CN106651770B - Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization - Google Patents
Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization Download PDFInfo
- Publication number
- CN106651770B CN106651770B CN201610832828.2A CN201610832828A CN106651770B CN 106651770 B CN106651770 B CN 106651770B CN 201610832828 A CN201610832828 A CN 201610832828A CN 106651770 B CN106651770 B CN 106651770B
- Authority
- CN
- China
- Prior art keywords
- image
- matrix
- similar block
- resolution
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 74
- 238000003384 imaging method Methods 0.000 title description 7
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 10
- 238000012545 processing Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 92
- 238000005070 sampling Methods 0.000 claims description 8
- 238000001228 spectrum Methods 0.000 claims description 5
- 230000017105 transposition Effects 0.000 claims description 4
- 230000008602 contraction Effects 0.000 claims description 3
- 230000003595 spectral effect Effects 0.000 description 14
- 238000002310 reflectometry Methods 0.000 description 11
- 238000002474 experimental method Methods 0.000 description 9
- 230000006870 function Effects 0.000 description 5
- 230000000052 comparative effect Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 102000002274 Matrix Metalloproteinases Human genes 0.000 description 1
- 108010000684 Matrix Metalloproteinases Proteins 0.000 description 1
- 230000006978 adaptation Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4007—Scaling of whole images or parts thereof, e.g. expanding or contracting based on interpolation, e.g. bilinear interpolation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
- G06T3/4076—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution using the original low-resolution images to iteratively correct the high-resolution images
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a kind of Image Super-resolution reconstructing method based on second order Laplace transform, mainly solving the problems, such as that the image reliability after the prior art reconstructs is not high and mistake occurs has color lump.Its technical solution is: 1. pairs of low-resolution images carry out interpolation processing and obtain initial pictures；2. being iterated according to the iterative formula of setting to initial pictures；3. the image after pair iteration carries out piecemeal and matching similar block；4. the similar block after matching is merged into each image array block；5. pair image array block carries out singular value decomposition, and updates similar block using iterative shrinkage method；6. merging similar block obtains updated image array block；6. iteration realizes the super-resolution reconstruct of image by updated each image array merged block at image.The present invention reduces mistakes color lump, effectively increases the confidence level and spatial resolution of reconstructed image, can be used for reconstructing high-definition picture from low-resolution image.
Description
Technical field
The invention belongs to technical field of image processing, in particular to a kind of multispectral super-resolution imaging reconstructing method is used for
High-definition picture is reconstructed from low-resolution image.
Background technique
The reconstruct of super-resolution imaging is to reconstruct high-definition picture by specific method to low-resolution image.It is logical
Often, the spatial high resolution and the high confidence level of multispectral image of image are obtained by the reconstruct of super-resolution imaging.Although existing
Super-resolution imaging reconstructing method can get relatively high spatial resolution image.But existing super-resolution imaging reconstructing method
Occur mistake in the high resolution gray figure reconstructed has color lump, as shown in attached drawing 4 (a), 4 (b), 4 (c).And for one
It is a little to require more accurate work, such as material analysis, identification identification, precise classification etc., it needs to improve spatial resolution
With multispectral image confidence level.
Currently, there is no very good solution method on improving multispectral confidence level.The method for improving spatial resolution is main
There are two types of effective methods.
First method is directly to carry out super resolution image reconstruct in spatial domain, reconstruct spatial high resolution figure with this
Picture.It is suboptimal solution by the solution after reconstruct known to mathematical analysis but since the reconstructing method has ignored the Spectral correlation of high-order,
Therefore this method cannot effectively improve the confidence level of image.
Second method is first to carry out the transformation of RGB to YUV to image in image reconstruction, then carries out space oversubscription
It distinguishes reconstruct, spatial high resolution image is reconstructed with this.But this method Structure adaptation between spectral coverage in terms of interband decorrelation
Property aspect Shortcomings, therefore the confidence level of image cannot be effectively improved, also, in the conversion process by RGB to YUV, meeting
It is aligned the corresponding spatial position of each component of R, G, B of image can not, finally occur mistake in grayscale image has color lump, causes
Image after reconstruct includes error message, and then cannot accurately realize the reconstruct of image.
Summary of the invention
It is a kind of based on Laplce l the purpose of the present invention is in view of the above shortcomings of the prior art, proposing1,2Norm canonical
The super-resolution reconstructing method of change improves the spatial discrimination of image to reduce the error message of existing method reconstructed image appearance
Rate and confidence level.
Embodiment of the present invention is completed in this way:
One kind being based on Laplce l1,2The Image Super-resolution reconstructing method of norm regularization, includes the following steps:
(1) bicubic interpolation is carried out to low-resolution image, obtains initial image X(0)；
(2) iterative formula X is set(l)=X(l-1)+δDT(Y-D(X(l-1))), wherein l=1,2 ..., L, L are greatest iteration time
Number, X(l)For the image after the l times iteration, δ is iteration regular parameter；
(3) to initial pictures X(0)Figure X after obtaining first time iteration using the iterative formula of above-mentioned setting(1)；
(4) by the image X after first time iteration(1)It is divided into M block, and obtains S using block matching method to i-th pieceiIt is a similar
Block matrix remembers x(i,j)For i-th piece of j-th of similar block matrix, then by this SiA similar block is merged into i-th of image array Xi,
Middle i=1,2 ..., M, j=1,2 ..., Si；
(5) to image array XiUtilize formula [U, Σ ', V]=SVD (Xi) singular value decomposition is carried out, obtain U, Σ ', V tri-
A split-matrix, wherein U be and image array XiRelevant left orthogonal matrix, Σ ' are comprising image array XiSingular value it is unusual
Matrix, V are and image array XiRelevant right orthogonal matrix；
(6) U obtained according to step (5), these three matrixes of Σ ', V utilize public Xi=USμ(∑′)VTUpdate image array
Xi, wherein Sμ(∑ ')=max (∑ '-μl|k1, 0) and it is soft-threshold operation to singular matrix Σ ', VTIt represents to right orthogonal matrix V
Transposition, μlTake the third-largest characteristic value, k in singular value matrix Σ '1For the parameter for setting first regular terms, max () expression pair
Its maximizing；
(7) similar block matrix x is updated using based on the method for reconstructing in full variational regularization(i,j)It obtains and updates matrix
(8) it calculates and updates matrixThe maximum value r of orderi,j；
(9) formula is utilizedTo update matrixSingular value decomposition is carried out, U is obtained1, Σ1, V1Three
Split-matrix, wherein U1It is and updates matrixRelevant left orthogonal matrix, Σ1It is comprising updating matrixSingular value it is unusual
Matrix, V1It is and updates matrixRelevant right orthogonal matrix；
(10) shrinkage operation formula is setTo similar block matrix x(i,j)It is updated, in formula,For with
Similar block matrix x(i,j)Apart from nearest contraction matrix, H (∑1) it is to make to constrain Rank (x(i,j))≤ri,jThe hard threshold set up
It is worth operation, V1 TRepresent to similar block matrix x(i,j)Relevant right orthogonal matrix V1Transposition, Rank (x(i,j)) represent similar block
Matrix x(i,j)Order；
(11) by i-th piece of SiA similar block matrix merges, and obtains i-th of image array Xi；
(12) by M image array XiMerge, obtains image X(1), return step (2) repeats the above steps, until passing through
Super-resolution reconstructed image X is exported after L iteration(L)。
Compared with the prior art, the present invention has the following advantages
First: improving the spatial resolution and confidence level of reconstructed image.
The present invention is on the basis of existing reconstructing method such as A+, BCSR, NCSR method, it is contemplated that the correlation on space structure
Property, similar block matrix is updated using iterative shrinkage method, so that this method is in Y-PSNR PSNR and structural dependence
SSIM is performed better than, and effectively improves the spatial resolution and confidence level of image.
Second: effectively reducing mistake has color lump.
The image that existing method reconstructs will appear the color lump that has of mistake, and the present invention is special using second order Laplacian space
Property, and to similar block matrix addition of constraints, can effectively reduce mistake has color lump.
Third: the image after reconstruct is more accurate.
There is any discrepancy for the spectral reflectivity curve of full resolution pricture and former full resolution pricture that existing method reconstructs, and this hair
Bright method can be fitted the spectral reflectivity curve of original image well, thus the image after reconstruct is more accurate.
Detailed description of the invention
Fig. 1 is implementation flow chart of the invention；
Fig. 2 is the original spectrum image that emulation uses；
Fig. 3 is to visualize the low-resolution image used when comparative experiments emulation；
Fig. 4 is that Fig. 3 is reconstructed with existing A+ method, BSSC method, NSCR method and the method for the present invention respectively
Simulation result comparison diagram；
The original image that Fig. 5 is used when being trust verification experiment simulation
Fig. 6 is in trust verification experiment, with existing A+ method, BSSC method, NSCR method and the method for the present invention
The spectral reflectivity curve figure drawn after Fig. 5 is reconstructed.
Specific embodiment
The present invention is described in detail with example with reference to the accompanying drawing
Referring to Fig.1, steps are as follows for realization of the invention:
Step 1, it initializes.
Down-sampling processing is carried out to original spectrum image shown in Fig. 2 and obtains low-resolution image Y as shown in Figure 3；
Bicubic interpolation is carried out to low-resolution image Y, i.e., the interlacing of low-resolution image matrix is extracted every column and is compressed
To sampling matrix, the sampling matrix of acquisition is utilized respectively in line direction and column directionCarry out three
Secondary interpolation obtains initial image X(0), in formula, n is interpolation point number, CkIt is the value of k-th of original function, h (x-xk) it is interpolation base
Function, the highest power of the Interpolation-Radix-Function are that three times, and the single order second dervative of basic function is continuous in domain.
Step 2, after carrying out first time iteration to initial pictures, successively make piecemeal, merging treatment, obtain image array.
(2a) sets iterative formula X(l)=X(l-1)+δDT(Y-D(X(l-1))), wherein l=1,2 ..., L, L are greatest iteration
Number, X(l)For the image after the l times iteration, δ is Iteration Regularized coefficient, value 0.22；D (x) expression makees down-sampling to x
The function of processing；DT(x) transposition of D (x) is indicated；
The initial pictures X that (2b) obtains step 1(0)It utilizes setting formula in (2a) to be iterated, obtains first time iteration
Image X afterwards(1)；
(2c) is by the image X after first time iteration(1)It is divided into M block, calculates current block and adjacent block most using block matching method
Close block obtains SiA similar block matrix remembers x(i,j)For i-th piece of j-th of similar block matrix, then by this SiA similar block is merged into
I-th of image array Xi, wherein i=1,2 ..., M, j=1,2 ..., Si。
Step 3, to image array XiSingular value decomposition is carried out, and updates the image array.
(3a) is to image array XiUtilize formula [U, Σ ', V]=SVD (Xi) singular value decomposition is carried out, obtain U, Σ ', V tri-
A split-matrix, wherein U be and image array XiRelevant left orthogonal matrix, Σ ' are comprising image array XiSingular value it is unusual
Matrix, V are and image array XiRelevant right orthogonal matrix；
(3b) utilizes formula X according to step 3a U obtained, these three matrixes of Σ ', Vi=USμ(∑′)VTMore new images
Matrix Xi, in which:
Sμ(∑ ')=max (∑ '-μl|k1, 0) and it is soft-threshold operation to singular matrix Σ ',
k1For set first regular terms parameter, value 0.5,
VTThe transposition to right orthogonal matrix V is represented,
μlThe third-largest characteristic value in singular value matrix Σ ' is taken, so that image array X obtainediLow-rank,
Max () is indicated to its maximizing.
Step 4, similar block matrix x is updated using the method for reconstructing based on full variational regularization(i,j)。
The specific implementation of this step is the formula in the method for reconstructing according to full variational regularization:
Minimum problems are solved, are obtained and similar block matrix x(i,j)Distance is most
Close update matrixIn formula:
X is independent variable；Argmin () indicates the function for making some functional obtain minimum value；ρlIt is weight coefficient, value is
1；k2For the parameter for setting second regular terms, value 0.59；It is the square for doing second order Laplace's operation to x and obtaining
Battle array；||·||1,2Represent l1,2Norm,Represent l2Square of norm.
Step 5, estimation updates matrixMaximum order.
Utilize inequality constraintsEstimation obtains updating matrixMaximum order ri,j；Wherein,
γkRepresent similar block matrix x(i,j)K-th of singular value；Γ is given threshold value, and value is the second largest singular value and the third-largest
The average value of singular value.
Step 6, to update matrixSingular value decomposition is carried out, and updates similar block matrix x(i,j)。
(6a) utilizes formulaTo update matrixSingular value decomposition is carried out, U is obtained1, Σ1, V1Three
A split-matrix, wherein U1It is and updates matrixRelevant left orthogonal matrix, Σ1It is comprising updating matrixSingular value surprise
Different matrix, V1It is and updates matrixRight orthogonal matrix；
(6b) sets shrinkage operation formulaTo similar block matrix x(i,j)It is updated, in which:
For to similar block matrix x(i,j)Apart from nearest contraction matrix,
H(∑1) it is to make to constrain Rank (x(i,j))≤ri,jThe hard -threshold operation set up,
Rank(x(i,j)) represent similar block matrix x(i,j)Order,
V1 TRepresent to similar block matrix x(i,j)Relevant right orthogonal matrix V1Transposition.
Step 7, similar block x is successively merged(i,j), image array Xi, L output super-resolution reconstructed image of iteration.
(7a) is by the image X in step 2(1)It is divided into M block, for i-th piece, recycles j times and obtain SiA similar block matrix
x(i,j), i-th of image array X is obtained after being mergedi, by M image array XiImage X is obtained after merging(1), wherein i=1,
2 ..., M, j=1,2 ..., Si；
(7b) is by image X(1)It substitutes into the iterative formula set in step 2 and obtains image X(2), to image X(2)Execute step 2
~(7a) obtains new image X(2), step 2 is returned again to, l iteration is successively executed, the image X after obtaining the l times iteration(l),
Wherein, l=1,2,3 ..., L, L value are 200；
(7c) judges image X(l)Whether iteration L times, if so, stop iteration, the full resolution pricture X after output reconstruct(L),
If it is not, then continuing to repeat step 2~(7a), until l=L.
Effect of the invention can be further illustrated by emulation below
1. simulated conditions
The hardware test platform of this experiment is: Intel Core i7CPU, dominant frequency 3.40GHz, memory 8GB；Software emulation
Platform are as follows: windows 7,64 bit manipulation systems and Matlab2013b.
2. emulation experiment
Emulation 1: comparative experiments,
In order to verify the validity of method in the present invention, using the first eight spectrum picture disclosed in Colombia as experiment institute
The original spectrum image needed, down-sampling is carried out to it and obtains eight low-resolution image A~H.
With the method for the present invention and existing A+ method, NCSR method, BSSC method respectively to low-resolution image A~H into
Row reconstruct obtains high-definition picture a~h after eight reconstruct, respectively in σ2=0 and σ2Under the noise of=25 different levels, than
The size of size and structural similarity SSIM compared with the signal-to-noise ratio PSNR of high-definition picture a~h after reconstruct, as a result in table
Shown in 1.
Table 1
As can be seen from Table 1: under the noise of different level, the size ratio side A+ of the signal-to-noise ratio PSNR of the method for the present invention
Method averagely can increase 0.67dB~1.08dB, averagely increases 0.7dB~2.22dB than BCSR method, averagely increase than NCSR method
0.24dB~0.96dB, structural dependence SSIM are from table it can also be seen that the structural dependence size of the method for the present invention compares
Other methods want high.
Emulation 2: visualization comparative experiments
For the advantages of more visually prominent present invention can effectively reduce color lump, down-sampling processing is carried out to Fig. 2
Obtain low-resolution image shown in Fig. 3, to Fig. 3 by the method for the invention with existing A+ method, NCSR method, BSSC method
It is reconstructed, the color image after obtaining the reconstruct indicated such as Fig. 4, as a result as shown in the figure, in which:
4 (a) be that rear color image is reconstructed by A+ method,
4 (b) be that rear color image is reconstructed by BSSC method,
4 (c) be that rear color image is reconstructed by NCSR method,
4 (d) be that rear color image is reconstructed by the method for the invention.
From 2 original image of Fig. 4 (a)~4 (d) comparison diagram, it can be seen that the full resolution pricture meeting of A+, BCSR, NCSR method reconstruct
Different degrees of wrong color lump is presented.And the high-definition picture after present invention reconstruct effectively reduces mistake and has color lump.
Emulation 3: trust verification experiment,
It is identical from Fig. 5 chosen material, spectral signature is identical in order to illustrate superiority of the method for the present invention in confidence level
3 points: i.e. 1: 1, second point 2, thirdly 3.
Down-sampling processing is carried out to Fig. 5 and obtains low-resolution image Y1, to low-resolution image Y1By the method for the invention and
Existing A+ method, NCSR method, BSSC method are reconstructed, the high-definition picture after being reconstructed.
The spectral reflectivity curve figure at high-definition picture midpoint 1, point 2, point 3 after drawing reconstruct, while drawing Fig. 5's
Spectral reflectivity curve figure, as a result as shown in Figure 6, wherein
Fig. 6 (a) indicates 1: 1 spectral reflectivity curve figure,
Fig. 6 (b) indicates the spectral reflectivity curve figure of second point 2,
Fig. 6 (c) indicates thirdly 3 spectral reflectivity curve figure.
Compare 3 spectral reflectivity curves shown in fig. 6, it can be clearly seen that the height of A+, BCSR, NCSR method reconstruct
The spectral reflectivity curve of image in different resolution and original image has biggish discrepancy, and the method for the present invention can be fitted original image very well
The curve of spectral reflectivity realizes accurate reconstruction, ensure that the confidence level of the high-definition picture of reconstruct.
Claims (3)
1. a kind of Image Super-resolution reconstructing method based on Laplce's norm regularization, includes the following steps:
(1) bicubic interpolation is carried out to low-resolution image, obtains initial image X(0)；
(2) iterative formula is setWhereinL is maximum number of iterations,It isImage after secondary iteration, δ are iteration regular coefficient, and Y is that original spectrum image carries out the low of down-sampling processing acquisition
Image in different resolution；
(3) to initial pictures X(0)Image X after obtaining first time iteration using the iterative formula of above-mentioned setting(1)；
(4) by the image X after first time iteration(1)It is divided into M block, and obtains S using block matching method to i-th pieceiA similar block square
Battle array remembers x(i,j)For i-th piece of j-th of similar block matrix, then by this SiA similar block is merged into i-th of image array Xi, wherein i=
1,2 ..., M, j=1,2 ..., Si；
(5) to image array XiUtilize formula [U, Σ ', V]=SVD (Xi) singular value decomposition is carried out, obtain U, tri- points of Σ ', V
Dematrix, wherein U be and image array XiRelevant left orthogonal matrix, Σ ' are comprising image array XiThe unusual square of singular value
Battle array, V are and image array XiRelevant right orthogonal matrix；
(6) U obtained according to step (5), these three matrixes of Σ ', V utilize formula Xi=USμ(∑′)VTUpdate image array Xi,
Wherein Sμ(∑ ')=max (∑ '-μl|k1, 0) and it is soft-threshold operation to singular matrix Σ ', VTIt represents to right orthogonal matrix V's
Transposition, μlTake the third-largest characteristic value, k in singular value matrix Σ '1For the parameter for setting first regular terms, max () is indicated to it
Maximizing；
(7) similar block matrix x is updated using based on the method for reconstructing in full variational regularization(i,j)It obtains and updates matrix
(8) it calculates and updates matrixThe maximum value r of orderi,j；
(9) formula is utilizedTo update matrixSingular value decomposition is carried out, U is obtained1, Σ1, V1Three decomposition
Matrix, wherein U1It is and updates matrixRelevant left orthogonal matrix, Σ1It is comprising updating matrixSingular value unusual square
Battle array, V1It is and updates matrixRelevant right orthogonal matrix；
(10) shrinkage operation formula is setTo similar block matrix x(i,j)It is updated, in formula,For to it is similar
Block matrix x(i,j)Apart from nearest contraction matrix, H (∑1) it is to make to constrain Rank (x(i,j))≤ri,jThe hard -threshold fortune set up
It calculates, V1 TRepresent to similar block matrix x(i,j)Relevant right orthogonal matrix V1Transposition, Rank (x(i,j)) represent similar block matrix
x(i,j)Order；
(11) by i-th piece of SiA similar block matrix merges, and obtains i-th of image array Xi；
(12) by M image array XiMerge, obtains image X(1), return step (2) repeats the above steps, until passing through L times repeatedly
Super-resolution reconstructed image X is exported after generation(L)。
2. according to the method described in claim 1, wherein being updated in step (7) using the method for reconstructing based on full variational regularization
Similar block matrix x(i,j), it is carried out by following formula:
In formula, x is independent variable；For updated matrix；Argmin () indicates the function for making some functional obtain minimum value；ρl
It is weight coefficient, value 1；k2For the parameter for setting second regular terms, value 0.59；Be x is done second order draw it is general
The matrix that Lars operation obtains；||·||1,2It representsNorm,It representsSquare of norm.
3. according to the method described in claim 1, wherein calculating similar block matrix x in step (8)(i,j)Maximum order ri,j, it is
Pass through inequality constraintsEstimation obtains, wherein ∑ indicates summation symbol；γkRepresent similar block
Matrix x(i,j)K-th of singular value；Γ is given threshold value, and value is being averaged for the second largest singular value and the third-largest singular value
Value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610832828.2A CN106651770B (en) | 2016-09-19 | 2016-09-19 | Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610832828.2A CN106651770B (en) | 2016-09-19 | 2016-09-19 | Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106651770A CN106651770A (en) | 2017-05-10 |
| CN106651770B true CN106651770B (en) | 2019-07-23 |
Family
ID=58853152
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610832828.2A Active CN106651770B (en) | 2016-09-19 | 2016-09-19 | Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106651770B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109685125A (en) * | 2018-12-14 | 2019-04-26 | 大连海事大学 | Daily behavior feature mining and calculation method based on frequent Sensor Events sequence |
| CN111257833B (en) * | 2019-12-24 | 2023-08-01 | 重庆大学 | Sound source identification method based on Laplace norm rapid iteration shrinkage threshold |
| CN111311729B (en) * | 2020-01-18 | 2022-03-11 | 西安电子科技大学 | Natural scene three-dimensional human body posture reconstruction method based on bidirectional projection network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156878A (en) * | 2011-04-08 | 2011-08-17 | 南京邮电大学 | Sparse embedding with manifold information-based human face identification method |
| CN103473740A (en) * | 2013-08-31 | 2013-12-25 | 西安电子科技大学 | Sparse representation and low-rank double restraints-based nonlocal denoising method |
| CN104200439A (en) * | 2014-09-05 | 2014-12-10 | 西安电子科技大学 | Image super-resolution method based on adaptive filtering and regularization constraint |
| CN105021280A (en) * | 2015-08-13 | 2015-11-04 | 西安电子科技大学 | Panchromatic-and-compressed-spectral-imaging-based image super-resolution system and method |
| CN105590296A (en) * | 2015-12-07 | 2016-05-18 | 天津大学 | Dual-dictionary learning-based single-frame image super-resolution reconstruction method |
-
2016
- 2016-09-19 CN CN201610832828.2A patent/CN106651770B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102156878A (en) * | 2011-04-08 | 2011-08-17 | 南京邮电大学 | Sparse embedding with manifold information-based human face identification method |
| CN103473740A (en) * | 2013-08-31 | 2013-12-25 | 西安电子科技大学 | Sparse representation and low-rank double restraints-based nonlocal denoising method |
| CN104200439A (en) * | 2014-09-05 | 2014-12-10 | 西安电子科技大学 | Image super-resolution method based on adaptive filtering and regularization constraint |
| CN105021280A (en) * | 2015-08-13 | 2015-11-04 | 西安电子科技大学 | Panchromatic-and-compressed-spectral-imaging-based image super-resolution system and method |
| CN105590296A (en) * | 2015-12-07 | 2016-05-18 | 天津大学 | Dual-dictionary learning-based single-frame image super-resolution reconstruction method |
Non-Patent Citations (2)
| Title |
|---|
| High-resolution spectral imaging based on coded dispersion;Yaohai Lin 等;《Europe PMC》;20130201;第194-200页 |
| 基于自学习的稀疏正则化图像超分辨率方法;李娟 等;《仪器仪表学报 》;20150311;第1041-1048页 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106651770A (en) | 2017-05-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liang et al. | Details or artifacts: A locally discriminative learning approach to realistic image super-resolution | |
| CN110119780B (en) | Hyper-spectral image super-resolution reconstruction method based on generation countermeasure network | |
| CN106952228B (en) | Super-resolution reconstruction method of single image based on image non-local self-similarity | |
| Zhang et al. | Learning multiple linear mappings for efficient single image super-resolution | |
| CN109035142B (en) | Satellite image super-resolution method combining countermeasure network with aerial image prior | |
| Ren et al. | Single image super-resolution using local geometric duality and non-local similarity | |
| CN105513033B (en) | A kind of super resolution ratio reconstruction method that non local joint sparse indicates | |
| CN106651770B (en) | Multispectral super-resolution imaging reconstructing method based on Laplce's norm regularization | |
| CN104657962B (en) | The Image Super-resolution Reconstruction method returned based on cascading linear | |
| Huang et al. | Single image super-resolution via multiple mixture prior models | |
| CN111640059B (en) | Multi-dictionary image super-resolution method based on Gaussian mixture model | |
| Ding et al. | Tgan: Deep tensor generative adversarial nets for large image generation | |
| Yang et al. | Image super-resolution based on deep neural network of multiple attention mechanism | |
| CN106067165B (en) | High spectrum image denoising method based on clustering sparse random field | |
| CN112184549A (en) | Super-resolution image reconstruction method based on space-time transformation technology | |
| Xue et al. | Research on gan-based image super-resolution method | |
| CN106296583B (en) | Based on image block group sparse coding and the noisy high spectrum image ultra-resolution ratio reconstructing method that in pairs maps | |
| Wen et al. | The power of complementary regularizers: Image recovery via transform learning and low-rank modeling | |
| US20240046602A1 (en) | Hyperspectral image distributed restoration method and system based on graph signal processing and superpixel segmentation | |
| Alonso-Fernandez et al. | Super-resolution for selfie biometrics: Introduction and application to face and iris | |
| Ji et al. | From patches to images: a nonparametric generative model | |
| Güven et al. | An augmented Lagrangian method for image reconstruction with multiple features | |
| Zhang et al. | Image super-resolution via RL-CSC: when residual learning meets convolutional sparse coding | |
| Wang et al. | Image super-resolution via self-similarity learning and conformal sparse representation | |
| Liu et al. | Fine-grained scale space learning for single image super-resolution |
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 |