CN106447629A - Ratio distance-based non-local-mean radar image coherent speckle suppression method - Google Patents
Ratio distance-based non-local-mean radar image coherent speckle suppression method Download PDFInfo
- Publication number
- CN106447629A CN106447629A CN201610810707.8A CN201610810707A CN106447629A CN 106447629 A CN106447629 A CN 106447629A CN 201610810707 A CN201610810707 A CN 201610810707A CN 106447629 A CN106447629 A CN 106447629A
- Authority
- CN
- China
- Prior art keywords
- similitude
- window
- ratio
- point
- image
- 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
- 230000001427 coherent effect Effects 0.000 title claims abstract description 26
- 238000000034 method Methods 0.000 title claims abstract description 25
- 230000001629 suppression Effects 0.000 title abstract description 10
- 238000003384 imaging method Methods 0.000 claims description 8
- 230000009467 reduction Effects 0.000 description 23
- 230000008901 benefit Effects 0.000 description 6
- 230000000694 effects Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000001914 filtration Methods 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 238000009795 derivation Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G06T5/70—
Landscapes
- Image Processing (AREA)
Abstract
The invention provides a ratio distance-based non-local-mean radar image coherent speckle suppression method. The method comprises the steps of reading a synthetic aperture radar coherent speckle image; setting corresponding search window and similar window sizes; determining a ratio distance between any two similar windows according to gray-level values in the any two similar windows; determining a gray-level value ratio of central points of the any two similar windows according to gray-level values of the central points of the any two similar windows; determining a space distance between the central points of the any two similar windows according to space coordinates of the central points of the any two similar windows; calculating weight values of pixel points in the two similar windows according to the ratio distance, the gray-level value ratio of the central points and the space distance between the central points; and estimating estimated values of the current pixel points through the weight values. According to the method, coherent speckle noises are effectively suppressed, the edges can be effectively kept, the performance index is improved, and the visibility is better.
Description
Technical field
The present embodiments relate to radar imaging coherent spot suppression field, more particularly, to a kind of non-office based on ratio distance
Portion's average radar imaging coherent spot suppressing method.
Background technology
In traditional non-local mean algorithm, the Euclidean distance of image block gray value is used as calculating similitude no longer
It is applied to the property taken advantage of model, and passes through knowable to strict theory deduction, ratio distance is more suitable for the property taken advantage of model, can keep pixel
Between true relevance.But when calculating similitude using ratio distance, the similitude that distance is closer to 1 is bigger, and from 1
More remote similitude is less, therefore can not directly using ratio distance as measurement two image block similarities module.
Traditional coherent speckle suppression method includes spatial domain class, small echo, anisotropy parameter and non-local mean method.
Spatial domain class speckle suppression method occurred smoothing in strong edge and texture region, caused rejection to reduce；Little
Can part in radar image and disturb block in ripple and anisotropy parameter Speckle reduction, and impact is visual；Non local at present equal
Value method only passes through ratio distance metric similitude, and Speckle reduction performance is not sufficiently stable.
Content of the invention
The embodiment of the present invention provides a kind of non-local mean radar imaging coherent spot suppressing method based on ratio distance, with
Overcome the problems referred to above.
A kind of non-local mean radar imaging coherent spot suppressing method based on ratio distance of the present invention, including：
Read synthetic aperture radar coherent spot image；
Set corresponding search window and similitude window size；
According to the gray value in described any two similitude window determine described any two similitude window ratio away from
From；
The gray value of the central point according to described any two similitude window determines described any two similitude window
The gray value ratio of central point；
The space coordinates of the central point according to described any two similitude window determines described any two similitude window
The space length of mouth central point；
Two phases are calculated according to described ratio distance, described central point gray value ratio and described central point space length
Weighted value like pixel in property window；
Estimate the estimate of current pixel point by described weighted value.
Further, described determined according to the gray value in similitude window size and described similitude window described two
Similitude window ratio distance, including：
Using formula
Calculate two image block ratio distances, wherein, described W is the size of similitude window, wherein, NiWith NjPoint
It is not with pixel QiWith QjCentered on image block, andWithIt is then two image blocks in observed image f, D1 (i, j)
For two image block ratio distances, (i, j) is respectively the position coordinates of any two image block central pixel point, and M is arbitrary figure
As pixel number in block.
Further, the gray value of the described central point according to described any two similitude window determines described any two
The gray value ratio of central point between individual similitude window, including：
Using formula
Calculate the gray value ratio between two central pixel point, wherein, described (x, y) and (p, q) is image block respectively
NiWith NjCentral pixel point position, D2 (i, j) is the gray value ratio between two central pixel point.
Further, the space coordinates of the central point according to described any two similitude window determines described any two
The space length of similitude window center point, including：
Using formula
D3 (i, j)=(x-p)2+(y-q)2(3)
Determine the space length of any two similitude window center point, wherein, (x, y) and (p, q) is two phases respectively
Like the center position coordinate of property window, D3 (i, j) is the space length of two similitude window center points.
Further, described according to described ratio distance, described central point gray value ratio and the described center space of points
Distance calculates weighted value, including：
Using formula
Calculate the weighted value of pixel in two similitude windows, wherein, described h1, h2 and h3 are to control filtering journey respectively
The variable of degree, the variable-definition controlling filter strength is h1=h2=h3=kL σ, and for image regarding number, k is constant to L, and σ is
The standard deviation of image, w (i, j) is the weighted value of pixel in two similitude windows.
The ratio distance of the present invention can robustly be adapted to the property taken advantage of model, by by this ratio distance with reference to image block
Central point gray value information and spatial positional information preferably to calculate the similitude between image block.The present invention has been not only able to
The suppression coherent speckle noise of effect, moreover it is possible to effectively keep edge, improves performance indications, and visual more preferable.
Brief description
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Have technology description in required use accompanying drawing be briefly described it should be apparent that, drawings in the following description are these
Some bright embodiments, for those of ordinary skill in the art, without having to pay creative labor, acceptable
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is the non-local mean radar imaging coherent spot suppressing method flow chart based on ratio distance for the present invention；
Fig. 2 is three kinds of position image block schematic diagrames of the present invention；
Fig. 3 is three kinds of position image block schematic diagrames of the present invention；
Fig. 4 is the Speckle reduction comparison diagram based on composograph for the present invention
Fig. 5 is the Speckle reduction comparison diagram based on true radar image for the present invention；
Fig. 6 is that the equivalent number of composograph of the present invention keeps angle value with h1 change curve with the edge based on ratio of averages
Figure；
Fig. 7 is that the equivalent number of the true radar image of the present invention keeps angle value to change with h1 with the edge based on ratio of averages
Curve map.
Specific embodiment
Purpose, technical scheme and advantage for making the embodiment of the present invention are clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described it is clear that described embodiment is
The a part of embodiment of the present invention, rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment being obtained under the premise of not making creative work, broadly falls into the scope of protection of the invention.
Fig. 1 is the non-local mean radar imaging coherent spot suppressing method flow chart based on ratio distance for the present invention, such as Fig. 1
Shown, the method for the present embodiment can include：
Step 101, reading synthetic aperture radar coherent spot image；
Step 102, the corresponding search window of setting and similitude window size；
Step 103, described any two similitude window is determined according to the gray value in described any two similitude window
Mouth ratio distance；
Step 104, described any two phase is determined according to the gray value of the central point of described any two similitude window
Gray value ratio like property window center point；
Step 105, described any two is determined according to the space coordinates of the central point of described any two similitude window
The space length of similitude window center point；
Step 106, according to described ratio distance, described central point gray value ratio and described central point space length meter
Calculate the weighted value of pixel in two similitude windows；
Step 107, estimate the estimate of current pixel point by described weighted value.
Further, described determined according to the gray value in similitude window size and described similitude window described two
Similitude window ratio distance, including：
Using formula
Calculate two image block ratio distances, wherein, described W is the size of similitude window, wherein, NiWith NjPoint
It is not with pixel QiWith QjCentered on image block, andWithIt is then two image blocks in observed image f, D1 (i, j)
For two image block ratio distances, (i, j) is respectively the position coordinates of any two image block central pixel point, and M is arbitrary figure
As pixel number in block.
Specifically, the meaning of defined formula (1) is, takes advantage of in order to can the range formula of analytical formula (1) be applied to
Property model, so we adjust the distance now, the mathematic expectaion of formula D1 (i, j) carries out mathematical derivation, and derivation is as follows：
Here our assumed conditions of employing are：We this
In formula (2) is derived following expression：
Equally, coherent speckle noise v is to obey average for 1, and variance isRandom noise, so,
Formula (4) is brought in formula (3), obtains：
Formula (5) presents whether in homogeneous region or in heterogeneous areas, observed image f and not polluted by coherent spot
True picture u between corresponding image block the mathematic expectaion of distance remain identical similitude between pixel, have more
Good robustness and validity.
Further, the gray value of the described central point according to described any two similitude window determines described any two
The gray value ratio of central point between individual similitude window, including：
Using formula
Calculate the gray value ratio between two central pixel point, wherein, described (x, y) and (p, q) is image block respectively
NiWith NjCentral pixel point position, D2 (i, j) is the gray value ratio between two central pixel point.
Further, the space coordinates of the central point according to described any two similitude window determines described any two
The space length of similitude window center point, including：
Using formula
D3 (i, j)=(x-p)2+(y-q)2(7)
Determine the space length of any two similitude window center point, wherein, (x, y) and (p, q) is two phases respectively
Like the center position coordinate of property window, D3 (i, j) is the space length of two similitude window center points.
Specifically, as image block NiWith image block NjWhen completely similar, new apart from D1 (i, j)=0.If two images
Block is more dissimilar, then bigger apart from D1 (i, j).Can be directly used for the weight computing of two image blocks apart from D1, but in order to improve
The accuracy of edge pixel point estimate, in addition increased two conditions：The gray value of central point between two similitude windows
The ratio D2 and space length D3 of two similitude window center points, is illustrated with emulating image below.
As shown in Fig. 2 in figure has two homogeneous regions of A, B, 1,2,3 represent three image blocks, wherein in the middle of each block
Round dot represents the central pixel point of this block, and No. 1 image block is completely in A homogeneous region, and No. 2 image blocks are completely in B homogeneity
In region, and No. 3 image blocks are located between two homogeneous regions of A, B, but the central pixel point of No. 3 image blocks is located at B homogeneity
In region.If calculating the distance between the distance between 1 and 3 and 2 and 3 using D1, two, apart from almost equal, have D1 (1,3)
≈D1(2,3).This is because the left side approximate similarity of image block 1 and image block 3, the right dissmilarity；And image block 2 and image block
3 situation situation is just contrary with the former, the right of image block 3 and image block 2 approximate similarity, and the left side is dissimilar.If according to
This distance is to measure out image block 1 and the difference to image block 3 for the image block 2, easily causes estimation marginal value
Inaccurate.So, in the present embodiment it is also contemplated that the gray value information of central point in image block, namely D2 in formula (2)
(i,j).Take the maximum that the central point gray value of two image blocks is compared, if two blocks are in same homogeneous region, D2 (i,
J) less；If two blocks are in two different regions, the pixel value meeting difference of two central points is very big, leads to D2 (i, j)
Also larger.Above-mentioned conclusion can be verified from Fig. 2, that is, the central pixel point of 1 and 3 image blocks not in same homogeneous region, institute
With, the central point grey value difference of two blocks larger it is possible to have the central spot of a block in another different homogeneity area
Near domain, so the D2 (i, j) obtaining is larger.But the central pixel point of image block 2 and 3 is in identical homogeneous region, institute
Smaller with D2 (i, j), tend to 0 possibility than larger.
If as shown in figure 3, image block NiOr NjHave one to contain the more rich texture information of ratio, and two image blocks it
Between apart distant, now two image blocks should have smaller similitude, preferably both similitudes are 0.With
Sample is taking emulating image as a example.
There are 1,2,3 three image blocks in Fig. 3, and three image blocks are located in same region of search, image block 2 and figure
As all comprising edge in block 3, and the central point of three image blocks is all located at same homogeneous region.If using D1 (i, j) and
D2 (i, j) calculates the distance that image block 1 is with image block 2 and the distance of image block 1 and image block 3, then can obtain D1 (1,2) ≈ D1
And D2 (1,2) ≈ D2 (1,3) (1,3).Comprise edge due in image block 3, and range image block 1 is farther out, so we are uncommon
Hope that image block 3 and image block 1 just have less similitude.So, herein further according to the locus of image block central point, build
A kind of space length measure formulas D3 (i, j), shown in its expression formula such as formula (7).
Further, described according to described ratio distance, described central point gray value ratio and the described center space of points
Distance calculates weighted value, including：
Using formula
Calculate the weighted value of pixel in two similitude windows, wherein, described h1, h2 and h3 are to control filtering journey respectively
The variable of degree, the variable-definition controlling filter strength is h1=h2=h3=kL σ, and for image regarding number, k is constant to L, and σ is
The standard deviation of image, w (i, j) is the weighted value of pixel in two similitude windows.
According to the estimate of the weighted value calculating pixel i of pixel in above-mentioned two similitude window it is：
In order to the performance of the property the taken advantage of non-local mean Speckle Reduction Algorithm each side based on ratio and space length is described,
Tested for two class images, one type is composograph, another kind of is true radar image.Respectively using spatial domain
Frost Speckle Reduction Algorithm is (hereinafter referred to as：Frost), anisotropy SRAD Speckle Reduction Algorithm is (hereinafter referred to as：
SRAD), the improvement SRAD Speckle Reduction Algorithm based on Gaussian-Gamma double window is (hereinafter referred to as：) and SAR- GGSRAD
Four kinds of relevant Restrainable algorithms of BM3D Speckle Reduction Algorithm and this paper algorithm carry out Experimental comparison.
First, composograph
As shown in figure 4, it can be seen that this paper algorithm whether suppresses the effect of coherent spot or edge keeps
Effect is all relatively good.
In wherein Fig. 4, (a) regards composograph artwork for 16, and size is that in 300 × 300, Fig. 5, (a) regards composograph for 8,
Size is 256 × 256.Fig. 4 b is the Frost algorithm process result for Fig. 4 a, and in Fig. 4, (c) is the SRAD algorithm for Fig. 4 a
Speckle reduction result, in Fig. 4, (d) is the GGSRAD algorithm Speckle reduction result for (a) in Fig. 4, in Fig. 4
E () is the radar-BM3D algorithm Speckle reduction result for (a) in Fig. 4, in Fig. 4, (f) is that the application method is directed to figure
(a) Speckle reduction result in 4.
For composograph, the parameter setting of various suppression coherent speckle noise algorithms is as follows：Frost Speckle reduction
The local window yardstick of algorithm is 7 × 7；SRAD algorithm, GGSRAD algorithm all adopt 3 × 3 local windows, and iterations is 100, step-length
For 0.1；In SAR-BM3D algorithm, search window is 21 × 21, and similitude window is 7 × 7, controls the variable h1=of filter strength
H2=h3=kL σ, k=3.
2nd, true picture
As shown in figure 5, it can be seen that the ratio that suppresses in terms of Speckle reduction of this paper algorithm more thoroughly, meanwhile,
Fuzzy or excess smoothness phenomenon yet in edge.
Wherein, in Fig. 5, (a) regards true radar image for 5, and image size is all 256 × 256.In Fig. 5, (b) is for Fig. 5
In (a) Frost algorithm Speckle reduction result, in Fig. 5, (c) is the SRAD algorithm Speckle reduction for (a) in Fig. 5
Result, in Fig. 5, (d) is the GGSRAD algorithm Speckle reduction result for (a) in Fig. 5, and in Fig. 5, (e) is to be directed to
Radar-BM3D algorithm Speckle reduction the result of (a) in Fig. 5, in Fig. 5, (f) is that the application method is directed to (a) phase in Fig. 5
Dry spot suppresses result.
For true radar image, the parameter setting of various suppression coherent speckle noise algorithms is as follows：Frost coherent spot
The local window yardstick of Restrainable algorithms is 7 × 7；SRAD algorithm, GGSRAD algorithm all adopt 3 × 3 local windows, and iterations is 80,
Step-length is 0.1；Parameter used in radar-BM3D algorithm is default value in bibliography；The search window of this paper algorithm is
21 × 21, similitude window is 7 × 7, controls the variable h1=h2=h3=kL σ, k=5 of filter strength.
(it is abbreviated as V using equivalent number hereinENL) and V (is abbreviated as based on the edge conservation degree of ratio of averagesEPD_ROA) two
Individual performance indications, the image containing coherent spot multiplicative noise to composograph and true radar image carries out pressing down spot Performance Evaluation.
Obtain table 1 and all kinds of algorithm suppression spot effect performance comparison result shown in table 2.Wherein, the V of calculatingENLValue is to select each image
In two pieces of homogeneous regions, respectively region 1 and region 2, the V calculatingENLValue is bigger, just represents removal noise immune strong；
VEPD_ROAIn HD and VD represent the V being calculated both horizontally and vertically in selected areas respectivelyEPD_ROAValue, this two
Value is closer to 1, more represents edge holding capacity strong.Table 1 is for composograph, this paper algorithm and contrast algorithm performance ratio
Relatively.
Table 1
In table 1, first is classified as the Image Name for radar image and contrast algorithm and this paper algorithm title, and second is classified as institute
Select the equivalent number value in two regions, be designated as VENL, the 3rd is classified as edge holding capacity, is designated as VEPD_ROA.Ratio due to this paper
Distance is applied to Multiplicative noise model, and also add the gray value information of central point and the constraint of spatial positional information, makes
Obtain image whether all relatively good in terms of edge holding or in terms of Speckle reduction.For composograph, herein
The ENL value of algorithm is more much higher than original classic algorithm.In addition to radar-BM3D algorithm, the edge of this paper keeps index
Also it is higher than other algorithms.From the point of view of synthetic image and performance indications, this paper algorithm can be realized suppressing keeping while coherent spot
Edge.Table 2 is for true radar image, and this paper algorithm is compared with contrast algorithm performance.
Table 2
In table 2, first is classified as input picture title and each Speckle Reduction Algorithm title, and second is classified as in selected areas
Equivalent number (the V calculatingENL), represent the ability that Speckle Reduction Algorithm suppresses noise；3rd edge being classified as calculating keeps energy
Power (VEPD_ROA), represent the marginal texture after filtering and keep degree.Because this paper algorithm is by ratio distance presented herein
Replace original Euclidean distance and calculate weight, this ratio distance can robustly be applied in Multiplicative noise model.Meanwhile, in order to
Protect edge information information, the structure that this paper algorithm is directed to edge has done corresponding process.So, whether directly observe image also
It is to compare performance indications, this paper algorithm generally will contrast algorithm better than other.
The parameter of setting manually is herein needed to have control filter strength variable h1, h2, h3.Due to for a class image
Each algorithm in parameter h1, h2, h3 be identical numerical value.So, next this section discusses this parameter to radar image
The impact of performance.It is analyzed for composograph and true radar image, wherein h1={ 0.1,1,3,5,7,9,11,15 },
The value of ENL is by calculating selected two pieces of homogeneous regions in image.
As shown in Figure 6 and Figure 7, being gradually increased with parameter h1, whether to composograph or true radar image
ENL value be all gradually increased, when, to after certain numerical value, the speed of increase becomes slow, and EPD_ROA value is then gradually to subtract
Little.Thus illustrate, h1 is bigger, Speckle reduction ability is strong, but edge holding capacity is weaker, conversely, h1 is less, coherent spot presses down
Ability processed is weak, and edge keeps effect fine.So, select a suitable h1 it is critical that.
The present invention ensure that by the identical similitude between the image of noise pollution and noise-free picture pixel.Meanwhile, draw
Enter the gray value information of the central point of image block and the spatial positional information of central point, process image block can be maintained at simultaneously
When comprising substantial amounts of texture information in edge or image block, Protect edge information and texture information.So, this paper algorithm is in phase
Edge is maintained well while dry spot suppression.
Finally it should be noted that：Various embodiments above only in order to technical scheme to be described, is not intended to limit；To the greatest extent
Pipe has been described in detail to the present invention with reference to foregoing embodiments, it will be understood by those within the art that：Its according to
So the technical scheme described in foregoing embodiments can be modified, or wherein some or all of technical characteristic is entered
Row equivalent；And these modifications or replacement, do not make the essence of appropriate technical solution depart from various embodiments of the present invention technology
The scope of scheme.
Claims (5)
1. a kind of non-local mean radar imaging coherent spot suppressing method based on ratio distance is it is characterised in that include：
Read synthetic aperture radar coherent spot image；
Set corresponding search window and similitude window size；
Described any two similitude window ratio distance is determined according to the gray value in described any two similitude window；
The gray value of the central point according to described any two similitude window determines described any two similitude window center
The gray value ratio of point；
The space coordinates of the central point according to described any two similitude window determines in described any two similitude window
The space length of heart point；
Two similitudes are calculated according to described ratio distance, described central point gray value ratio and described central point space length
The weighted value of pixel in window；
Estimate the estimate of current pixel point by described weighted value.
2. the method according to right 1 it is characterised in that described according in similitude window size and described similitude window
Gray value determine described two similitude window ratio distances, including：
Using formula
Calculate two image block ratio distances, wherein, described W is the size of similitude window, wherein, NiWith NjIt is respectively
With pixel QiWith QjCentered on image block, andWithIt is then two image blocks in observed image f, D1 (i, j) is two
Individual image block ratio distance, (i, j) is respectively the position coordinates of any two image block central pixel point, and M is any image block
Interior pixel number.
3. method according to claim 1 is it is characterised in that the described center according to described any two similitude window
The gray value of point determines the gray value ratio of central point between described any two similitude window, including：
Using formula
Calculate the gray value ratio between two central pixel point, wherein, described (x, y) and (p, q) is image block N respectivelyiWith Nj
Central pixel point position, D2 (i, j) is the gray value ratio between two central pixel point.
4. method according to claim 1 is it is characterised in that according to the central point of described any two similitude window
Space coordinates determines the space length of described any two similitude window center point, including：
Using formula
D3 (i, j)=(x-p)2+(y-q)2(3)
Determine the space length of any two similitude window center point, wherein, (x, y) and (p, q) is two similitudes respectively
The center position coordinate of window, D3 (i, j) is the space length of two similitude window center points.
5. method according to claim 1 it is characterised in that described according to described ratio distance, described central point gray scale
Value ratio and described central point space length calculate weighted value, including：
Using formula
Calculate the weighted value of pixel in two similitude windows, wherein, described h1, h2 and h3 are to control filter strength respectively
Variable, the variable-definition controlling filter strength is h1=h2=h3=kL σ, and for image regarding number, k is constant to L, and σ is image
Standard deviation, w (i, j) is the weighted value of pixel in two similitude windows.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610810707.8A CN106447629B (en) | 2016-09-08 | 2016-09-08 | Non-local mean radar imaging coherent spot suppressing method based on ratio distance |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610810707.8A CN106447629B (en) | 2016-09-08 | 2016-09-08 | Non-local mean radar imaging coherent spot suppressing method based on ratio distance |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106447629A true CN106447629A (en) | 2017-02-22 |
| CN106447629B CN106447629B (en) | 2019-07-09 |
Family
ID=58165376
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610810707.8A Expired - Fee Related CN106447629B (en) | 2016-09-08 | 2016-09-08 | Non-local mean radar imaging coherent spot suppressing method based on ratio distance |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106447629B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005189099A (en) * | 2003-12-25 | 2005-07-14 | National Institute Of Information & Communication Technology | Method and device for removing noise in sar data processing |
| EP2146315A1 (en) * | 2008-07-16 | 2010-01-20 | Galileian Plus s.r.l. | Method of filtering SAR images from speckle noise and related device. |
| CN102073989A (en) * | 2010-11-09 | 2011-05-25 | 西安电子科技大学 | Speckle suppression method for polarized SAR (Synthetic Aperture Radar) data based on non-local mean value fused with PCA (Polar Cap Absorption) |
| CN103020919A (en) * | 2013-01-09 | 2013-04-03 | 西安电子科技大学 | Polarimetric SAR (synthetic aperture radar) phase speckled noise suppression method based on non-local Lee |
| CN103325090A (en) * | 2012-09-10 | 2013-09-25 | 中国科学院电子学研究所 | Method and device for restraining speckles |
| CN104680536A (en) * | 2015-03-09 | 2015-06-03 | 西安电子科技大学 | Method for detecting SAR image change by utilizing improved non-local average algorithm |
| CN104867120A (en) * | 2015-06-02 | 2015-08-26 | 西安电子科技大学 | Ratio distribution-based sar image non-local despeckling method |
-
2016
- 2016-09-08 CN CN201610810707.8A patent/CN106447629B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005189099A (en) * | 2003-12-25 | 2005-07-14 | National Institute Of Information & Communication Technology | Method and device for removing noise in sar data processing |
| EP2146315A1 (en) * | 2008-07-16 | 2010-01-20 | Galileian Plus s.r.l. | Method of filtering SAR images from speckle noise and related device. |
| CN102073989A (en) * | 2010-11-09 | 2011-05-25 | 西安电子科技大学 | Speckle suppression method for polarized SAR (Synthetic Aperture Radar) data based on non-local mean value fused with PCA (Polar Cap Absorption) |
| CN103325090A (en) * | 2012-09-10 | 2013-09-25 | 中国科学院电子学研究所 | Method and device for restraining speckles |
| CN103020919A (en) * | 2013-01-09 | 2013-04-03 | 西安电子科技大学 | Polarimetric SAR (synthetic aperture radar) phase speckled noise suppression method based on non-local Lee |
| CN104680536A (en) * | 2015-03-09 | 2015-06-03 | 西安电子科技大学 | Method for detecting SAR image change by utilizing improved non-local average algorithm |
| CN104867120A (en) * | 2015-06-02 | 2015-08-26 | 西安电子科技大学 | Ratio distribution-based sar image non-local despeckling method |
Non-Patent Citations (2)
| Title |
|---|
| HONGXIAO FENG ET AL: "SAR Image Despeckling Based on Local Homogeneous-Region Segmentation by Using Pixel-Relativity Measurement", 《IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING》 * |
| 王爽: "基于双边滤波的极化SAR相干斑抑制", 《雷达学报》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106447629B (en) | 2019-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103353985B (en) | A kind of Measurement Method of image Gaussian Blur | |
| CN103369209B (en) | Vedio noise reduction device and method | |
| CN103020906B (en) | A kind of preprocess method of star sensor measuring star in daytime image | |
| CN108681992A (en) | The image interpolation algorithm of laser facula is measured for detector array method | |
| CN102129704A (en) | SURF operand-based microscope image splicing method | |
| CN103119623A (en) | Pupil detection device and pupil detection method | |
| DE102013211930A1 (en) | Binocular broad base line object comparison method using a minimal cost flownet | |
| CN104252708B (en) | A kind of x-ray chest radiograph image processing method and system | |
| CN105989347A (en) | Intelligent marking method and system of objective questions | |
| US11093778B2 (en) | Method and system for selecting image region that facilitates blur kernel estimation | |
| CN107784651A (en) | A kind of blurred picture quality evaluating method based on fuzzy detection weighting | |
| CN109064418A (en) | A kind of Images Corrupted by Non-uniform Noise denoising method based on non-local mean | |
| CN103177451A (en) | Three-dimensional matching algorithm between adaptive window and weight based on picture edge | |
| CN108550145A (en) | A kind of SAR image method for evaluating quality and device | |
| CN103826032A (en) | Depth map post-processing method | |
| CN103996029A (en) | Expression similarity measuring method and device | |
| CN104778684A (en) | Method and system thereof for automatically measuring, representing and classifying heterogeneous defects on surface of steel | |
| CN110378853A (en) | Depth map treating method and apparatus | |
| CN106169173A (en) | A kind of image interpolation method | |
| CN106204454A (en) | High accuracy rapid image interpolation method based on texture edge self-adaption data fusion | |
| CN105354863A (en) | Adaptive scale image sequence target tracking method based on feature filtering and fast motion detection template prediction | |
| US20080131002A1 (en) | Rapid and high precision centroiding method and system for spots image | |
| CN108921170A (en) | A kind of effective picture noise detection and denoising method and system | |
| CN107194902A (en) | The method that wave filter parameter is automatically determined in image co-registration | |
| CN110084818A (en) | Dynamic down-sampled images dividing 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 | ||
| GR01 | Patent grant | ||
| 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: 20190709 Termination date: 20200908 |