CN102447817B - Image processing device and space image noise eliminating method - Google Patents
Image processing device and space image noise eliminating method Download PDFInfo
- Publication number
- CN102447817B CN102447817B CN201010507037.5A CN201010507037A CN102447817B CN 102447817 B CN102447817 B CN 102447817B CN 201010507037 A CN201010507037 A CN 201010507037A CN 102447817 B CN102447817 B CN 102447817B
- Authority
- CN
- China
- Prior art keywords
- pixel
- noise
- value
- central point
- variation value
- 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.)
- Active
Links
Landscapes
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
- Picture Signal Circuits (AREA)
Abstract
The embodiment of the invention discloses an image processing device and a space image noise eliminating method. The image processing device comprises an estimator, a noise detector and a space noise filter, wherein the estimator is used for estimating the local variance of each pixel of an input image signal and generating a noise critical value according the status of the local mutation value; the noise detector is used for determining which pixels in the pixels are noises or images according to the noise critical value; and the space noise filter is used for filtering the pixels which are noises and generating an output image signal.
Description
Technical field
The present invention relates to a kind of image processing apparatus and method, particularly a kind of image processing apparatus and the method that can eliminate noise and can retain image detail.
Background technology
Fig. 1 shows the schematic diagram of conventional images processing unit.General image processing apparatus 10 utilizes a low pass filter to eliminate noise.After this low pass filter reception received image signal Vi, by the noise filtering of high frequency, only retain the signal of low frequency, to produce an output image signal Vo.
Yet on spatial image noise is eliminated, the signal processing mode of low pass filter also can thicken image except eliminating noise simultaneously, and prior art cannot reach the effect of eliminating noise simultaneously and retaining image detail, so image is second-rate.
Summary of the invention
One object of the present invention, is to provide a kind of image processing apparatus and space image noise eliminating method, and it can be eliminated the picture noise on space and retain image detail.
One embodiment of the invention provide a kind of image processing apparatus, comprise an estimator, a noise detector and a spatial noise filter.This estimator is in order to estimate the part variation value of each pixel of received image signal, and produces a noise critical value according to the state of those local variation values.Noise detector determines according to noise critical value in those pixels, which pixel is noise or is image.And those pixels that are noise of spatial noise filter filtering, to produce an output image signal.
Another embodiment of the present invention provides a kind of image processing apparatus, comprises an estimator, a noise detector and a spatial noise filter.This estimator receives a received image signal, and received image signal comprises a plurality of pixels.Estimator is got to each this pixel a plurality of pixels that this pixel is adjacent and is calculated the part variation value of this pixel, and determines a noise critical value according to the state of those local variation values.Noise detector receives received image signal, and according to the comparative result of the part variation value of noise critical value and each pixel, the flag value of each pixel of mark, to produce a flag distribution signal.And spatial noise filter receives received image signal and flag distribution signal, according to flag distribution signal, set the weight of each pixel of received image signal, to produce an output image signal.
Another embodiment of the present invention provides a kind of image processing apparatus, comprises an estimator, a noise detector and a spatial noise filter.This estimator receives a received image signal, and received image signal comprises a plurality of pixels.Estimator is divided into a plurality of first areas by the pixel region of one first preset number, calculates respectively the part variation value of each pixel of each first area, and determines a noise critical value according to the state of those local variation values.Noise detector receives received image signal, the pixel region of one second preset number is divided into a plurality of second areas, calculate respectively the part variation value of each pixel of each second area, comparative result according to the part variation value of noise critical value and each pixel, each pixel is determined whether to mark flag, to produce a flag distribution signal.Spatial noise filter receives received image signal and flag distribution signal, sets respectively the weight of each pixel of each second area, to produce an output image signal according to flag distribution signal.
One embodiment of the invention provide a kind of space image noise eliminating method, comprise the following steps: first, estimate the part variation value of each this pixel of received image signal, and produce a noise critical value according to the state of those local variation values.Then, according to noise critical value, determine in those pixels, which pixel is noise or is image.Afterwards, those pixels that are noise of filtering, produce an output image signal.
Image processing apparatus of the present invention and space image noise eliminating method utilize each pixel to be adjacent the part variation value of pixel, understand the degree of variation between whole image pixel value, recycling this degree of variation does to judge to image, obtain may for the pixel on edge or border be the pixel of noise, correspondingly adjust again the weight of pixel with edge or the border of filtering noise, strengthening image, and then the problem of solution prior art, reach filtering image noise and the effect that retains image detail simultaneously.
Accompanying drawing explanation
Fig. 1 shows the schematic diagram of the image processing apparatus of prior art;
Fig. 2 A shows the schematic diagram of the image processing apparatus of one embodiment of the invention;
The image processing apparatus that Fig. 2 B shows Fig. 2 A is the schematic diagram of a treatment step wherein;
The image processing apparatus that 2C illustrates Fig. 2 A is the schematic diagram of a treatment step wherein; And
Fig. 3 shows the flow chart of the space image noise eliminating method of one embodiment of the invention.
Main element symbol description
10,20 image processing apparatus
21 estimators
22 noise detectors
23 spatial noise filters
Embodiment
Fig. 2 A shows the schematic diagram of the image processing apparatus of one embodiment of the invention.This image processing apparatus 20 is applicable to the noise of spatial image and eliminates, when noise is eliminated, and the details of also possessing image.
Image processing apparatus 20 comprises an estimator (Estimator) 21, a noise detector (Noise detector) 22, one spatial noise filter (Spatial De-noising Filter) 23.
This estimator 21 can be a local variation value estimator (Local VarianceEstimator).Estimator 21 for example, in order to estimate the part variation value (Local variance) of each pixel of received image signal Vi (whole picture frame (frame)), and the state computation of the local variation value of foundation goes out a noise critical value Th.In one embodiment, the part of received image signal Vi variation state of value can reference section or all the local variation value of maximum in pixel and/or partly or entirely the average local variation of pixel be worth to learn.
One embodiment, as shown in Figure 2, the mode that estimator 11 calculates local variation value is in picture signal Vi, for each pixel, gets the part variation value that its neighbor calculates this pixel.The sum of this pixel and those neighbors can be n (n is that positive integer and n are greater than 2).Nine pixels (n=9) of for example take in figure are carried out the part variation value of calculating pixel P5 as a region a.Must note, local variation value can be used to represent the degree of variation of pixel in this region a centered by pixel P5.
Estimator 21 treatment steps (a): estimator 21 calculates local variation value and obtains the coherent reference value that can show local variation state of value:
The pixel value of nine the pixel P1～P9 of region a in this figure is added, to obtain a summation sum=P1+P2+P3+P4+P5+P6+P7+P8+P9.
Nine multiple value of calculating each pixel deduct summation sum and take absolute value, to obtain nine pixel absolute differences:
v_tr＝abs(9*P1-sum)；
v_tm＝abs(9*P2-sum)；
v_tl＝abs(9*P3-sum)；
v_mr＝abs(9*P4-sum)；
v_mm＝abs(9*P5-sum)；
v_ml＝abs(9*P6-sum)；
v_br＝abs(9*P7-sum)；
v_bm＝abs(9*P8-sum)；
v_bl＝abs(9*P9-sum)；
Then, each pixel absolute difference is added, to obtain the part variation value var of the region a centered by pixel P5, var=v_tr+v_tm+v_tl+v_mr+v_mm+v_ml+v_br+v_bm+v_bl.
According to aforesaid way, calculate the part variation value var of each pixel in whole figure of received image signal Vi afterwards.
Obtain the coherent reference value of the local variation of performance state of value: an embodiment, in obtaining received image signal Vi after the part variation value of each pixel, in the part variation value of whole pixels, select maximum local variation value, and whole part variation Zhi Jia General Logistics Departments is on average obtained to average local variation value.
Estimator 21 treatment steps (b): estimator 21 produces noise critical value (NoiseThreshold) NTh, and noise critical value NTh can be produced by one of following two kinds of modes.
One, can insert default numerical value by a default software; Or
Two, can be according to trying to achieve after the local variation value of above-mentioned maximum and average variation value.
Above-mentioned two kinds of account forms of trying to achieve maximum local variation value and average variation value are as follows:
NTh＝CP_En_Sw_Noise_Thl？CP_Noise_Thl
：((CP_Max_Local_Var/CP_Mean_Local_Var ＜＝CP_Var_Ratio_Thl)
？(CP_Max_Local_Var+CP_Mean_Local_Var)/2
： 2*CP_Mean_Local_Var)；
...(1)
Wherein, NThl represents that default value, CP_Noise_Thl that noise critical value, CP_En_Sw_Noise_Thl represent above-mentioned first kind of way and inserted by software represent that the above-mentioned second way is via the preset value of obtaining after judgement relatively.
And the relatively judgement of the second way is that maximum variation value CP_Max_Local_Var is compared with a default temporary value CP_Var_Ratio_Thl divided by average resulting the first ratio of variation value CP_Mean_Local_Var (ratio).
When if the first ratio is less than or equal to this default temporary value CP_Var_RatioT_hl, therefore the picture numerical value that represents each neighbor is more approaching, will select mean value (i.e. two numerical value additions remove 2) that maximum variation value (CP_Max_Local_Var) adds average variation value (CP_Mean_Local_Var) as this noise critical value NTh.
On the contrary, if when the first ratio is more than or equal to this default temporary value CP_Var_Ratio_Thl, represent that the numerical value variation of each neighbor is larger, therefore using the twice of the average variation value CP_Mean_Local_Var of selection as this noise critical value NTh.
According to above-mentioned judgment mode, why can obtain the rank, position (Level) of current noise.The edge (edge) that common maximum variation value can be image frame, therefore more approaching when average variation value and maximum variation value, represent that the numerical value of noise and pixel is more approaching, so that the less noise critical value NTh of needs setting.For example, using maximum variation value with the mean value of average variation value as this noise critical value NTh; And differ larger when average variation value and maximum variation value, and represent that the numerical value of noise and pixel differs far away, need the noise critical value NTh that setting is larger.For example, using the multiple of the average value of variation as noise critical value.When the pixel value below noise critical value is noise; Above pixel value is edge (edge) or border (boundary) of image.
Please also refer to Fig. 2 A, Fig. 2 C, the noise critical value NTh that noise detector (Noise detector) 22 reception received image signal Vi and estimator 21 calculate, and for each pixel, get the part variation value that its neighbor calculates this pixel, according to noise critical value NTh and this part variation value, determine the flag value of each pixel in picture signal Vi or determine which element marking flag value which is not marked.Wherein, the sum of this pixel and those neighbors can be n (n is that positive integer and n are greater than 2).27 pixels (n=27) of for example take in Fig. 2 C are carried out the part variation value of calculating pixel P25 as a region b.The account form of local variation value can be same as the mode of above-mentioned estimator 21, therefore repeats no more its details.
The effect of pixel flag value be for mark pixel with pixel value whether authentic representative edge (edge) or border (boundary), rather than noise.Do not have flag target pixel or mark flag value be zero or the pixel value of a preset value can be considered noise.When the flag of marked pixels, noise detector 22 is got the image of a predeterminable area in the picture of received image signal Vi, and the area size that generally this region can adopt the local variation value of noise detector 22 calculating to adopt is processed.
One embodiment, supposes that each pixel can get the scope of a 3*9, as the region b in figure, take scheme in pixel P25 be example, can capture element P11～P39 as a region.Next the flag of marked region b pixel:
Flag marker step (a): calculate preposition pixel flag pre_pixel_flag
With central point x (pixel P25), according to following formula, calculate the preposition pixel flag pre_pixel_flag of pixel P25.
if(x＜x_center)
Pre_Pixel_Flag(x)＝var(x)＞NTh&&var(x-1)＞NThl；
Else?if(x＞x_center)
Pre_Pixel_Flag(x)＝var(x)＞NTh&&var(x+1)＞NTh；
Else?Pre_Pixel_Flag(x)＝0；
...(2)
Wherein, Pre_Pixel_Flag (x) represents the part variation value of pixel x neighbor for preposition pixel flag, the x_center of pixel x represents pixel that central point pixel, x represent that x_center is adjacent, part variation value, var (x-1) or var (x+1) that var (x) represents pixel x.
Above formula represents, in pixel P11～39, if the part variation value (var (x)) of central point pixel P25 (x_center) first left pixel P24 (x < x_center) is greater than the part variation value (var (x-1)) of noise critical value and second left pixel P23 and is also greater than noise critical value, pixel P24 (x < x_center) can put on preposition pixel flag pre pixel flag.If when the part variation value (var (x)) of (x > x_center) of first pixel 26 of central point 25 (x_center) the right and the part variation value (var (x+1)) of second right pixel P27 are all greater than noise critical value, represent that pixel P26 (x > x_center) is for can put on preposition pixel flag Pre_Pixel_Flag.If above-mentioned two kinds of judgements are all false, the flag of pixel P24 and P26 is all denoted as 0 or mark flag not.According to whole scan line Y (middle this P21～P29) is obtained to each pixel, whether be preposition pixel flag Pre Pixel Flag afterwards.
Flag marker step (b): calculate pixel_flag
With central point x, according to following formula, calculate pixel flag Pixel Flag
If(x＜x_center)
Pixel_Flag(x)＝max(Pre_Pixel_Flag(x’))，x’＝x～x_center-1；
Else?iF(x＞x_center)
Pixel_Flag(x)＝max(Pre_Pixel_Flag(x’))，x’＝x_center+1～x；
Else?Pixel_Flag(x)＝0；
...(3)
Wherein, Pixel_Flag (x) is for putting the neighbor of pixel centered by pixel flag, x～x_center-1 and x_center+1～x.
Above formula represents, whether in the pixel on the central point pixel P25 left side, for example judge in P24～P23 (x～x_center-1) is real pixel but not noise, if pixel P24 has pre_Pixel_Flag=1, the pixel flag of pixel P23, P24 is 1, if now P23 has pre_Pixel Flag=1, and the pre_Pixel Flag of P24 equals 0, presentation video edge or border are during to P24 discontinuous (having broken); Whether the pixel on central point pixel P25 the right is for example judged in pixel P26～P27 (x_center+1～x) is real pixel but not which pixel of noise has maximum preposition pixel flag Pre_Pixel_Flag value, if pixel P26 has pre_Pixel_flag=1, the pixel flag of pixel P26, P27 is all made as 1, if pixel P27 has pre_Pixel_fiag=1, the pre_Pixel Flag of P26 equals 0, just discontinuous while representing edge or border to pixel 26.
Take central point pixel P25 as example explanation, when pixel P24 and pixel P26 all put on preposition pixel flag Pre Pixel Flag, and when the standard that pixel P24 and pixel P26 meet flag marker step (b) is maximum, can determine that central point pixel P25 is real pixel value, is not noise, therefore can the suitable flag value of mark one, represent that this pixel is not noise.
Then, according to aforesaid way, mark the flag value of each pixel of picture signal Vi, or indicate and for the pixel of noise, not to be denoted as the pixel of noise, to produce a flag distribution signal Fs to spatial noise filter 23 processing.
Should be noted, why above-mentioned is to adopt the mode of calculating 2 points (pixel), because only have the probability that a pixel is edge or border very low, therefore can adopt two or more pixels to compare.If two or more pixels are all set up, can determine that this central point pixel P25 is edge or border, be not noise.If only have to the left or to the right a bit (pixel) to set up, the pixel value of central point pixel P25 may be noise, is therefore judged as noise.The embodiment of the present invention is to adopt left and the state of observing central point pixel neighbor to the right; Certainly, in another embodiment, also can adopt upwards with downward observation or adopt other angle to observe the state of central point pixel neighbor, for example miter angle.
Then, spatial noise filter 23 receives the flag distribution signal Fs of received image signal Vi and 22 generations of noise detector, according to those flag value of flag distribution signal Fs, change the pixel value of pixel in received image signal Vi, to produce an output image signal Vo.
One embodiment, spatial noise filter 23 can be chosen a predeterminable area in the image of received image signal Vi, for example, can adopt the predeterminable area b of Fig. 2 C, and respectively for each region, carries out filtering processing.
Operational mode with subregional example explanation spatial noise filter 23:
Filter step (a): the image of received image signal Vi is taken out to default big or small region, for example 3 * 9 regions in Fig. 2 C;
Filter step (b): put pixel centered by pixel P25, produce the pixel flag Pixel_Flag of flag distribution signal Fs institute mark according to noise detector 22, change weight (weight) the W size of each pixel in the b of region.
Filter step (c): each pixel of region b is multiplied by its weights W, to produce output pixel data (Output data), an embodiment, this output pixel data (output pixeldata) can be obtained by following equation:
Wherein, outputpixel (i, j) is that output pixel data, W (i+k, j+l) are that weighted value, pixel (i+k, j+l) are pixel value.
Filter step (d): get back to filter step (a) until calculate the output pixel value of whole pixels, produce output image signal Vo.
Should be noted, the design of embodiment of the present invention spatial noise filter 23 weight sizes, the pixel value that calculates P25 of take is example, if be considered as the pixel on edge or border in b region, its weight is made as 0; And if the pixel close with P25 pixel value, its weight can be established larger.For example, pixel P24 and pixel P23 are the words on edge or border, its weight of taking advantage of is less, for example be multiplied by 0, and the pixel close with P25 pixel value, its weight can be established larger, for example be multiplied by 0.5, finally calculate the due pixel value of P25, allow pixel noise eliminate and to allow the pixel energy with edge or border more obviously present, and can reach the effect that retains image detail and filtering noise.
Should be noted that the pixel value of above-mentioned pixel can be the existing formatted numerical value such as Ycrcb, Yprpb, RGB....Certainly, in another embodiment, pixel value also can be the various forms that future development goes out.
Fig. 3 shows the flow chart of a kind of space image noise eliminating method of one embodiment of the invention.The method comprises the following steps:
Step S302: start.
Step S304: noise critical value produces step, estimates the part variation value of each this pixel of received image signal, and produces a noise critical value according to the state of those local variation values.
Step S306: noise determining step, according to noise critical value, determine in those pixels, which pixel is noise or is image.
Step S308: noise filtering step, those pixels that are noise of filtering, produce an output image signal.
Step S310: finish.
Should be noted, one embodiment, noise critical value produces step S304 and also comprises the following steps: to take out a maximum local variation value and calculate an average variation value in those local variation values, and by ratio and a preset value comparison of the local variation value of this maximum and this average local variation value, to judge the state of this part variation value.
One embodiment, noise determining step S306 also comprises the following steps: each pixel to be made as central point pixel, and by the part variation value of adjacent at least continuous two pixels of central point pixel and the comparison of noise critical value, according to comparative result judgement central point pixel, be noise or be image.
Another embodiment, noise determining step S306 also can comprise the following steps: each pixel to be made as central point pixel, judge that whether the part variation value that approaches most the pixel of central point in adjacent at least continuous two pixels of central point pixel is greater than the part variation value of other pixel, is noise or is image according to comparative result judgement central point pixel.
Image processing apparatus of the present invention and space image noise eliminating method utilize each pixel to be adjacent the part variation value of pixel, understand the degree of variation between whole image pixel value, recycling this degree of variation does to judge to image, obtain may for the pixel on edge or border be the pixel of noise, correspondingly adjust again the weight of pixel with edge or the border of filtering noise, strengthening image, and then the problem of solution prior art, reach filtering image noise and the effect that retains image detail simultaneously.
Though the present invention is described with embodiment above, therefore do not limit scope of the present invention, only otherwise depart from main idea of the present invention, those skilled in the art can carry out various distortion or change.
Claims (10)
1. an image processing apparatus, comprising:
One estimator, in order to estimate the part variation value of each pixel of received image signal, and produces a noise critical value according to the state of these local variation values;
One noise detector, is coupled to described estimator, according to described noise critical value, determines in described pixel, which pixel is noise or is image; And
One spatial noise filter, is coupled to described noise detector, those pixels that are noise of filtering, and to produce an output image signal,
Wherein, described local variation state of value judged by the average local variation value of the local variation value of maximum of the part or all of pixel of described received image signal and the part or all of pixel of described input picture,
Wherein, described estimator will be tried to achieve one first ratio after the local variation value of described maximum and described average local variation value, and by described the first ratio and a preset value comparison, to utilize described comparative result to determine the size of described noise critical value.
2. image processing apparatus according to claim 1, wherein, the degree of variation of described local variation value representation a plurality of pixels that pixel is adjacent described in each.
3. image processing apparatus according to claim 1, wherein, described noise detector is made as central point pixel by pixel described in each, and by the part variation value of adjacent at least continuous two pixels of described central point pixel and the comparison of described noise critical value, according to comparative result, determine that described central point pixel is noise or is image.
4. image processing apparatus according to claim 3, wherein, when variation value in the part of described two pixels is all greater than described noise critical value, is considered as image by described central point pixel; When the variation of the part of described two pixels is worth one of them while being less than described noise critical value, described central point pixel is considered as to noise.
5. image processing apparatus according to claim 1, wherein, described noise detector is made as central point pixel by pixel described in each, judge that whether the part variation value that approaches most the pixel of central point in adjacent at least continuous two pixels of described central point pixel is greater than the part variation value of other pixel, determines that according to comparative result described central point pixel is noise or is image.
6. image processing apparatus according to claim 5, wherein, when approaching most the part variation value of the pixel of central point and be greater than the part variation value of other pixel, is considered as image by described central point pixel; When approaching most the part variation value of the pixel of central point and be less than the part variation value of other pixel, described central point pixel is considered as to noise.
7. image processing apparatus according to claim 1, wherein, described spatial noise filter is noise according to described pixel or the pixel value of described pixel is multiplied by a default weighted value for image.
8. a space image noise eliminating method, comprising:
Noise critical value produces step, estimates the part variation value of each pixel of received image signal, and produces a noise critical value according to the state of described local variation value;
Noise determining step, determines according to described noise critical value in described pixel, which pixel is noise or is image; And
Noise filtering step, those pixels that are noise of filtering, produce an output image signal,
Wherein, described noise critical value generation step also comprises:
In described local variation value, take out a maximum local variation value and calculate an average variation value, and by the ratio of the local variation value of described maximum and described average local variation value and a preset value comparison, to judge the state of described local variation value.
9. method according to claim 8, wherein, described noise determining step also comprises:
Pixel described in each is made as to central point pixel, and by the part variation value of adjacent at least continuous two pixels of described central point pixel and the comparison of described noise critical value, according to comparative result, judges that described central point pixel is noise or is image.
10. method according to claim 8, wherein, described noise determining step also comprises:
Pixel described in each is made as to central point pixel, judge that whether the part variation value that approaches most the pixel of central point in adjacent at least continuous two pixels of described central point pixel is greater than the part variation value of other pixel, judges that according to comparative result described central point pixel is noise or is image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010507037.5A CN102447817B (en) | 2010-09-30 | 2010-09-30 | Image processing device and space image noise eliminating method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201010507037.5A CN102447817B (en) | 2010-09-30 | 2010-09-30 | Image processing device and space image noise eliminating method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102447817A CN102447817A (en) | 2012-05-09 |
| CN102447817B true CN102447817B (en) | 2014-11-26 |
Family
ID=46009876
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201010507037.5A Active CN102447817B (en) | 2010-09-30 | 2010-09-30 | Image processing device and space image noise eliminating method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102447817B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103945090B (en) * | 2013-01-17 | 2017-06-30 | 联咏科技股份有限公司 | Noise estimating apparatus and its method |
| CN105991900B (en) * | 2015-02-05 | 2019-08-09 | 扬智科技股份有限公司 | Noise detecting method and denoising method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5461655A (en) * | 1992-06-19 | 1995-10-24 | Agfa-Gevaert | Method and apparatus for noise reduction |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101087365B (en) * | 2006-06-10 | 2010-08-18 | 中兴通讯股份有限公司 | A method for filtering image mixed noise |
| CN100454970C (en) * | 2006-09-30 | 2009-01-21 | 四川长虹电器股份有限公司 | Spatial domain pixel data processing method |
| US7983501B2 (en) * | 2007-03-29 | 2011-07-19 | Intel Corporation | Noise detection and estimation techniques for picture enhancement |
| KR101481551B1 (en) * | 2008-06-03 | 2015-01-13 | 엘지전자 주식회사 | Apparatus and method for image noise reduction |
| WO2010093709A2 (en) * | 2009-02-10 | 2010-08-19 | Anchor Bay Technologies, Inc. | Block noise detection and filtering |
-
2010
- 2010-09-30 CN CN201010507037.5A patent/CN102447817B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5461655A (en) * | 1992-06-19 | 1995-10-24 | Agfa-Gevaert | Method and apparatus for noise reduction |
Non-Patent Citations (1)
| Title |
|---|
| Qiu，Fang 等.Speckle Noise Reduction in SAR Imagery Usinga Local Adaptive Median Filter.《GIScience and Remote sensing》.2004,第41卷(第3期),第244-266页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102447817A (en) | 2012-05-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US8224085B2 (en) | Noise reduced color image using panchromatic image | |
| US7792384B2 (en) | Image processing apparatus, image processing method, program, and recording medium therefor | |
| EP2089848B1 (en) | Noise reduction of panchromatic and color image | |
| WO2009130820A1 (en) | Image processing device, display, image processing method, program, and recording medium | |
| EP2359604B1 (en) | Modifying color and panchromatic channel cfa image | |
| US8594451B2 (en) | Edge mapping incorporating panchromatic pixels | |
| EP2164040B1 (en) | System and method for high quality image and video upscaling | |
| US8564863B2 (en) | Image processing device that reduces a non-target image to prevent the non-target image from being read | |
| US20070132865A1 (en) | Filtered noise reduction in digital images | |
| US20060244861A1 (en) | Method for detecting bisection pattern in deinterlacing | |
| EP2138975A2 (en) | Apparatus, method and computer-readable recording medium for pixel interpolation | |
| US6175659B1 (en) | Method and apparatus for image scaling using adaptive edge enhancement | |
| CN103312963A (en) | Image processing device and image processing method | |
| US20110243472A1 (en) | Image processing device and spatial noise reducing method | |
| US20090226097A1 (en) | Image processing apparatus | |
| CN101976436A (en) | Pixel-level multi-focus image fusion method based on correction of differential image | |
| JP3309941B2 (en) | Noise detection circuit, noise elimination circuit and contour enhancement circuit | |
| US20090022402A1 (en) | Image-resolution-improvement apparatus and method | |
| CN102447817B (en) | Image processing device and space image noise eliminating method | |
| US10158790B2 (en) | Image processing apparatus, image processing system, and image processing method | |
| CN112435182B (en) | Image noise reduction method and device | |
| US7697775B2 (en) | Image signal processing apparatus and method, and program and recording medium used therewith | |
| US20140147056A1 (en) | Depth image noise removal apparatus and method based on camera pose | |
| US20170053176A1 (en) | Image processing apparatus and image processing method | |
| US6449389B1 (en) | Method and apparatus for single channel color image segmentation using local context based adaptive weighting |
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 |