CN106920222A - A kind of image smoothing method and device - Google Patents
A kind of image smoothing method and device Download PDFInfo
- Publication number
- CN106920222A CN106920222A CN201710146581.3A CN201710146581A CN106920222A CN 106920222 A CN106920222 A CN 106920222A CN 201710146581 A CN201710146581 A CN 201710146581A CN 106920222 A CN106920222 A CN 106920222A
- Authority
- CN
- China
- Prior art keywords
- image
- smoothed
- default
- original image
- function
- 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
- 238000000034 method Methods 0.000 title claims abstract description 154
- 238000003706 image smoothing Methods 0.000 title claims abstract description 49
- 238000001914 filtration Methods 0.000 claims abstract description 81
- 238000009499 grossing Methods 0.000 claims abstract description 50
- 230000002146 bilateral effect Effects 0.000 claims abstract description 34
- 230000004087 circulation Effects 0.000 claims description 18
- 238000011084 recovery Methods 0.000 claims description 12
- 238000010276 construction Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 27
- 238000012545 processing Methods 0.000 abstract description 18
- 230000008901 benefit Effects 0.000 abstract description 7
- 239000000470 constituent Substances 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 69
- 239000013598 vector Substances 0.000 description 11
- 238000005516 engineering process Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000008859 change Effects 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- 238000000605 extraction Methods 0.000 description 4
- 239000011159 matrix material Substances 0.000 description 4
- 239000000203 mixture Substances 0.000 description 4
- 230000004075 alteration Effects 0.000 description 3
- 230000001965 increasing effect Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000004321 preservation Methods 0.000 description 2
- 230000002633 protecting effect Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000021615 conjugation Effects 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000012163 sequencing technique Methods 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000002834 transmittance Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20048—Transform domain processing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
A kind of image smoothing method is the embodiment of the invention provides, original image is smoothed first with the bilateral filtering in local smoothing method method and transform domain filtering, obtain navigational figure；Then least square model is constructed to original image, navigational figure and default smoothed image using least square method, adds the constraint function to default smoothed image to control the degree of rarefication of smoothed image, obtain smoothed energy object function；Finally using half secondary split method and alternately, fixed variable method solves the function, so as to obtain original image by the smoothed image after smoothing processing.Consider global characteristics and local feature, enhance the protection of constituent in original image, while having recovered the details of some high-contrasts, the image smoothing effect for obtaining is conducive to improving the accuracy rate of image recognition.Additionally, the embodiment of the present invention is additionally provided realizes device accordingly, further such that methods described has more practicality, described device has corresponding advantage.
Description
Technical field
The present embodiments relate to technical field of image processing, more particularly to a kind of image smoothing method and device.
Background technology
With the fast development of computer image processing technology, Image Smoothing Skill is in order to meet present image treatment technology
The high request in field, has also obtained developing faster.Image Smoothing Skill refer to for protrude the enlarged regions of image, low frequency into
Point, trunk portion or suppress picture noise and interference radio-frequency component, make the gentle gradual change of brightness of image, reduce mutation gradient, improve
The image processing techniques of picture quality, is widely used in image segmentation, denoising, details enhancing, target classification, edge extracting etc.
In field.
Image is often inevitably subject to noise, unessential details (especially high in acquisition and transmittance process
Contrast) interference so that there is the unconspicuous problem of contour feature in target to be identified, and difficulty is brought to identification.Utilize
Image Smoothing Skill carries out treatment to image can to a certain extent evade interference, so as to improve the success rate of image recognition.
In the prior art, image smoothing method is generally local smoothing method method and global smoothing method.Local smoothing method side
Method, such as gaussian filtering, bilateral filtering, medium filtering, transform domain filtering etc., refer to that the regional area or patch of image are carried out
Treatment., only with respect to the local features of image, the smooth effect to image local is preferable, but easily makes for local smoothing method method
Into structural fuzzy problem.Global exponential smoothing, such as full variation smoothing algorithm, weighted least mean square smoothing algorithm, L0Gradient minimisation
Smoothing algorithm etc., is that all regions simultaneously to whole image are processed.The structure of global exponential smoothing keeps the excellent of bound term
Change framework more flexible, be better than local smoothing method method, the particularly background portion to image for the global characteristics of image and grade not
Important details, but smooth effect of the global smoothing algorithm often to local high-contrast noise is poor.
In sum, local smoothing method method and global smoothing method respectively have quality, when being smoothed to image,
How integrated application global characteristics and local feature, evade the inferior position of local smoothing method method and global smoothing method, to carry
The robustness of hi-vision smoothing algorithm, the image smoothing effect for having obtained is those skilled in the art's problem demanding prompt solution.
The content of the invention
The purpose of the embodiment of the present invention is to provide a kind of image smoothing method and device, to improve the Shandong of smooth algorithm
Rod, the image smoothing effect for having obtained.
In order to solve the above technical problems, the embodiment of the present invention provides following technical scheme：
On the one hand the embodiment of the present invention provides a kind of image smoothing method, including：
Bilateral filtering and the transform domain filtering of preset times are carried out to original image circulation, to obtain navigational figure；
Least square is constructed to the original image, the navigational figure and default smoothed image according to least square method
Model；
The image pixel intensities and gradient of the default smoothed image are defined using norm, to obtain the default smoothed image
Constraint function；
Smoothed energy object function is obtained according to the least square model and the constraint function；
The smoothed energy object function is solved using half secondary split method and alternately fixed variable method, it is smooth to obtain
Image.
Optionally, it is described according to least square method to the original image, the navigational figure and default smoothed image
Constructing least square model is：
According to the least square method and 2- norms square to the original image, the navigational figure and default
Smoothed image constructs least square model：
In formula, S is the default smoothed image, and I is the original image, and G is the navigational figure, and α is detail recovery
The factor.
Optionally, the image pixel intensities and gradient of the utilization norm definition default smoothed image are：
The image pixel intensities and gradient of the default smoothed image are defined using 0- norms.
Optionally, it is described that smoothed energy object function is obtained according to the least square model and the constraint function
For：
Obtaining smoothed energy object function according to the least square model and the constraint function is：
In formula, E (S) is the smoothed energy object function, and S is the default smoothed image, and I is the original image, G
It is the navigational figure, α is the detail recovery factor, and λ is smoothing factor, | | ▽ S | |0It is the constraint function.
Optionally, it is described according to least square method to the original image, the navigational figure and default smoothed image
Constructing least square model is：
According to the least square method to the original image by medium filtering, the navigational figure and default smooth figure
As construction least square model.
Optionally, bilateral filtering and the transform domain filtering that preset loop number of times is carried out to original image circulation, with
Obtaining navigational figure includes：
S1：Obtain the original image, constant value image, space weighted value and scope weighted value；
S2：It is initial guide function with the constant value image, according to transform domain filter method to the original image and institute
Stating initial guide function carries out transform domain filtering, obtains new guidance function；
S3：Bilateral filtering is carried out to the original image and the new guidance function according to bilateral filtering method, one is obtained
Secondary guidance function；
S4：The preset times are performed to S2 and S3 circulations, to obtain the navigational figure.
Optionally, the preset times are 3 times.
Optionally, the space weighted value and scope weighted value are：
The space weighted value is 3；
The scope weighted value is 0.01.
Optionally, it is described to utilize half secondary split method and alternately the fixed variable method solution smoothed energy target letter
Number, is included with obtaining smoothed image：
It is introduced into the constraint function that auxiliary variable is replaced in the smoothed energy object function；
Minimum treatment is carried out to the smoothed energy object function replaced using the half secondary split method, error is added
Penalty term, obtains smooth minimum model；
The smooth minimum model is solved according to the alternately fixed variable method, to obtain the smoothed image.
On the other hand the embodiment of the present invention provides a kind of image smoothing device, including：
Filtering module, bilateral filtering and the transform domain filtering for carrying out preset times to original image circulation,
To obtain navigational figure；
Model module is set up, for according to least square method is to the original image, the navigational figure and presets flat
Sliding image configuration least square model；
Smoothed image module is obtained, image pixel intensities and gradient for defining the default smoothed image using norm,
To obtain the constraint function of the default smoothed image；Smoothed according to the least square model and the constraint function
Energy object function；The smoothed energy object function is solved using half secondary split method and alternating fixed variable method, to obtain
Obtain smoothed image.
A kind of image smoothing method is the embodiment of the invention provides, first with the bilateral filtering in local smoothing method method and change
Change domain filtering to be smoothed original image, obtain navigational figure；Then using least square method to original image, guiding
Image and default smoothed image construction least square model, add to the constraint function for presetting smoothed image to control smooth figure
The degree of rarefication of picture, obtains smoothed energy object function；Finally being solved using half secondary split method and alternately fixed variable method should
Function, so as to obtain original image by the smoothed image after smoothing processing.
The technical scheme that the application is provided, considers global characteristics and local feature, first carries out office to original image
Portion's smoothing processing, then carries out global smoothing processing, has evaded the inferior position of local smoothing method method and global smoothing method, effectively
Make use of the advantage of the two.By controlling difference and control smoothed image and guiding figure between smoothed image and original image
Aberration is different, enhances the protection to constituent in original image, remains the structure of image, in removal detail textures feature
The details of some high-contrasts, the image smoothing effect for obtaining are recovered simultaneously；Additionally, having carried out effective making an uproar to image
Sound is filtered, and strengthens the intensity of boundary pixel, is conducive to the extraction of image outline, so as to be conducive to improving the accurate of image recognition
Rate and efficiency.
Additionally, the embodiment of the present invention is provided also directed to image smoothing method realizes device accordingly, further such that institute
Method is stated with more practicality, described device has corresponding advantage.
Brief description of the drawings
For the clearer explanation embodiment of the present invention or the technical scheme of prior art, below will be to embodiment or existing
The accompanying drawing to be used needed for technology description is briefly described, it should be apparent that, drawings in the following description are only this hair
Some bright embodiments, for those of ordinary skill in the art, on the premise of not paying creative work, can be with root
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1-1 is the original image of an illustrative example provided in an embodiment of the present invention；
Fig. 1-2 is the image after the smoothed treatment of original image in Fig. 1-1 provided in an embodiment of the present invention；
Fig. 2 is a kind of schematic flow sheet of image smoothing method provided in an embodiment of the present invention；
Fig. 3 is the original image of another illustrative example provided in an embodiment of the present invention；
Fig. 4 is that original image is smoothed by the different space weighted values of input in Fig. 3 provided in an embodiment of the present invention
The image that treatment is obtained；
Fig. 5 is that original image is smoothed by the different scope weighted values of input in Fig. 3 provided in an embodiment of the present invention
The image that treatment is obtained；
Fig. 6 is that original image is put down by the different detail recovery factors of input in Fig. 3 provided in an embodiment of the present invention
The image that sliding treatment is obtained；
Fig. 7 is that original image carries out smooth place by the different smoothing factors of input in Fig. 3 provided in an embodiment of the present invention
The image that reason is obtained；
Fig. 8 is a kind of implementation method structure chart of image smoothing device provided in an embodiment of the present invention；
Fig. 9 is the original image of another illustrative example provided in an embodiment of the present invention；
Figure 10 is that original image is smoothed by RGF and BLF algorithms and obtains in Fig. 9 provided in an embodiment of the present invention
Image；
Figure 11 is that original image is smoothed the figure for obtaining by RTV algorithms in Fig. 9 provided in an embodiment of the present invention
Picture；
Original image is smoothed by NLGRTV algorithms and obtains in Figure 12 Fig. 9 provided in an embodiment of the present invention
Image；
Original image is smoothed by SSPTF algorithms and obtains in Figure 13 Fig. 9 provided in an embodiment of the present invention
Image；
Figure 14 in Fig. 9 provided in an embodiment of the present invention original image be smoothed by the algorithm that the application is provided
The image for obtaining；
Figure 15 is the original image of another illustrative example provided in an embodiment of the present invention；
Figure 16 is that original image carries out denoising and obtains by RGF and BLF algorithms in Figure 15 provided in an embodiment of the present invention
The image for arriving；
Figure 17 in Figure 15 provided in an embodiment of the present invention original image carried out at denoising by the algorithm that the application is provided
The image that reason is obtained；
Figure 18 is the original image of another illustrative example provided in an embodiment of the present invention；
Figure 19 in Figure 19 provided in an embodiment of the present invention original image carry out image increasing by the algorithm that the application is provided
The image that strength reason is obtained.
Specific embodiment
In order that those skilled in the art more fully understand the present invention program, with reference to the accompanying drawings and detailed description
The present invention is described in further detail.Obviously, described embodiment is only a part of embodiment of the invention, rather than
Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creative work premise
Lower obtained every other embodiment, belongs to the scope of protection of the invention.
Term " first ", " second ", " the 3rd " " in the description and claims of this application and above-mentioned accompanying drawing
Four " it is etc. for distinguishing different objects, rather than for describing specific order.In addition term " comprising " and " having " and
Their any deformations, it is intended that covering is non-exclusive to be included.For example contain the process of series of steps or unit, method,
System, product or equipment are not limited to the step of having listed or unit, but may include the step of not listing or unit.
Present inventor has found by research, local feature is often only considered in the prior art or only considers global special
Levy, or the image structure for causing smoothing processing to be crossed more is obscured, or being exactly bad treatment of details, local noise is larger, always
It, image smoothing effect is difficult to meet the requirement of technical field of image processing.In consideration of it, the application is by considering global spy
Levy and local feature, local smoothing method treatment first is carried out to original image, then carry out global smoothing processing, evaded local flat
The inferior position of sliding method and global smoothing method, effectively make use of the advantage of the two, the image smoothing effect for obtaining.
Technical scheme based on the embodiments of the present invention, first below with reference to Fig. 1 and Fig. 2 to the embodiment of the present invention
Some possible application scenarios that technical scheme is related to carry out citing introduction, and Fig. 1-1 is original graph provided in an embodiment of the present invention
Picture, Fig. 1-2 processed for the method provided by the application after image.
Smooth place is carried out to original image (Fig. 1-1) first with the bilateral filtering in local smoothing method method and transform domain filtering
Reason, obtains navigational figure；Then original image, navigational figure and default smoothed image are constructed using least square method minimum
Two multiply model, add the constraint function to default smoothed image to control the degree of rarefication of smoothed image, obtain smoothed energy target
Function；Finally using half secondary split method and alternately, fixed variable method solves the function, so as to obtain original image by flat
Smoothed image (Fig. 1-2) after sliding treatment.Comparison diagram 1-1 and Fig. 1-2 can be seen that the technical scheme of the application offer to figure
Be smoothed after, not only remain the structure of image, to have recovered some height right while detail textures feature is removed
Than the details of degree, good smooth effect has been obtained.
It should be noted that above-mentioned application scenarios are for only for ease of the thought and principle that understand the application and show, this
The implementation method of application is unrestricted in this regard.Conversely, presently filed embodiment can apply to it is applicable any
Scene.
After the technical scheme for describing the embodiment of the present invention, the various non-limiting reality of detailed description below the application
Apply mode.
Referring first to Fig. 2, Fig. 2 is a kind of schematic flow sheet of image smoothing method provided in an embodiment of the present invention, this hair
Bright embodiment may include herein below：
S201：Bilateral filtering and the transform domain filtering of preset times are carried out to original image circulation, to obtain guiding figure
Picture.
Bilateral filtering belongs to local smoothing method method with transform domain filtering.
Bilateral filtering is a kind of nonlinear filtering method, is the spatial neighbor degree and pixel value similarity for combining image
A kind of compromise, while considering spatial information (si) and grey similarity, reaches the purpose for protecting side denoising.With simple, non-iterative,
Local the characteristics of.The advantage of bilateral filtering can be to do Edge preservation, and general Wiener filtering or gaussian filtering in the past is gone
Noise reduction, all can relatively significantly fuzzy edge, the protecting effect for high frequency detail is obvious.Two-sided filter compares as its name suggests
The many Gauss variances of gaussian filtering, it is the Gaussian filter function based on spatial distribution, so in adjacent edges, from compared with
Remote pixel will not too much have influence on the pixel value on edge, this ensures that there the preservation of adjacent edges pixel value.But by
In excessive high-frequency information is saved, for the high-frequency noise in coloured image, what two-sided filter can not be clean filters, only
Can preferably be filtered for low-frequency information.
Find that when using only bilateral filtering, the navigational figure of acquisition can damage the side of primary structure by many experiments
Angle, but gained image smoothing effect is preferable；And when being filtered using only transform domain, the navigational figure of acquisition can preferably protect master
The corner of structure is wanted, but gained image smoothing effect is poor.Therefore the application is filtered to figure using bilateral filtering with transform domain simultaneously
As being processed, idiographic flow can be as follows：
S2011：Obtain original image, constant value image, space weighted value and scope weighted value；
S2012：It is initial guide function with constant value image, according to transform domain filter method to original image and initial guide
Function carries out transform domain filtering, obtains new guidance function；
S2013：Bilateral filtering is carried out to original image and new guidance function according to bilateral filtering method, one-step boot is obtained
Function；
S2014：The preset times are performed to S2012 and S2013 circulations, to obtain navigational figure.
Original image is treats smoothing processing image, can be the image of arbitrary format, such as tiff, tif, bmp, gif
Deng this does not influence the realization of technical scheme.
Constant value image is the image for being all constant 0.
Space weight σsAnd scope weight σrIt is the parameter used when being filtered to image, is typically using conversion
When domain filters, settable space weight is 1.5 times of bilateral filtering space weight, and scope weight can be set to bilateral filtering space right
3 times of weight.Certainly, can not be also configured according to above-mentioned parameter, those skilled in the art can be matched somebody with somebody according to actual conditions
Parameter is put, this does not influence the realization of the application.
Value (space weight σ on space weight and scope weights＞ 1, scope weight σr＞ 0), image is put down
The influence of sliding effect and resolution ratio, reference can be made to Fig. 3-5, Fig. 3 is pending original image.From Fig. 4-5, in other ginsengs
When number is constant, with σsSubtract big, image is increasingly obscured；When other specification is constant, with σrIncrease, image is increasingly
It is fuzzy.As seen from the figure, optionally, in space weighted value σs=3, scope weighted value σrWhen=0.01, image is more clear, and tool
There is preferable smooth effect.Certainly, depending on those skilled in the art can be according to real needs and image resolution ratio, the application is to this
Any restriction is not done.
It should be noted that bilateral filtering can be first passed through, then filtered by transform domain, then circulation is performed n times；Also may be used
Transform domain filtering is first passed through, then by bilateral filtering, then circulation is performed n times；Certainly n filter also can be carried out using bilateral filtering
Ripple, recycles transform domain filtering to carry out n filtering；Or n filtering is carried out first with transform domain filtering, recycle bilateral filtering
N filtering is carried out, this does not influence the realization of the application.But, found by many experiments, first passing through transform domain filtering is carried out
Treatment, then treatment is filtered by bilateral filtering, then circulation is performed n times, and the image smoothing effect for obtaining is best.Therefore, can
Choosing, can use first carries out transform domain filtering to original image, then carries out bilateral filtering, and then circulation performs preset times.
Once complete filtering is that transform domain filtering is first carried out to original image, then carries out bilateral filtering, and circulation execution is
Finger carries out multi-pass operation to once complete filtering, for example, the navigational figure that once complete filtering is obtained is once to draw
Derived function, then perform once that complete filtering obtains is secondary boot function；Circulation n times, what is obtained is n guiding letter
Number.
Analyzed according to many experiments, circulate preset times just only trickle change more than 3 times, and increase number of times to experiment
Result does not have any effect, and the time of image procossing can be increased on the contrary, causes image processing efficiency relatively low.Therefore, optionally, preset
Number of times can value be 3 times.
Aforesaid operations are performed by circulation, the details and small yardstick noise of original image can be obscured, while also retaining former
The constituent of beginning image.
S202：The original image, the navigational figure and default smoothed image are constructed most according to least square method
A young waiter in a wineshop or an inn multiplies model.
Least square method (also known as least squares method) is a kind of Mathematics Optimization Method.It is by minimizing the quadratic sum of error
The optimal function matching of data is found, unknown data can be easily tried to achieve using least square method, and cause that these are tried to achieve
Data and real data between error quadratic sum for minimum.
Smoothed image be original image after the image smoothing method of the application is smoothed, the image for obtaining.
Because smoothed image is unknown, and the solution that can realize to unknown data using the model that least square method is constructed, therefore can be false
If smoothed image is default smoothed image, brought into least square model as unknown data.For example in mathematics unitary is once
The concept of equation, for solving unknown quantity, typically can first assume variable x, then known quantity be brought into solution, so as to obtain x's
Value.
Norm, be have " length " concept function its be all vectors in vector space assign non-zero positive length or
Size.Semi-norm can be on the contrary that the vector of non-zero assigns zero-length.For example, can in the euclidean geometry space R of two dimension
Define Euclidean Norm.Element in this vector space is usually painted as one from origin in Cartesian coordinate system
The directed line segment with arrow.The Euclidean Norm of each vector is exactly the length of directed line segment.
Norm generally can be divided into vector norm and matrix norm, and in vector norm, 0- norms refer to non-zero entry in vector
The number of element；1- norms are vector element absolute value sum；The quadratic sum of 2- norm vector element absolute values evolution again.In matrix
In norm, 0- norms refer to the number of nonzero element in matrix；1- norms are the maximum of all rectangular array absolute value of a vector sums
Value；2- norms, are also spectral norm, refer to the extraction of square root of the eigenvalue of maximum of matrix, that is, the mould on ordinary meaning.
Can be according to least square method and 2- norms square to original image, navigational figure and default smoothed image structure
Making least square model is：
In formula, S is default smoothed image, and I is original image, and G is navigational figure, and α is the detail recovery factor.
It should be noted that the usual values of α are α ∈ [0,1].
The least square model of construction has Double Data fidelity, and one is used to control between smoothed image and original image
Difference is (i.e.), another is used to control smoothed image and navigational figure differencePass through
This two controls, can both strengthen the protection to constituent in original image, while also giving particulars the recovery factor to recover one
The details of a little high-contrasts.
It should be noted that comparing in the prior art using a data fidelity (i.e. for controlling smoothed image with original
Difference between beginning image), two of the Double Data fidelity that the application is used remain primary structure in artwork, and difference is one
Individual retains details (Section 1), and another obscures and weakens details (Section 2).Using Double Data fidelity can more added with
Effect ground retains primary structure, while high contrast detail composition can in varying degrees be retained according to selection different recovery parameters.
If using only first data fidelity, high contrast detail, and fragile primary structure cannot be effectively filtered out, if only making
With second data fidelity, then all small compositions (useful component may be included) are arbitrarily deleted.
S203：The image pixel intensities and gradient of the default smoothed image are defined using norm, it is described default flat to obtain
The constraint function of sliding image.
Using the image pixel intensities and gradient of the default smoothed image of 0- norms definition, certainly, 1- norms, 2- can be also used
Norm is defined, and this does not influence the realization of the application.But, by many experiments and analysis, relative to 1- norms and 2-
Norm, 0- norms can obtain more preferable effect.
0- norms refer to the number being not zero in variable, and S is default smoothed image, and ▽ S are the gradient of default smoothed image
Figure, the image pixel intensities and gradient that the default smoothed image is defined using 0- norms are | | ▽ S | |0, wherein, ▽ S ∈ RM×N。
Represent the number being not zero in ▽ S.
By structure constraint function, can be used to control the degree of rarefication of smoothed image.
S204：Smoothed energy object function is obtained according to the least square model and the constraint function.
Obtaining smoothed energy object function according to least square model and the constraint function is：
In formula, E (S) is smoothed energy object function, and S is default smoothed image, and I is original image, and G is navigational figure, α
It is the detail recovery factor, λ is smoothing factor, | | ▽ S | | are constraint function.
When constraint function is that when being defined using 0- norms, above-mentioned smoothed energy object function is：
In formula, E (S) is smoothed energy object function, and S is default smoothed image, and I is original image, and G is navigational figure, α
It is the detail recovery factor, λ is smoothing factor, | | ▽ S | |0It is constraint function.
On the detail recovery factor and smoothing factor (α ∈ [0,1], λ >=0), smooth effect and resolution to image
The influence of rate, reference can be made to Fig. 6 and Fig. 7.As seen from the figure, when other specification is constant, with the reduction of α, image is increasingly obscured；
When other specification is constant, with the increase of λ, image is increasingly obscured.As seen from the figure, optionally, in α=1, λ=0.005
When, image is more clear, with preferable smooth effect.Certainly, those skilled in the art can be according to real needs and image point
Depending on resolution, the application does not do any restriction to this.
S205：The smoothed energy object function is solved using half secondary split method and alternating fixed variable method, to obtain
Obtain smoothed image.
Specifically may include：
It is introduced into the constraint function that auxiliary variable is replaced in the smoothed energy object function；
Minimum treatment is carried out to the smoothed energy object function replaced using the half secondary split method, error is added
Penalty term, obtains smooth minimum model；
The smooth minimum model is solved according to the alternately fixed variable method, to obtain the smoothed image.
Due to that cannot realize directly minimizing smoothed energy object function, this is non-convex optimization problem, therefore needs to introduce auxiliary
Help variable to carry out surrogating constraint function, it is approached minimum value as far as possible.
For example, when smoothed energy object function is：
Introduce auxiliary variable g=(gx, gy)T| | ▽ S | | in instead of bound term0；
Minimum treatment, original smooth figure are carried out to the smoothed energy object function replaced using half secondary split method
Transform is：
Add error penalty term on above-mentioned smoothed image function, constituting final smooth minimum model is：
Wherein, β is auto-adaptive parameter, to control the similarity of g and ▽ S.
Alternately fixed variable method is generally and fixes an amount, solves another amount, is an iterative process.Need to hand over
For g and S is solved, smoothed image S is finally tried to achieve.Idiographic flow can be as follows：
Default smoothed image, auto-adaptive parameter β, iterations i are initialized；
Calculating is iterated using following calculation relational expressions：
β=k β；
Until β ＞ βmax, S is exported, the as final smoothed image for obtaining.
Wherein, default smoothed image is initialized as original image, and auto-adaptive parameter β is initialized as β0, k is gaining rate.
F-1() represent inverse discrete Fourier transform operator, F () represent complex conjugation operator, F (1) represent δ functions from
Dissipate Fourier transform.Above-mentioned all of operator, plus, multiplication and division by unit usually operate.By Fourier transformation, accelerate S's
Solving speed, is conducive to improving the efficiency of general image smoothing processing.
Optionally, β0=λ, βmax=105, k=2.Certainly, those skilled in the art can be according to real needs and image resolution
Depending on rate, the application does not do any restriction to this.
Comprehensive to understand, the picture smooth treatment method that the application is provided, S201 is local smoothing method, and S202-S205 is the overall situation
It is smooth.Small high-contrast composition can effectively be obscured using local smoothing method method, retain big structure；Using global exponential smoothing, can
The small structure that effectively removal is obscured by part filter, also retains big structure, ultimately results in big result in artwork and retains, small
Details noise remove.
From the foregoing, it will be observed that the embodiment of the present invention considers global characteristics and local feature, office first is carried out to original image
Portion's smoothing processing, then carries out global smoothing processing, has evaded the inferior position of local smoothing method method and global smoothing method, effectively
Make use of the advantage of the two.By controlling difference and control smoothed image and guiding figure between smoothed image and original image
Aberration is different, enhances the protection to constituent in original image, remains the structure of image, in removal detail textures feature
The details of some high-contrasts, the image smoothing effect for obtaining are recovered simultaneously；Additionally, having carried out effective making an uproar to image
Sound is filtered, and strengthens the intensity of boundary pixel, is conducive to the extraction of image outline, so as to be conducive to improving the accurate of image recognition
Rate and efficiency.
When image contains a large amount of high-contrast noises, when least square model is constructed, because original function is made an uproar
Sound is too big, and noise jamming is too big to cause the model for constructing to be received, and the smoothed image for solving has very big deviation, it is difficult to ensure flat
The accuracy rate of image after cunning.Therefore, the application additionally provides one embodiment based on above-described embodiment.
Before least square model is constructed, medium filtering first is carried out to original image.
Median filtering method is a kind of nonlinear smoothing technology, and a kind of based on sequencing statistical theory can effectively suppress noise
Nonlinear signal processing technology, the gray value of each pixel is set to all pixels point ash in the point neighborhood window for it
The intermediate value of angle value.The general principle of medium filtering is a neighborhood with the point the value of any in digital picture or Serial No.
In the Mesophyticum of each point value replace, the actual value for making the pixel value of surrounding close, so as to eliminate isolated noise spot.Method is to use certain
The two-dimentional sleiding form of structure is planted, pixel in plate is ranked up according to the size of pixel value, generation monotone increasing (or decline)
Be 2-D data sequence.
I.e. according to least square method to the original image by medium filtering, the navigational figure and default smoothed image
Construction least square model.Specifically, the implementation method with above-described embodiment is identical, herein, just do not repeating.
After being processed original image using medium filtering, the high-contrast that can effectively filter in original image is made an uproar
Sound, improves image smoothing effect, lifts the accuracy rate of smoothed image.
The embodiment of the present invention is provided also directed to image smoothing method and realizes device accordingly, further such that methods described
With more practicality.Image smoothing device provided in an embodiment of the present invention is introduced below, image smoothing described below
Device can be mutually to should refer to above-described image smoothing method.
Referring to Fig. 8, Fig. 8 is a kind of structure chart of image smoothing device provided in an embodiment of the present invention, and the device may include：
Filtering module 801, bilateral filtering and the transform domain filter for carrying out preset times to original image circulation
Ripple, to obtain navigational figure.
Model module 802 is set up, for according to least square method is to the original image, the navigational figure and presets
Smoothed image constructs least square model.
Smoothed image module 803 is obtained, image pixel intensities and ladder for defining the default smoothed image using norm
Degree, to obtain the constraint function of the default smoothed image；Obtained according to the least square model and the constraint function
Smoothed energy object function；The smoothed energy object function is solved using half secondary split method and alternating fixed variable method,
To obtain smoothed image.
In a kind of specific embodiment, the acquisition smoothed image module 803 be according to the least square method and
2- norms square to the original image, the navigational figure and default smoothed image construct least square model mould
Block, the least square model is：
In formula, S is the default smoothed image, and I is the original image, and G is the navigational figure, and α is detail recovery
The factor.
The function of each functional module of embodiment of the present invention described image smoothing apparatus can be according in above method embodiment
Method implement, it implements the associated description that process is referred to above method embodiment, and here is omitted.
From the foregoing, it will be observed that the embodiment of the present invention considers global characteristics and local feature, office first is carried out to original image
Portion's smoothing processing, then carries out global smoothing processing, has evaded the inferior position of local smoothing method method and global smoothing method, effectively
Make use of the advantage of the two.By controlling difference and control smoothed image and guiding figure between smoothed image and original image
Aberration is different, enhances the protection to constituent in original image, remains the structure of image, in removal detail textures feature
The details of some high-contrasts, the image smoothing effect for obtaining are recovered simultaneously；Additionally, having carried out effective making an uproar to image
Sound is filtered, and strengthens the intensity of boundary pixel, is conducive to the extraction of image outline, so as to be conducive to improving the accurate of image recognition
Rate and efficiency.
In order to verify the image smoothing effect that the technical scheme that the application is provided has had, this application provides specific reality
Example is applied, Fig. 9-14 are referred to, Fig. 9 is pending original image, and Figure 10-13 is the image that other algorithm process are crossed, Tu14Wei
The image of the application treatment, the figure in figure in square frame is the partial enlarged drawing of corresponding square frame.As seen from the figure, by RTV
The image of (Relative Total Variation, correlation total variation) algorithm process, close two lines bar is relatively obscured,
None- identified；Other algorithms (such as RGF (Rolling Guidance Filter circulate guiding filtering) and BLF
(Bilateral filter, bilateral filtering) algorithm, NLGRTV (Nonlocal version of Generalized
Relative Total Variation, the general correlation total variation of non local version) algorithm and SSPTF (Scale-
Aware Structure-Preserving Texture Filtering, yardstick concern structure preserves texture filtering) algorithm) it is right
The smooth effect of image detail is more coarse.It can be seen that, the image smoothing effect that the present processes have had.
In order to verify that the technical scheme that the application is provided has effective denoising effect, implement this application provides specific
Example, refers to Figure 15-17, and Figure 15 is pending original image, and Figure 16 is the image crossed by RGF and BLF algorithm process, figure
17 is the image of the application treatment, still there is fuzzy mixed and disorderly lines as seen from the figure, in Figure 16, and the method that the application is provided is effective
Eliminate mixed and disorderly lines in original image, obtain destination object (object of rectangular structure).It can be seen that, what the application was provided
Technical scheme can effective filtering image noise, it is to avoid the interference of picture noise.
In order to verify that the technical scheme that the application is provided has the effect for strengthening image, implement this application provides specific
Example, it is original image to refer to Figure 18 and 19, Figure 18, and Figure 19 is the image of the application treatment, as seen from the figure, by the application
The method of offer, original image pixels are remarkably reinforced, especially the intensity of edge pixel, are conducive to extracting the profile of image, from
And be conducive to improving the accuracy rate and efficiency of image recognition.
Each embodiment is described by the way of progressive in this specification, and what each embodiment was stressed is and other
The difference of embodiment, between each embodiment same or similar part mutually referring to.For being filled disclosed in embodiment
For putting, because it is corresponded to the method disclosed in Example, so description is fairly simple, related part is referring to method part
Illustrate.
Professional further appreciates that, with reference to the unit of each example of the embodiments described herein description
And algorithm steps, can be realized with electronic hardware, computer software or the combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software, generally describes the composition and step of each example according to function in the above description.These
Function is performed with hardware or software mode actually, depending on the application-specific and design constraint of technical scheme.Specialty
Technical staff can realize described function to each specific application using distinct methods, but this realization should not
Think beyond the scope of this invention.
The step of method or algorithm for being described with reference to the embodiments described herein, directly can be held with hardware, processor
Capable software module, or the two combination is implemented.Software module can be placed in random access memory (RAM), internal memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In field in known any other form of storage medium.
A kind of image smoothing method provided by the present invention and device are described in detail above.It is used herein
Specific case is set forth to principle of the invention and implementation method, and the explanation of above example is only intended to help and understands
The method of the present invention and its core concept.It should be pointed out that for those skilled in the art, not departing from this
On the premise of inventive principle, some improvement and modification can also be carried out to the present invention, these are improved and modification also falls into the present invention
In scope of the claims.
Claims (10)
1. a kind of image smoothing method, it is characterised in that including：
Bilateral filtering and the transform domain filtering of preset times are carried out to original image circulation, to obtain navigational figure；
Least square mould is constructed to the original image, the navigational figure and default smoothed image according to least square method
Type；
The image pixel intensities and gradient of the default smoothed image are defined using norm, to obtain the pact of the default smoothed image
Beam function；
Smoothed energy object function is obtained according to the least square model and the constraint function；
The smoothed energy object function is solved using half secondary split method and alternating fixed variable method, to obtain smooth figure
Picture.
2. image smoothing method according to claim 1, it is characterised in that it is described according to least square method to described original
Image, the navigational figure and default smoothed image construct least square model：
According to the least square method and 2- norms square to the original image, the navigational figure and default smooth
Image configuration least square model is：
In formula, S is the default smoothed image, and I is the original image, and G is the navigational figure, and α is the detail recovery factor.
3. image smoothing method according to claim 1, it is characterised in that the utilization norm defines described default smooth
The image pixel intensities and gradient of image are：
The image pixel intensities and gradient of the default smoothed image are defined using 0- norms.
4. image smoothing method according to claim 3, it is characterised in that it is described according to the least square model and
The constraint function obtains smoothed energy object function：
Obtaining smoothed energy object function according to the least square model and the constraint function is：
In formula, E (S) is the smoothed energy object function, and S is the default smoothed image, and I is the original image, and G is institute
Navigational figure is stated, α is the detail recovery factor, and λ is smoothing factor, | | ▽ S | |0It is the constraint function.
5. the image smoothing method according to Claims 1-4 any one, it is characterised in that described according to least square
Method constructs least square model to the original image, the navigational figure and default smoothed image：
According to the least square method to the original image by medium filtering, the navigational figure and default smoothed image structure
Make least square model.
6. image smoothing method according to claim 5, it is characterised in that described that default following is carried out to original image circulation
Bilateral filtering and the transform domain filtering of ring number of times, are included with obtaining navigational figure：
S1：Obtain the original image, constant value image, space weighted value and scope weighted value；
S2：Be initial guide function with the constant value image, according to transform domain filter method to the original image and it is described just
Beginning guidance function carries out transform domain filtering, obtains new guidance function；
S3：Bilateral filtering is carried out to the original image and the new guidance function according to bilateral filtering method, is once drawn
Derived function；
S4：The preset times are performed to S2 and S3 circulations, to obtain the navigational figure.
7. image smoothing method according to claim 6, it is characterised in that the preset times are 3 times.
8. image smoothing method according to claim 7, it is characterised in that the space weighted value and scope weighted value
For：
The space weighted value is 3；
The scope weighted value is 0.01.
9. image smoothing method according to claim 8, it is characterised in that described to utilize half secondary split method and alternating
Fixed variable method solves the smoothed energy object function, is included with obtaining smoothed image：
It is introduced into the constraint function that auxiliary variable is replaced in the smoothed energy object function；
Minimum treatment is carried out to the smoothed energy object function replaced using the half secondary split method, error punishment is added
, obtain smooth minimum model；
The smooth minimum model is solved according to the alternately fixed variable method, to obtain the smoothed image.
10. a kind of image smoothing device, it is characterised in that including：
Filtering module, bilateral filtering and the transform domain filtering for carrying out preset times to original image circulation, to obtain
Obtain navigational figure；
Model module is set up, for scheming to the original image, the navigational figure and default smoothing according to least square method
As construction least square model；
Smoothed image module is obtained, image pixel intensities and gradient for defining the default smoothed image using norm, to obtain
Obtain the constraint function of the default smoothed image；Smoothed energy is obtained according to the least square model and the constraint function
Object function；The smoothed energy object function is solved using half secondary split method and alternately fixed variable method, it is flat to obtain
Sliding image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710146581.3A CN106920222B (en) | 2017-03-13 | 2017-03-13 | A kind of image smoothing method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710146581.3A CN106920222B (en) | 2017-03-13 | 2017-03-13 | A kind of image smoothing method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106920222A true CN106920222A (en) | 2017-07-04 |
| CN106920222B CN106920222B (en) | 2019-11-08 |
Family
ID=59460670
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710146581.3A Active CN106920222B (en) | 2017-03-13 | 2017-03-13 | A kind of image smoothing method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106920222B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108897044A (en) * | 2018-07-25 | 2018-11-27 | 梧州市兴能农业科技有限公司 | A kind of device visually showing geological anomalous body |
| CN109345478A (en) * | 2018-09-27 | 2019-02-15 | 深圳市牧月科技有限公司 | A kind of picture smooth treatment method based on gradient minimisation |
| CN109389565A (en) * | 2018-10-19 | 2019-02-26 | 山东大学 | A kind of holding edge and the adjustable image smoothing method of smoothness |
| CN109598680A (en) * | 2018-10-19 | 2019-04-09 | 浙江工业大学 | Shearing wave conversion medicine CT image denoising method based on quick non-local mean and TV-L1 model |
| CN111882810A (en) * | 2020-07-31 | 2020-11-03 | 广州市微智联科技有限公司 | Fire identification and early warning method and system |
| WO2021036442A1 (en) * | 2019-08-29 | 2021-03-04 | 北京迈格威科技有限公司 | Cyclic edge-preserving smooth filtering method and apparatus, and electronic device |
| CN112967195A (en) * | 2021-03-04 | 2021-06-15 | 浙江大华技术股份有限公司 | Image denoising method and device and computer readable storage medium |
| CN113436163A (en) * | 2021-06-23 | 2021-09-24 | 四川大学 | Method for identifying and processing flow field characteristics of schlieren image in impeller mechanical blade lattice test |
| CN116977227A (en) * | 2023-09-22 | 2023-10-31 | 福建晟哲自动化科技有限公司 | Image smoothing method and device based on local structure variation |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103606132A (en) * | 2013-10-31 | 2014-02-26 | 西安电子科技大学 | Multiframe digital image denoising method based on space domain and time domain combination filtering |
| CN104217405A (en) * | 2014-09-23 | 2014-12-17 | 闽江学院 | Salt-pepper noise filter method for image fusing local information and global information |
| CN105023245A (en) * | 2015-05-05 | 2015-11-04 | 苏州大学 | Image smoothing method under strength and gradient sparsity constraint |
-
2017
- 2017-03-13 CN CN201710146581.3A patent/CN106920222B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103606132A (en) * | 2013-10-31 | 2014-02-26 | 西安电子科技大学 | Multiframe digital image denoising method based on space domain and time domain combination filtering |
| CN104217405A (en) * | 2014-09-23 | 2014-12-17 | 闽江学院 | Salt-pepper noise filter method for image fusing local information and global information |
| CN105023245A (en) * | 2015-05-05 | 2015-11-04 | 苏州大学 | Image smoothing method under strength and gradient sparsity constraint |
Non-Patent Citations (2)
| Title |
|---|
| QI ZHANG ETC.: ""Rolling guidance filter"", 《EUROPEAN CONFERENCE ON COMPUTER VISION.SPRINGER》 * |
| 崔艳萌: ""基于空间域和变换域的图像降噪方法研究"", 《通讯世界》 * |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108897044A (en) * | 2018-07-25 | 2018-11-27 | 梧州市兴能农业科技有限公司 | A kind of device visually showing geological anomalous body |
| CN109345478A (en) * | 2018-09-27 | 2019-02-15 | 深圳市牧月科技有限公司 | A kind of picture smooth treatment method based on gradient minimisation |
| CN109598680B (en) * | 2018-10-19 | 2021-11-23 | 浙江工业大学 | Shear wave transformation medical CT image denoising method based on rapid non-local mean value and TV-L1 model |
| CN109389565A (en) * | 2018-10-19 | 2019-02-26 | 山东大学 | A kind of holding edge and the adjustable image smoothing method of smoothness |
| CN109598680A (en) * | 2018-10-19 | 2019-04-09 | 浙江工业大学 | Shearing wave conversion medicine CT image denoising method based on quick non-local mean and TV-L1 model |
| WO2021036442A1 (en) * | 2019-08-29 | 2021-03-04 | 北京迈格威科技有限公司 | Cyclic edge-preserving smooth filtering method and apparatus, and electronic device |
| CN111882810B (en) * | 2020-07-31 | 2022-07-01 | 广州市微智联科技有限公司 | Fire identification and early warning method and system |
| CN111882810A (en) * | 2020-07-31 | 2020-11-03 | 广州市微智联科技有限公司 | Fire identification and early warning method and system |
| CN112967195A (en) * | 2021-03-04 | 2021-06-15 | 浙江大华技术股份有限公司 | Image denoising method and device and computer readable storage medium |
| CN112967195B (en) * | 2021-03-04 | 2024-04-23 | 浙江大华技术股份有限公司 | Image denoising method, device and computer readable storage medium |
| CN113436163A (en) * | 2021-06-23 | 2021-09-24 | 四川大学 | Method for identifying and processing flow field characteristics of schlieren image in impeller mechanical blade lattice test |
| CN113436163B (en) * | 2021-06-23 | 2023-06-09 | 四川大学 | Impeller machinery cascade test schlieren image flow field characteristic identification processing method |
| CN116977227A (en) * | 2023-09-22 | 2023-10-31 | 福建晟哲自动化科技有限公司 | Image smoothing method and device based on local structure variation |
| CN116977227B (en) * | 2023-09-22 | 2023-12-15 | 福建晟哲自动化科技有限公司 | Image smoothing method and device based on local structure variation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106920222B (en) | 2019-11-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106920222B (en) | A kind of image smoothing method and device | |
| Jain et al. | A survey of edge-preserving image denoising methods | |
| Ta et al. | Nonlocal PDEs-based morphology on weighted graphs for image and data processing | |
| CN102246204B (en) | Devices and methods for processing images using scale space | |
| CN110210524B (en) | Training method of image enhancement model, image enhancement method and device | |
| CN105335947A (en) | Image de-noising method and image de-noising apparatus | |
| CN106296675A (en) | A kind of dividing method of the uneven image of strong noise gray scale | |
| Kiefel et al. | Permutohedral lattice cnns | |
| CN101739670B (en) | Non-local mean space domain time varying image filtering method | |
| Lou et al. | Nonlocal similarity image filtering | |
| CN105023245A (en) | Image smoothing method under strength and gradient sparsity constraint | |
| CN109741258B (en) | Image super-resolution method based on reconstruction | |
| Zhao et al. | Constant time texture filtering | |
| WO2009047208A1 (en) | Method of generating a multiscale contrast enhanced image | |
| Xu et al. | Affine non-local Bayesian image denoising algorithm | |
| Sun et al. | Edge guidance filtering for structure extraction | |
| CN101957912B (en) | Method for obtaining human face illumination invariant images based on multiscale anisotropic diffusion | |
| CN111932472B (en) | Image edge protection filtering method based on soft clustering | |
| Zhang et al. | Image denoising using hybrid singular value thresholding operators | |
| CN112241938A (en) | Image restoration method based on smooth Tak decomposition and high-order tensor Hank transformation | |
| CN111784610A (en) | Clustering-based side window filter optimization method | |
| CN115131226B (en) | Image restoration method based on wavelet tensor low-rank regularization | |
| Yang et al. | A nonlinear anisotropic diffusion model with non-standard growth for image segmentation | |
| CN115937302A (en) | Hyperspectral image sub-pixel positioning method combined with edge preservation | |
| CN113160069B (en) | Hyperspectral image denoising method based on image signal |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |