, after the component software of described processor 520 references to storage 530, carry out the instruction of following process:

According to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

Further, after the component software of described processor 520 references to storage 530, carry out the instruction of following process:

Obtain first filter value of tell the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Further, after the component software of described processor 520 references to storage 530, carry out the instruction of following process:

According to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

Further, after the component software of described processor 520 references to storage 530, carry out the instruction of following process:

Described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels;

After the software module of described processor 520 references to storage 530, the instruction of carrying out the first pixel average process of the first block of pixels in present image of obtaining is specially the instruction of carrying out following process:

Obtain the first pixel average of the first block of pixels of preliminary treatment described in described present image;

After the software module of described processor 520 references to storage 530, carry out to determine in described present image that the instruction of the pixel value difference piece process of the second block of pixels in the first block of pixels and reference picture is specially the instruction of the following process of execution:

Determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

Therefore, the Image Denoising by Use device provided by the application embodiment of the present invention, at first terminal utilizes the first pixel average of the first block of pixels, obtains the first pixel variance yields, utilize the second pixel average of pixel value difference piece, obtain the second pixel variance yields, then, piece between the first block of pixels and the second block of pixels distance is compared with the first distance threshold, when the piece distance is less than the first default distance threshold, and when the first pixel variance yields is greater than default variance threshold values, choose in the first pixel variance yields and the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in the first block of pixels, finally, utilize piece distance and noise level, determine the first de-noising weight that the first pixel is carried out to the time domain de-noising, determine the second de-noising weight of carrying out the spatial domain de-noising according to the first de-noising weight, avoided in prior art, because the value of random noise has uncertainty, cause computing block apart from there being certain error, form time domain, the weight allocation in spatial domain is inaccurate, the image streaking that causes the undesirable and dynamic area of de-noising effect, the stagnant zone noise is eliminated sordid problem.And then realized adjusting according to the pixel information self-adapting of the block of pixels of processing in present image the de-noising weight in time domain, spatial domain, and effectively removing under the prerequisite of stagnant zone random noise, avoid the image streaking of dynamic area.

The professional should further recognize, unit and the algorithm steps of each example of describing in conjunction with embodiment disclosed herein, can realize with electronic hardware, computer software or the combination of the two, for the interchangeability of hardware and software clearly is described, composition and the step of each example described according to function in the above description in general manner.These functions are carried out with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.The professional and technical personnel can specifically should be used for realizing described function with distinct methods to each, but this realization should not thought and exceeds scope of the present invention.

The software module that the method for describing in conjunction with embodiment disclosed herein or the step of algorithm can use hardware, processor to carry out, or the combination of the two is implemented.Software module can be placed in the storage medium of any other form known in random asccess memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field.

Above-described embodiment; purpose of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the foregoing is only the specific embodiment of the present invention; the protection range be not intended to limit the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (12)

1. a Denoising method of images, is characterized in that, described method comprises:

Obtain the first pixel average of the first block of pixels in present image, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, described the first block of pixels comprises a plurality of pixels;

Determine in described present image the pixel value difference piece of the second block of pixels in the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

According to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Determine the piece distance between described the first block of pixels and described the second block of pixels;

When described distance is less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

2. Denoising method of images according to claim 1, it is characterized in that, after the described piece distance of determining between described the first block of pixels and described the second block of pixels and described described distance and the described noise level of utilizing, determine that the first de-noising weight that described the first pixel is carried out to the time domain de-noising also comprises before:

When described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

3. Denoising method of images according to claim 1, is characterized in that, described described distance and the described noise level of utilizing determines that the first de-noising weight that described the first pixel is carried out to the time domain de-noising specifically comprises:

Judge whether described distance is greater than default second distance threshold value;

If described distance is greater than described second distance threshold value, described the first de-noising weight is 0;

If described distance is less than described second distance threshold value, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

If described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, and described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

If described distance is less than described second distance threshold value, and described distance be less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

4. Denoising method of images according to claim 1, is characterized in that, described described the second pixel average and the described pixel noise value utilized determined described the first pixel is carried out also comprising after the first de-noising weight of time domain de-noising:

According to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

5. Denoising method of images according to claim 4, is characterized in that, described the first pixel average of obtaining the first block of pixels in present image also comprises before:

Obtain first filter value of described the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Described according to described the first de-noising weight, determine described the first pixel is carried out also comprising after the second de-noising weight of spatial domain de-noising:

According to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

6. picture noise method according to claim 5, is characterized in that, described first filter value of described the first block of pixels after time-domain filtering that obtain also comprises before:

Described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels;

Describedly obtain the first pixel average of the first block of pixels in present image and specifically comprise: the first pixel average of obtaining the first block of pixels of preliminary treatment described in described present image;

Describedly determine in described present image in the first block of pixels and reference picture that the pixel value difference piece of the second block of pixels specifically comprises:

Determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

7. an Image Denoising by Use device, is characterized in that, described device comprises:

Acquiring unit, for obtaining the first pixel average of present image the first block of pixels, and utilize described the first pixel average to obtain the first pixel variance yields of described the first block of pixels, and described the first block of pixels comprises a plurality of pixels;

Determining unit, for determining the pixel value difference piece of the second block of pixels in described present image the first block of pixels and reference picture, described the second block of pixels is the block of pixels with described the first block of pixels correspondence position;

Described acquiring unit also for, according to described pixel value difference piece, obtain the second pixel average of described pixel value difference piece, and utilize described the second pixel average to obtain the second pixel variance yields of described pixel value difference piece;

Described determining unit also for, determine the piece distance between described the first block of pixels and described the second block of pixels;

Choose unit, for when described distance, being less than the first default distance threshold, and, when described the first pixel variance yields is greater than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the maximum pixel variance yields as the noise level of the first pixel in described the first block of pixels;

Described determining unit also for, utilize described distance and described noise level, determine the first de-noising weight that described the first pixel is carried out to the time domain de-noising.

8. Image Denoising by Use device according to claim 7, it is characterized in that, described choose unit also for, when described distance is greater than the first default distance threshold, and, when described the first pixel variance yields is less than default variance threshold values, choose in described the first pixel variance yields and described the second pixel variance yields the minimum pixel variance yields as the noise level of described the first pixel.

9. Image Denoising by Use device according to claim 7, is characterized in that, described determining unit also comprises:

Whether judgment sub-unit, be greater than default second distance threshold value for judging described distance;

First processes subelement, if be greater than described second distance threshold value for described distance, described the first de-noising weight is 0;

Second processes subelement, if be less than described second distance threshold value for described distance, and described distance be greater than the 3rd default distance threshold, and described the first de-noising weight is set to α _b;

Wherein, described α _bbe specially from interval (0, Max α _b) in arbitrary value of choosing;

Described second process subelement also for, if described distance is less than described second distance threshold value, and described distance be less than the 4th default distance threshold, described the first de-noising weight is set to α _a;

Wherein, described α _abe specially from interval (Max α _a, arbitrary value of choosing in ∞);

Described second process subelement also for, if described distance is less than described second distance threshold value, and described distance is less than default described the 3rd distance threshold, is greater than default described the 4th distance threshold, and described the first de-noising weight AlphaIIR determines by distance function;

Wherein, described the first de-noising weight is directly proportional to described distance in described distance function, with described noise level, is inversely proportional to.

10. Image Denoising by Use device according to claim 7, is characterized in that, described determining unit also for, according to described the first de-noising weight, determine the second de-noising weight that described the first pixel is carried out to the spatial domain de-noising.

11. Image Denoising by Use device according to claim 10, is characterized in that, described acquiring unit also for, obtain first filter value of tell the first block of pixels after time-domain filtering;

Obtain second filter value of described the first block of pixels after airspace filter;

Described determining unit also for, according to described the first filter value, described the second filter value, described the first de-noising weight and described the second de-noising weight, determine the output valve P_out that described the first pixel is carried out to the time-space domain de-noising;

Described output valve is specially:

P_out=AlphaIIR*Pout_time+(1-AlphaIIR)*Pout_spa；

Wherein, described AlphaIIR is the first de-noising weight; Described (1-AlphaIIR) is the second de-noising weight; Described Pout_time is that described the first filter value, described Pout_spa are the second filter value.

12. picture noise device according to claim 11, is characterized in that, described device also comprises:

Pretreatment unit, for described the first block of pixels and described the second block of pixels are carried out respectively to the filtering processing, obtain preliminary treatment the first block of pixels and preliminary treatment the second block of pixels;

Described acquiring unit specifically for, obtain the first pixel average of the first block of pixels of preliminary treatment described in described present image;

Described determining unit specifically for, determine the described pixel value difference piece of preliminary treatment the second block of pixels described in the first block of pixels of preliminary treatment described in described present image and described reference picture.

Images