CN101527037A - Method for denoising stationary wavelet image based on neighborhood windowing - Google Patents
Method for denoising stationary wavelet image based on neighborhood windowing Download PDFInfo
- Publication number
- CN101527037A CN101527037A CN200910068457A CN200910068457A CN101527037A CN 101527037 A CN101527037 A CN 101527037A CN 200910068457 A CN200910068457 A CN 200910068457A CN 200910068457 A CN200910068457 A CN 200910068457A CN 101527037 A CN101527037 A CN 101527037A
- Authority
- CN
- China
- Prior art keywords
- detail coefficient
- coefficient
- image
- denoising
- stationary wavelet
- 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.)
- Granted
Links
Images
Abstract
The invention belongs to the technical field of image processing of a computer and relates to a method for denoising stationary wavelet image based on neighborhood windowing, which comprises the following steps of: carrying out monoscale transform of a two-dimensional stationary wavelet to a noisy image f (x, y), decomposing the image to be one layer, respectively obtaining four sub-band coefficients of low-frequency coefficient A1, horizontal detail coefficient H1, vertical detail coefficient V1 and diagonal detail coefficient D1; keeping A1 to be constant; adopting a vertical linear filter template, a horizontal linear filter template and a diagonal direction filter template to filter the horizontal detail coefficient H1, the vertical detail coefficient V1 and the diagonal detail coefficient D1 into H1, V1 and D1; carrying out stationary wavelet reconstruction to A1 and high-frequency sub-bands of H1, V1 and D1 after being filtered; and then obtaining the denoised image with *=A1+H1+V1+D1. By adopting the method, the purpose of restoring the original image and improving denoising performance to the image can be achieved better.
Description
Technical field
The invention belongs to technical field of image processing, be specifically related to a kind of method for denoising stationary wavelet image based on neighborhood window addition.
Background technology
Traditional denoising method mainly is to utilize noise to carry out with signal different characteristic distributions on frequency domain, adopt the low-pass filtering method will mainly be in the noise filtering of high-frequency region, but the details of image also mainly is distributed in high-frequency region simultaneously, though therefore the low-pass filtering method can reach the reduction anti noise, has also destroyed image detail.So keeping image detail as much as possible when reducing picture noise becomes an emphasis problem of studying in the image denoising.Because wavelet transformation has adopted the multiresolution analysis method, so application is wider in image denoising, as the wavelet threshold method that picture noise suppresses that is applicable to of people such as Donoho proposition.But this method is easy to generate vibration to picture signal reconstructed image edge after the detail coefficients behind the orthogonal wavelet transformation is carried out threshold process, causes the edge of image distortion.
Summary of the invention
The present invention be directed to the above-mentioned deficiency of prior art, a kind of method for denoising stationary wavelet image based on neighborhood window addition is provided.This method has been utilized the information redundancy of stationary wavelet conversion, is more conducive to find in the yardstick and the dependence between the wavelet coefficient between yardstick, thereby the parameter variance estimated accuracy that is based upon on the wavelet coefficient neighborhood is greatly improved.And also taken into full account correlativity in the layer of wavelet coefficient on this basis, selected the different window filtering templates that add according to its characteristic for each sub-band coefficients, recovered original image better thereby reach, improvement is to the purpose of the denoising performance of image.
Method for denoising stationary wavelet image based on neighborhood window addition of the present invention comprises the following steps:
Step 1: (x y) carries out the conversion of two-dimentional stationary wavelet single scale, obtains four sub-band coefficients respectively: low frequency coefficient A with noisy image f
1, level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1
Step 2: with A
1Remain unchanged, to level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Adopt veritcal linearity filtering template, the linear filtering template of level and diagonal angle trend pass filtering template to carry out mean filter respectively, after the filtering be
Step 3: with A
1With filtered high-frequency sub-band
Carry out stationary wavelet reconstruct, can obtain image after the denoising
Above-mentioned step 2 can be carried out according to following method:
1) at first respectively according to level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Frequency characteristic select corresponding mean filter template w respectively
ζζ=H, V, D;
2) carry out windowing respectively for the filtering template;
3) carry out filtering operation:
ζ=H,V,D
Wherein S is so that (x y) is the set of the neighborhood mid point at center, and M is counting in the S, and (x y) is sub-band coefficients after the filtering to g.
In the above-mentioned step 1), for level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Adopt respectively
As the filtering template.
Method for denoising stationary wavelet image based on neighborhood window addition provided by the invention, made full use of the information redundancy of stationary wavelet conversion, and the characteristics of the interior correlativity of layer of wavelet coefficient, and as adding window template denoising method based on what stationary wavelet decomposed according to providing a kind of, reach higher Y-PSNR, had better image denoising effect.
Description of drawings
Fig. 1 the present invention is based on the overview flow chart of the method for denoising stationary wavelet image of neighborhood window addition.
Fig. 2 stationary wavelet conversion synoptic diagram.Fig. 2 (a) stationary wavelet filtering; Fig. 2 (b) is the stationary wavelet interpolation process.
Fig. 3 denoising master drawing of the present invention.Fig. 3 (a) is the former figure of denoising master drawing; Fig. 3 (b) adds the image of making an uproar for the former figure of master drawing; Fig. 3 (c)~(g) is an image after the use denoising method denoising of the present invention, and wherein, (c) window area is 5, adds rectangular window; (d) window area is 5, adds Hanning window; (e) window area is 5, adds quarter window; (f) window area is 5, adds the hamming window; (g) window area is 5, adds the Brackman window.
Embodiment
Below by drawings and Examples the present invention is further described.
1. image transformation
(x y) carries out the stationary wavelet conversion, obtains the wavelet coefficient matrix, decomposes 1 layer, and wavelet basis is the sym8 small echo will to add the image f that makes an uproar.
The stationary wavelet conversion all produces the wavelet coefficient of similar number at each yardstick, digital picture f (its decomposition formula is (synoptic diagram is as shown in Figure 2) for x, two-dimentional stationary wavelet conversion y):
Wherein j is a decomposition scale, { h
kAnd { g
kBe respectively low pass and Hi-pass filter, h
0 ↑ 2jAnd g
0 ↑ 2jBe illustrated in h
0, g
0Insert 2j-1 individual zero between 2.J-1 yardstick tomographic image A
J-1Result after one deck wavelet decomposition is: low frequency coefficient
The level detail coefficient
The vertical detail coefficient
With the diagonal detail coefficient
Corresponding restructing algorithm is:
Here (x y) carries out the stationary wavelet conversion, decomposes 1 layer, obtains low frequency coefficient A respectively to image f
1With high frequency detail coefficients H
1, V
1And D
1
2. neighborhood window addition filtering
Wavelet transformation can recursively use low pass and Hi-pass filter to realize by the low frequency coefficient to same subband, means that wavelet coefficient is correlated with in a small neighbourhood, is called the interior correlativity of layer of wavelet coefficient.In the neighborhood of the bigger wavelet coefficient of value, may have one group of bigger wavelet coefficient.
To the wavelet coefficient individual processing in each subband, treatment step is as follows:
1) with A
1Remain unchanged, respectively according to H
1, V
1And D
1Frequency characteristic select corresponding mean filter template, after the filtering be
H wherein
1The low-frequency information of image signal level direction and the high-frequency information of vertical direction have been comprised, and Gaussian noise is bigger at high frequency region noise energy proportion, carry out filtering so selected veritcal linearity filtering template, suc as formula (2), so both eliminate the noise signal of vertical direction, kept image edge information again largely; V
1Then comprised the high-frequency information of image signal level direction and the low-frequency information of vertical direction, therefore selected the linear filtering template of level for use, suc as formula (4); Comprised high-frequency information, therefore adopted diagonal angle trend pass filtering template, as the formula (6) the angular direction.
2) carry out windowing for the filtering template;
For improving the filtering performance of template, the filtering template is carried out windowing, select rectangular window, Hanning window, quarter window, hamming window and Blackman window here.The characteristic of various window function w (n) is as follows:
◆ rectangular window (Rectangular Window)
Frequency domain characteristic is:
◆ Hanning window
The time domain form can be expressed as:
Frequency domain characteristic is:
Wherein, W
R(ω) be the amplitude-frequency characteristic function of rectangular window function.
◆ quarter window
Quarter window is the simplest frequency spectrum function W (e
J ω) be non-negative a kind of window function.The time domain form of quarter window function can be expressed as:
When n is odd number
When n is even number
Frequency domain characteristic is:
◆ the hamming window function
The time domain form can be expressed as
Frequency domain characteristic is
Wherein, W
R(ω) be the amplitude-frequency characteristic function of rectangular window function.
◆ the Blacknam window function
The time domain form can be expressed as
Frequency domain characteristic is
Wherein, W
R(ω) be the amplitude-frequency characteristic function of rectangular window function.
4) carry out filtering operation.Use corresponding filtering template that each subband is carried out filtering operation.
Wherein S is so that (x y) is the set of the neighborhood mid point at center, and M is counting in the S, and (x y) is sub-band coefficients after the filtering to g.
3. image reconstruction
4. experimental result
In order to verify the validity of denoising method of the present invention, (shown in Fig. 3 (a)) tests to concrete picture.Adopt the sym8 small echo to carry out Flame Image Process in the experiment, to image in addition standard variance be 15 noise, image is decomposed 1 layer with stationary wavelet, and uses the filtering template of different windowings to carry out mean filter adding the image of making an uproar.As the criterion of anti-acoustic capability quality, experimental result is as shown in table 1 with PSNR (PeakSignal to Noise Ratio).
Use the PSNR/db of various window area filtering to compare under the different windowing situations of table 1
The data that provide from table 1 use the method for denoising stationary wavelet image based on neighborhood window addition that provides among the present invention can obtain higher Y-PSNR as can be seen.Simultaneously from Fig. 3 (c)~(g) handle the back image also as can be seen this method obtained denoising effect preferably.
Claims (3)
1. the method for denoising stationary wavelet image based on neighborhood window addition comprises the following steps:
Step 1: (x y) carries out the conversion of two-dimentional stationary wavelet single scale, decomposes 1 layer, obtains four sub-band coefficients respectively: low frequency coefficient A with noisy image f
1, level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1
Step 2: with A
1Remain unchanged, to level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Adopt veritcal linearity filtering template, the linear filtering template of level and diagonal angle trend pass filtering template to carry out mean filter respectively, after the filtering be
2. the method for denoising stationary wavelet image based on neighborhood window addition according to claim 1, step 2 wherein can be carried out according to following method:
1) at first respectively according to level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Frequency characteristic select corresponding mean filter template w respectively
ζζ=H, V, D;
2) carry out windowing respectively for the filtering template;
3) carry out filtering operation:
ζ=H,V,D
Wherein S is so that (x y) is the set of the neighborhood mid point at center, and M is counting in the S, and (x y) is sub-band coefficients after the filtering to g.
3. the method for denoising stationary wavelet image based on neighborhood window addition according to claim 2 is in the step 1) wherein, for level detail coefficient H
1, vertical detail coefficient V
1With diagonal detail coefficient D
1Adopt respectively
As the filtering template.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2009100684575A CN101527037B (en) | 2009-04-13 | 2009-04-13 | Method for denoising stationary wavelet image based on neighborhood windowing |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2009100684575A CN101527037B (en) | 2009-04-13 | 2009-04-13 | Method for denoising stationary wavelet image based on neighborhood windowing |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101527037A true CN101527037A (en) | 2009-09-09 |
CN101527037B CN101527037B (en) | 2011-05-04 |
Family
ID=41094894
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2009100684575A Expired - Fee Related CN101527037B (en) | 2009-04-13 | 2009-04-13 | Method for denoising stationary wavelet image based on neighborhood windowing |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101527037B (en) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509271A (en) * | 2011-11-21 | 2012-06-20 | 洪涛 | Image restoration method based on multi-dimensional decomposition, iteration enhancement and correction |
CN104995885A (en) * | 2013-02-05 | 2015-10-21 | 交互数字专利控股公司 | Pulse-shaped orthogonal frequency division multiplexing |
CN110766627A (en) * | 2019-10-16 | 2020-02-07 | 北京信息科技大学 | Speckle interference image noise reduction method and device |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6785411B1 (en) * | 1999-08-05 | 2004-08-31 | Matsushita Electric Industrial Co., Ltd. | Image analyzing apparatus and image analyzing method |
JP2003204436A (en) * | 2001-10-29 | 2003-07-18 | Victor Co Of Japan Ltd | Image coding equipment and program thereof |
CN1141639C (en) * | 2002-05-09 | 2004-03-10 | 宣国荣 | Digital watermark method based on integer wavelet without damage to image |
-
2009
- 2009-04-13 CN CN2009100684575A patent/CN101527037B/en not_active Expired - Fee Related
Cited By (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102509271A (en) * | 2011-11-21 | 2012-06-20 | 洪涛 | Image restoration method based on multi-dimensional decomposition, iteration enhancement and correction |
CN104995885A (en) * | 2013-02-05 | 2015-10-21 | 交互数字专利控股公司 | Pulse-shaped orthogonal frequency division multiplexing |
US10523475B2 (en) | 2013-02-05 | 2019-12-31 | Idac Holdings, Inc. | Pulse-shaped orthogonal frequency division multiplexing |
CN110766627A (en) * | 2019-10-16 | 2020-02-07 | 北京信息科技大学 | Speckle interference image noise reduction method and device |
Also Published As
Publication number | Publication date |
---|---|
CN101527037B (en) | 2011-05-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103854264B (en) | A kind of Wavelet Transformation of Image Denoising method based on modified threshold function table | |
CN101739671B (en) | Method and system for reducing noise of noisy images | |
CN107274908A (en) | Small echo speech de-noising method based on new threshold function table | |
Verma et al. | Performance analysis of wavelet thresholding methods in denoising of audio signals of some Indian Musical Instruments | |
CN101482617B (en) | Synthetic aperture radar image denoising method based on non-down sampling profile wave | |
CN101477681B (en) | Wavelet image denoising process based on adaptive sliding window adjacent region threshold | |
CN102663695B (en) | DR image denoising method based on wavelet transformation and system thereof | |
CN101477680A (en) | Wavelet image denoising process based on sliding window adjacent region data selection | |
CN101719267B (en) | A kind of denoising noise image and system | |
CN101719268B (en) | Generalized Gaussian model graph denoising method based on improved Directionlet region | |
CN102393423A (en) | Lamb wave denoising method based on adaptive threshold value orthogonal wavelet transform | |
CN103489157A (en) | Filtering method and system for enhancing synthetic aperture sonar interferogram quality | |
CN102624349B (en) | Harmonic noise and white-noise interference eliminating method with low distortion to initial data | |
CN104849757A (en) | System and method for eliminating random noise in seismic signals | |
CN101430788B (en) | Lunette local Wiener filtering method based on second generation Curvelet transformation | |
CN101527037B (en) | Method for denoising stationary wavelet image based on neighborhood windowing | |
CN104732498B (en) | A kind of thresholded image denoising method based on non-downsampling Contourlet conversion | |
CN101527036B (en) | Lifting wavelet image de-noising method based on neighborhood windowing | |
Rao et al. | Selective neighbouring wavelet coefficients approach for image denoising | |
CN106023103B (en) | A kind of adaptive orthogonal wavelet image de-noising method based on the modeling of accurate local variance priori | |
CN101510301A (en) | Method for removing noise of wavelet image based on neighborhood window addition | |
CN109242804A (en) | A kind of image de-noising method based on level set curvature and wavelet transformation | |
CN102509268B (en) | Immune-clonal-selection-based nonsubsampled contourlet domain image denoising method | |
CN104182944A (en) | Optical image denoising method based on serial connection of curvelet transform and wavelet transform | |
CN101561925A (en) | Wavelet image denoising method based on full-phase neighborhood windowing |
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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110504 Termination date: 20210413 |