CN109064414A

CN109064414A - A kind of image de-noising method and device

Info

Publication number: CN109064414A
Application number: CN201810737255.4A
Authority: CN
Inventors: 李俊
Original assignee: Vivo Mobile Communication Co Ltd
Current assignee: Vivo Mobile Communication Co Ltd
Priority date: 2018-07-06
Filing date: 2018-07-06
Publication date: 2018-12-21
Anticipated expiration: 2038-07-06
Also published as: CN109064414B

Abstract

The present invention provides a kind of image de-noising method and devices, are related to technical field of image processing.The described method includes: carrying out down-sampling according to the first predetermined manner and the second predetermined manner to original image respectively, obtaining the first down-sampled images and the second down-sampled images；First down-sampled images and second down-sampled images are denoised respectively, obtain the first denoising image and the second denoising image；It is up-sampled based on the first denoising image and the second denoising image, obtains denoising image corresponding with the original image.The embodiment of the present invention, two different predetermined manner down-samplings are used to original image, then it is denoised respectively, obtain the first denoising image and the second denoising image, it is then based on two kinds of denoising picture up-samplings, the corresponding denoising image of the original image is obtained, the denoising effect to original image is improved.

Description

A kind of image de-noising method and device

Technical field

The present embodiments relate to technical field of image processing, more particularly to a kind of image de-noising method and device.

Background technique

Currently, the continuous promotion with people to digital photographing effect demand, seems for the denoising of raw image data It is more and more important.

In first technology, traditional Image denoising algorithm is to carry out single method down-sampling for original image and go mostly It makes an uproar.

Inventor has found that above-mentioned prior art has the following disadvantages: for raw image data in the course of the research, Down-sampling and denoising are carried out using single method, denoising effect is poor.

Summary of the invention

The embodiment of the present invention provides a kind of image de-noising method and device, it is intended to solve asking for prior art denoising effect difference Topic.

In a first aspect, the embodiment of the invention provides a kind of image de-noising methods, comprising:

According to the first predetermined manner and the second predetermined manner, down-sampling is carried out to original image respectively, is adopted under obtaining first Sampled images and the second down-sampled images；

First down-sampled images and second down-sampled images are denoised respectively, obtain the first denoising image With the second denoising image；

It is up-sampled, is obtained and the original image pair based on the first denoising image and the second denoising image The denoising image answered.

Second aspect, the embodiment of the invention also provides a kind of image denoising devices, comprising:

Down sample module, for adopt to original image respectively according to the first predetermined manner and the second predetermined manner Sample obtains the first down-sampled images and the second down-sampled images；

Image collection module is denoised, for distinguishing first down-sampled images and second down-sampled images Denoising obtains the first denoising image and the second denoising image；

Image generation module is denoised, for adopt based on the first denoising image and the second denoising image Sample obtains denoising image corresponding with the original image.

The first sub- down sample module third aspect of sub- down sample module second, the embodiment of the invention also provides a kind of movements Terminal, the mobile terminal include processor, memory and are stored on the memory and can run on the processor Computer program, the computer program realize the step of image de-noising method of the present invention when being executed by the processor Suddenly.

Fourth aspect, the embodiment of the invention also provides a kind of computer readable storage medium, the computer-readable storages Computer program is stored on medium, the computer program realizes image de-noising method of the present invention when being executed by processor The step of.

In embodiments of the present invention, according to the first predetermined manner and the second predetermined manner, original image is carried out down respectively Sampling, obtains the first down-sampled images and the second down-sampled images；To first down-sampled images and second down-sampling Image is denoised respectively, obtains the first denoising image and the second denoising image；Based on the first denoising image and described the Two denoising images are up-sampled, and denoising image corresponding with the original image is obtained.In the embodiment of the present invention, to original graph As using two different predetermined manner down-samplings, it is denoised respectively then, obtains the first denoising image and the second denoising figure Picture is then based on two kinds of denoising picture up-samplings, obtains the corresponding denoising image of the original image, improve going for original image It makes an uproar effect.

Detailed description of the invention

Fig. 1 shows the flow chart of the image de-noising method provided in the embodiment of the present invention one；

Fig. 2 shows the flow charts of the image de-noising method provided in the embodiment of the present invention two；

Fig. 3 shows the Baeyer format original image pixels distribution schematic diagram provided in the embodiment of the present invention two；

Fig. 4 shows and adopts under first for Bayer format original image shown in Fig. 3 provided in the embodiment of the present invention two Sampled images schematic diagram；

Fig. 5 is shown in the Bayer format original image provided in the embodiment of the present invention two, and mutually independent every group first Bayer formatted data schematic diagram；

Fig. 6 shows and adopts under second for Bayer format original image shown in Fig. 3 provided in the embodiment of the present invention two Sampled images schematic diagram；

Fig. 7 shows the step flow chart of the second down-sampled images of acquisition provided in the embodiment of the present invention two；

Fig. 8 shows the second denoising image schematic diagram provided in the embodiment of the present invention two；

Fig. 9 shows the first denoising image schematic diagram provided in the embodiment of the present invention two；

Figure 10 shows the intermediate up-sampling schematic diagram data provided in the embodiment of the present invention two；

Figure 11 show provided in the embodiment of the present invention two to centre up-sampling data in the 5th pixel normalized Schematic diagram data afterwards；

Figure 12 show provided in the embodiment of the present invention two for Fig. 9 and Figure 11, the 4th obtained Bayer formatted data Schematic diagram；

Figure 13 shows the pixel color value in Figure 12 that is directed to provided in the embodiment of the present invention two and is rounded, obtained figure The corresponding denoising image schematic diagram of original image shown in 3；

Figure 14 shows the structural block diagram for the image denoising device that according to embodiments of the present invention three provide；

Figure 15 shows the structural block diagram for another image denoising device that according to embodiments of the present invention three provide；

Figure 16 is the hardware structural diagram of one of embodiment of the present invention mobile terminal.

Specific embodiment

Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts Example, shall fall within the protection scope of the present invention.

Embodiment one

Referring to Fig.1, the flow chart of the image de-noising method provided in the embodiment of the present invention one is provided, can specifically include Following steps:

Step 101, according to the first predetermined manner and the second predetermined manner, down-sampling is carried out to original image respectively, is obtained First down-sampled images and the second down-sampled images.

In embodiments of the present invention, which can be digital photographing equipment just acquired image data, In the embodiment of the present invention, this is not especially limited.

In embodiments of the present invention, the first predetermined manner for carrying out down-sampling to original image may is that closest sampling Method, quadratic interpolattion, bicubic convolution method etc., in sampling process, can according to the image attributes information etc. of the original image, Suitable sample mode is set, in embodiments of the present invention, this is not especially limited.

In embodiments of the present invention, the second predetermined manner for carrying out down-sampling to original image is also possible to: closest to adopt Sample method, quadratic interpolattion, bicubic convolution method etc. can be according to the image attributes information of the original image in sampling process Deng suitable sample mode is arranged, in embodiments of the present invention, is not especially limited to this.

It should be noted that during carrying out down-sampling to same original image, using different the first predetermined manner and Second predetermined manner carries out, and can be concerned about the different factors of the original image, and then subsequent obtained original image is corresponding Denoising image, also paid close attention to the different factors of the original image, compared with the prior art only consider single factors for, promoted The denoising effect of original image.

Step 102, first down-sampled images and second down-sampled images are denoised respectively, obtains first Denoise image and the second denoising image.

In embodiments of the present invention, above-mentioned first down-sampled images and the second down-sampled images are denoised respectively, obtains the One denoising image and the second denoising image.In embodiments of the present invention, the first down-sampled images and the second down-sampled images are gone Method for de-noising can be identical, can not also be identical.This is not especially limited in the embodiment of the present invention.

For example, being directed to the first down-sampled images, can be denoised using non-local mean denoising, based on discrete cosine transform, At least one of denoising methods such as mean denoising are denoised.For the second down-sampled images, non-office can also be equally used Portion's mean denoising is denoised based at least one of discrete cosine transform denoising, mean denoising etc. denoising method method.

After the denoising of above-mentioned first down-sampled images, the first denoising image is obtained.Above-mentioned second down-sampled images are denoised Afterwards, the second denoising image is obtained.

In embodiments of the present invention, the first down-sampled images of above-mentioned original image, paid close attention to original image first are special Property, the first characteristic information of the original image, therefore the first denoising image obtained based on the first down-sampled images are remained, together Sample has paid close attention to the first characteristic information of the original image；Second down-sampled images of the original image, have paid close attention to original image Second characteristic remains the second characteristic information of the original image, therefore the second denoising obtained based on the second down-sampled images Image, focus attentions equally on the second characteristic information of the original image.For only considering single factors compared with the prior art, promoted The denoising effect of original image.

Step 103, it is up-sampled, is obtained and the original based on the first denoising image and the second denoising image The corresponding denoising image of beginning image.

In embodiments of the present invention, it is up-sampled, is somebody's turn to do based on above-mentioned first denoising image and the second denoising image The denoising image of original image.

In concrete application, using interpolation algorithm, the second denoising image can be inserted on the basis of the first denoising image Enter, and then obtains the denoising image of the original image.The pixel quantity that the denoising image of the original image includes, with the original graph As the pixel quantity for including can be identical.In embodiments of the present invention, this is not especially limited.

In concrete application, using interpolation algorithm, the first denoising image can also be inserted on the basis of the second denoising image Enter, and then obtains the denoising image of the original image.The pixel quantity that the denoising image of the original image includes, can be with the original The pixel quantity that beginning image includes is identical.In embodiments of the present invention, this is not especially limited.

Above-mentioned interpolation algorithm can be the difference based on edge image, image difference based on region etc., of the invention real It applies in example, this is not especially limited.

In embodiments of the present invention, the first denoising image of above-mentioned original image, has paid close attention to the first characteristic of original image, Remain the first characteristic information of the original image；Second denoising image of the original image, has paid close attention to the second of original image Characteristic remains the second characteristic information of the original image, therefore based on the first denoising image and the second denoising image, carries out Sampling obtains the corresponding denoising image of the original image, focus attentions equally on the first characteristic information and the second characteristic of the original image Information.For only considering single factors compared with the prior art, the denoising effect of original image is improved.

In embodiments of the present invention, according to the first predetermined manner and the second predetermined manner, original image is carried out down respectively Sampling, obtains the first down-sampled images and the second down-sampled images；To first down-sampled images and second down-sampling Image is denoised respectively, obtains the first denoising image and the second denoising image；Based on the first denoising image and described the Two denoising images are up-sampled, and denoising image corresponding with the original image is obtained.In the embodiment of the present invention, to original graph As using two different predetermined manner down-samplings, it is denoised respectively then, obtains the first denoising image and the second denoising figure Picture is then based on two kinds of denoising picture up-samplings, obtains the corresponding denoising image of the original image, obtaining going for original image It makes an uproar during image, has fully considered Multiple factors, for only considering single factors compared with the prior art, improve original The denoising effect of image.

Embodiment two

Referring to Fig. 2, the flow chart of the image de-noising method provided in the embodiment of the present invention two, the original image are shown It may include: the original image of Bayer format.The method can specifically include following steps:

It step 201, is one by each first Bayer formatted data down-sampling in the original image of the Bayer format Second pixel obtains the first down-sampled images；The first Bayer formatted data includes four the first pictures of 2*2 matrix form Element.

In embodiments of the present invention, referring to shown in Fig. 3, Fig. 3 shows Bayer format original image pixels distribution schematic diagram. The Bayer format original image may include: the first Bayer formatted data.In Fig. 3, first Bayer formatted data can be as Oval 11 frames select shown.One the first Bayer formatted data can be one on the acquisition terminal of original image acquisition The collected data of photosensitive element.In embodiments of the present invention, this is not especially limited.

In embodiments of the present invention, during being the first down-sampling by the Bayer format original image down-sampling, each First Bayer formatted data can be mutually indepedent.It, will be under four the first pixels of oval 11 frames choosing for example, referring to shown in Fig. 3 After being sampled as second pixel, can choose pixel coordinate in the Fig. 3 be (1,3), (Isosorbide-5-Nitrae), (2,3), (2,4) composition the One Bayer formatted data carries out down-sampling again, obtains next second pixel.It is each i.e. in the original image of the Bayer format Pixel can be only primary by selection.That is, unfolded pixel between each first Bayer formatted data.

In the Bayer format original image, pixel coordinate point can be with, has only included a color value, for example, referring to Shown in Fig. 3, pixel coordinate is that the color value that the pixel coordinate point of (1,1) includes is G color value.

In embodiments of the present invention, a first Bayer formatted data may include four the first pictures of 2*2 matrix form Element.Specifically, the first Bayer formatted data can referring to a first Bayer formatted data for selecting oval 11 frames in Fig. 3 To include four the first pixels of 2*2 matrix form, it is specifically as follows pixel coordinate are as follows: (1,1), (1,2), (2,1), (2,2) Four the first pixels, a first Bayer formatted data may include corresponding first pixel of two G color values, a R face Corresponding first pixel of color value, corresponding first pixel of a B color value.In one the first Bayer formatted data, two G colors Being worth corresponding first pixel does not have adjacent, and corresponding first pixel of R color value and corresponding first pixel of B color value do not have phase It is adjacent.

For example, referring to a first Bayer formatted data shown in ellipse 11 in Fig. 3, the G color value in the upper left corner is corresponding The first pixel, the first pixel corresponding with the G color value in the lower right corner is adjacent.The G color value corresponding first in the upper left corner Pixel can the first pixel corresponding with R color value, positioned at mutually going together, adjacent column.Corresponding first picture of the G color value in the lower right corner Element can the first pixel corresponding with the R color value, be located at same column, adjacent rows, corresponding first pixel of B color value and upper left Corresponding first pixel of the G color value at angle is located at same column, and the first pixel corresponding with the G color value in the lower right corner is mutually gone together.

It in embodiments of the present invention, can will be mutually independent each in the Bayer format Bayer format original image First Bayer formatted data down-sampling is second pixel, obtains the first down-sampled data.

In concrete application, the pixel value for four the first pixels that each first Bayer formatted data includes can be averaging Value, can also obtain second pixel in the hope of weighted mean etc., in embodiments of the present invention, be not especially limited to this.

For example, the pixel value for four the first pixels that each first Bayer formatted data includes can be sought arithmetic average Value obtains second pixel, i.e. the pixel value of second pixel can be flat for the arithmetic of the pixel value of four first pixels Mean value.The pixel value for four the first pixels that each first Bayer formatted data includes can also be sought geometrical mean, obtained One the second pixel, the i.e. pixel value of second pixel can be the geometrical mean of the pixel value of four first pixels.Also The pixel value for four the first pixels that each first Bayer formatted data includes can be sought into harmonic-mean, obtain one second Pixel, the i.e. pixel value of second pixel can be the harmonic-mean of the pixel value of four first pixels.It can will be each The pixel value for four the first pixels that first Bayer formatted data includes is averaged according to certain weight, obtains one Two pixels, the i.e. pixel value of second pixel can be the weighted average of the pixel value of four first pixels.In the present invention In embodiment, this is not especially limited.

In embodiments of the present invention, it in above-mentioned Bayer format original image, will be adopted under each first Bayer formatted data Sample is second pixel, obtains the first down-sampled images.

In embodiments of the present invention, which is four Color Channels that each first Bayer formatted data includes Average value or weighted mean of pixel value etc., then two pixel is by the original image of the Bayer format, each first Bayer lattice Four Color Channels of formula data are associated with, and are primarily upon the luminance information of the original image of the Bayer format.Therefore, this is visitd It is that four Color Channels are associated as a result, the major embodiment Bayer format that ear format original image, which obtains the first down-sampled images, The luminance information of original image.

For example, referring to shown in Fig. 4, Fig. 4 shows the first down-sampled images for the original image of Bayer format shown in Fig. 3 Schematic diagram.In Fig. 4, the second pixel 12 at pixel coordinate (1,1) can be the first Bayer formatted data 11 include in Fig. 3 four The second pixel after a first pixel down-sampling.And so on, in Fig. 4, the second pixel at pixel coordinate (1,2) can be figure (2,4) four pixel coordinate (1,3), pixel coordinate (1,4), pixel coordinate (2,3), pixel coordinate the first pixels form in 3 The second pixel after first Bayer formatted data down-sampling.In Fig. 4, the second pixel at pixel coordinate (2,1) can be in Fig. 3 Pixel coordinate (3,1), pixel coordinate (3,2), pixel coordinate (4,1), (4,2) four the first pixels of pixel coordinate composition first The second pixel after Bayer formatted data down-sampling.In embodiments of the present invention, this is not especially limited.

In Fig. 4, the pixel value of the second pixel 12 at pixel coordinate (1,1) can be original for Bayer format shown in Fig. 3 The R face of the G color value of the pixel coordinate (1,1) that the first Bayer formatted data shown in 11 includes in image, pixel coordinate (1,2) Color value, the B color value of pixel coordinate (2,1), arithmetic mean of instantaneous value of G color value of pixel coordinate (2,2) etc..Implement in the present invention In example, this is not especially limited.

For example, the pixel value of the second pixel 12 at pixel coordinate (1,1) can be (the G of pixel coordinate (1,1) in Fig. 4 The R color value of color value+pixel coordinate (1,2)+pixel coordinate (2,1) B color value+pixel coordinate (2,2) G color value)/ 4。

In concrete application, to above-mentioned Bayer format original image, by above-mentioned down-sampling, it is equivalent to above-mentioned Bayer format Four the first pixels that mutually independent each first Bayer formatted data includes in original image, are replaced by above-mentioned second pixel Generation, therefore, the pixel quantity which includes are a quarter of the Bayer format original image.By above-mentioned Down-sampling greatly reduces subsequent operand, accelerates denoising speed.

Meanwhile in embodiments of the present invention, by above-mentioned Bayer format original image, mutually independent each first Bayer Formatted data down-sampling is second pixel, obtains the first down-sampled images of the Bayer format original image.Being equivalent to will The color value of four pixels converts a pixel value for four channels, has paid close attention to the relevance in four channels, and then obtains The part luma information of the Bayer format original image.That is the first down-sampled images of the Bayer format original image, concern The relevance in four channels, remains the part luma information of the Bayer format original image.

Step 202, in the original image of the Bayer format, using every four the first Bayer formatted datas as one group, And by every group of first Bayer formatted data down-sampling be a second Bayer formatted data, obtain the second down-sampled images；Wherein, The matrix that every group of first Bayer formatted data is 2*2.

In embodiments of the present invention, using every four the first Bayer formatted datas as one group, by every group of first Bayer format Data down-sampling is a second Bayer formatted data, obtains the second down-sampled data.Every four first Bayer formatted datas For 2*2 matrix form.

In embodiments of the present invention, every group of first Bayer formatted data is mutually indepedent, that is, every two group of first Bayer format number According to not being overlapped between two group of first Bayer formatted data different or every the location of in the Bayer format original image.

In concrete application, referring to Figure 5, Fig. 5 is shown in Bayer format original image, and mutually independent every group first Bayer formatted data schematic diagram.The 15 of big oval frame choosing are mutually independent one group of first Bayer formatted data in Fig. 5.The phase Mutual independent one group of first Bayer formatted data is 2*2 matrix form.Mutually independent one group of first Bayer formatted data can By the first Bayer formatted data 2 of the first Bayer formatted data 1 of small oval 11 frames choosing, small oval 13 frames choosing, small ellipse First Bayer formatted data 3 of circle 14 frames choosing, and with small oval 14 and small oval 13 adjacent oval institute's frame choosing the One Bayer formatted data 4 composition.In Fig. 5, selected by the first Bayer formatted data 1 of small oval 11 frames choosing, small oval 13 frames The first Bayer formatted data 2, the first Bayer formatted data 3 of small oval 14 frames choosing, and with small oval 14 and small ellipse First Bayer formatted data 4 of 13 adjacent oval institute's frame choosings, mutual unfolded pixel.

In embodiments of the present invention, every two group of first Bayer formatted data be mutually indepedent or every two group of first Bayer lattice Formula data do not have between two group of first Bayer formatted data different or every the location of in the Bayer format original image Overlapping, refers to the unfolded pixel of every two group of first Bayer formatted data.For example, referring to shown in Fig. 5, big ellipse 15 in Fig. 5 Frame selects one group of first Bayer formatted data, selects one group of first Bayer formatted data, unfolded pixel with big oval 30 frame.

Referring to shown in Fig. 6, Fig. 6 shows the second down-sampled images signal for the original image of Bayer format shown in Fig. 3 Figure.

In Fig. 6, elliptical region 16 frames choosing can be with to be visitd for one group first shown in big oval frame selects in Fig. 5 15 The second Bayer formatted data that ear formatted data down-sampling obtains.

In Fig. 6, a second Bayer formatted data of 16 frame of elliptical region choosing, may include four third pixels.Fig. 6 Middle pixel coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) can be big oval frame in Fig. 5 The second Bayer formatted data that mutually independent four the first Bayer formatted data down-samplings shown in the 15 of choosing obtain is wrapped The four third pixels included.Four third pixels are similarly 2*2 matrix form.Four third pixels are belonging respectively to four face Chrominance channel.Specifically, in Fig. 6, four third pictures included by a second Bayer formatted data of 16 frame of elliptical region choosing In element, pixel coordinate (1,1) belongs to the first G Color Channel, pixel coordinate (1,2) belongs to R Color Channel, pixel coordinate (2,1) Belong to B color channel, pixel coordinate (2,2) may belong to the 2nd G Color Channel.

In Fig. 6, pixel coordinate (1,3), pixel coordinate (Isosorbide-5-Nitrae), pixel coordinate (2,3), pixel coordinate (2,4) can be, The second Bayer formatted data that 30 1 group of first Bayer formatted data down-sampling of big oval frame choosing obtains in Fig. 5.

In embodiments of the present invention, in Bayer format original image, using every four the first Bayer formatted datas as one Every group of first Bayer formatted data down-sampling is a second Bayer formatted data, obtains the second down-sampled images by group, and every group First Bayer formatted data is the matrix of 2*2.Therefore, in embodiments of the present invention, which also belongs to visit Ear format.

In embodiments of the present invention, in Bayer format original image, using every four the first Bayer formatted datas as one Every group of first Bayer formatted data down-sampling is a second Bayer formatted data, obtains the second down-sampled images, specifically by group In, to above-mentioned initial data, by above-mentioned down-sampling, it is equivalent to every four the first Bayer lattice in above-mentioned initial data Formula data are substituted by said one the second Bayer formatted data, and therefore, the pixel quantity which includes can be with For a quarter of the Bayer format original image.By above-mentioned down-sampling, subsequent operand is greatly reduced, is accelerated It makes an uproar speed.

In embodiments of the present invention, in Bayer format original image, using every four the first Bayer formatted datas as one Every group of first Bayer formatted data down-sampling is a second Bayer formatted data, obtains the second down-sampled images, quite by group In by four the first Bayer formatted datas, conversion one is the second Bayer formatted data, which also retains Four channels, and then remain the part colours information of the Bayer format original image.Therefore, second down-sampled images, Four channels are relatively independent, remain the part colours information of the Bayer format original image.

In embodiments of the present invention, optionally, referring to shown in Fig. 7, Fig. 7 is shown, and obtains the step of the second down-sampled images Rapid flow chart.Specifically, described in the original image of the Bayer format, using every four the first Bayer formatted datas as one Group, and be a second Bayer formatted data by every group of first Bayer formatted data down-sampling, obtain the second down-sampled images Can include: with step

Every four the first Bayer formatted datas are divided into one group in the original image by sub-step S21；Wherein, often The matrix that the first Bayer formatted data of group is 2*2.

In embodiments of the present invention, in original image, every four the first Bayer formatted datas are divided into one group, every group First Bayer formatted data is the matrix of 2*2.

For example, referring to shown in Fig. 5, the 15 of big oval frame choosing are to be divided into four the first Bayer formatted datas in Fig. 5 One group.Group the first Bayer formatted data is 2*2 matrix form.The first Bayer of group formatted data can be by small oval 11 institute First Bayer formatted data 1 of frame choosing, the first Bayer formatted data 2 of small oval 13 frames choosing, small oval 14 frames choosing the One Bayer formatted data 3, and the first Bayer formatted data 4 selected with small oval 14 and small oval 13 adjacent oval institute's frame Composition.

In Fig. 5, the first Bayer formatted data 1, the first Bayer formatted data 2, the first Bayer formatted data 3, the first Bayer Formatted data 4, one group of first Bayer formatted data of composition are 2*2 matrix form.

Sub-step S22, the first pixel that will include in every group of first Bayer formatted data, according to face belonging to each pixel Chrominance channel, down-sampling are four third pixels.

In the original image of Bayer format, every 16 the first pictures included by every group of first Bayer formatted data are obtained Element, four the first pixels that each the first data of Bayer format include are belonging respectively to four Color Channels.

For example, referring to shown in Fig. 5, the 15 of big oval frame choosing are one group of first Bayer formatted data in Fig. 5.This group of Bayer Four the first Bayer formatted datas that the first data of format include can be by the first Bayer formatted data of small oval 11 frames choosing 1, the first Bayer formatted data 2 of small oval 13 frames choosing, the first Bayer formatted data 3, Yi Jiyu of small oval 14 frames choosing Small oval 14 and small oval 13 adjacent oval institute's frame choosing the first Bayer formatted data 4 composition.It is mutually independent in Fig. 5 First Bayer formatted data 1, the first Bayer formatted data 2, the first Bayer formatted data 3, the first Bayer formatted data 4 are 2*2 Matrix form.

In embodiments of the present invention, this four the first Bayer formatted datas are mutually indepedent, refer to this four the first Bayer lattice The unfolded pixel of formula data.Referring for example to shown in Fig. 5, in Fig. 5, by the first Bayer formatted data of small oval 11 frames choosing 1, the first Bayer formatted data 2 of small oval 13 frames choosing, the first Bayer formatted data 3, Yi Jiyu of small oval 14 frames choosing Small oval 14 and small oval 13 adjacent oval institute's frame choosing the first Bayer formatted data 4, mutual unfolded picture Element.

In embodiments of the present invention, a first Bayer formatted data includes four the first pixels.Four first pixels It is belonging respectively to four Color Channels.For example, by the first Bayer formatted data 1 of small oval 11 frames choosing, that is, including in Fig. 5 Four the first pixels are respectively as follows: the first pixel of pixel coordinate (1,1), the first pixel of pixel coordinate (1,2), pixel coordinate (2,1) the first pixel of the first pixel, pixel coordinate (2,2).First pixel of pixel coordinate (1,1) belongs to the first G color Channel, the first pixel of pixel coordinate (1,2) belong to R Color Channel, and it is logical that the first pixel of pixel coordinate (2,1) belongs to B color First pixel in road, pixel coordinate (2,2) belongs to the 2nd G Color Channel.

In embodiments of the present invention, 16 the first pixels that every group of first Bayer formatted data includes are obtained.For example, ginseng According to shown in Fig. 5, big oval frame choosing 15 shown in one group of first Bayer formatted data be the first picture for including pixel coordinate (1,1) Element, the first pixel of pixel coordinate (1,2), the first pixel of pixel coordinate (1,3), pixel coordinate (1,4) the first pixel, as The first pixel, the first pixel of pixel coordinate (2,2), the first pixel of pixel coordinate (2,3), the pixel of plain coordinate (2,1) are sat Mark first pixel of (2,4), the first pixel of pixel coordinate (3,1), the first pixel of pixel coordinate (3,2), pixel coordinate (3, 3) the first pixel, the first pixel of pixel coordinate (3,4), the first pixel of pixel coordinate (4,1), pixel coordinate (4,2) First pixel, the first pixel of pixel coordinate (4,3), the first pixel of pixel coordinate (4,4), totally 16 the first pixels.

In Fig. 5, the first pixel, the first pixel of pixel coordinate (Isosorbide-5-Nitrae) of pixel coordinate (1,3), pixel coordinate (2,3) First pixel of the first pixel, pixel coordinate (2,4).Belong to the first Bayer formatted data 2 of small oval 13 frames choosing, pixel is sat First pixel of mark (1,3) belongs to the first G Color Channel, and the first pixel of pixel coordinate (Isosorbide-5-Nitrae) belongs to R Color Channel, pixel First pixel of coordinate (2,3) belongs to B color channel, and the first pixel of pixel coordinate (2,4) belongs to the 2nd G Color Channel.

In Fig. 5, the first pixel, the first pixel of pixel coordinate (3,2) of pixel coordinate (3,1), pixel coordinate (4,1) First pixel of the first pixel, pixel coordinate (4,2).Belong to the first Bayer formatted data 3 of small oval 14 frames choosing, pixel is sat First pixel of mark (3,1) belongs to the first G Color Channel, and the first pixel of pixel coordinate (3,2) belongs to R Color Channel, pixel First pixel of coordinate (4,1) belongs to B color channel, and the first pixel of pixel coordinate (4,2) belongs to the 2nd G Color Channel.

In Fig. 5, the first pixel, the first pixel of pixel coordinate (3,4) of pixel coordinate (3,3), pixel coordinate (4,3) First pixel of the first pixel, pixel coordinate (4,4).Belong to the first Bayer formatted data 4, the first picture of pixel coordinate (3,3) Element belongs to the first G Color Channel, and the first pixel of pixel coordinate (3,4) belongs to R Color Channel, and the first of pixel coordinate (4,3) Pixel belongs to B color channel, and the first pixel of pixel coordinate (4,4) belongs to the 2nd G Color Channel.

In embodiments of the present invention, in Bayer format original image, include by every group of first Bayer formatted data ten Six the first pixels, according to channel belonging to each first pixel, down-sampling is four third pixels, obtains the second down-sampling figure Picture.

In concrete application, for example, referring to shown in Fig. 6, in Fig. 6, the third picture of pixel coordinate (1,1) can be the first G face Chrominance channel, can in every four the first Bayer formatted datas as included by above-mentioned one group of first Bayer formatted data four The average value of one G Color Channel.For example, the first pixel of pixel coordinate (1,1) belongs to the first G Color Channel, pixel in Fig. 5 First pixel of coordinate (1,3) belongs to the first G Color Channel, and it is logical that the first pixel of pixel coordinate (3,1) belongs to the first G color First pixel in road, pixel coordinate (3,3) belongs to the first G Color Channel, then, in Fig. 6, the third pixel of pixel coordinate (1,1) Can be, in Fig. 5, the first pixel of pixel coordinate (1,1), the first pixel of pixel coordinate (1,3), pixel coordinate (3,1) First pixel, the first pixel of pixel coordinate (3,3), the mean value of this four pixel values for belonging to the first G Color Channel.

In embodiments of the present invention, mean value herein can be arithmetic mean of instantaneous value, geometrical mean, harmonic-mean, add Mean value etc. is weighed, in embodiments of the present invention, this is not especially limited.When mean value herein is arithmetic mean of instantaneous value, then Fig. 6 In, the pixel value of the third pixel of pixel coordinate (1,1)=(in Fig. 5, the pixel value of the first pixel of pixel coordinate (1,1)+as Pixel value+pixel coordinate (3,3) of first pixel of pixel value+pixel coordinate (3,1) of the first pixel of plain coordinate (1,3) The pixel value of first pixel)/4.In Fig. 6, the third pixel of pixel coordinate (1,1) also belongs to the first G Color Channel.

Likewise, in Fig. 6, the pixel value of the third pixel of pixel coordinate (1,2)=(in Fig. 5, pixel coordinate (1,2) The picture of first pixel of pixel value+pixel coordinate (3,2) of the first pixel of pixel value+pixel coordinate (1,4) of the first pixel The pixel value of first pixel of plain value+pixel coordinate (3,4))/4.In Fig. 5, the first pixel, the pixel of pixel coordinate (1,2) are sat Mark first pixel of (1,4), the first pixel of pixel coordinate (3,2), the first pixel of pixel coordinate (3,4) belong to R color Channel,

In Fig. 6, the third pixel of pixel coordinate (1,2) belongs to R Color Channel.

Likewise, in Fig. 6, the pixel value of the third pixel of pixel coordinate (2,1)=(in Fig. 5, pixel coordinate (2,1) The pixel of the first pixel of the pixel value of first pixel of pixel value+pixel coordinate (2,3) of the first pixel+element coordinate (4,1) The pixel value of first pixel of value+pixel coordinate (4,3))/4.In Fig. 5, the first pixel of pixel coordinate (2,1), pixel coordinate (2,3) it is logical that the first pixel of the first pixel, pixel coordinate (4,1), the first pixel of pixel coordinate (4,3) belongs to B color Road, in Fig. 6, the third pixel of pixel coordinate (2,1) belongs to B color channel.

Likewise, in Fig. 6, the pixel value of the third pixel of pixel coordinate (2,2)=(in Fig. 5, pixel coordinate (2,2) The picture of first pixel of pixel value+pixel coordinate (4,2) of the first pixel of pixel value+pixel coordinate (2,4) of the first pixel The pixel value of first pixel of plain value+pixel coordinate (4,4))/4.In Fig. 5, the first pixel, the pixel of pixel coordinate (2,2) are sat Mark first pixel of (2,4), the first pixel of pixel coordinate (4,2), the first pixel of pixel coordinate (4,4) belong to the 2nd G Color Channel, in Fig. 6, the third pixel of pixel coordinate (2,2) belongs to the 2nd G Color Channel.

Based on same method, 16 the first pixels for including by every group of first Bayer formatted data, according to each first Channel belonging to pixel, down-sampling are a second Bayer formatted data.

In embodiments of the present invention, four third pixels are belonging respectively to four different Color Channels.For example, in Fig. 6, as The third pixel of plain coordinate (1,1) also belongs to the first G Color Channel, and it is logical that the third pixel of pixel coordinate (1,2) belongs to R color Road, the third pixel of pixel coordinate (2,1) belong to B color channel, and the third pixel of pixel coordinate (2,2) belongs to the 2nd G color Channel.

Sub-step S23 forms the second Bayer formatted data with four third pixels, to obtain under described second Sampled images.

In embodiments of the present invention, aforementioned four third pixel forms a second Bayer formatted data.In concrete application, For example, referring to shown in Fig. 6, in Fig. 6, the third pixel of pixel coordinate (1,1), the third pixel of pixel coordinate (1,2), pixel are sat Third pixel, the third pixel of pixel coordinate (2,2) of (2,1) are marked, that is, forms a second Bayer formatted data.It is ellipse in Fig. 6 What 16 frames of border circular areas selected can be for one group of first Bayer formatted data down-sampling shown in the 15 of big oval frame choosing in Fig. 5 The second obtained Bayer formatted data.

In embodiments of the present invention, four third pixels are formed into the second Bayer formatted data, to obtain the second down-sampling Image.

In embodiments of the present invention, four third pixels are belonging respectively to four different Color Channels.For example, in Fig. 6, as The third pixel of plain coordinate (1,1) also belongs to the first G Color Channel, and it is logical that the third pixel of pixel coordinate (1,2) belongs to R color Road, the third pixel of pixel coordinate (2,1) belong to B color channel, and the third pixel of pixel coordinate (2,2) belongs to the 2nd G color Channel, and then second down-sampled images equally include four Color Channels.Referring to shown in Fig. 6, Fig. 6 is shown for Fig. 3 institute Show the second down-sampled images schematic diagram of Bayer format original image.

In concrete application, to above-mentioned initial data, by above-mentioned down-sampling, it is equivalent to every four in above-mentioned initial data A first Bayer formatted data is substituted by said one the second Bayer formatted data, and therefore, which includes Pixel quantity can be a quarter of the Bayer format original image.By above-mentioned down-sampling, subsequent fortune is greatly reduced Calculation amount accelerates denoising speed.

In embodiments of the present invention, in Bayer format original image, using every four the first Bayer formatted datas as one The first Bayer formatted data of group, 16 the first pixels for including by every group of first Bayer format, according to each first pixel institute The channel of category, down-sampling are a second Bayer formatted data, obtain the second down-sampled images, are equivalent to four the first Bayers Formatted data is converted into a second Bayer formatted data, which also retains four channels, and is four A channel it is relatively independent sample, and then remain the part colours information of the Bayer format original image.Therefore, this Two down-sampled images, four channels are relatively independent, remain the part colours information of the Bayer format original image.

In embodiments of the present invention, the first down-sampled images of above-mentioned Bayer format original image and the Bayer format are former Second down-sampled images of beginning image, including pixel quantity it is identical, be the pixel number that the Bayer format original image includes The a quarter of amount.Subsequently through to first down-sampled images and the second down-sampled images, during being denoised, compare In the Bayer format original image, since pixel quantity is reduced, denoising speed is fast.

In embodiments of the present invention, the first down-sampled images of above-mentioned Bayer format original image have paid close attention to four channels Relevance, remain the part luma information of the Bayer format original image；It is adopted under the second of the Bayer format original image Sampled images, four channels are relatively independent, remain the part colours information of the Bayer format original image.

Step 203, first down-sampled images are denoised, obtains the first denoising image.

In the embodiment of the present invention, the first down-sampled images are removed dryness, first is obtained and removes dryness image.It, can be in concrete application Using non-local mean denoising, based at least one of denoising such as discrete cosine transform denoising, mean denoising, to adopting under first Sampled images are denoised.In embodiments of the present invention, this is not especially limited.

In embodiments of the present invention, optionally the acquisition terminal of the Bayer format original image can repeatedly be tried It tests, mainly for the camera of the acquisition terminal, obtains under each intensity of illumination, it is corresponding between different pixels brightness and noise Relationship.

Specifically, can be fixed by the acquisition terminal itself, acquisition target is also remained unchanged, acquisition end is changed The locating intensity of illumination in end specifically can be and change the intensity of illumination that the camera of the acquisition terminal receives, in each illumination Under intensity, the corresponding relationship between different pixels brightness and noise is obtained.Under one intensity of illumination, a different picture can be corresponded to The corresponding relationship of plain brightness and noise.In embodiments of the present invention, this is not especially limited.

It in embodiments of the present invention, can be with the Bayer format original image during to the denoising of the first down-sampled images The corresponding intensity of illumination of acquisition terminal, the corresponding relationship between pixel intensity and noise, for reference.In the embodiment of the present invention In, this is not especially limited.

In concrete application, the intensity of illumination of the available Bayer format original image determines the first down-sampled images Pixel intensity determines making an uproar for first down-sampled images with the corresponding relationship between above-mentioned intensity of illumination, pixel intensity and noise Sound.The intensity of illumination of Bayer format original image can be attribute of the Bayer format original image itself, and the Bayer format is former Beginning image determines that corresponding intensity of illumination determines that.The intensity of illumination of the Bayer format original image can visit to acquire this When ear format original image, intensity of illumination that acquisition terminal receives.In embodiments of the present invention, this is not especially limited.

In concrete application, the Bayer format original graph can be obtained from attribute information of the Bayer format original image etc. The intensity of illumination of picture.For example, the Bayer format can be obtained from the property files such as the header file of the Bayer format original image The intensity of illumination of original image.In embodiments of the present invention, this is not especially limited.

Step 204, based on the first denoising image, second down-sampled images are denoised, second is obtained and goes It makes an uproar image.

In embodiments of the present invention, based on the first denoising image, the second down-sampled images is removed dryness, second is obtained and removes dryness figure Picture.

In concrete application, it can be denoised using non-local mean denoising, based on discrete cosine transform denoising, mean denoising etc. At least one of, the second down-sampled images are denoised.In embodiments of the present invention, this is not especially limited.

For example, can be denoised using based on discrete cosine transform, which is denoised.This second During down-sampled images denoising, it can be carried out based on the corresponding relationship between above-mentioned intensity of illumination, pixel intensity and noise Denoising.In embodiments of the present invention, this is not especially limited.

In embodiments of the present invention, optionally, when to second down-sampled images using non-local mean denoise, it is described Based on the first denoising image, the step of being denoised to second down-sampled images, obtain the second denoising image, packet It includes: determining similarity between first piece of the first denoising image；According to the first denoising image and second down-sampling The corresponding relationship of image determines similarity between second piece of the second down-sampled images based on similarity between described first piece；It is based on Similarity between described second piece denoises second down-sampled images using non-local mean, obtains the second denoising image.

Specifically, if using non-local mean during being denoised to the second down-sampled images and denoising, it can be with base In above-mentioned first denoising image, which is denoised.Specifically, when being gone to the second down-sampled images During making an uproar, when using non-local mean denoising, similarity between first piece of the first denoising image can be determined, then root According to the corresponding relationship of the first denoising image and the second down-sampled images, similarity between being based on this first piece determines this under second Similarity between second piece of sampled images denoises second down-sampled images based on similarity between this second piece, obtains second Denoise image.

Specifically, according to the corresponding relationship of the first denoising image and the second down-sampled images, based on phase between this first piece Like degree, similarity between second piece of second down-sampled images is determined, can be based on the first denoising image corresponding first Similarity between this first piece is converted the second down-sampled images by down-sampled images, the corresponding relationship with the second down-sampled images Second piece between similarity this is not especially limited in embodiments of the present invention.

In embodiments of the present invention, between second piece of the second down-sampled images has been determined after similarity, further according to second The self-information of down-sampled images determines the corresponding second denoising image of the second down-sampled images.It is right based on the first denoising image During second down-sampled images are denoised, and the first denoising image is the denoising figure of the Bayer format original image Picture is equivalent to the denoising image based on Bayer format original image, denoises to the second down-sampled images, to a certain extent to this Original image is denoised twice, it can be understood as is denoised twice to the Bayer format original image, denoises effect More preferably.

In embodiments of the present invention, the pixel quantity that the first denoising image, the second denoising image include, is similarly the Bayer The a quarter for the pixel quantity that format original image includes.In embodiments of the present invention, above-mentioned Bayer format original image First down-sampled images have paid close attention to the relevance in four channels, remain the part luma information of the Bayer format original image, Therefore the first denoising image obtained based on the first down-sampled images, the focus attentions equally on relevance in four channels remain this The part luma information of Bayer format original image；Second down-sampled images of the Bayer format original image, four channel phases To independence, the part colours information of the Bayer format original image, therefore obtained based on the second down-sampled images are remained Two denoising images, same four channels are relatively independent, remain the part colours information of the Bayer format original image.

Step 205, with the third Bayer formatted data in the second denoising image, the first denoising figure is replaced respectively Four the 4th pixels as in obtain intermediate up-sampling data；Four the 4th pixels are in the first denoising image The 5th pixel in position, with the third Bayer formatted data, the position in the second denoising image are corresponding.

In embodiments of the present invention, with the third Bayer formatted data in the second denoising image, the first denoising is replaced respectively Four the 4th pixels in image obtain intermediate up-sampling data；Position of four the 4th pixels in the first denoising image It sets, with the 5th pixel in third Bayer formatted data, the position in the second denoising image is corresponding.

In embodiments of the present invention, which may include third Bayer formatted data.The third Bayer lattice Formula data equally include four the 5th pixels.For example, referring to shown in Fig. 8, Fig. 8 shows the second denoising image schematic diagram.

In Fig. 8, what elliptical region 17 was indicated is third Bayer formatted data.It is wrapped in the third Bayer formatted data It has included four the 5th pixels, i.e. the third Bayer formatted data that is indicated of elliptical region 17, that is, has included pixel coordinate (1,1) The 5th pixel, the 5th pixel of pixel coordinate (1,2), the 5th pixel of pixel coordinate (2,1), pixel coordinate (2,2) Five pixels.

In embodiments of the present invention, available above-mentioned third Bayer formatted data, the corresponding color value in each channel.Example Such as, referring to shown in Fig. 8, in the third Bayer formatted data that elliptical region 17 is indicated, pixel coordinate (1,1) corresponding first The color value in the channel G is 60, the color value in the corresponding channel R of pixel coordinate (1,2) is 70, the corresponding B of pixel coordinate (2,1) is logical The color value in road is 80, the color value in corresponding 2nd channel G of pixel coordinate (2,2) is 80.

In embodiments of the present invention, which includes four the 4th pixels.For example, referring to shown in Fig. 9, Fig. 9, Show the first denoising image schematic diagram.In Fig. 9, pixel coordinate (1,1) that border circular areas 18 is indicated can with for this first Denoise the 4th pixel in image.In Fig. 9, pixel coordinate (1,1) that border circular areas 19 is indicated, pixel coordinate (1,2), Pixel coordinate (2,1), pixel coordinate (2,2), four the 4th pixels are 2*2 matrix form.

In the present invention is implemented, the corresponding pixel value of available 4th pixel, for example, referring to shown in Fig. 9, border circular areas 19 pixel coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), the corresponding pixel values of pixel coordinate (2,2) indicated It respectively can be 120,130,140,150.

In embodiments of the present invention, with second denoising image a third Bayer formatted data, replace respectively this first Four the 4th pixels in image are denoised, centre up-sampling data are obtained.That is, by the picture that border circular areas 19 is indicated in Fig. 9 Four plain coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) the 4th pixels, with above-mentioned Fig. 8 In, the third Bayer formatted data replacement that elliptical region 17 is indicated is equivalent to and replaces the 4th pixel of each of Fig. 9 For 4 pixels of a third Bayer formatted data.

For example, referring to shown in Figure 10, Figure 10 is shown, centre up-sampling schematic diagram data.Elliptical region 20 in Figure 10 The region of frame choosing is that, by Fig. 8, the third Bayer formatted data that elliptical region 17 is indicated is replaced in Fig. 9, circle After domain 18 indicates four the 4th pixels, corresponding intermediate up-sampling data.

In Figure 10, the color value in corresponding first channel G of pixel coordinate (1,1) is 60, the corresponding R of pixel coordinate (1,2) The color value in channel is 70, the color value of the corresponding channel B of pixel coordinate (2,1) is 80, pixel coordinate (2,2) corresponding second The color value in the channel G is 80；The color value in corresponding first channel G of pixel coordinate (1,3) is 60, and pixel coordinate (Isosorbide-5-Nitrae) is corresponding The channel R color value be 70, the color value of the corresponding channel B of pixel coordinate (2,3) is 80, pixel coordinate (2,4) is corresponding The color value in the 2nd channel G is 80；The color value in corresponding first channel G of pixel coordinate (3,1) is 60, pixel coordinate (3,2) The color value in the corresponding channel R is 70, the color value of the corresponding channel B of pixel coordinate (4,1) is 80, pixel coordinate (4,2) is right The color value in the 2nd channel G answered is 80；The color value in corresponding first channel G of pixel coordinate (3,3) is 60, pixel coordinate (3,4) color value in the corresponding channel R be 70, the color value of the corresponding channel B of pixel coordinate (4,3) be 80, pixel coordinate (4, 4) color value in corresponding 2nd channel G is 80.

In embodiments of the present invention, four the 4th pixels remove dryness position and above-mentioned third Bayer in image above-mentioned first Formatted data, at this, second to remove dryness position in image corresponding, refers to four the 4th pixels, remove dryness with before down-sampling, upper State corresponding 16 the first pixels in original image, and the third Bayer formatted data, remove dryness with before down-sampling, in above-mentioned original 16 the first pixels included by corresponding four the first Bayer formatted datas, respectively correspond identical in beginning image.

For example, in Fig. 9, pixel coordinate (1,1) that border circular areas 19 is indicated, pixel coordinate (1,2), pixel coordinate (2, 1), pixel coordinate (2,2), 16 the first pixels before removing dryness and sampling, in the raw Bayer format-pattern are as follows: in Fig. 3, First pixel of pixel coordinate (1,1), the first pixel of pixel coordinate (1,2), the first pixel, the pixel of pixel coordinate (1,3) First pixel of coordinate (1,4), the first pixel of pixel coordinate (2,1), the first pixel of pixel coordinate (2,2), pixel coordinate (2,3) the first pixel, the first pixel of pixel coordinate (2,4), the first pixel of pixel coordinate (3,1), pixel coordinate (3,2) The first pixel, the first pixel of pixel coordinate (3,3), the first pixel of pixel coordinate (3,4), pixel coordinate (4,1) One pixel, the first pixel of pixel coordinate (4,2), the first pixel of pixel coordinate (4,3), pixel coordinate (4,4) the first picture Element.In Fig. 8, third Bayer formatted data that elliptical region 17 is indicated.Before removing dryness and sampling, in the raw Bayer format In image, referring to Figure 5, the first Bayer formatted data 1, the first Bayer formatted data 2, the first Bayer formatted data 3, first 16 the first pixels that 4, four the first Bayer formatted datas of Bayer formatted data include respectively correspond identical.

In embodiments of the present invention, pixel quantity included by centre up-sampling data, for the first denoising image Four times, while being also second denoising four times of image.Meanwhile second denoising image be made of third Bayer formatted data, wrap The independence and colouring information in four channels of Bayer format original image have been included, and, above-mentioned first denoising image considers this and visits The relevance and Bayer format original image luminance information in four channels of ear format original image, and then above-mentioned intermediate up-sampling number According to the independence and colouring information for not only including four channels of Bayer format original image, the association in four channels is further comprised Property and Bayer format original image luminance information.

In embodiments of the present invention, the pixel quantity as included by the first denoising image is Bayer format original image A quarter is then equivalent to by above-mentioned up-sampling and denoises the pixel quantity quadruplication for including in image for first, and then in this Between the sampled data pixel quantity that includes it is identical as the pixel quantity that the Bayer format original image includes.Meanwhile second denoises Image is made of third Bayer formatted data, includes the independence and colouring information in four channels of Bayer format original image, And above-mentioned first denoising image considers the relevance and Bayer format original image in four channels of Bayer format original image Luminance information, and then above-mentioned intermediate up-sampling data not only include the independence and face in four channels of Bayer format original image Color information further comprises the relevance and Bayer format original image luminance information in four channels.

Step 206, by the up-sampling data of centre, four the 5th pixels of each third Bayer formatted data carry out normalizing Change processing, obtains four the 6th pixels.

It in embodiments of the present invention, can be to included by each third Bayer formatted data in third Bayer formatted data Four the 5th pixels are normalized, and obtain four the 6th pixels.

In embodiments of the present invention, data are up-sampled in centre, every two third Bayer formatted data does not have lap.

In embodiments of the present invention, the third Bayer formatted data in centre up-sampling data, is referred to such as Figure 10 It is shown, Tu10Zhong, 20 frame of elliptical region choosing region in, the color value in corresponding first channel G of pixel coordinate (1,1) is 60, the color value in the corresponding channel R of pixel coordinate (1,2) is 70, the color value of the corresponding channel B of pixel coordinate (2,1) is 80, The color value in corresponding 2nd channel G of pixel coordinate (2,2) is 80, the third Bayer format number as in the intermediate samples data According to.

In concrete application, four the 5th pixels that can include by the third Bayer formatted data are normalized, Obtain four the 6th pixels.

It in embodiments of the present invention, is 20 frame of elliptical region choosing in Figure 10 for example, if the third Bayer formatted data Pixel coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) are formed in region, then the third Bayer formatted data includes four the 5th pixels, which is 2*2 matrix form.Such as, elliptical region in Figure 10 In the region of 20 frames choosing, the color value in corresponding first channel G of pixel coordinate (1,1) is 60, the corresponding R of pixel coordinate (1,2) The color value in channel is 70, the color value of the corresponding channel B of pixel coordinate (2,1) is 80, pixel coordinate (2,2) corresponding second The color value in the channel G is 80, and four the 5th pixels which includes are normalized, then can be with It obtains, referring to Fig.1 shown in 1, Figure 11 is shown, and is illustrated to the data after the 5th pixel normalized in intermediate up-sampling data Figure.

In embodiments of the present invention, place is normalized to four the 5th pixels that above-mentioned third Bayer formatted data includes Reason, can be the mean value for four the 5th pixels for taking above-mentioned third Bayer formatted data to include, with the third Bayer formatted data The color value in each channel divided by four the 5th pixels in the third Bayer formatted data mean value.For example, referring in Figure 11, 21 frame choosings of elliptical region up-sample the data in data after the 5th pixel normalized to centre, as according to the party Method is in above-mentioned Figure 10, and the color value in corresponding first channel G of pixel coordinate (1,1) is 60, the corresponding R of pixel coordinate (1,2) The color value in channel is 70, the color value of the corresponding channel B of pixel coordinate (2,1) is 80, pixel coordinate (2,2) corresponding second The color value in the channel G is the processing result of third Bayer formatted data corresponding to 80.

Pixel coordinate (1,1) in Figure 11, pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) are, right In above-mentioned Figure 10, the color value in corresponding first channel G of pixel coordinate (1,1) is 60, the corresponding channel R of pixel coordinate (1,2) Color value be 70, the color value of the corresponding channel B of pixel coordinate (2,1) is 80, corresponding 2nd G of pixel coordinate (2,2) is logical The color value in road is third Bayer formatted data corresponding to 80, four the 6th pixels obtained after normalization.

Step 207, four the 6th pixels are multiplied with the 4th pixel of corresponding position respectively, obtain one the 4th and visits Ear formatted data, to obtain denoising image corresponding with the original image.

In embodiments of the present invention, four the 6th pixels are multiplied with the 4th pixel of above-mentioned corresponding position respectively, are obtained One the 4th Bayer formatted data, to obtain denoising image corresponding with the original image.

In concrete application, referring to Fig.1 shown in 1, then the pixel coordinate (1,1) that elliptical region 21 is indicated, pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) can be for four the 5th pixels to above-mentioned third Bayer formatted data After first normalized, obtained four the 6th pixels.

By aforementioned four 6th pixel, it is multiplied respectively with the 4th pixel of above-mentioned corresponding position, that is, by Figure 11, pixel The corresponding G color value of coordinate (1,1), corresponding 4th pixel value is multiplied with pixel coordinate in Fig. 9 (1,1), by Figure 11, pixel The corresponding R color value of coordinate (1,2), corresponding 4th pixel value is multiplied with pixel coordinate in Fig. 9 (1,2), by Figure 11, pixel The corresponding B color value of coordinate (2,1), corresponding 4th pixel value is multiplied with pixel coordinate in Fig. 9 (2,1), by Figure 11, pixel The corresponding G color value of coordinate (2,2), corresponding 4th pixel value is multiplied with pixel coordinate in Fig. 9 (2,2), respectively obtains Figure 12 In, the corresponding G color value of pixel coordinate (1,1), the corresponding R color value of pixel coordinate (1,2), the corresponding B of pixel coordinate (2,1) Color value, the corresponding G color value of pixel coordinate (2,2).

In Figure 12, pixel coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) constituted One the 4th Bayer formatted data.According to example shown in above-mentioned Fig. 9 and Figure 11, referring to Fig.1 shown in 2, Figure 12, which is shown, to be directed to Fig. 9 and Figure 11, the 4th obtained Bayer formatted data schematic diagram.

In embodiments of the present invention, it if aforementioned four 6th pixel is multiplied with the 4th pixel of corresponding position respectively, obtains To the 4th Bayer formatted data in, if the color value of each pixel has score or decimal, such as

In Figure 12, pixel coordinate (1,1), pixel coordinate (1,2), pixel coordinate (2,1), shown in pixel coordinate (2,2), then Can be in above-mentioned 4th Bayer formatted data, the color value of pixel is rounded for the color value of decimal or score.Similarly For in the 4th Bayer formatted data, the color value of pixel is rounded for all colours value of decimal or score, and then To the denoising image of the original image.

It is rounded for example, referring to shown in Figure 13, Figure 13 is shown for pixel color value in Figure 12, obtained Fig. 3 institute Show the corresponding denoising image schematic diagram of original image.In Figure 13, the pixel coordinate (1,1) of oval 32 frames choosing, pixel coordinate (1,2), pixel coordinate (2,1), pixel coordinate (2,2) are to be directed in Figure 12, pixel coordinate (1,1), pixel coordinate (1,2), Pixel coordinate (2,1), the pixel color value of pixel coordinate (2,2), the result of rounding.

In Figure 13, the region of 33 frames of elliptical region choosing, i.e., for according to the third that in above-mentioned Fig. 8,17 frames are selected Bayer formatted data, in Fig. 9, four the 4th pixels of 19 frames choosing are up-sampled, 15 frames of big ellipse choosing in Fig. 5 Four Bayer format Bayer format original images, corresponding denoising image.In embodiments of the present invention, and so on, it is somebody's turn to do The denoising image of Bayer format original image.

In embodiments of the present invention, optionally, data can also be up-sampled to centre after obtaining intermediate up-sampling data In third Bayer formatted data in four the 5th pixels without normalization, directly will be in the third Bayer formatted data Four the 5th pixels are multiplied with the 4th pixel of corresponding position respectively, obtain a Bayer formatted data.If by third Bayer lattice Four the 5th pixels in formula data, are multiplied with the 4th pixel of corresponding position respectively, and product can multiply this if more than 255 Product is taken as being 255.If the same product is score or decimal, also the product is rounded, and then obtain the original image Corresponding denoising image is in embodiments of the present invention not especially limited this.

In embodiments of the present invention, above-mentioned first denoising image considers the pass in four channels of Bayer format original image Connection property and Bayer format original image luminance information, the second denoising image have examined the only of four channels of Bayer format original image Vertical property and colouring information, and then the second denoising for denoising image and the second down-sampled images by the first of the first down-sampled images Image, up-sampling, during obtaining the denoising image of Bayer format original image, has fully considered the pass between four channels The independence of the luminance information and four channels of connection property and Bayer format original image and the color of Bayer format original image Information, and then improve the denoising effect of Bayer format original image.

It should be noted that for simple description, therefore, it is stated as a series of action groups for embodiment of the method It closes, but those skilled in the art should understand that, the embodiment of the present application is not limited by the described action sequence, because according to According to the embodiment of the present application, some steps may be performed in other sequences or simultaneously.Secondly, those skilled in the art also should Know, the embodiments described in the specification are all preferred embodiments, and related movement might not all be the application reality It applies necessary to example.

Embodiment three

Referring to Fig.1 shown in 4, for the structural block diagram for the image denoising device 300 that the embodiment of the present invention three provides, above-mentioned image Denoising device 300 can specifically include:

Down sample module 301, for being carried out down to original image respectively according to the first predetermined manner and the second predetermined manner Sampling, obtains the first down-sampled images and the second down-sampled images；

Image collection module 302 is denoised, for carrying out to first down-sampled images and second down-sampled images It denoises respectively, obtains the first denoising image and the second denoising image；

Image generation module 303 is denoised, for carrying out based on the first denoising image and the second denoising image Sampling, obtains denoising image corresponding with the original image.

Optionally, referring to Fig.1 shown in 5, on the basis of above-mentioned Figure 14, the original image includes: the original of Bayer format Beginning image；The down sample module 301 includes:

First sub- down sample module 3011, in the original image of the Bayer format, by each first Bayer lattice Formula data down-sampling is second pixel, obtains the first down-sampled images；The first Bayer formatted data includes 2*2 matrix Four the first pixels of form；

Second sub- down sample module 3012, in the original image of the Bayer format, with every four the first Bayers Formatted data is a second Bayer formatted data as one group, and by every group of first Bayer formatted data down-sampling, obtains the Two down-sampled images；Wherein, the matrix that every group of first Bayer formatted data is 2*2.

Optionally, the described second sub- down sample module 3012 may include:

Grouped element 30121, in the original image, every four the first Bayer formatted datas to be divided into one Group；

Downsampling unit 30122, the first pixel for that will include in every group of first Bayer formatted data, according to each picture Color Channel belonging to element, down-sampling are four third pixels；

Second down-sampled images acquiring unit 30123, for forming the second Bayer lattice with four third pixels Formula data, to obtain second down-sampled images.

Optionally, the denoising image collection module 302 may include:

First denoising image acquisition unit 3021 obtains the first denoising figure for denoising to first down-sampled images Picture；

Second denoising image acquisition unit 3022, for denoising image based on described first, to the second down-sampling figure As being denoised, the second denoising image is obtained.

Optionally, the second denoising image acquisition unit 3022 may include:

Similarity determines subelement between first piece, similarity between first piece for determining the first denoising image；

Similarity determines subelement between second piece, for according to the first denoising image and second down-sampled images Corresponding relationship similarity between second piece of the second down-sampled images is determined based on similarity between described first piece；

Second denoising image obtains subelement, for similarity between being based on described second piece, to the second down-sampling figure Picture is denoised using non-local mean, obtains the second denoising image.

Optionally, the denoising image generation module 303 may include:

Centre up-sampling data capture unit 3031, for denoising the third Bayer format number in image with described second According to replacing four the 4th pixels in the first denoising image respectively, obtain and intermediate up-sample data；Four the 4th pictures Position of the element in the first denoising image is gone with the 5th pixel in the third Bayer formatted data described second The position made an uproar in image is corresponding；

Normalization unit 3032, for centre to be up-sampled in data, four the 5th of each third Bayer formatted data Pixel is normalized, and obtains four the 6th pixels；

Image generation unit 3033 is denoised, for by four the 6th pixels the 4th pixel phase with corresponding position respectively Multiply, obtain a 4th Bayer formatted data, to obtain denoising image corresponding with the original image.

Image denoising device provided in an embodiment of the present invention can be realized image denoising in the embodiment of the method for Fig. 1 to Figure 13 Each process that device is realized, to avoid repeating, which is not described herein again.

In this way, in embodiments of the present invention, according to the first predetermined manner and the second predetermined manner, respectively to original image into Row down-sampling obtains the first down-sampled images and the second down-sampled images；To under first down-sampled images and described second Sampled images are denoised respectively, obtain the first denoising image and the second denoising image；Based on the first denoising image and institute It states the second denoising image to be up-sampled, obtains denoising image corresponding with the original image.In the embodiment of the present invention, to original Beginning image uses two different predetermined manner down-samplings, then denoises to it respectively, obtains the first denoising image and second and go It makes an uproar image, is then based on two kinds of denoising picture up-samplings, obtains the corresponding denoising image of the original image, obtaining original image Denoising image during, fully considered Multiple factors, compared with the prior art only consider single factors for, improve The denoising effect of original image.

The hardware structural diagram of Figure 16 one of each embodiment mobile terminal to realize the present invention, the mobile terminal 700 include but is not limited to: radio frequency unit 701, network module 702, voice output unit 703, input unit 704, sensor 705, display unit 706, user input unit 707, interface unit 708, memory 709, processor 710 and power supply 711 etc. Component.It will be understood by those skilled in the art that mobile terminal structure shown in Figure 16 does not constitute the restriction to mobile terminal, Mobile terminal may include perhaps combining certain components or different component layouts than illustrating more or fewer components.? In the embodiment of the present invention, mobile terminal includes but is not limited to mobile phone, tablet computer, laptop, palm PC, vehicle-mounted end End, wearable device and pedometer etc..

Wherein, processor 710, for being carried out to original image respectively according to the first predetermined manner and the second predetermined manner Down-sampling obtains the first down-sampled images and the second down-sampled images；

The embodiment of the present invention carries out down-sampling to original image respectively according to the first predetermined manner and the second predetermined manner, Obtain the first down-sampled images and the second down-sampled images；To first down-sampled images and second down-sampled images into Row denoises respectively, obtains the first denoising image and the second denoising image；Based on the first denoising image and second denoising Image is up-sampled, and denoising image corresponding with the original image is obtained.In the embodiment of the present invention, original image is used Two different predetermined manner down-samplings, then denoise it respectively, obtain the first denoising image and the second denoising image, then Based on two kinds of denoising picture up-samplings, the corresponding denoising image of the original image is obtained, in the denoising image for obtaining original image During, it has fully considered Multiple factors, for only considering single factors compared with the prior art, has improved original image Denoise effect.

It should be understood that the embodiment of the present invention in, radio frequency unit 701 can be used for receiving and sending messages or communication process in, signal Send and receive, specifically, by from base station downlink data receive after, to processor 710 handle；In addition, by uplink Data are sent to base station.In general, radio frequency unit 701 includes but is not limited to antenna, at least one amplifier, transceiver, coupling Device, low-noise amplifier, duplexer etc..In addition, radio frequency unit 701 can also by wireless communication system and network and other set Standby communication.

Mobile terminal provides wireless broadband internet by network module 702 for user and accesses, and such as user is helped to receive It sends e-mails, browse webpage and access streaming video etc..

Voice output unit 703 can be received by radio frequency unit 701 or network module 702 or in memory 709 The voice data of storage is converted into voice signal and exports to be sound.Moreover, voice output unit 703 can also be provided and be moved The relevant sound output of the specific function that dynamic terminal 700 executes is (for example, call signal receives sound, message sink sound etc. Deng).Voice output unit 703 includes loudspeaker, buzzer and receiver etc..

Input unit 704 is for receiving sound or vision signal.Input unit 704 may include graphics processor (Graphics Processing Unit, GPU) 7041 and microphone 7042, graphics processor 7041 is in video acquisition mode Or the image data of the static images or video obtained in image capture mode by image capture apparatus (such as camera) carries out Reason.Treated, and picture frame may be displayed on display unit 706.Through graphics processor 7041, treated that picture frame can be deposited Storage is sent in memory 709 (or other storage mediums) or via radio frequency unit 701 or network module 702.Mike Wind 7042 can receive sound, and can be voice data by such acoustic processing.Treated voice data can be The format output that mobile communication base station can be sent to via radio frequency unit 701 is converted in the case where telephone calling model.

Mobile terminal 700 further includes at least one sensor 705, such as optical sensor, motion sensor and other biographies Sensor.Specifically, optical sensor includes ambient light sensor and proximity sensor, wherein ambient light sensor can be according to environment The light and shade of light adjusts the brightness of display panel 7061, and proximity sensor can close when mobile terminal 700 is moved in one's ear Display panel 7061 or backlight.As a kind of motion sensor, accelerometer sensor can detect (generally three in all directions Axis) acceleration size, can detect that size and the direction of gravity when static, it is (such as horizontal to can be used to identify mobile terminal posture Vertical screen switching, dependent game, magnetometer pose calibrating), Vibration identification correlation function (such as pedometer, tap) etc.；Sensor 705 can also include fingerprint sensor, pressure sensor, iris sensor, molecule sensor, gyroscope, barometer, humidity Meter, thermometer, infrared sensor etc., details are not described herein.

Display unit 706 is for showing information input by user or being supplied to the information of user.Display unit 706 can wrap Display panel 7061 is included, liquid crystal display (Liquid Crystal Display, LCD), Organic Light Emitting Diode can be used Forms such as (Organic Light-Emitting Diode, OLED) configure display panel 7061.

User input unit 707 can be used for receiving the number or character information of input, and generate the use with mobile terminal Family setting and the related key signals input of function control.Specifically, user input unit 707 include touch panel 7071 and Other input equipments 7072.Touch panel 7071, also referred to as touch screen collect the touch operation of user on it or nearby (for example user uses any suitable objects or attachment such as finger, stylus on touch panel 7071 or in touch panel 7071 Neighbouring operation).Touch panel 7071 may include both touch detecting apparatus and touch controller.Wherein, touch detection Device detects the touch orientation of user, and detects touch operation bring signal, transmits a signal to touch controller；Touch control Device processed receives touch information from touch detecting apparatus, and is converted into contact coordinate, then gives processor 710, receiving area It manages the order that device 710 is sent and is executed.Furthermore, it is possible to more using resistance-type, condenser type, infrared ray and surface acoustic wave etc. Seed type realizes touch panel 7071.In addition to touch panel 7071, user input unit 707 can also include other input equipments 7072.Specifically, other input equipments 7072 can include but is not limited to physical keyboard, function key (such as volume control button, Switch key etc.), trace ball, mouse, operating stick, details are not described herein.

Further, touch panel 7071 can be covered on display panel 7061, when touch panel 7071 is detected at it On or near touch operation after, send processor 710 to determine the type of touch event, be followed by subsequent processing device 710 according to touching The type for touching event provides corresponding visual output on display panel 7061.Although in Figure 16, touch panel 7071 and aobvious Show that panel 7061 is the function that outputs and inputs of realizing mobile terminal as two independent components, but in some embodiments In, can be integrated by touch panel 7071 and display panel 7061 and realize the function that outputs and inputs of mobile terminal, it is specific this Place is without limitation.

Interface unit 708 is the interface that external device (ED) is connect with mobile terminal 700.For example, external device (ED) may include having Line or wireless head-band earphone port, external power supply (or battery charger) port, wired or wireless data port, storage card end Mouth, port, the port voice input/output (I/O), video i/o port, earphone end for connecting the device with identification module Mouthful etc..Interface unit 708 can be used for receiving the input (for example, data information, electric power etc.) from external device (ED) and By one or more elements that the input received is transferred in mobile terminal 700 or can be used in 700 He of mobile terminal Data are transmitted between external device (ED).

Memory 709 can be used for storing software program and various data.Memory 709 can mainly include storing program area The storage data area and, wherein storing program area can (such as the sound of application program needed for storage program area, at least one function Sound playing function, image player function etc.) etc.；Storage data area can store according to mobile phone use created data (such as Voice data, phone directory etc.) etc..In addition, memory 709 may include high-speed random access memory, it can also include non-easy The property lost memory, a for example, at least disk memory, flush memory device or other volatile solid-state parts.

Processor 710 is the control centre of mobile terminal, utilizes each of various interfaces and the entire mobile terminal of connection A part by running or execute the software program or module that are stored in memory 709, and calls and is stored in memory Data in 709 execute the various functions and processing data of mobile terminal, to carry out integral monitoring to mobile terminal.Processing Device 710 may include one or more processing units；Preferably, processor 710 can integrate application processor and modulation /demodulation processing Device, wherein the main processing operation system of application processor, user interface and application program etc., modem processor is mainly located Reason wireless communication.It is understood that above-mentioned modem processor can not also be integrated into processor 110.

Mobile terminal 700 can also include the power supply 711 (such as battery) powered to all parts, it is preferred that power supply 711 Can be logically contiguous by power-supply management system and processor 710, to realize management charging by power-supply management system, put The functions such as electricity and power managed.

In addition, mobile terminal 700 includes some unshowned functional modules, details are not described herein.

Preferably, the embodiment of the present invention also provides a kind of mobile terminal, including processor 710, and memory 709 is stored in On memory 709 and the computer program that can run on above-mentioned processor 710, the computer program are executed by processor 710 Each process of the above-mentioned image de-noising method embodiment of Shi Shixian, and identical technical effect can be reached, to avoid repeating, here It repeats no more.

Based on the hardware configuration of above-mentioned mobile terminal, various embodiments of the present invention are described in detail below.

The embodiment of the present invention also provides a kind of computer readable storage medium, and meter is stored on computer readable storage medium Calculation machine program, the computer program realize each process of above-mentioned image de-noising method embodiment, and energy when being executed by processor Reach identical technical effect, to avoid repeating, which is not described herein again.Wherein, above-mentioned computer readable storage medium, such as only Read memory (Read-Only Memory, abbreviation ROM), random access memory (Random Access Memory, abbreviation RAM), magnetic or disk etc..

It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row His property includes, so that the process, method, article or the device that include a series of elements not only include those elements, and And further include other elements that are not explicitly listed, or further include for this process, method, article or device institute it is intrinsic Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do There is also other identical elements in the process, method of element, article or device.

Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art The part contributed out can be embodied in the form of software products, which is stored in a storage medium In (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal (can be mobile phone, computer, service Device, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.

The embodiment of the present invention is described with above attached drawing, but the invention is not limited to above-mentioned specific Embodiment, the above mentioned embodiment is only schematical, rather than restrictive, those skilled in the art Under the inspiration of the present invention, without breaking away from the scope protected by the purposes and claims of the present invention, it can also make very much Form, all of these belong to the protection of the present invention.

Claims

1. a kind of image de-noising method, which is characterized in that the described method includes:

According to the first predetermined manner and the second predetermined manner, down-sampling is carried out to original image respectively, obtains the first down-sampling figure Picture and the second down-sampled images；

First down-sampled images and second down-sampled images are denoised respectively, obtain the first denoising image and Two denoising images；

It is up-sampled, is obtained corresponding with the original image based on the first denoising image and the second denoising image Denoise image.

2. the method according to claim 1, wherein the original image includes: the original image of Bayer format； It is described according to the first predetermined manner and the second predetermined manner, respectively to original image carry out down-sampling, obtain the first down-sampling figure The step of picture and the second down-sampled images includes:

In the original image of the Bayer format, it is second pixel by each first Bayer formatted data down-sampling, obtains Obtain the first down-sampled images；The first Bayer formatted data includes four the first pixels of 2*2 matrix form；

In the original image of the Bayer format, using every four the first Bayer formatted datas as one group, and by every group first Bayer formatted data down-sampling is a second Bayer formatted data, obtains the second down-sampled images；Wherein, every group of first Bayer Formatted data is the matrix of 2*2.

3. according to the method described in claim 2, it is characterized in that, described in the original image of the Bayer format, with every Four the first Bayer formatted datas are a second Bayer format as one group, and by every group of first Bayer formatted data down-sampling Data, obtain the second down-sampled images the step of include:

In the original image, every four the first Bayer formatted datas are divided into one group；

The first pixel that will include in every group of first Bayer formatted data, according to Color Channel belonging to each pixel, down-sampling For four third pixels；

The second Bayer formatted data is formed with four third pixels, to obtain second down-sampled images.

4. the method according to claim 1, wherein described under first down-sampled images and described second Sampled images are denoised respectively, obtain the first denoising image and second denoising image the step of include:

First down-sampled images are denoised, the first denoising image is obtained；

Based on the first denoising image, second down-sampled images are denoised, obtain the second denoising image.

5. according to the method described in claim 2, it is characterized in that, described gone based on the first denoising image and described second The step of image of making an uproar is up-sampled, and denoising image corresponding with the original image is obtained include:

With the third Bayer formatted data in the second denoising image, four in the first denoising image the are replaced respectively Four pixels obtain intermediate up-sampling data；Position of four the 4th pixels in the first denoising image, with described the The 5th pixel in three Bayer formatted datas, the position in the second denoising image are corresponding；

By in the up-sampling data of centre, four the 5th pixels of each third Bayer formatted data are normalized, and obtain Four the 6th pixels；

Four the 6th pixels are multiplied with the 4th pixel of corresponding position respectively, obtain a 4th Bayer formatted data, To obtain denoising image corresponding with the original image.

6. a kind of image denoising device, which is characterized in that described device includes:

Down sample module, for carrying out down-sampling to original image respectively, obtaining according to the first predetermined manner and the second predetermined manner To the first down-sampled images and the second down-sampled images；

Image collection module is denoised, for being gone respectively to first down-sampled images and second down-sampled images It makes an uproar, obtains the first denoising image and the second denoising image；

Image generation module is denoised, for being up-sampled based on the first denoising image and the second denoising image, is obtained To denoising image corresponding with the original image.

7. device according to claim 6, which is characterized in that the original image includes: the original image of Bayer format； The down sample module includes:

First sub- down sample module will be under each first Bayer formatted data in the original image of the Bayer format It is sampled as second pixel, obtains the first down-sampled images；The first Bayer formatted data includes the four of 2*2 matrix form A first pixel；

Second sub- down sample module, in the original image of the Bayer format, with every four the first Bayer formatted datas It is a second Bayer formatted data as one group, and by every group of first Bayer formatted data down-sampling, obtains the second down-sampling Image；Wherein, the matrix that every group of first Bayer formatted data is 2*2.

8. device according to claim 7, which is characterized in that the second sub- down sample module includes:

Grouped element, in the original image, every four the first Bayer formatted datas to be divided into one group；

Downsampling unit, the first pixel for that will include in every group of first Bayer formatted data, according to belonging to each pixel Color Channel, down-sampling are four third pixels；

Second down-sampled images acquiring unit, for forming the second Bayer formatted data with four third pixels, with Obtain second down-sampled images.

9. a kind of mobile terminal, which is characterized in that including processor, memory and be stored on the memory and can be described The computer program run on processor is realized when the computer program is executed by the processor as in claim 1 to 5 The step of described in any item image de-noising methods.

10. a kind of computer readable storage medium, which is characterized in that store computer journey on the computer readable storage medium Sequence, the image de-noising method as described in any one of claims 1 to 5 is realized when the computer program is executed by processor Step.