CN101742086B - Method for eliminating image noise - Google Patents
Method for eliminating image noise Download PDFInfo
- Publication number
- CN101742086B CN101742086B CN 200810174849 CN200810174849A CN101742086B CN 101742086 B CN101742086 B CN 101742086B CN 200810174849 CN200810174849 CN 200810174849 CN 200810174849 A CN200810174849 A CN 200810174849A CN 101742086 B CN101742086 B CN 101742086B
- Authority
- CN
- China
- Prior art keywords
- image
- data
- noise
- low
- pass filtering
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Landscapes
- Image Processing (AREA)
- Color Image Communication Systems (AREA)
- Processing Of Color Television Signals (AREA)
- Picture Signal Circuits (AREA)
Abstract
The invention discloses a method for eliminating image noise and an image processing device. The method for eliminating the image noise comprises the following steps: receiving an image and performing a first stage processing on the image so as to acquire brightness information Y corresponding to a pixel array and color information Cb and/or Cr which are separated under a YCbCr coordinate space (domain); aiming at the brightness information Y, performing a second stage processing so as to at least reduce brightness noise; aiming at the color information Cb and/or Cr, performing a third stage processing so as to at least reduce color noise; and combining the brightness information Y with the color information Cb and/or Cr.
Description
Technical field
The present invention relates to a kind of image processing technique, particularly relate to a kind of image compression technology to monochrome information Y and a color information Cb and Cr difference filtering (filtering).
Background technology
Image processing technique for example requires more effectively treatment mechanism along with the development of digital camera in recent years.Ordinary consumer, need to be taken with low-light level for the requirement that has anti-hand to shake at the photochrome of digital camera, its also make have high sensitivity (High ISO) to eliminate the requirement of noise effect more and more higher.Generally speaking, lower sensitivity, its noise is less.Higher sensitivity, its noise are also just larger.The image noise that how to suppress producing due to high sensitivity is a problem.Wherein noise section, especially make us more being difficult to accepting with chromatic noise.
Generally speaking, be processed into the photo of a digital camera, the digitized video processor, generally speaking need to have several large significant element: sensing interface module (Sensor Interface Module), image path module (Image Pipeline Module), Zoom module (Scalar Module), and compression module (Jpeg Module).Image just can obtain general image compression file through the processing of one-level one-level.
Generally speaking, be the image of pel array by the image of sensor output, may be the pattern of the Bayer pattern (pattern) of RGGB or complementary color CMYG and so on.The image path module of rear end image processing can transfer to RGB transfer to again YCbCr through Zoom module dwindling or amplifies after do jpg file known to being compressed into generally to compression module.
Yet in traditional image processing, brightness noise and color noise are not done effective inhibition.
Summary of the invention
The invention provides a kind of image processing mechanism, under the coordinate space of YCbCr (domain), do the filtering processing for monochrome information Y and a color information CbCr respectively.
The present invention proposes a kind of method for eliminating image noise, comprise receiving an image and this image being carried out a phase I process, to obtain isolating a monochrome information Y and color information Cb and the Cr corresponding to a pel array under YCbCr coordinate space (domain).For this monochrome information Y, carry out a second stage and process, to reduce at least by a brightness noise.For this color information Cb and Cr, carry out a phase III and process, to reduce at least by a color noise.With this monochrome information Y and this color information Cb and Cr combination, wherein this second stage processes that to process with this phase III be all being undertaken by a low-pass filtering mode, and this low-pass filtering mode comprises:
Each serial data on a first direction of the row of this pel array and row is divided into first's data and second portion data, these first's data are these serial datas, these second portion data are these first's data relative shift n pixel on this first direction relatively, n 〉=1, wherein these second portion data n pixel of exceeding a borderline region of these first's data is set to a predetermined value;
These first's data and this second portion data are average, obtain a low-pass filtering image; And
According to the aforesaid step of this low-pass filtering image and this low-pass filtering mode, carry out pulling over (recurrence) processes, to reach a desired filtering exponent number.
According to an embodiment, in described method for eliminating image noise, also comprise each serial data on a second direction of the row of this pel array and row is divided into first's data and second portion data, these first's data are these serial datas, these second portion data are these first's data relative shift m pixel on this second direction relatively, m 〉=1, wherein these second portion data m pixel of exceeding a borderline region of these first's data is set to a predetermined value;
These first's data and this second portion data are average, obtain a low-pass filtering image; And
According to the aforesaid step of this low-pass filtering image and this low-pass filtering mode, carry out the processing of pulling over, reaching this desired filtering exponent number, and realize that an images filter of two dimension processes.
According to an embodiment, in described method for eliminating image noise, this desired filtering exponent number can be selected by the outside.
According to an embodiment, in described method for eliminating image noise, comprise that also an edge judgment mechanism determines a borderline region, and for this desired filtering exponent number of this borderline region adjustment.
According to an embodiment, in described method for eliminating image noise, can be adjusted by the outside at this desired filtering exponent number of borderline region.
According to an embodiment, in described method for eliminating image noise, the low-pass filtering image after processing through this low-pass filtering mode comprises that also a high-pass filtering image of this low-pass filtering image and this image is done a weight again to be processed.
For above and other purpose of the present invention, feature and advantage can be become apparent, preferred embodiment cited below particularly, and be described with reference to the accompanying drawings as follows.
Description of drawings
Fig. 1 shows according to the embodiment of the present invention, the method schematic diagram of image processing.
Fig. 2 shows the distribution of Y, Cb, Cr respective pixel array.
Fig. 3 shows the leg-of-mutton Relationship of Coefficients of Pascal.
Fig. 4 shows according to the embodiment of the present invention, and Pascal triangle low pass filter is at the filtering mechanism schematic diagram of x direction.
Fig. 5 shows according to the embodiment of the present invention, and Pascal triangle low pass filter is at the filtering mechanism schematic diagram of y direction.
The reference numeral explanation
100～116: step
The 120:Y array
The 122:Cb array
The 124:Cr array
140,240: the original pixels array
142,242: the first part data
144,242: the second part data
146,246: the first rank filtering images
148,248: the first part data
150,250: the second part data
152,252: second-order filtering image
Embodiment
The invention provides a kind of image processing mechanism, under the coordinate space of YCbCr (domain), the utilization processor accelerator that counts is realized the adjustable filter of a kind of exponent number, and utilize different rank correspond to different frequencies realize on the YCbCr space be adjustable exponent number and be the low passband ripple device that the edge judgement is arranged, can effectively reduce the brightness chromatic noise of image for different frequencies like this, and strengthen image edge by the distortion computing simultaneously.
Basically, the present invention is divided into Y brightness and C chroma with image.Generally speaking, such strong of the variation that the variation of chroma can't image brightness that is to say that the spatial frequency of chroma is much lower compared to the spatial frequency of brightness.The present invention utilizes so image characteristics, uses respectively different rank for brightness and chroma, proposes to have more the low pass filter of object edge judgement.Chroma partly can be the higher low pass filter of exponent number, and brightness can be the lower filter of exponent number, improves brightness and the chroma noise of image, to promote the video vision effect, can keep the demonstration at object edge, can be by excessive obfuscation.
Fig. 1 shows according to the embodiment of the present invention, the method schematic diagram of image processing.Consult Fig. 1, the method for image processing of the present invention can be arranged on an image processor.Image processor is for example digital camera, and device for image or the treatment system of computer etc. are by hardware or the processing of software, to reach the image of better quality.In step 100, for example obtain the original file data of image by the image induction module of camera.Step 102 carries out the image of obtaining is done the processing of phase I, for example comprises that white balance, pixel interpolating, noise are eliminated, gamma (Gamma) colour correction, size scaling etc.Step 104 carries out obtaining the Y monochrome information of respective pixel array in step 106 in the Y brightness in YCbCr space and the separation of C chroma, obtains C chroma information in step 108.Fig. 2 shows the distribution of Y, Cb, Cr respective pixel array.Y array 120 is luminance pixel values.Cb array 122 is Cb pixel values.Cr array 124 is Cr pixel values.Again or Cb, Cr also can synthesize CbCr.Each pel array needs corresponding filtering to process.
In step 110, do at least the inhibition of noise for the Y monochrome information.For example can coordinate again the image display of considering the object edge, carry out simultaneously the boostfiltering (Edge Enhancement Filter) at edge.In step 112, do at least the inhibition of noise for the CbCr color information.Also for example can coordinate the image display of considering the object edge, carry out simultaneously the boostfiltering at edge.After can being described in than detailed mechanism of step 110 and step 112.
Generally for example carry out compression step 114 after recombinant when the Y monochrome information of finishing dealing with respectively and CbCr color information, be compressed into the image file of general Jpeg/Jpg.After completing, compression just can export 116.Compression step 114 is only general image processing step.Yet the present invention has done noise effective filtering with image, to promote image quality.
The Y monochrome information is different from the image effect of CbCr color information.For example in conventional method, the color information at object edge is easily fallen by fuzzy.The present invention is in step 110 and step 112, and proposition can simple realization, and the filtering exponent number is adjustable filter, can be described as Pascal triangle low pass filter.
So-called Pascal triangle, its relation as shown in Figure 3.The leg-of-mutton relation of Pascal is the mathematics general knowledge known to general, can rule extrapolate the coefficient of every single order.The low pass filter that one embodiment of the invention propose, the structure of its filtered version along with the increase of exponent number also can change just like the leg-of-mutton rule of Pascal, therefore be called Pascal triangle low pass filter.It is adjustable also therefore allowing the filtering exponent number, and it can have initial set value (default value), but also can be adjusted from the outside by the user, for user's selection, to meet desired image effect.
Below describe the adjustable Pascal triangle low pass filter of exponent number for an embodiment, but it not unique selection of the present invention.Fig. 4 shows according to the embodiment of the present invention, and Pascal triangle low pass filter is at the filtering mechanism schematic diagram of x direction.Consult Fig. 4, represent the position of parameter x and y respective pixel for the pel array 140 of an image with A (x, y).The pixel value of one pel array of A (x, y) typical example such as PxQ resolution.
The image of A (x, y) need to be eliminated noise.Therefore, the filtering for inferior single order is namely that { exponent number of 11} can be used as first part data A (x, y) 142 with original A (x, y) data 140.First just the positive direction of x is carried out the situation that noise eliminates and describe, yet if the negative direction of x is carried out the noise elimination, its mechanism is still similar.In addition, the resolution of x direction is take 0 to p-1 p pixel as example.One second portion image B (x+n, y) the 144th, take from original A (x, y) data, but belong to the data after n pixel of displacement, be namely n pixel to the data of p-1 pixel, as the 0th pixel of second portion image B to p-1-n pixel.And because the resolution of x direction has p pixel, so p-n the pixel of second portion image B be a borderline region (boundary region) to p-1 pixel, can insert a set point.Borderline region for example can be inserted all set points of formed objects, and it more for example gets the pixel value of last pixel of A (x, y) data.
Then, for for the filtering on 11} rank, with first part data A with second partly data B do and on average obtain at { first rank filtering image C 146, the namely C=(A+B)/2 on 11} rank.Then, (recur) the aforesaid mode of pulling over can obtain time single order { filtering of 121}.With image C 146 beginnings that the first rank filtering produces, tell once again first part data C (x, y) 148 and the second part data C (x+n, y) 150 according to principle of identity.{ image on 121} rank is namely second-order filtering image 152 can to obtain the after then average.The filter of other exponent numbers can be pulled over desired number of times and be obtained by same way as.
Above-mentioned is again the filtering mode of the one dimension of x direction, and identical mode can be done filtering in the y direction.Fig. 5 shows according to the embodiment of the present invention, and Pascal triangle low pass filter is at the filtering mechanism schematic diagram of y direction.Original A (x, y) data 240 are used as first part data A (x, y) 242.Carry out the situation of noise elimination for the positive direction of y direction and describe, yet if the negative direction of y is carried out noise to be eliminated, its mechanism is still similar.In addition, the resolution of y direction is take 0 to q-1 q pixel as example.One second portion image B (x, y+m) the 244th, take from original A (x, y) data, but belong to the data after m pixel of displacement, be namely m pixel to the data of q-1 pixel, as the 0th pixel of second portion image B to q-1-m pixel.And because the resolution of y direction has q pixel, so q-m the pixel of second portion image B be the borderline region on the y direction to q-1 pixel, can insert a set point as the mode of x direction.
Then, for for the filtering on 11} rank, with first part data A 242 with second partly data B 244 do and on average obtain at { first rank filtering image C 246, the namely C=(A+B)/2 on 11} rank.Similarly, can obtain image C (x, y) 248, C (x, y+m) 250 by the aforesaid mode of pulling over, and then { the filtering image of 121} is second-order filtering image 252 to obtain time single order.The filter of other exponent numbers can be pulled over desired number of times and be obtained by same way as.
Above-mentioned is the filtering mode of one dimension.With regard to filter, it for example also can be simplified to processes a string pixel data, is for example the data of a pixel column (pixel column) or a pixel column (pixel row).
Yet be the pel array of two dimension due to image, image generally can need the filter effect of two dimension.The filter effect of two dimension can according to above-mentioned mode, after for example completing the filtering on x direction or a direction of y direction, then be done filtering to other direction.Again for example, the method for two-dimensional filtering can first after a direction is completed the filtering of the exponent number of wanting, just be carried out the filtering of second direction at other direction again.Yet this method is only the one of multiple choices mode.In addition for example, the mode of two-dimensional filtering also can in a direction after pull over each time or several times, just be changed another direction and do the filtering action.
Two-dimensional filtering be coefficient, take 121} rank are as example, and the distribution of its two-dimentional coefficient is as follows:
121
242
121。
Again take 1331} rank are as example, and the distribution of two-dimentional coefficient is as follows:
1331
3993
3993
1331。
Other exponent numbers also can obtain according to same way as.In addition, the filtering exponent number of x direction and y direction can be identical or different.
In addition, by the judgment mechanism at object edge (object edge), for example the image of this block can be done the filtering of other exponent numbers.In other words, for example the different blocks of a whole image, can do separately respectively the filtering of suitable exponent number and process, and need not be all the filtering of identical exponent number.Also for example, after a whole image is first done the filtering of identical exponent number, then for the object marginal block that need consider, then do further filtering and process.In other words, according to identical Pascal triangle filtering mechanism, it is applied to the filtering arrangement of variety of way.
On hardware is realized, because of hardware item framework and cost consideration, normally limited for the exponent number of finite length frequency response (Finite Impulse Response, FIR).The present invention proposes the implementation method of the exponent number that can infinitely increase.The method image arithmetic accelerator of for example arranging in pairs or groups.This image arithmetic accelerator can be to two image source A, and B does arithmetic operator, and result is deposited to C again.C=A/2+B/2 for example.Utilize B (x, y)=A (x+n, y) or B (x, y)=A (x, y+m) to come to do computing with A (x, y), reach can the capable of regulating exponent number level and the low passband ripple device of vertical direction.Parameter n can equate with m or be unequal.N=1 for example, 2 ...; M=1,2 ....
How below lift a computing example describes mat and has and on average obtain desired filtering exponent number with the mode of pulling over.A (x, y) refers at x, the image block that the y coordinate begins, take the block of PxP as example, and get Fig. 4 described in, the pixel that is shifted is namely the situation of n=1.x＝0～P-1，y＝0～P-1。
Select A1 (x, y)=A (x, y) as first's data, select in addition B 1 (x, y)=A (x+1, y).Then do the computing of C=(A1+B1)/2.So, C1 (x, y)=(A (x, y)+A (x+1, y))/2 so can obtain { the Gaussian blur filter on 11} rank.
Then with C1 (x, y) beginning, repeat above-mentioned computing, get B2 (x, y)=C1 (x+1, y), A2 (x, y)=C 1 (x, y).B2 (x, y) and A2 (x, y) are done on average,
C2 (x, y)=(A2+B2)/2=(C 1 (x, y)+C 1 (x+1))/2=(A (x, y)+2*A (x+1, y)+A (x+2, y))/4, its coefficient are the corresponding { filters on 121} rank.
So again with C2 (x, y) beginning, repeat above-mentioned computing and can obtain { the filter on 1331} rank.Once just can obtain { the filter on 14641} rank if pull over again.In other words, the filter of hardware needn't essence increase, and just can reach the filter of arbitrary exponent number.Therefore the number of times of pulling over can adjust at any time.
For example, just with whole algorithm flow process, can define the mathematics formula as follows:
I[n] be the input image, can be Y brightness or chroma Cb or chroma Cr.
I_Mod1[n] be the input image of revising.
I_Mod2[n] be the input image of revising.
I_Mod3[n] be the input image of revising.
LPF[I[n]] be Gaussian Blur low pass filter A, can utilize exponent number to adjust the selection different frequency.
G_ForEdge[n] be Gaussian Blur low pass filter B, can utilize exponent number to adjust the selection different frequency.
EdgeMap{.} is that parameter is selected in the outside of input, to remove unnecessary noise jamming.
EhnEdgeMap{.} is another outside parameter of selecting of input, to remove unnecessary noise jamming.
G_ForEdgeEhn[n] be Gaussian Blur low pass filter C, can utilize exponent number to adjust the selection different frequency.
De_Edge[n] be that the input image edge detects output, and be standardized between 0～1.This high frequency filter can by adjusting the low pass filter of selecting different frequency, reach the edge that detects different frequency.
EE_Edge[n] be the part that the edge strengthens.This high frequency filter can by the low pass filter of adjusting different frequency, reach the edge that detects different frequency.
O[n] be image output, can be the output respectively of Y brightness or chroma Cb or chroma Cr.
Det_Edge[n]
＝EdgeMap{abs(I_Mod1[n]-G_ForEdge[n])}。
EE_Edge[n]
＝EhnEdgeMap(I_Mod2[n]-G_ForEdgeEhn[n])。
O[n]＝(Det_Edge0_2[n]*(I_Mod3[n]+EE_Edge[n]))
+((1-Det_Edge[n])*LPF[I[n])。
This computing is to adjust output by the mode of weight (weighting).Det_Edge0_2[n] * (I_Mod3[n]+EE_Edge[n])) be the part that former image adds enhancing.((1-Det_Edge[n]) * LPF[I[n]) be the part that filters noise.Utilize such mathematics formula, for example can reach simultaneously, eliminate the noise of brightness and chroma different frequency, and strengthen image different frequency edge strength.Yet so computing is only the one that the present invention uses.
The present invention preferably utilizes the Gaussian blur filter of different rank, correspond to different frequencies, also just can produce the low pass filter of corresponding different frequency, again by the low pass filter of different frequency, utilize some simple conversions, HPF{X[n for example] }=X[n]-LPF1{X[n], the high frequency filter of corresponding different frequency just can be obtained.By the adjustment of different frequency, can make us eliminate in (Noise Reduction) control more flexible by noise.
The method that the present invention proposes, for example can realize that (implement) is in having the device of image processing, for example be implemented in digital camera or camera, can effectively eliminate different frequency brightness on image with the noise of chroma, especially in color noise part, except eliminating noise, also can also can strengthen the algorithm of edge strength simultaneously.The device of image processing can be for example also the computer system that need to do image processing in addition.
Although the present invention discloses as above with preferred embodiment; so it is not to limit the present invention; those skilled in the art can do some changes and retouching under the premise without departing from the spirit and scope of the present invention, so protection scope of the present invention is as the criterion with claim of the present invention.
Claims (9)
1. method for eliminating image noise comprises:
Receive an image;
This image is carried out a phase I process, to obtain isolating a monochrome information Y and color information Cb and the Cr corresponding to a pel array under YCbCr coordinate space;
For this monochrome information Y, carry out a second stage and process, to reduce at least by a brightness noise;
For this color information Cb and Cr, carry out a phase III and process, to reduce at least by a color noise; And
With this monochrome information Y and this color information Cb and Cr combination,
Wherein this second stage processes that to process with this phase III be all being undertaken by a low-pass filtering mode, and wherein this low-pass filtering mode comprises:
Each serial data on a first direction of the row of this pel array and row is divided into first's data and second portion data, these first's data are these serial datas, these second portion data are these first's data relative shift n pixel on this first direction relatively, n 〉=1, wherein these second portion data n pixel of exceeding a borderline region of these first's data is set to a predetermined value;
These first's data and this second portion data are average, obtain a low-pass filtering image; And
According to the aforesaid step of this low-pass filtering image and this low-pass filtering mode, carry out the processing of pulling over, to reach a desired filtering exponent number.
2. method for eliminating image noise as claimed in claim 1, also comprise each serial data on a second direction of the row of this pel array and row is divided into first's data and second portion data, these first's data are these serial datas, these second portion data are these first's data relative shift m pixel on this second direction relatively, m 〉=1, wherein these second portion data m pixel of exceeding a borderline region of these first's data is set to a predetermined value;
These first's data and this second portion data are average, obtain a low-pass filtering image; And
According to the aforesaid step of this low-pass filtering image and this low-pass filtering mode, carry out the processing of pulling over, reaching this desired filtering exponent number, and realize that an images filter of two dimension processes.
3. method for eliminating image noise as claimed in claim 2, n=1 wherein, m=1.
4. method for eliminating image noise as claimed in claim 2, wherein n=m.
5. method for eliminating image noise as claimed in claim 1, wherein n=1.
6. method for eliminating image noise as claimed in claim 1, wherein this desired filtering exponent number can be selected by the outside.
7. method for eliminating image noise as claimed in claim 1, comprise that also an edge judgment mechanism determines a borderline region, and for this desired filtering exponent number of this borderline region adjustment.
8. method for eliminating image noise as claimed in claim 7, wherein can be adjusted by the outside at this desired filtering exponent number of borderline region.
9. method for eliminating image noise as claimed in claim 1, wherein pass through the low-pass filtering image after this low-pass filtering mode is processed, and comprises that also a high-pass filtering image of this low-pass filtering image and this image is done a weight again to be processed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810174849 CN101742086B (en) | 2008-11-07 | 2008-11-07 | Method for eliminating image noise |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200810174849 CN101742086B (en) | 2008-11-07 | 2008-11-07 | Method for eliminating image noise |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101742086A CN101742086A (en) | 2010-06-16 |
| CN101742086B true CN101742086B (en) | 2013-05-15 |
Family
ID=42464919
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200810174849 Expired - Fee Related CN101742086B (en) | 2008-11-07 | 2008-11-07 | Method for eliminating image noise |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101742086B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TWI571856B (en) * | 2016-01-08 | 2017-02-21 | 友達光電股份有限公司 | Display method |
| CN108108733A (en) * | 2017-12-19 | 2018-06-01 | 北京奇艺世纪科技有限公司 | A kind of news caption detection method and device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1183696A (en) * | 1996-09-30 | 1998-06-03 | 三星电子株式会社 | Image quality enhancement circuit and method therefor |
| CN1335579A (en) * | 2000-07-27 | 2002-02-13 | 诺日士钢机株式会社 | Image processing method, apparatus and recording medium for recording & executing the same method program |
| US20070097391A1 (en) * | 2005-11-01 | 2007-05-03 | Hwang Sung-Hyun | Apparatus and method for improving image quality of image sensor |
| CN101009771A (en) * | 2006-01-23 | 2007-08-01 | 株式会社理光 | Image processing apparatus, imaging apparatus, image processing method, and computer program product |
| CN101119448A (en) * | 2006-07-31 | 2008-02-06 | 株式会社理光 | Image processing apparatus, imaging apparatus, image processing method, and computer program product |
-
2008
- 2008-11-07 CN CN 200810174849 patent/CN101742086B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1183696A (en) * | 1996-09-30 | 1998-06-03 | 三星电子株式会社 | Image quality enhancement circuit and method therefor |
| CN1335579A (en) * | 2000-07-27 | 2002-02-13 | 诺日士钢机株式会社 | Image processing method, apparatus and recording medium for recording & executing the same method program |
| US20070097391A1 (en) * | 2005-11-01 | 2007-05-03 | Hwang Sung-Hyun | Apparatus and method for improving image quality of image sensor |
| CN101009771A (en) * | 2006-01-23 | 2007-08-01 | 株式会社理光 | Image processing apparatus, imaging apparatus, image processing method, and computer program product |
| CN101119448A (en) * | 2006-07-31 | 2008-02-06 | 株式会社理光 | Image processing apparatus, imaging apparatus, image processing method, and computer program product |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101742086A (en) | 2010-06-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI387312B (en) | Image noise reduction method and processor | |
| US11625815B2 (en) | Image processor and method | |
| US8948506B2 (en) | Image processing device, image processing method, and program | |
| EP2055093B1 (en) | Adaptive spatial image filter for filtering image information | |
| CN103854259B (en) | Image processing equipment and the method for handling image | |
| JP4526445B2 (en) | Imaging device | |
| JP4395789B2 (en) | Image processing apparatus, imaging apparatus, image processing method, and program | |
| EP2162863B1 (en) | Non-linear transformations for enhancement of images | |
| JP2002077645A (en) | Image processor | |
| CN110730336B (en) | Demosaicing method and device | |
| WO2007139343A1 (en) | Color interpolation method and device | |
| JP2014110633A (en) | Apparatus and method for processing image | |
| JP2005520442A (en) | Method and apparatus for processing sensor images | |
| WO2004110071A1 (en) | Image processor and image processing program | |
| JP2002077928A (en) | Image processor | |
| CN110517206B (en) | Method and device for eliminating color moire | |
| JP5076186B2 (en) | Image enlargement method | |
| CN108122201A (en) | A kind of Bayer interpolation slide fastener effect minimizing technology | |
| US7609300B2 (en) | Method and system of eliminating color noises caused by an interpolation | |
| KR101136688B1 (en) | Single-pass image resampling system and method with anisotropic filtering | |
| JPWO2005013622A1 (en) | Image processing apparatus, image processing program, electronic camera, and image processing method for processing image in which color components are mixed and arranged | |
| US7269295B2 (en) | Digital image processing methods, digital image devices, and articles of manufacture | |
| CN1965570A (en) | Image processing device, method, and program | |
| US20050025379A1 (en) | Digital imaging systems, articles of manufacture, and digital image processing methods | |
| CN101742086B (en) | Method for eliminating image noise |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130515 Termination date: 20141107 |
|
| EXPY | Termination of patent right or utility model |