CN107833182A - The infrared image super resolution ratio reconstruction method of feature based extraction - Google Patents
The infrared image super resolution ratio reconstruction method of feature based extraction Download PDFInfo
- Publication number
- CN107833182A CN107833182A CN201711158675.9A CN201711158675A CN107833182A CN 107833182 A CN107833182 A CN 107833182A CN 201711158675 A CN201711158675 A CN 201711158675A CN 107833182 A CN107833182 A CN 107833182A
- Authority
- CN
- China
- Prior art keywords
- image
- infrared image
- resolution
- super
- feature extraction
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 63
- 238000000605 extraction Methods 0.000 title claims abstract description 23
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 12
- 230000000877 morphologic effect Effects 0.000 claims abstract description 5
- 238000010586 diagram Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 7
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 230000007797 corrosion Effects 0.000 claims description 4
- 238000005260 corrosion Methods 0.000 claims description 4
- 102100037651 AP-2 complex subunit sigma Human genes 0.000 claims description 3
- 101000806914 Homo sapiens AP-2 complex subunit sigma Proteins 0.000 claims description 3
- 230000006740 morphological transformation Effects 0.000 claims description 3
- 230000002708 enhancing effect Effects 0.000 claims description 2
- 230000000717 retained effect Effects 0.000 claims description 2
- 230000003628 erosive effect Effects 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 5
- 238000003384 imaging method Methods 0.000 description 6
- 230000007547 defect Effects 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 238000012549 training Methods 0.000 description 4
- 230000000007 visual effect Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 2
- 238000003331 infrared imaging Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000004660 morphological change Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 239000013589 supplement Substances 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
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared image
-
- 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
-
- 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
- G06T2207/20036—Morphological image processing
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The present invention discloses a kind of infrared image super resolution ratio reconstruction method of feature based extraction, including:The first step, nonsingular decomposition is carried out to the original infrared image of collection, filters out the noise jamming in original infrared image;Second step, image after denoising is used feature extraction is carried out based on Gabor filter, and the feature after extraction is further enhanced with based on morphologic method;3rd step, enhanced characteristic pattern is subjected to Its Sparse Decomposition, sparse coefficient is solved using method of Lagrange multipliers, and coefficient reconstruction required by use goes out the super-resolution image of original infrared image.The present invention can effectively avoid picture noise when carrying out super-resolution rebuilding to infrared image to rebuilding the influence of effect, meanwhile, because the characteristic pattern that is extracted has preferably openness, also further improve the quality of reconstruction image.
Description
Technical Field
The invention belongs to the technical field of image processing, and particularly relates to a super-resolution reconstruction method for an infrared image.
Background
The infrared imaging technology is characterized by sensing the infrared ray radiation degree of an object and then imaging, can effectively overcome the defect that the target cannot be observed due to objective factors such as rain fog, shielding, insufficient illumination and the like, can effectively monitor for 24 hours in all weather, and is widely applied.
In the imaging process, images can only obtain relatively low-resolution images due to the influence of a series of interference factors such as optical lens distortion, atmospheric blurring, sensor blurring, optical blurring, motion blurring and noise. Conventional methods for improving infrared image quality are often limited to making image details visible by improving contrast. However, as digital images are increasingly applied to various fields, the demand for high resolution images is increasing. Such as medical images, satellite remote sensing, video monitoring, etc. The higher density of pixels in a high resolution image provides more detail than a low resolution image of the same size, which is essential in many practical applications, and which cannot be obtained with only improved contrast. However, due to the limitations of the technical process level, the development cost, the volume requirement, and the like, there are many limitations on improving the resolution of the infrared image from the aspect of hardware, and the resolution of the existing passive infrared detector is generally low. Therefore, the resolution of the infrared image should be improved from a software perspective, and the improvement of the quality of the infrared image is also one of important ways.
The super-resolution reconstruction concept was proposed by Harris and Goodman in the last 60 th century. The method has the advantages that the image resolution is improved in a software mode, the limitation of inherent information of the image and system hardware conditions is broken through, the spatial resolution of original data is improved on the premise of not replacing the conventional imaging equipment and not increasing the system cost, the resolution distortion caused by image degradation and image discretization is overcome, the defect of the spatial resolution of the original image is made up, the visual effect of the image is improved, and the method is favorable for further analyzing, processing and identifying the image. Therefore, when the super-resolution reconstruction technology is applied to the imaging process of the infrared system, the hardware level limitation of the existing infrared imaging system can be effectively broken through, the imaging spatial resolution of the observation equipment is improved, the characteristic information such as image edges is enhanced, the image identification capability and the identification precision of the observation equipment are further improved, and the application of the super-resolution reconstruction technology to various fields of military affairs and civil use is facilitated.
Since the theoretical basis of the super-resolution reconstruction technology established by Harris in 1964, scholars at home and abroad put forward a plurality of methods in the field.
In 2006, a method based on MAP estimation is proposed by f.sroubek and j.flusser, and multi-frame low-resolution images are fused to realize super-resolution reconstruction, so that the information defect of the acquired low-resolution images and the unpredictable error between frames are overcome.
In 2007, Carlos Miravet et al propose a super-resolution reconstruction method based on learning, design a structure training sampling image combining a neural network and a genetic method, and then obtain a high-resolution image through linear filtering.
In 2010, Liuyang et al propose a super-resolution image restoration algorithm based on wiener filtering and POCS mixing, and effectively suppress the defects of ringing effect and amplified noise easily generated in the classical POCS algorithm.
In 2013, Yuzhang Chen et al propose a MAP-based method for underwater target observation, and simultaneously combine a point spread function and regularization to further improve the image resolution and the robustness of the algorithm.
In 2015, Kwok-Wai Hung et al propose a FIR-based wiener filtering method, which only performs super-resolution reconstruction on a single-frame low-resolution image, and although parameters in the algorithm adopt empirical values, the method can achieve good effects when applied in various environments.
In 2016, y.tao et al proposed a GPT method based on non-redundant sub-pixels of the original image and applied to remote sensing images to improve resolution by a factor of five.
However, among many theoretical methods, there are few studies on super-resolution reconstruction techniques for infrared images, and the quality of the image to be reconstructed is generally good. In recent years, learning-based reconstruction methods have been developed more rapidly, which are believed to supplement the missing information of the original low-resolution images. The super-resolution reconstruction of the image is realized by establishing a super-complete dictionary set of the high-low resolution image by taking sparse representation as a representative. However, the reconstruction effect is easily affected by information redundancy due to the adoption of the over-complete dictionary in the reconstruction process.
Disclosure of Invention
The invention aims to provide a super-resolution reconstruction method based on feature extraction, which aims to solve the problems that an infrared image is easily interfered by noise and the imaging quality is poor in the prior art. The method provided by the invention has the advantages that the mode of combining feature extraction and sparse representation is utilized, and the image features are extracted, so that the sparsity is improved, the influence of redundant information is reduced, and the super-resolution reconstruction effect of the infrared image is improved.
In order to achieve the purpose, the invention adopts the following technical scheme:
the infrared image super-resolution reconstruction method based on feature extraction comprises the following steps:
performing singular decomposition on an acquired original infrared image, and filtering noise interference in the original infrared image;
secondly, extracting the features of the denoised image by adopting a Gabor-based filter, and further enhancing the extracted features by using a morphology-based method;
and thirdly, performing sparse decomposition on the enhanced characteristic diagram, solving a sparse coefficient by adopting a Lagrange multiplier method, and reconstructing a super-resolution image of the original infrared image by using the solved coefficient.
Further, in the first step, the size of the original infrared image Y is mxn; after Singular Value Decomposition (SVD), obtaining a diagonal singular value matrix sigma-diag (sigma 1, sigma 2, …, sigma L) consisting of L singular values, wherein L is the rank of a graph Y, and the singular values are arranged from large to small; setting a threshold value sigmathRemoving the singular value lower than the threshold value, and reconstructing an image by using the retained singular value and the corresponding singular vector to realize denoising;
further, an adaptive threshold σthIs selected by
The reconstructed image after being denoised is Y'm×n≈Um×m∑′m×nVn×n TWherein the singular value matrix is sigma'm×n=diag(σ1,...σth)。
Further, in the second step, when a Gabor filter is used for feature extraction, multi-directional feature extraction is carried out; meanwhile, information covered by each characteristic diagram is considered when the characteristic diagrams are fused, and different weight coefficients are adopted for fusion.
Further, the two-dimensional gabor filter is
Where (i, j) is the spatial pixel, σiAnd σjIs the standard deviation, f0Is a spatial domain frequency; if the original coordinate system i 'icos theta + jssin theta and j' isin theta + jcos theta are transformed, a plurality of directional gabor filters are obtained by adjusting the angle parameter theta.
Further, 6 different angle directions are selected to construct a gabor filter, and convolution operation is performed on the gabor filter and the denoised image respectively to obtain a series of characteristic images F (i, j) ═ Fθ(i,j)|θ=0°,30°,60°,90°,120°,150°};
Wherein, Fθ(i,j)=Y'(i,j)*gθ(i,j),gθIs a gabor filter at an angle θ; the 6 different angles are 0 °, 30 °, 60 °, 90 °, 120 ° and 150 °, respectively.
Furthermore, in the second step, the image characteristics contained in the characteristic diagrams in different directions are different, and different weight coefficients are adopted during characteristic fusion; coefficient of
Wherein, M (F)θ(i, j)) is a feature map Fθ(ii) the mean square error of (i, j);
the feature map fusion result is
Furthermore, in the second step, the image is processed by adopting expansion and corrosion operation
Wherein,and Θ represents expansion, corrosion operations; b is a morphological operation template, u and v are intra-template elements; meanwhile, gradient information is considered in the morphological transformation process, wherein F ″ (i-u, j-v) ═ F + Δ F') (i-u, j-v) is used for adding the gradient information in the original image; f "(i + u, j + v) ═ F- Δ F') (i + u, j + v) is to subtract the gradient information from the original image;
the enhanced signature is F ".
Further, the third step specifically includes:
for the enhanced feature map F' in the low-resolution dictionary DlDecomposing; by using L1The norm is solved as follows:
min||αl||1s.t.||Dlαl-F”||≤ξ
wherein, αlFor sparse coding from the low resolution profile, s.t. represents constraints, ξ is the first positive real number set.
The above formula is optimized to ensure the continuity and consistency of the reconstructed images
min||α||1s.t.||Dlα-F”||≤ξ1
||PDhα-Q||≤ξ2
Wherein Q is the part overlapped with the previous reconstructed image block, P is the size of a dictionary which is extracted and is adaptive to the size of the overlapped part, α is the sparse code to be solved, ξ1ξ for a set second positive real number2Is a set third positive real number; dhFor a high-resolution dictionary, a Lagrange multiplier method is introduced for solving to obtain the following formula
Wherein,λ ═ 0.1 is the relaxation factor;
solving for the coefficient α, via x ═ Dhα reconstructing high-resolution image blocks and combining the image blocks to form a complete high-resolution image.
Compared with the prior art, the invention has the following beneficial effects:
the method takes a single-frame image as input, adopts a method based on a visual mechanism to extract the characteristics of an original low-resolution infrared image, and improves the image sparsity; meanwhile, the extracted characteristic diagram is enhanced by adopting a morphology combined gradient method, and key information in the image is further highlighted. The enhanced feature map is used for learning a training set, the influence of redundant information in an over-complete dictionary in the reconstruction process is reduced, the lack of prior knowledge is overcome, an additional sample base is not needed, and the quality of the reconstructed high-resolution image is improved.
Drawings
FIG. 1 is an overall flow chart of a super-resolution reconstruction method based on feature extraction;
fig. 2 is a specific flowchart of a super-resolution reconstruction method based on feature extraction.
Detailed Description
Reference will now be made in detail to the present method for further illustrating the embodiments of the present invention with reference to the accompanying drawings.
Referring to fig. 1 and 2, the method for reconstructing super-resolution infrared image based on feature extraction of the present invention includes:
firstly, denoising an acquired original infrared image: the collected original infrared image often contains a large amount of noise, and the image is denoised firstly.
Singular Value Decomposition (SVD) is a matrix decomposition method in linear algebra, can reflect the intrinsic algebra essence of images, and can be used as an important method for mode classification.
Decomposing original noisy low-resolution infrared image Y into Y through SVDm×n=Um×m∑m×nVn×n T。
Wherein the size of the image Y is mxn; after singular value decomposition, a diagonal singular value matrix sigma-diag (sigma 1, sigma 2, …, sigma L) composed of L singular values is obtained, wherein L is the rank of the graph Y, and the singular values are arranged from large to small. The size of the singular value reflects different components and characteristics of the image, and the larger singular value is related to a signal in the image; in contrast, smaller singular values correspond to noise in the image. By setting a threshold value sigmathAnd eliminating the singular value lower than the threshold value, and reconstructing the image by using the reserved singular value and the corresponding singular vector to realize denoising. Compared with other denoising methods, the method can better retain the edge characteristics of image information while denoising.
Adaptive threshold σthIs selected by
Reserve greater than or equal to σthReconstructing an image to realize denoising according to the singular value of the image; the reconstructed image after being denoised is Y'm×n≈Um×m∑′m×nVn×n TWherein the singular value matrix is sigma'm×n=diag(σ1,...σth)。
And secondly, extracting the characteristics of the denoised original low-resolution image. The invention adopts a method combining a Gabor filter and morphological change to realize the process; the method specifically comprises the following steps:
2.1), the characteristics of the Gabor filter are similar to the human visual mechanism, and the Gabor filter is used for extracting the image characteristics.
The two-dimensional gabor filter is
Where (i, j) is the spatial pixel, σiAnd σjIs the standard deviation, f0Is the spatial domain frequency. If the original coordinate system i 'icos θ + jssin θ is transformed, and j' isin θ + jcos θ, a plurality of directional gabor filters can be obtained by adjusting the angle parameter θ.
The method selects 6 different angle directions to construct a gabor filter, and performs convolution operation on the gabor filter and the denoised image respectively to obtain a series of characteristic graphs F (i, j) ═ Fθ(i,j)|θ=0°,30°,60°,90°,120°,150°}。
Wherein Fθ(i,j)=Y'(i,j)*gθ(i,j),gθA gabor filter at an angle theta.
And fusing the characteristic graphs in different directions, and adopting different weight coefficients.
Coefficient of
Wherein, M (F)θ(i, j)) is a feature map Fθ(i, j) mean square error.
The feature map fusion result is
2.2) combining a morphological method to further enhance the extracted characteristic map.
The feature map is further enhanced using a morphology-based approach.
The method comprises
Wherein,and Θ represents expansion, corrosion operations; b is a morphological operation template, u and v are intra-template elements; meanwhile, gradient information is considered in the morphological transformation process, wherein F ″ (i-u, j-v) ═ F + Δ F') (i-u, j-v) is used for adding the gradient information in the original image; f "(i + u, j + v) ═ F- Δ F') (i + u, j + v) is to subtract the gradient information from the original image;
the enhanced signature is F ".
And thirdly, training by using a feature map of the low-resolution image to obtain a sparse expression coefficient, and training and learning by using an over-complete dictionary to obtain a high-resolution image.
For the feature map F' after enhancement in the low-resolution dictionary DlThe decomposition is carried out. Since the feature map F' has sufficient sparsity, L is adopted1The norm is solved as follows.
min||αl||1s.t.||Dlαl-F”||≤ξ
Wherein, αlFor sparse coding from the low resolution profile, s.t. represents constraints, ξ is the first positive real number set.
The above formula is optimized to ensure the continuity and consistency of the reconstructed images
min||α||1s.t.||Dlα-F”||≤ξ1
||PDhα-Q||≤ξ2
Wherein Q is the part overlapped with the previous reconstructed image block, P is the size of a dictionary which is extracted and is adaptive to the size of the overlapped part, α is the sparse code to be solved, ξ1ξ for a set second positive real number2Is a set third positive real number; dhFor a high-resolution dictionary, a Lagrange multiplier method is introduced for solving to obtain the following formula
Wherein,λ ═ 0.1 is the relaxation factor;
solving for the coefficient α, via x ═ Dhα reconstructing high-resolution image blocks and combining the image blocks to form a complete high-resolution image.
The above detailed description is intended to illustrate the objects, aspects and advantages of the present invention, and it should be understood that the above detailed description is only exemplary of the present invention and is not intended to limit the scope of the present invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. The infrared image super-resolution reconstruction method based on feature extraction is characterized by comprising the following steps:
performing singular decomposition on an acquired original infrared image, and filtering noise interference in the original infrared image;
secondly, extracting the features of the denoised image by adopting a Gabor-based filter, and further enhancing the extracted features by using a morphology-based method;
and thirdly, performing sparse decomposition on the enhanced characteristic diagram, solving a sparse coefficient by adopting a Lagrange multiplier method, and reconstructing a super-resolution image of the original infrared image by using the solved coefficient.
2. The infrared image super-resolution reconstruction method based on feature extraction as claimed in claim 1, wherein: in the first step, the size of an original infrared image Y is mxn; obtaining a diagonal singular value matrix sigma-diag (sigma 1, sigma 2, …, sigma L) consisting of L singular values after singular value decomposition, wherein L is the rank of a graph Y, and the singular values are arranged from large to small; setting a threshold value sigmathRemoving the singular value lower than the threshold value, and reconstructing an image by using the retained singular value and the corresponding singular vector to realize denoising;
adaptive threshold σthIs selected by
The reconstructed image after being denoised is Y'm×n≈Um×m∑′m×nVn×n TWherein the singular value matrix is sigma'm×n=diag(σ1,...σth)。
3. The infrared image super-resolution reconstruction method based on feature extraction as claimed in claim 1, wherein: in the second step, when a Gabor filter is used for feature extraction, multi-directional feature extraction is carried out; meanwhile, information covered by each characteristic diagram is considered when the characteristic diagrams are fused, and different weight coefficients are adopted for fusion.
4. The feature extraction-based infrared image super-resolution reconstruction method according to claim 3, characterized in that: the two-dimensional gabor filter is
Where (i, j) is the spatial pixel, σiAnd σjIs the standard deviation, f0Is a spatial domain frequency; if the original coordinate system i 'icos theta + jssin theta and j' isin theta + jcos theta are transformed, a plurality of directional gabor filters are obtained by adjusting the angle parameter theta.
5. The feature extraction-based infrared image super-resolution reconstruction method according to claim 3, characterized in that: selecting 6 different angle directions to construct a gabor filter, and performing convolution operation on the gabor filter and the denoised image respectively to obtain a series of characteristic graphs F (i, j) ═ Fθ(i,j)|θ=0°,30°,60°,90°,120°,150°};
Wherein, Fθ(i,j)=Y'(i,j)*gθ(i,j);gθIs a gabor filter at an angle θ; the 6 different angles are 0 °, 30 °, 60 °, 90 °, 120 ° and 150 °, respectively.
6. The feature extraction-based infrared image super-resolution reconstruction method according to claim 3, characterized in that: in the second step, the image characteristics contained in the characteristic diagrams in different directions are different, and different weight coefficients are adopted when the characteristics are fused, wherein the coefficient is
Wherein, M (F)θ(i, j)) is a feature map Fθ(ii) the mean square error of (i, j);
the feature map fusion result is
7. The infrared image super-resolution reconstruction method based on feature extraction as claimed in claim 1, wherein: in the second step, the image is processed by expansion and erosion operations, the method being
⊕ and theta represent expansion and corrosion operations, B is a morphological operation template, u and v are elements in the template, and gradient information is considered in the morphological transformation process, wherein F (i-u, j-v) ═ F + DeltaF ') (i-u, j-v) is to add gradient information in the original image, and F (i + u, j + v) ═ F') (i + u, j + v) is to subtract gradient information from the original image;
f' is a characteristic diagram obtained after enhancement.
8. The infrared image super-resolution reconstruction method based on feature extraction as claimed in claim 1, wherein: the third step specifically comprises:
for the enhanced feature map F' in the low-resolution dictionary DlDecomposing; by using L1The norm is solved as follows:
min||αl||1s.t.||Dlαl-F”||≤ξ
wherein, αlS.t. represents constraint, ξ is a first positive real number set;
the above formula is optimized to ensure the continuity and consistency of the reconstructed images
min||α||1s.t.||Dlα-F”||≤ξ1
||PDhα-Q||≤ξ2
Wherein Q is the part overlapped with the previous reconstructed image block, P is the size of a dictionary which is extracted and is adaptive to the size of the overlapped part, α is the sparse code to be solved, ξ1ξ for a set second positive real number2Is a set third positive real number; dhFor a high-resolution dictionary, a Lagrange multiplier method is introduced to solve to obtain the following formula:
<mrow> <munder> <mi>min</mi> <mi>&alpha;</mi> </munder> <mo>|</mo> <mo>|</mo> <msup> <mi>D</mi> <mo>&prime;</mo> </msup> <mi>&alpha;</mi> <mo>-</mo> <mi>G</mi> <mo>|</mo> <msubsup> <mo>|</mo> <mn>2</mn> <mn>2</mn> </msubsup> <mo>+</mo> <mi>&lambda;</mi> <mo>|</mo> <mo>|</mo> <mi>&alpha;</mi> <mo>|</mo> <msub> <mo>|</mo> <mn>1</mn> </msub> <mo>;</mo> </mrow>
wherein,λ ═ 0.1 is the relaxation factor;
solving for the coefficient α, via x ═ Dhα reconstructing high-resolution image blocks and combining the image blocks to form a complete high-resolution image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711158675.9A CN107833182A (en) | 2017-11-20 | 2017-11-20 | The infrared image super resolution ratio reconstruction method of feature based extraction |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711158675.9A CN107833182A (en) | 2017-11-20 | 2017-11-20 | The infrared image super resolution ratio reconstruction method of feature based extraction |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107833182A true CN107833182A (en) | 2018-03-23 |
Family
ID=61652238
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711158675.9A Pending CN107833182A (en) | 2017-11-20 | 2017-11-20 | The infrared image super resolution ratio reconstruction method of feature based extraction |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107833182A (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108648143A (en) * | 2018-04-17 | 2018-10-12 | 中国科学院光电技术研究所 | Image resolution enhancement method using sequence image |
CN108898557A (en) * | 2018-05-30 | 2018-11-27 | 商汤集团有限公司 | Image recovery method and device, electronic equipment, computer program and storage medium |
CN109214989A (en) * | 2018-09-04 | 2019-01-15 | 四川大学 | Single image super resolution ratio reconstruction method based on Orientation Features prediction priori |
CN109697691A (en) * | 2018-12-27 | 2019-04-30 | 重庆大学 | A kind of limited view projection method for reconstructing based on the optimization of the biregular item of L0 norm and singular value threshold decomposition |
CN112508786A (en) * | 2020-12-03 | 2021-03-16 | 武汉大学 | Satellite image-oriented arbitrary-scale super-resolution reconstruction method and system |
US11346938B2 (en) | 2019-03-15 | 2022-05-31 | Msa Technology, Llc | Safety device for providing output to an individual associated with a hazardous environment |
CN116205806A (en) * | 2023-01-28 | 2023-06-02 | 荣耀终端有限公司 | Image enhancement method and electronic equipment |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7809155B2 (en) * | 2004-06-30 | 2010-10-05 | Intel Corporation | Computing a higher resolution image from multiple lower resolution images using model-base, robust Bayesian estimation |
CN102722865A (en) * | 2012-05-22 | 2012-10-10 | 北京工业大学 | Super-resolution sparse representation method |
CN104063856A (en) * | 2014-05-28 | 2014-09-24 | 北京大学深圳研究生院 | Rapid super-resolution image reconstruction method and device |
CN104766273A (en) * | 2015-04-20 | 2015-07-08 | 重庆大学 | Infrared image super-resolution reestablishing method based on compressed sensing theory |
JP2016035512A (en) * | 2014-08-04 | 2016-03-17 | 国立研究開発法人物質・材料研究機構 | Optical super-resolution medium, hyper lens, method for producing the same and hyper lens array |
CN106934766A (en) * | 2017-03-15 | 2017-07-07 | 西安理工大学 | A kind of infrared image super resolution ratio reconstruction method based on rarefaction representation |
-
2017
- 2017-11-20 CN CN201711158675.9A patent/CN107833182A/en active Pending
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7809155B2 (en) * | 2004-06-30 | 2010-10-05 | Intel Corporation | Computing a higher resolution image from multiple lower resolution images using model-base, robust Bayesian estimation |
CN102722865A (en) * | 2012-05-22 | 2012-10-10 | 北京工业大学 | Super-resolution sparse representation method |
CN104063856A (en) * | 2014-05-28 | 2014-09-24 | 北京大学深圳研究生院 | Rapid super-resolution image reconstruction method and device |
JP2016035512A (en) * | 2014-08-04 | 2016-03-17 | 国立研究開発法人物質・材料研究機構 | Optical super-resolution medium, hyper lens, method for producing the same and hyper lens array |
CN104766273A (en) * | 2015-04-20 | 2015-07-08 | 重庆大学 | Infrared image super-resolution reestablishing method based on compressed sensing theory |
CN106934766A (en) * | 2017-03-15 | 2017-07-07 | 西安理工大学 | A kind of infrared image super resolution ratio reconstruction method based on rarefaction representation |
Non-Patent Citations (2)
Title |
---|
ZENG, JIAN 等: "The infrared image closely spaced objects super resolution method based on sparse reconstruction under the noise environment", 《FOURTH INTERNATIONAL CONFERENCE ON OPTICAL AND PHOTONICS ENGINEERING》 * |
刘春生: "人脸特征检测与疲劳状态识别研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108648143A (en) * | 2018-04-17 | 2018-10-12 | 中国科学院光电技术研究所 | Image resolution enhancement method using sequence image |
CN108898557A (en) * | 2018-05-30 | 2018-11-27 | 商汤集团有限公司 | Image recovery method and device, electronic equipment, computer program and storage medium |
CN109214989A (en) * | 2018-09-04 | 2019-01-15 | 四川大学 | Single image super resolution ratio reconstruction method based on Orientation Features prediction priori |
CN109214989B (en) * | 2018-09-04 | 2019-08-13 | 四川大学 | Single image super resolution ratio reconstruction method based on Orientation Features prediction priori |
CN109697691A (en) * | 2018-12-27 | 2019-04-30 | 重庆大学 | A kind of limited view projection method for reconstructing based on the optimization of the biregular item of L0 norm and singular value threshold decomposition |
CN109697691B (en) * | 2018-12-27 | 2022-11-25 | 重庆大学 | Dual-regularization-term-optimized finite-angle projection reconstruction method based on L0 norm and singular value threshold decomposition |
US11346938B2 (en) | 2019-03-15 | 2022-05-31 | Msa Technology, Llc | Safety device for providing output to an individual associated with a hazardous environment |
CN112508786A (en) * | 2020-12-03 | 2021-03-16 | 武汉大学 | Satellite image-oriented arbitrary-scale super-resolution reconstruction method and system |
CN112508786B (en) * | 2020-12-03 | 2022-04-29 | 武汉大学 | Satellite image-oriented arbitrary-scale super-resolution reconstruction method and system |
CN116205806A (en) * | 2023-01-28 | 2023-06-02 | 荣耀终端有限公司 | Image enhancement method and electronic equipment |
CN116205806B (en) * | 2023-01-28 | 2023-09-19 | 荣耀终端有限公司 | Image enhancement method and electronic equipment |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107833182A (en) | The infrared image super resolution ratio reconstruction method of feature based extraction | |
Lau et al. | Restoration of atmospheric turbulence-distorted images via RPCA and quasiconformal maps | |
Lefkimmiatis et al. | Hessian Schatten-norm regularization for linear inverse problems | |
AU2020100462A4 (en) | Edge-preserving image super-resolution via low rank and total variation model | |
Chierchia et al. | A nonlocal structure tensor-based approach for multicomponent image recovery problems | |
CN109509160A (en) | Hierarchical remote sensing image fusion method utilizing layer-by-layer iteration super-resolution | |
CN107067380B (en) | High-resolution image reconstruction method based on low-rank tensor and hierarchical dictionary learning | |
Li et al. | Underwater image high definition display using the multilayer perceptron and color feature-based SRCNN | |
CN102243711A (en) | Neighbor embedding-based image super-resolution reconstruction method | |
CN107292819A (en) | A kind of infrared image super resolution ratio reconstruction method protected based on edge details | |
Zhang et al. | Dense haze removal based on dynamic collaborative inference learning for remote sensing images | |
CN105590296A (en) | Dual-dictionary learning-based single-frame image super-resolution reconstruction method | |
Wang et al. | Medical image super-resolution analysis with sparse representation | |
CN105608674B (en) | A kind of image enchancing method based on image registration, interpolation and denoising | |
Zhou et al. | Multi-scale dilated convolution neural network for image artifact correction of limited-angle tomography | |
Li et al. | Ruiesr: Realistic underwater image enhancement and super resolution | |
Hua et al. | Iterative residual network for image dehazing | |
Zhang et al. | Multisensor Infrared and Visible Image Fusion via Double Joint Edge Preservation Filter and Nonglobally Saliency Gradient Operator | |
Zhang | A selection of image processing techniques: from fundamentals to research front | |
CN109615584B (en) | SAR image sequence MAP super-resolution reconstruction method based on homography constraint | |
Cengiz et al. | The Effect of Super Resolution Method on Classification Performance of Satellite Images | |
Jung et al. | Multispectral fusion of rgb and nir images using weighted least squares and convolution neural networks | |
Shin et al. | LoGSRN: Deep super resolution network for digital elevation model | |
Tun et al. | Joint Training of Noisy Image Patch and Impulse Response of Low-Pass Filter in CNN for Image Denoising | |
Kandula et al. | Distortion Disentanglement and Knowledge Distillation for Satellite Image Restoration |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180323 |
|
WD01 | Invention patent application deemed withdrawn after publication |