CN107274343A - Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework - Google Patents
Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework Download PDFInfo
- Publication number
- CN107274343A CN107274343A CN201710403557.3A CN201710403557A CN107274343A CN 107274343 A CN107274343 A CN 107274343A CN 201710403557 A CN201710403557 A CN 201710403557A CN 107274343 A CN107274343 A CN 107274343A
- Authority
- CN
- China
- Prior art keywords
- spectrum
- library
- spectra
- remote sensing
- spectral
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
- G06T3/4053—Super resolution, i.e. output image resolution higher than sensor resolution
Abstract
Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework, first, obtains multi-spectral remote sensing image；Secondly, selection can cover the library of spectra of the image observation region atural object classification；Then, provide the band class information for intending rebuilding high-spectrum remote sensing and carry out wave band with library of spectra and match；Then, dictionary training is carried out using the library of spectra data after matching, obtained complete spectrum dictionary；Again by multispectral image under the conditions of without nonnegativity restrictions rarefaction representation, obtain rarefaction representation coefficient；Finally, the high spectrum image after spectrum super-resolution is obtained using spectrum dictionary, sparse coefficient；The present invention carries out spectrum super-resolution rebuilding under framework of sparse representation merely with a width multispectral image, obtains the high spectrum image under Same Scene, reduces data acquisition difficulty；Meanwhile, deficiency of the multispectral image to spectral details descriptive power is compensate for, the precision and validity of spectrum acquisition is effectively increased.
Description
Technical field
Obtained the invention belongs to remote sensing images and processing technology field, it is adaptable to the spectrum super-resolution of multi-spectral remote sensing image
Rate is rebuild, the multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under more particularly to a kind of sparse framework.
Background technology
High-spectrum remote sensing is in Objects recognition with being widely applied in classification, environmental monitoring.Spectral resolution
Beam splitting system of the lifting dependent on imaging spectrometer, because its optical texture is complicated, volume is big, quality weight, is mounted in satellite more
Or on air remote sensing platform, this causes the convenience and economy of high-spectral data acquisition to be rather limited；Meanwhile, spectrum point
The lifting of resolution causes every spectral bandwidth to narrow, and larger instantaneous field of view (Instantaneous must be used during imaging
Field of View, IFOV) enough light quantum could be accumulated to maintain the signal to noise ratio of imaging, and instantaneous field of view and space
Resolution ratio is two technical indicators mutually restricted, and the increase of instantaneous field of view can cause the reduction of spatial resolution.But permitted
Many remote sensing application fields, higher spatial resolution and spectral resolution are all indispensable, therefore how to keep higher
Spectrum super-resolution image is obtained on the basis of spatial resolution to have important practical significance.
At this stage, the Remote sensing image fusion based on Decomposition of Mixed Pixels is by merging the multispectral figure under Same Scene
Picture and high spectrum image, so as to obtain the remote sensing images with high spectral resolution and maintain higher spatial resolution.It is mixed
Close pixel analysis and remote sensing images are decomposed into various atural object compositions (end member), then solved by multispectral image under nonnegativity restrictions
Ratio (abundance) shared by each composition.But the determination of end member number and the extraction of pure end member are all deposited in Decomposition of Mixed Pixels
In certain difficulty, therefore there is spectrum problem of dtmf distortion DTMF in resulting fused images；Meanwhile, phase light more while under Same Scene
Spectrogram picture is often difficult to obtain with high spectrum image, and this also causes this method to be difficult to promote.
In recent years, framework of sparse representation presents huge potentiality in Remote sensed image super-resolution reconstruction field, and it will be distant
Feel the product that graphical representation is dictionary and sparse coefficient.End member need not be extracted during rarefaction representation, and sparse coefficient is nothing but
Negative limitation, therefore this method can overcome the shortcoming of image fusion technology at this stage so that contain not in multi-source Remote Sensing Images
It can be given full expression to information.In addition, set of the library of spectra as object light modal data in large quantities, in remote sensing images information solution
Translate, classify with being widely used in terms of identification, can effectively provide high-resolution spectral information, it is to avoid to Same Scene simultaneously
The dependence of phase multi-source Remote Sensing Images.Framework of sparse representation is introduced into the spectrum super-resolution rebuilding of remote sensing images by the present invention, by
Library of spectra data training after wave band matching obtains spectrum dictionary, sparse coefficient is solved by multi-spectral remote sensing image, so as to obtain
Obtain the high-spectrum remote sensing after spectrum super-resolution.
The content of the invention
In order to overcome the shortcoming of above-mentioned prior art, it is an object of the invention to provide spectrum is based under a kind of sparse framework
The multi-spectral remote sensing image spectrum super-resolution method in storehouse, the library of spectra data training matched first with wave band obtains spectrum word
Allusion quotation.Secondly, rarefaction representation coefficient is solved by multi-spectral remote sensing image.Finally, light is obtained by spectrum dictionary and sparse coefficient
Compose the high-spectrum remote sensing of super-resolution.The method can obtain spectrum super-resolution on the basis of spatial resolution is kept
High-spectrum remote sensing, and avoid the dependence to Same Scene with phase multi-source Remote Sensing Images.
To achieve these goals, the technical solution adopted by the present invention is：
A kind of multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework, step is as follows：
Step 1, the multi-spectral remote sensing image of the low spectral resolution of a width is obtained；
Step 2, the library of spectra for covering the image observation of this in step 1 region atural object classification is obtained；
Step 3, in step 2 in library of spectra wavelength band, provide intend rebuild high-spectrum remote sensing wave band number and
Its positional information；
Step 4, the wave band in library of spectra is matched with the band class information that step 3 is obtained；
Step 5, using the library of spectra data after being matched in step 4, dictionary training is carried out under the pattern of non-decomposition, is obtained
Cross complete spectrum dictionary D；
Step 6, the excessively complete spectrum dictionary D that the multi-spectral remote sensing image and step 5 obtained using step 1 is obtained,
Without solution rarefaction representation coefficient A under conditions of nonnegativity restrictions；
Step 7, pass throughObtain the high-spectrum remote sensing after spectrum super-resolution.
The multi-spectral remote sensing image Y of the low spectral resolution obtained in the step 1L∈Rb×NWhile obtain spectrum ring
Wave band and space pixel number that function L, b and N are respectively multi-spectral remote sensing image are answered, R is real number space.
Library of spectra match referring to the band class information obtained in step 3 in the step 4, by the following method：
WΩ=min||PWl-Wx||2 (1)
Obtain the band class information W of both matchingsΩ∈RB×1, wherein P ∈ RB×M, it is selection matrix, Wl∈RM×1With Wx∈RB×1
The respectively band class information of library of spectra and spectrum super-resolution remote sensing images, M >=B, respectively library of spectra rebuild EO-1 hyperion with intending
The wave band number of remote sensing images.
Excessively complete spectrum dictionary D utilizes the spectroscopic data matched in library of spectra in the step 5, by K-SVD dictionary learnings
Method is tried to achieve by optimizing following minimization problem：
Wherein, Y ∈ RB×nIt is that band class information is WΩLibrary of spectra data,For sparse coefficient matrix, n is light
Spectrum number in storehouse is composed, k is spectrum dictionary columns, λ1For regularization coefficient.
Rarefaction representation coefficient A solves following 1 norm constraint problem by alternating direction multiplier method (ADMM) in the step 6
Obtain:
Wherein, L ∈ Rb×B, it is spectral response functions, B ＞ ＞ b, λ2For regularization coefficient.
Compared with prior art, the beneficial effects of the invention are as follows：
1) present invention carries out spectrum super-resolution rebuilding merely with a width multispectral image, it is to avoid to Same Scene similarly hereinafter
The dependence of phase multi-source Remote Sensing Images, reduces data acquisition difficulty.
2) present invention introduces the non-decomposition of the spectrum dictionary based on library of spectra under wave band matching strategy under framework of sparse representation
Method for solving, compensate for deficiency of the multispectral image to spectral details descriptive power, improve spectrum super-resolution rebuilding precision.
3) present invention is by without nonnegativity restrictions and non-iterative method solution rarefaction representation coefficient, improving spectrum super-resolution
The accuracy and validity of rate reconstruction image.
Brief description of the drawings
Fig. 1 is the multi-spectral remote sensing image spectrum super-resolution method flow based on library of spectra under sparse framework of the invention
Figure.
Fig. 2 is the 10th wave band EO-1 hyperion reference picture in the embodiment of the present invention.
Fig. 3 is the spectral transfer function that the embodiment of the present invention is used.
Fig. 4 is the spectrum super-resolution result of a certain pixel in the embodiment of the present invention.
Embodiment
Describe embodiments of the present invention in detail with reference to the accompanying drawings and examples.
As shown in figure 1, the multi-spectral remote sensing image spectrum super-resolution based on library of spectra under a kind of sparse framework of the present invention
Method, comprises the following steps：
Step 1, the multi-spectral remote sensing image of the low spectral resolution of a width is obtained；
In the present embodiment, in order to carry out referring to quantitative assessment entirely, multi-spectral remote sensing image Y hereL∈Rb×NBy bloom
Spectrum reference picture drop spectral resolution is obtained, and acquisition process is modeled as into YL=LX.Wherein X ∈ RB×NFor high-spectrum remote-sensing figure
Picture, L ∈ Rb×BFor spectral response functions, respectively B ＞ ＞ b, the wave band number of two images, N is that remote sensing images Spatial Dimension is included
Pixel number, R is real number space；In order to carry out referring to quantitative assessment entirely, spectrometer is surveyed from airborne visible ray/infrared imaging
(Airborne Visible/Infrared Imaging Spectrometer, AVIRIS) gathers high-spectrum remote sensing
97 wave bands higher with library of spectra wave band matching degree are per band image size as EO-1 hyperion reference picture in (see Fig. 2)
300*300, spectral transfer function is L ∈ R4×97(see Fig. 3), therefore, the multispectral image obtained in the present embodiment only have four
Wave band；Multi-spectral remote sensing image under practical application scene is directly collected.
Step 2, the library of spectra for covering the image observation of this in step 1 region atural object classification is obtained；
In the present embodiment, the library of spectra for covering atural object classification corresponding to multi-spectral remote sensing image in step 1 is used
Mineral spectra in United States Geological Survey (United States Geological Survey, UGRS) library of spectra, its wavelength
Scope is 0.4~2.5 μm, it is seen that light resolution ratio is 0.2nm, and near-infrared resolution ratio is 0.5nm, totally 481 curves of spectrum.
Step 3, in step 2 in library of spectra wavelength band, provide intend rebuild high-spectrum remote sensing wave band number and
Its positional information；
In the present embodiment, in order to carry out referring to quantitative assessment entirely, the band class information for intending rebuilding high-spectrum remote sensing is adopted
With the band position information W of EO-1 hyperion reference picture in step 1x∈R97×1, containing 97 wave bands, wave-length coverage is 0.4~1.6
μm。
Step 4, the wave band in library of spectra is matched with the band class information that step 3 is obtained；
In the present embodiment, selection matrix P often row has and only one of which " 1 " element, and remaining is all " 0 " element, is used
Euclidean distance weighs the matching degree of band class information.
Step 5, using the library of spectra data after being matched in step 4, dictionary training is carried out under the pattern of non-decomposition, is obtained
Cross complete spectrum dictionary D；
In the present embodiment, it is to the parameter that the library of spectra data after matching carry out K-SVD dictionary training：Degree of rarefication is
10, dictionary columns k=481
Step 6, the excessively complete spectrum dictionary D that the multi-spectral remote sensing image and step 5 obtained using step 1 is obtained,
Without solution rarefaction representation coefficient A under conditions of nonnegativity restrictions；
In this example, formula (3) is solved using augmented vector approach, its parameter is：Iterations T=1, canonical
Change coefficient lambda2=10-6, LaGrange parameter μ=10-2。
Step 7, pass throughObtain the high-spectrum remote sensing after spectrum super-resolution.
In the present embodiment, the PSNR of the high-spectrum remote sensing after spectrum super-resolution is that 40.91, MSE is 5.27, spectrum
Angle SAM is 1.21, it is seen that present invention obtains high-quality high-spectrum remote sensing.
To sum up, the present invention is under framework of sparse representation, and the spectral information provided using library of spectra is in non-negative and non-decomposition mould
Spectrum dictionary and sparse coefficient are solved under formula, spectrum super-resolution rebuilding is carried out merely with a width multispectral image, is obtained same
High spectrum image under scene, reduces data acquisition difficulty；Meanwhile, the spectrum dictionary based on library of spectra under wave band matching strategy
Build, compensate for deficiency of the multispectral image to spectral details descriptive power, effectively increase spectrum acquisition precision and effectively
Property.
Claims (5)
1. a kind of multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework, it is characterised in that bag
Include following steps：
Step 1, the multi-spectral remote sensing image of the low spectral resolution of a width is obtained；
Step 2, the library of spectra for covering the image observation of this in step 1 region atural object classification is obtained；
Step 3, in step 2 in library of spectra wavelength band, the wave band number for intending rebuilding high-spectrum remote sensing and its position are provided
Confidence ceases；
Step 4, the wave band in library of spectra is matched with the band class information that step 3 is obtained；
Step 5, using the library of spectra data after being matched in step 4, dictionary training is carried out under the pattern of non-decomposition, obtained complete
Standby spectrum dictionary D；
Step 6, the excessively complete spectrum dictionary D that the multi-spectral remote sensing image and step 5 obtained using step 1 is obtained, nothing but
Break a promise and rarefaction representation coefficient A is solved under conditions of beam；
Step 7, pass throughObtain the high-spectrum remote sensing after spectrum super-resolution.
2. the multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework according to claim 1,
Characterized in that, obtaining the multi-spectral remote sensing image Y of low spectral resolution in the step 1L∈Rb×NWhile obtain spectrum ring
Wave band and space pixel number that function L, b and N are respectively multi-spectral remote sensing image are answered, R is real number space.
3. the multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework according to claim 1,
Characterized in that, being logical by with the band class information obtained match the wave band in library of spectra in step 3 in the step 4
Cross formula：
WΩ=min | | PWl-Wx||2
Obtain the band class information W of both matchingsΩ∈RB×1, wherein P ∈ RB×M, it is selection matrix, Wl∈RM×1With Wx∈RB×1Respectively
For library of spectra and the band class information for intending rebuilding high-spectrum remote sensing, M >=B, respectively library of spectra rebuild high-spectrum remote-sensing with intending
The wave band number of image.
4. the multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework according to claim 1,
Characterized in that, excessively complete spectrum dictionary D utilizes the spectroscopic data matched in library of spectra in the step 5, by K-SVD dictionaries
Learning method is tried to achieve by optimizing following minimization problem：
<mrow>
<mi>D</mi>
<mo>=</mo>
<mi>arg</mi>
<mi>min</mi>
<mo>|</mo>
<mo>|</mo>
<mi>Y</mi>
<mo>-</mo>
<mi>D</mi>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>1</mn>
</msub>
<mo>|</mo>
<mo>|</mo>
<mover>
<mi>A</mi>
<mo>^</mo>
</mover>
<mo>|</mo>
<msub>
<mo>|</mo>
<mn>0</mn>
</msub>
</mrow>
Wherein, Y ∈ RB×nIt is that band class information is WΩLibrary of spectra data,For sparse coefficient matrix, n is in library of spectra
Spectrum number, k is spectrum dictionary columns, λ1For regularization coefficient.
5. the multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under sparse framework according to claim 1,
Characterized in that, rarefaction representation coefficient A solves following 1 norm constraint by alternating direction multiplier method (ADMM) in the step 6
Problem is obtained：
<mrow>
<mi>A</mi>
<mo>=</mo>
<mi>argmin</mi>
<mo>|</mo>
<mo>|</mo>
<msub>
<mi>Y</mi>
<mi>L</mi>
</msub>
<mo>-</mo>
<mi>L</mi>
<mi>D</mi>
<mi>A</mi>
<mo>|</mo>
<msubsup>
<mo>|</mo>
<mi>F</mi>
<mn>2</mn>
</msubsup>
<mo>+</mo>
<msub>
<mi>&lambda;</mi>
<mn>2</mn>
</msub>
<mo>|</mo>
<mo>|</mo>
<mi>A</mi>
<mo>|</mo>
<msub>
<mo>|</mo>
<mn>1</mn>
</msub>
</mrow>
Wherein, L ∈ Rb×B, it is spectral response functions, B ＞ ＞ b, λ2For regularization coefficient.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710403557.3A CN107274343A (en) | 2017-06-01 | 2017-06-01 | Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710403557.3A CN107274343A (en) | 2017-06-01 | 2017-06-01 | Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107274343A true CN107274343A (en) | 2017-10-20 |
Family
ID=60065806
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710403557.3A Pending CN107274343A (en) | 2017-06-01 | 2017-06-01 | Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107274343A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818546A (en) * | 2017-11-01 | 2018-03-20 | 淮海工学院 | A kind of positron emission fault image super-resolution rebuilding method based on rarefaction representation |
| CN108596819A (en) * | 2018-03-28 | 2018-09-28 | 广州地理研究所 | A kind of inland optics Complex water body bloom spectrum reconstruction method based on sparse expression |
| CN110852950A (en) * | 2019-11-08 | 2020-02-28 | 中国科学院微小卫星创新研究院 | Hyperspectral image super-resolution reconstruction method based on sparse representation and image fusion |
| CN111353937A (en) * | 2020-02-28 | 2020-06-30 | 南京航空航天大学 | Super-resolution reconstruction method of remote sensing image |
| CN111861885A (en) * | 2020-07-15 | 2020-10-30 | 中国人民解放军火箭军工程大学 | Super-pixel sparse representation method for hyperspectral super-resolution reconstruction |
| CN112102218A (en) * | 2020-09-25 | 2020-12-18 | 北京师范大学 | Fusion method for generating high-spatial-resolution multispectral image |
| CN113723348A (en) * | 2021-09-10 | 2021-11-30 | 中国石油大学(华东) | Hyperspectral mixed pixel decomposition method based on abundance sparsity constraint |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102324047A (en) * | 2011-09-05 | 2012-01-18 | 西安电子科技大学 | High spectrum image atural object recognition methods based on sparse nuclear coding SKR |
| CN103247034A (en) * | 2013-05-08 | 2013-08-14 | 中国科学院光电研究院 | Sparse-spectrum-dictionary hyperspectral image reconstruction method by using compressed sensing |
| CN105528623A (en) * | 2016-01-09 | 2016-04-27 | 北京工业大学 | Imaging spectrum image sparse representation method based on ground object class classification redundant dictionary |
| CN105761234A (en) * | 2016-01-28 | 2016-07-13 | 华南农业大学 | Structure sparse representation-based remote sensing image fusion method |
| CN106780424A (en) * | 2017-01-12 | 2017-05-31 | 清华大学 | A kind of high spectrum image acquisition methods based on only a few optimum choice wave band |
-
2017
- 2017-06-01 CN CN201710403557.3A patent/CN107274343A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102324047A (en) * | 2011-09-05 | 2012-01-18 | 西安电子科技大学 | High spectrum image atural object recognition methods based on sparse nuclear coding SKR |
| CN103247034A (en) * | 2013-05-08 | 2013-08-14 | 中国科学院光电研究院 | Sparse-spectrum-dictionary hyperspectral image reconstruction method by using compressed sensing |
| CN105528623A (en) * | 2016-01-09 | 2016-04-27 | 北京工业大学 | Imaging spectrum image sparse representation method based on ground object class classification redundant dictionary |
| CN105761234A (en) * | 2016-01-28 | 2016-07-13 | 华南农业大学 | Structure sparse representation-based remote sensing image fusion method |
| CN106780424A (en) * | 2017-01-12 | 2017-05-31 | 清华大学 | A kind of high spectrum image acquisition methods based on only a few optimum choice wave band |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107818546A (en) * | 2017-11-01 | 2018-03-20 | 淮海工学院 | A kind of positron emission fault image super-resolution rebuilding method based on rarefaction representation |
| CN108596819A (en) * | 2018-03-28 | 2018-09-28 | 广州地理研究所 | A kind of inland optics Complex water body bloom spectrum reconstruction method based on sparse expression |
| CN110852950A (en) * | 2019-11-08 | 2020-02-28 | 中国科学院微小卫星创新研究院 | Hyperspectral image super-resolution reconstruction method based on sparse representation and image fusion |
| CN110852950B (en) * | 2019-11-08 | 2023-04-07 | 中国科学院微小卫星创新研究院 | Hyperspectral image super-resolution reconstruction method based on sparse representation and image fusion |
| CN111353937A (en) * | 2020-02-28 | 2020-06-30 | 南京航空航天大学 | Super-resolution reconstruction method of remote sensing image |
| CN111353937B (en) * | 2020-02-28 | 2023-09-29 | 南京航空航天大学 | Super-resolution reconstruction method of remote sensing image |
| CN111861885A (en) * | 2020-07-15 | 2020-10-30 | 中国人民解放军火箭军工程大学 | Super-pixel sparse representation method for hyperspectral super-resolution reconstruction |
| CN112102218A (en) * | 2020-09-25 | 2020-12-18 | 北京师范大学 | Fusion method for generating high-spatial-resolution multispectral image |
| CN112102218B (en) * | 2020-09-25 | 2023-07-07 | 北京师范大学 | Fusion method for generating high-spatial-resolution multispectral image |
| CN113723348A (en) * | 2021-09-10 | 2021-11-30 | 中国石油大学(华东) | Hyperspectral mixed pixel decomposition method based on abundance sparsity constraint |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107274343A (en) | Multi-spectral remote sensing image spectrum super-resolution method based on library of spectra under a kind of sparse framework | |
| Bioucas-Dias et al. | Hyperspectral remote sensing data analysis and future challenges | |
| CN104867124B (en) | Multispectral and panchromatic image fusion method based on the sparse Non-negative Matrix Factorization of antithesis | |
| US10192288B2 (en) | Method and system for generating high resolution worldview-3 images | |
| CN106780345B (en) | Hyperspectral image super-resolution reconstruction method based on coupling dictionary and space conversion estimation | |
| Lanaras et al. | Hyperspectral super-resolution with spectral unmixing constraints | |
| Zhang et al. | Image fusion employing adaptive spectral-spatial gradient sparse regularization in UAV remote sensing | |
| Yu et al. | CapViT: Cross-context capsule vision transformers for land cover classification with airborne multispectral LiDAR data | |
| CN115984155A (en) | Hyperspectral, multispectral and panchromatic image fusion method based on spectrum unmixing | |
| CN106780424A (en) | A kind of high spectrum image acquisition methods based on only a few optimum choice wave band | |
| CN116091833A (en) | Attention and transducer hyperspectral image classification method and system | |
| Li et al. | Hyperspectral image super-resolution with rgb image super-resolution as an auxiliary task | |
| Rudrapal et al. | Land cover classification using support vector machine | |
| Fortuna et al. | Multivariate image fusion: A pipeline for hyperspectral data enhancement | |
| Guo et al. | Multispectral and hyperspectral image fusion based on regularized coupled non-negative block-term tensor decomposition | |
| CN106780423B (en) | Spectral reconstruction method based on minority-band high-resolution image | |
| Yang et al. | Deep‐learning based on‐chip rapid spectral imaging with high spatial resolution | |
| CN111126463A (en) | Spectral image classification method and system based on local information constraint and sparse representation | |
| CN107655571A (en) | A kind of spectrum imaging system obscured based on dispersion and its spectrum reconstruction method | |
| Wen et al. | A novel spatial fidelity with learnable nonlinear mapping for panchromatic sharpening | |
| CN105067116B (en) | The joining method and system of a kind of Frame projection imaging spectrometer data | |
| Wang et al. | Hyperspectral image super-resolution via knowledge-driven deep unrolling and transformer embedded convolutional recurrent neural network | |
| Malleswara Rao et al. | Hyperspectral and multispectral data fusion using fast discrete curvelet transform for urban surface material characterization | |
| CN108280486A (en) | A kind of high spectrum image solution mixing method based on end member cluster | |
| CN111091113A (en) | Hyperspectral image data fusion method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171020 |
|
| RJ01 | Rejection of invention patent application after publication |