CN100505832C - Image de-noising process of multi-template mixed filtering - Google Patents

Image de-noising process of multi-template mixed filtering Download PDF

Info

Publication number
CN100505832C
CN100505832C CNB2006100656798A CN200610065679A CN100505832C CN 100505832 C CN100505832 C CN 100505832C CN B2006100656798 A CNB2006100656798 A CN B2006100656798A CN 200610065679 A CN200610065679 A CN 200610065679A CN 100505832 C CN100505832 C CN 100505832C
Authority
CN
China
Prior art keywords
image
filter
denoising
numbers
variance
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
Application number
CNB2006100656798A
Other languages
Chinese (zh)
Other versions
CN101043581A (en
Inventor
王锡贵
李鹏
韩冀中
韩承德
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Computing Technology of CAS
Original Assignee
Institute of Computing Technology of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Institute of Computing Technology of CAS filed Critical Institute of Computing Technology of CAS
Priority to CNB2006100656798A priority Critical patent/CN100505832C/en
Publication of CN101043581A publication Critical patent/CN101043581A/en
Application granted granted Critical
Publication of CN100505832C publication Critical patent/CN100505832C/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Processing (AREA)

Abstract

The disclosed picture removing-noise method based on multi-template mixed filter comprises: dividing the objective picture into MXN blocks without superposition to contain every pixel in some block; defining one group of filters including the mean filter and mid-value filter, every with different removing-noise strength; defining a numerical sequence with number standing for picture information size; obtaining the noise variance of the picture, calculating the uniformity level for the picture block to thereby select one filter for finishing the removing-noise. This invention simplifies calculation, and reduces real running time.

Description

A kind of image de-noising method of process of multi-template mixed filtering
Technical field
The present invention relates to Digital Image Processing, particularly the image de-noising method in the Digital Image Processing.
Background technology
The existence of noise has caused image distortion to a certain extent, if noise is crossed the use that can hinder at most image, must carry out denoising this moment to image.Can cause the original image loss of detail inevitably in the image denoising process, thus noise suppressed and details to keep be a pair of contradiction, various denoising methods all will be selected balance point between the two.Existing denoising method respectively has its pluses and minuses, selects which kind of method denoising to depend on the demand of application-specific and the feature of used data.
Denoising method can be divided into two kinds of spatial domain and frequency domains by its working field.Existing space territory self adaptation denoising method at each pixel, is used the template operator of fixed size, utilizes local mean value, variance to regulate filtering strength.But they can not fully utilize the characteristic of medium filtering and mean filter as a kind of adaptive average or medium filtering.Mean filter and medium filtering respectively have its pluses and minuses: mean filter is smoothed image effectively, improves image correlation, but is easy to make edge, details to fog, and can not keep picture structure.Medium filtering can effectively be removed noise, keeps image detail, but the image smoothing effect is not so good as mean filter filter factor calculation of complex in addition, and is limited to the adjusting of filtering strength.In addition, mostly fixedly filtering template size of existing denoising method is only regulated intensity by changing filtering parameter.In the practical application, so not only filtering complexity of filtering, and also the filtering strength excursion is limited, do not have the smooth region of image detail at some, 3 * 3 templates are too small, and filtering strength is not enough, and in the abundant zone of details, but seem excessive, grievous injury image detail.The frequency domain denoising method is based on wavelet threshold denoising, and the wavelet transformation amount of calculation is big, and elapsed time is more.
Summary of the invention
The objective of the invention is to overcome fixedly filtering template size of existing space territory self adaptation denoising method, only regulate the defective of intensity by changing filtering parameter, can when removing noise, keep image detail preferably thereby provide a kind of, and have the image de-noising method of high efficiency spatial domain.
To achieve these goals, the invention provides a kind of image de-noising method of process of multi-template mixed filtering, comprise following steps:
1), be the piece of the non-overlapping copies of M * N size with the image division that will carry out denoising, make each pixel of image all be in certain piece;
2), the definition one group of filter, described filter comprises mean filter and median filter, the filter in the described bank of filters has different removing-noise strength;
3), one group of ordered series of numbers of definition, what of the contained amount of information of image are the numeral in the described ordered series of numbers be used to describe;
4), from imaging system, obtain to carry out the noise variance σ of the image of denoising ν
5), each image block that step 1) is divided calculates variance Var (Z) and average Z, according to the variance of image block and the result of calculation of average, the value of the even matter degree S of computed image piece, S = Var ( Z ) / Z ‾ ；
6), judge the even matter degree of each image block, the even matter degree S, the image noise variance σ that step 4) obtains that obtain according to step 5) νAnd defined ordered series of numbers in the step 3), for the some filters in the image block selective filter group, finish denoising to image block;
7), repeated execution of steps 6), all images piece in the image that will carry out denoising has all been finished denoising operation.
In the technique scheme, in described step 1), the size of described M and N is between 8 to 16.
In the technique scheme, in described step 2) in, the filter in the described bank of filters has different characteristics, different window size and different weight coefficients.
In the technique scheme, in described step 3), described ordered series of numbers is an ascending series, and first number of ordered series of numbers is 0, and the number of the numeral in the ordered series of numbers is than step 2) in the number of filter in the defined bank of filters Duo one.
In the technique scheme, in described step 4), the noise variance σ of described image νDirectly provide, or choose polylith smooth region in the image, calculate the mean variance ratio of each piece image, all mean variance ratios are averaged, as the noise variance of image by described imaging system.
In the technique scheme, in described step 6), when judging the even matter degree of image block, the image noise variance σ that step 4) is obtained at first νEach digital multiply in the ordered series of numbers that obtains with step 3) obtains a plurality of intervals, the corresponding a kind of filter in interval, and the value of the even matter degree S that obtains according to step 5) is selected the interval at place, thereby is selected concrete filter to do Filtering Processing.
The present invention fully utilizes the characteristic of medium filtering and mean filter, calculates simply, is beneficial to hardware and realizes that actual run time is short.
Description of drawings
Fig. 1 is the schematic diagram of the mean filter f1 that adopted among the embodiment;
Fig. 2 is the schematic diagram of the mean filter f2 that adopted among the embodiment;
Fig. 3 is the schematic diagram of the mean filter f3 that adopted among the embodiment;
The ENL value curve chart of Fig. 4 for image after the denoising is carried out quantitative analysis;
The EEI value curve chart of Fig. 5 for image after the denoising is carried out quantitative analysis;
The FPI value curve chart of Fig. 6 for image after the denoising is carried out quantitative analysis;
Fig. 7 is the flow chart of the image de-noising method of process of multi-template mixed filtering of the present invention.
The drawing explanation
Implication in accompanying drawing 4～legend employed in figure 6 is as follows:
Figure C200610065679D00061
Former figure
Figure C200610065679D00062
The present invention
Figure C200610065679D00063
Kuan
Figure C200610065679D00064
E.Lee
Frost
Figure C200610065679D00066
E.Frost
Gamma(MAP)
L.Sigma
Figure C200610065679D00068
The soft threshold values of small echo
Embodiment
Describe below in conjunction with the image de-noising method of the drawings and specific embodiments process of multi-template mixed filtering of the present invention.
As shown in Figure 7, in the present embodiment, be example with denoising to the ground remote sensing images, the specific implementation step of image denoising is described:
Step 1, being the piece of the non-overlapping copies of M * N size with the image division that will carry out denoising, making each pixel of image all be in certain piece, can make an exception in image border wherein.In the present embodiment, the size of image division piece selects length and width between 8 to 16, should not be too big or too little.The piece that image is divided is too big, and then included homogeneous area is less, and denoising effect is not good.The piece that image is divided is too small, and then regional probability statistics are poor, and randomness is strong.In addition, also can image division be become piece according to the resolution of image.The image resolution ratio height can become bigger piece with image division, otherwise then image division be become less piece.
Step 2, one group of filter of definition, these filters have different characteristics, and intermediate value and the mean filter be made up of different windows size, different weights coefficient constitute.By selecting window size and design weight coefficient can obtain the filter of different removing-noise strength.In the present embodiment, designed four kinds of different filters, described filter is represented with f1, f2, f3 and f4 respectively.It in Fig. 1, Fig. 2 and Fig. 3 the schematic diagram of mean filter f1, f2 and f4.Represented filter f1 is one 8 * 8 mean filter among Fig. 1, and the weight coefficient of filter all is 1.Not selecting 7 * 7 or 9 x, 9 odd sized windows traditional in the image processing in the present embodiment for use is in order to avoid division by displacement, to help hardware and realize.Represented filter f2 among Fig. 2, it is the heavy mean filter of a cum rights.The center is current filtering pixel location, and the coefficient in the template is the image pixel weighted value of relevant position.Filter f4 and filter f2 are similar, but its filtering window is littler, carry out filtering in littler scope.Filter f3 is 3 * 3 Sigma (Sigma) median filter.To a certain filtering pixel, 9 pixels in 3 * 3 the window ranges that with it is the center, obtain the mean variance of these 9 pixels, utilize normal distribution 3-Sigma (3-σ) principle then, the pixel value that surpasses 3 times of variances with mean distance is considered as " bad value " removes.If remaining value is less than 3, then current filtering pixel is adopted aforesaid f3 filter filtering, otherwise, remaining pixel value is pressed the increasing or decreasing rank order, get median at last and substitute current filtered pixel value.
Step 3, experimental result provide one group to describe image and comprise what ordered series of numbers of amount of information, and this ordered series of numbers can be used for describing the even matter degree of step 1 a divided image piece.To the multiplicative noise model, its even matter degree can recently be delineated with the mean variance of pixel in this image block.In the present embodiment, the value of described ordered series of numbers is { k1, k2, k3, k4, k5}={0,1,1.1,1.25 ,+∞ }.
Step 4, from imaging system, obtain to carry out the noise variance σ of the image of denoising νThe variance of noise is determined by imaging system, belongs to systematic error.For an all images that imaging system became, its noise variance is constant.This noise variance is provided by system, to the multiplicative noise image, also can calculate as follows from image.Judge according to human eye, choose polylith smooth region in the image, calculate the mean variance ratio of each piece image, to all mean variances than the noise variance of averaging as final image.
Step 5, each image block that step 1 is divided calculate variance Var (Z) and average Z.According to the variance of image block and the result of calculation of average, the even matter degree of computed image piece S = Var ( Z ) / Z ‾ Value.In this step, be ripe prior art to the calculating of variance and average, in the present embodiment, how it is not realized doing detailed description.
Step 6, judge the even matter degree of each image block,, utilize the ordered series of numbers of definition in the step 3, select suitable filters to finish denoising according to even matter degree in zone and image applications requirement.In the present embodiment, the σ that obtains according to step 4 νWith the S that step 5 obtains, the ordered series of numbers of being divided in the integrating step 3, institute's divided image piece in the step 1 is judged according to following condition:
1), to a certain image block, as 0＜S≤σ νThe time, experiment shows this image block without any details,, filtering operation can not reduce picture quality to not influence of image detail.Can adopt 8 * 8 very strong mean filter f1 of smoothing capability to the denoising of this image block.
2), to a certain image block, work as σ ν＜S≤1.1 σ νThe time, experiment shows that the variation in this represented zone of image block is comparatively slow, contains texture, details hardly, based on noise.The processing of this image block based on denoising, is selected for use the stronger mean filter f2 of smoothing capability.
3), to a certain image block, as 1.1 σ ν＜S≤1.25 σ νThe time, experiment shows that the represented regional details of this image block, texture are very many, should therefore adopt 3 * 3 Sigma (Sigma) median filter to keep details.
4), to a certain image block, work as S〉1.25 σ νThe time, experiment shows that this image block region all is edge, details, even experiment shows that the medium filtering of 3 * 3 windows also can lose details to the filtering in such zone, makes image blur; Therefore use the very weak filter f4 of smoothing capability.
Step 7, repeated execution of steps 6, all images piece in the image that will carry out denoising have all been finished the denoising operation.
In above-mentioned steps, described partitioned image piece of step 1 and step 2, step 3, the described definition filter of step 4 and ordered series of numbers be the precedence relationship on the life period not, can carry out simultaneously, often carries out by a certain order serial in the practical operation.
Present embodiment be with the ground remote sensing image as process object, and the ground remote sensing image is a kind of multiplicative noise image.Method of the present invention not only is applicable to the multiplicative noise image, and those of ordinary skill in the art should be easily be applied to the method described in the present embodiment in other noise model, as the additive noise image.
Evaluation to image denoising effect comprises two aspects, and one is the noise suppressed degree, and one is the image detail reserving degree.(Equivalent Number of Looks ENL) is used for the noise level of evaluation map picture to equivalent number, and the high more explanation picture noise of ENL value is few more, and just the noise removing degree is high more; The edge strengthens index, and (Edge EnhancingIndex EEI) can be used for estimating after the denoising image with respect to the edge feature reserve capability of original image; Feature keeps index (Feature Preserving Index, FPI) can be used for estimating after the denoising image with respect to the lines feature reserve capability of original image, FPI and EEI are used for estimating after the denoising image jointly with respect to the details reserving degree of original image, their value all between 0 and 1, be worth big more explanation details keep good more.Because denoising and reservation image detail feature are a pair of contradiction, for same width of cloth image, the increase of ENL value means reducing of EEI and FPI value.
The classical Denoising Algorithm of method of the present invention and some is tested on the different image of three width of cloth, and the classification of this three width of cloth image is respectively Canadian radar remote sensing satellite (RadarSAT) image, European long-range remote sensing satellite (ERS) image and airborne radar image.The result of test can be with reference to figure 4, Fig. 5 and Fig. 6.Fig. 4, be each algorithm to three width of cloth test pattern denoisings after the equivalent number value of image, Fig. 5, Fig. 6 make an uproar that edge of image strengthens index and feature keeps desired value.
Method of the present invention also has higher efficient, RadarSAR image for one 1024 * 1024 size, the filtering window size of other algorithm gets 3 * 3, when carrying out denoising with computer, the actual run time of various algorithms is as shown in table 1, can see from this table, and the inventive method is done the needed time of denoising to an image, want much less than the needed time of other algorithm, method of the present invention has very high efficient.
Table 1
Algorithm The present invention Kuan Lee E.Lee Frost E.Frost Gamma(MAP) Local_Sigma The soft threshold values of small echo
To (a unit: second) 1.2 2.4 3.4 4.0 12.4 12.0 3.7 10.3 3.6

Claims (6)

1, a kind of image de-noising method of process of multi-template mixed filtering comprises following steps:
1), be the piece of the non-overlapping copies of M * N size with the image division that will carry out denoising, make each pixel of image all be in certain piece;
2), the definition one group of filter, described filter comprises mean filter and median filter, the filter in the described bank of filters has different removing-noise strength;
3), one group of ordered series of numbers of definition, what of the contained amount of information of image are the numeral in the described ordered series of numbers be used to describe;
4), from imaging system, obtain to carry out the noise variance σ of the image of denoising v
5), each image block that step 1) is divided calculates variance Var (Z) and average Z, according to the variance of image block and the result of calculation of average, the value of the even matter degree S of computed image piece, S = Var ( Z ) / Z ‾ ；
6), judge the even matter degree of each image block, the even matter degree S, the image noise variance σ that step 4) obtains that obtain according to step 5) vAnd defined ordered series of numbers in the step 3), for the some filters in the image block selective filter group, finish denoising to image block;
7), repeated execution of steps 6), all images piece in the image that will carry out denoising has all been finished denoising operation.
2, the image de-noising method of process of multi-template mixed filtering according to claim 1 is characterized in that, in described step 1), the size of described M and N is between 8 to 16.
3, the image de-noising method of process of multi-template mixed filtering according to claim 1 is characterized in that, in described step 2) in, the filter in the described bank of filters has different characteristics, different window size and different weight coefficients.
4, the image de-noising method of process of multi-template mixed filtering according to claim 1, it is characterized in that, in described step 3), described ordered series of numbers is an ascending series, first number of ordered series of numbers is 0, and the number of the numeral in the ordered series of numbers is than step 2) in the number of filter in the defined bank of filters Duo one.
5, the image de-noising method of process of multi-template mixed filtering according to claim 1 is characterized in that, in described step 4), and the noise variance σ of described image vDirectly provide, or choose polylith smooth region in the image, calculate the mean variance ratio of each piece image, all mean variance ratios are averaged, as the noise variance of image by described imaging system.
6, the image de-noising method of process of multi-template mixed filtering according to claim 1 is characterized in that, in described step 6), and when judging the even matter degree of image block, the image noise variance σ that step 4) is obtained at first vEach digital multiply in the ordered series of numbers that obtains with step 3) obtains a plurality of intervals, the corresponding a kind of filter in interval, and the value of the even matter degree S that obtains according to step 5) is selected the interval at place, thereby is selected concrete filter to do Filtering Processing.
CNB2006100656798A 2006-03-21 2006-03-21 Image de-noising process of multi-template mixed filtering Expired - Fee Related CN100505832C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2006100656798A CN100505832C (en) 2006-03-21 2006-03-21 Image de-noising process of multi-template mixed filtering

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2006100656798A CN100505832C (en) 2006-03-21 2006-03-21 Image de-noising process of multi-template mixed filtering

Publications (2)

Publication Number Publication Date
CN101043581A CN101043581A (en) 2007-09-26
CN100505832C true CN100505832C (en) 2009-06-24

Family

ID=38808754

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2006100656798A Expired - Fee Related CN100505832C (en) 2006-03-21 2006-03-21 Image de-noising process of multi-template mixed filtering

Country Status (1)

Country Link
CN (1) CN100505832C (en)

Families Citing this family (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101472055B (en) * 2007-12-24 2012-10-10 深圳迈瑞生物医疗电子股份有限公司 Medium value filtering device and method, and ultrasound imaging system applying the device
CN101472058B (en) * 2007-12-29 2011-04-20 比亚迪股份有限公司 Apparatus and method for removing image noise
CN101770636B (en) * 2008-12-30 2012-09-26 比亚迪股份有限公司 Image noise removing method and device
CN101655977B (en) * 2009-09-17 2011-08-03 浙江大学 Method for eliminating image impulse noise based on differential image detection and filtration by multiple windows
CN102170519A (en) * 2010-06-22 2011-08-31 上海盈方微电子有限公司 High-efficient image denoising algorithm and hardware realization apparatus of algorithm
CN102222319B (en) * 2011-04-08 2015-08-12 中兴通讯股份有限公司 A kind of image processing method and device
JP5399578B2 (en) * 2012-05-16 2014-01-29 シャープ株式会社 Image processing apparatus, moving image processing apparatus, video processing apparatus, image processing method, video processing method, television receiver, program, and recording medium
CN103218781B (en) * 2013-04-07 2018-07-31 小米科技有限责任公司 A kind of method and apparatus that space filtering processing is carried out to image
CN104143177A (en) * 2013-05-07 2014-11-12 江南大学 Method for eliminating image interference of line-scan digital camera
CN103324123B (en) * 2013-06-27 2015-04-29 重庆大学 Vehicle-mounted data pre-processing method based on embedded system
CN104182935B (en) * 2014-08-11 2017-02-15 同济大学 Image denoising method for selecting wavelet basis on the basis of analytic hierarchy process
US9262810B1 (en) * 2014-09-03 2016-02-16 Mitsubishi Electric Research Laboratories, Inc. Image denoising using a library of functions
CN104933686A (en) * 2015-07-02 2015-09-23 山西省交通科学研究院 Smoothing method for image processing and mode identification through computer
CN106018201B (en) * 2016-05-26 2018-08-21 天津大学 Mixing field particle size measurement procedure based on mean filter
CN106651783B (en) * 2016-09-29 2019-10-22 厦门美图之家科技有限公司 Image filtering method and device
JP2019106590A (en) * 2017-12-11 2019-06-27 シャープ株式会社 Abnormality determination device, abnormality determination method, and control program
CN109712078A (en) * 2018-07-23 2019-05-03 永康市巴九灵科技有限公司 Cabinet security protection control platform
CN109660864A (en) * 2018-08-07 2019-04-19 任阿毛 Key frame movement range detection method
CN109241960A (en) * 2018-08-08 2019-01-18 任元华 Take medicine window prescription scanning platform
CN109034098A (en) * 2018-08-13 2018-12-18 明世中 App parameter setting platform
CN109639998A (en) * 2018-08-15 2019-04-16 吴伟锋 Data frame automatically grabs method
CN110006915A (en) * 2018-08-30 2019-07-12 永康市缘匠贸易有限公司 Glass blocks bar detection mechanism
CN109002816A (en) * 2018-08-30 2018-12-14 朱如兴 Film violence rank discrimination method
CN109151563A (en) * 2018-08-31 2019-01-04 李叶 Audio intensity modifies platform automatically
CN110174559A (en) * 2018-09-20 2019-08-27 永康市巴九灵科技有限公司 Real-time remaining capacity measuring mechanism
CN109635776A (en) * 2018-12-23 2019-04-16 广东腾晟信息科技有限公司 Pass through the method for procedure identification human action
CN110097518B (en) * 2019-04-28 2022-12-27 东软医疗系统股份有限公司 Image denoising method and device and terminal equipment
CN111445424B (en) * 2019-07-23 2023-07-18 广州市百果园信息技术有限公司 Image processing method, device, equipment and medium for processing mobile terminal video
CN112801882B (en) * 2019-11-14 2022-11-08 RealMe重庆移动通信有限公司 Image processing method and device, storage medium and electronic equipment
CN112381727B (en) * 2020-11-03 2023-12-15 中航航空电子有限公司 Image denoising method and device, computer equipment and storage medium
CN113299245B (en) * 2021-05-11 2022-07-19 深圳创维－Rgb电子有限公司 Method and device for adjusting local backlight of display equipment, display equipment and storage medium

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001144964A (en) * 1999-11-15 2001-05-25 Fuji Photo Film Co Ltd Noise eliminating method and unit, and scanner
US20050117807A1 (en) * 2003-12-01 2005-06-02 School Foundation Of Nippon Institute Of Technology Method and apparatus for reduction mosquito noise in decoded images
CN1633159A (en) * 2005-01-26 2005-06-29 北京中星微电子有限公司 A method for removing image noise
CN1668068A (en) * 2005-04-08 2005-09-14 杭州国芯科技有限公司 An adaptive picture noise suppression method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001144964A (en) * 1999-11-15 2001-05-25 Fuji Photo Film Co Ltd Noise eliminating method and unit, and scanner
US20050117807A1 (en) * 2003-12-01 2005-06-02 School Foundation Of Nippon Institute Of Technology Method and apparatus for reduction mosquito noise in decoded images
CN1633159A (en) * 2005-01-26 2005-06-29 北京中星微电子有限公司 A method for removing image noise
CN1668068A (en) * 2005-04-08 2005-09-14 杭州国芯科技有限公司 An adaptive picture noise suppression method

Also Published As

Publication number Publication date
CN101043581A (en) 2007-09-26

Similar Documents

Publication Publication Date Title
CN100505832C (en) Image de-noising process of multi-template mixed filtering
CN101847257B (en) Image denoising method based on non-local means and multi-level directional images
CN103295191B (en) Multiple scale vision method for adaptive image enhancement and evaluation method
CN102611910B (en) Objective evaluation method of no-reference video quality based on key frame image quality weight
CN102663708B (en) Ultrasonic image processing method based on directional weighted median filter
DE10319118A1 (en) Method and device for parameterized sharpening and smoothing
CN102256048B (en) Density-adaptive image salt-pepper noise switching filtering method
CN106169181A (en) A kind of image processing method and system
DE102013212730A1 (en) Entrauching device, system and method
CN102036098B (en) Full-reference type image quality evaluation method based on visual information amount difference
CN101655974A (en) Improved image enhancing method based on partial histogram equalization method
CN101505367B (en) Method for adaptive image enhancement
CN100417191C (en) Method of reducing noise for combined images
CN104680485A (en) Method and device for denoising image based on multiple resolutions
CN102306378B (en) Image enhancement method
CN113808042B (en) SAR image denoising method based on wavelet transformation and generation countermeasure network
CN101996321A (en) Fingerprint recognition pretreatment method and device
CN104331863A (en) Image filtering and denoising method
CN107945122A (en) Infrared image enhancing method and system based on self-adapting histogram segmentation
CN101625759B (en) Image quality evaluation method
CN102750700A (en) Fast robust fuzzy C-means image segmentation method combining neighborhood information
CN110634141A (en) Image segmentation method based on improved intuitionistic fuzzy c-means clustering and storage medium
CN115937231B (en) Infrared image iterative denoising method and system constrained by frequency spectrum structure
CN103269436A (en) Key frame selection method in 2D-3D video conversion
CN111311512A (en) Random noise point correction method

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: 20090624

Termination date: 20190321

CF01 Termination of patent right due to non-payment of annual fee