CN114719980A - End-to-end spectrum reconstruction method and system - Google Patents
End-to-end spectrum reconstruction method and system Download PDFInfo
- Publication number
- CN114719980A CN114719980A CN202210343128.2A CN202210343128A CN114719980A CN 114719980 A CN114719980 A CN 114719980A CN 202210343128 A CN202210343128 A CN 202210343128A CN 114719980 A CN114719980 A CN 114719980A
- Authority
- CN
- China
- Prior art keywords
- spectral
- template
- reconstruction
- hyperspectral image
- reconstruction method
- 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
- 238000001228 spectrum Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000003595 spectral effect Effects 0.000 claims abstract description 44
- 238000013461 design Methods 0.000 claims abstract description 19
- 238000013528 artificial neural network Methods 0.000 claims abstract description 16
- 238000012549 training Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 27
- 230000006870 function Effects 0.000 claims description 13
- 230000003044 adaptive effect Effects 0.000 claims description 12
- 238000005457 optimization Methods 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 4
- 238000004590 computer program Methods 0.000 claims description 3
- 238000000605 extraction Methods 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 230000003287 optical effect Effects 0.000 abstract description 5
- 238000000701 chemical imaging Methods 0.000 description 10
- 230000007547 defect Effects 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000011478 gradient descent method Methods 0.000 description 2
- 238000004611 spectroscopical analysis Methods 0.000 description 2
- 241000949473 Correa Species 0.000 description 1
- 230000004075 alteration Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000002591 computed tomography Methods 0.000 description 1
- 230000008602 contraction Effects 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000002059 diagnostic imaging Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 239000002360 explosive Substances 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000001303 quality assessment method Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J3/2823—Imaging spectrometer
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T11/00—2D [Two Dimensional] image generation
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01J—MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY
- G01J3/00—Spectrometry; Spectrophotometry; Monochromators; Measuring colours
- G01J3/28—Investigating the spectrum
- G01J3/2823—Imaging spectrometer
- G01J2003/2826—Multispectral imaging, e.g. filter imaging
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses an end-to-end spectrum reconstruction method, which comprises the following steps: s1, reconstructing a hyperspectral image; s2, taking the obtained hyperspectral image as the input of an optimized neural network, training neural network parameters according to an end-to-end spectral reconstruction model, and calculating a loss function; and S3, reversely transmitting the error to the coding template to be designed, changing the design parameters of the coding template, and continuously training the model until a high-quality hyperspectral image is reconstructed. By adopting the technical scheme, the design of the optical system is combined with the reconstruction method, and the self-adaptive coding template is additionally arranged on the basis of the coding aperture snapshot system according to the end-to-end spectrum reconstruction method to replace the traditional random coding template, so that the quality of spectrum reconstruction is improved.
Description
Technical Field
The invention relates to the technical field of hyperspectral image processing, in particular to an end-to-end spectrum reconstruction method and an end-to-end spectrum reconstruction system.
Background
Spectral imaging captures the spectral power distribution of a scene or object as a three-dimensional data cube, comprising multiple two-dimensional images of the same scene measured at different wavelengths. The hyperspectral images generally contain dozens or even hundreds of spectral channels, and the images on different spectral bands contain different spatial information and spectral information. Therefore, the application fields of the method are very wide and comprise medical imaging, remote sensing, national defense and monitoring, food quality assessment and the like.
Due to the explosive development of computational reconstruction, snapshot spectral imagers have also been developed further. Such techniques, including computed tomography spectroscopy (CTIS), prism mask spectroscopy video imaging systems (PMVIS), can capture a complete hyperspectral image with a single shot. Based on compressive sensing theory, Wagadarikar et al propose Coded Aperture Snapshot Spectral Imaging (CASSI) (e.g., Ashwin Wagadarikar, Renu John, RebeccaWillett, and David Brady, "Single discrete design for coded aperture snapshot spectral imaging," applied. Opt., vol.47, No.10, pp.B44-B51, Apr 2008).
In addition, coded aperture snapshot spectral imaging has grown rapidly in recent years. In the halftone field, Ulichney proposes a uniformly distributed random unbiased noise model, the blue noise model. This model can be applied to coded aperture templates. The only component that changes in the coded aperture snapshot spectral imaging system is the coded aperture template. And the coded aperture template determines the number of measurements acquired by the sensor and the quality of spectral reconstruction. Therefore, designing a reasonable coded aperture template will further improve the spectral reconstruction quality. In the paper "spatial blue noise coded aperture design for multi-shot compressive spectral imaging" (Claudia V Correa, HenryArguello, and Gonzalo RArce, "spatial blue noise coded aperture design for multi-shot compressive spectral imaging," JOSAA, vol.33, No.12, pp.2312-2322,2016), Correa et al designed spatial blue noise coding apertures based on the finite equidistant nature of the spectral imaging sensing matrix of the coding snapshot apertures, improving the quality of the spectral imaging. Meanwhile, the convolutional neural network is widely applied to spectrum reconstruction, and compared with traditional sparse recovery and dictionary learning, the convolutional neural network has excellent computing capability and feature learning mapping. At present, a new research direction is to replace iterative optimization in compressed sensing by a deep neural network, and the method uses data driving to replace empirical design, so that the quality of spectral reconstruction is further improved. However, the data-driven optimization of the spectral reconstruction process is only from an algorithmic level, and does not take into account the joint optimization of the optical system design.
Through the above analysis, the problems and defects of the prior art are as follows: at present, most of coded aperture snapshot spectral imaging adopts a fixed random coded aperture template, and the coded aperture template with a more reasonable structure is not designed. In addition, the reconstruction quality of the spectral image is improved only from the algorithm level, but the joint optimization of the whole optical system is not considered, so that the reconstruction quality of the spectral image is influenced.
The difficulty in solving the above problems and defects is: how to design a coding aperture template with a more reasonable structure; how to design joint optimization of the whole optical system.
Disclosure of Invention
The invention aims to provide an end-to-end spectrum reconstruction method and system aiming at the defects of the prior art, a coded aperture template and a neural network are optimally designed, learnable parameters are the coded aperture template and the neural network parameters, the deviation between a real image and a reconstructed image can be reduced to the maximum extent, and the quality of spectrum image reconstruction is further improved.
In order to solve the technical problems, the technical scheme of the invention is as follows:
a method of end-to-end spectral reconstruction comprising the steps of:
s1 reconstruction of hyperspectral image
S1-1, constructing a sensing model of the adaptive coding aperture snapshot system through the adaptive coding template:
y=Hf+n
y is an original image acquired by a sensor, H represents a system observation matrix, f represents an original hyperspectral image to be restored, and n is sensor noise;
s1-2, according to the model, preliminarily reconstructing a hyperspectral image;
s2, taking the obtained hyperspectral image as the input of an optimized neural network, training neural network parameters according to an end-to-end spectral reconstruction model, and calculating a loss function;
and S3, reversely transmitting the error to the coding template to be designed, changing the design parameters of the coding template, and continuously training the model until a high-quality hyperspectral image is reconstructed.
Preferably, in step S1, the system observation matrix H is obtained by the following formula: h ═ TPD
Wherein T is equal to {0,1} to represent the state of the encoding template, 0 represents closing, and 1 represents opening; p represents a discretization model of the dispersive element; d represents the spatial extraction related to the size of the sensor pixel, and the discretization model of the dispersive element is specifically expressed as P epsilon RM(N+L-1)×MNL。
Preferably, the sensor captures raw imagesThe system observation matrixThe original hyperspectral image f to be restored belongs to RMNLThe sensing noiseWhere R represents the dimension, M represents the length of the original image in space, N represents the width of the original image in space, and L represents the number of spectral bands.
Preferably, in step S1-2, the method for preliminarily reconstructing the hyperspectral image includes constraining a solution space by using image priors as regularization, and solving a minimization problem to obtain the hyperspectral image:
where τ is a balance parameter, μ is a penalty parameter, e ∈ RMNLAs an auxiliary variable, the number of variables,the middle subscript 2 represents the L2 norm, the upper right 2 represents the square, and the solution of the equation can be divided into two sub-questionsThus, find the minimum f and e.
Preferably, the optimized neural network in the step S2 is a CA-uet network.
Preferably, the end-to-end spectral reconstruction model in step S2 is:
the method comprises the steps of obtaining a spectrum reconstruction network, determining a system observation matrix, a code template, a system observation matrix, a whole spectrum reconstruction network, and alpha R (t), wherein theta represents optimization parameters of the spectrum reconstruction network and comprises design parameters of the code template, H is the system observation matrix and is used for optimizing the spectrum reconstruction network, M (-) represents the whole spectrum reconstruction network, and alpha R (t) represents regularization to realize binarization of the code template.
Preferably, said r (t) is represented by:
wherein, ti,jE {0,1} represents the binarization of the encoding template.
Preferably, the loss function is expressed as:
wherein the content of the first and second substances,for the high spectral value of each pixel point reconstructed, the spectral loss minimizes the angle between the reconstructed image and the true spectral feature.
Preferably, in step S3, the encoding template design parameter is a finite equidistant property constant δ of the system sensing matrix, δ being expressed as a function of the random encoding aperture pattern structure: t ═ T (δ), where T (·) represents the mapping function of the finite equidistant property constant δ to the system observation matrix.
The invention also discloses an end-to-end spectrum reconstruction system, which comprises an objective lens, an adaptive coding template, a relay lens, a prism and a gray camera, wherein the adaptive coding template comprises a memory, a processor and a computer program which is stored in the memory and can execute the end-to-end spectrum reconstruction method on the processor.
The invention has the following characteristics and beneficial effects:
by adopting the technical scheme, the design of the optical system is combined with the reconstruction method, and the self-adaptive coding template is additionally arranged on the basis of the coding aperture snapshot system according to the end-to-end spectrum reconstruction method to replace the traditional random coding template, so that the quality of spectrum reconstruction is improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of an end-to-end spectral reconstruction method according to the present invention.
FIG. 2 is a diagram of an end-to-end spectral reconstruction system according to the present invention.
Detailed Description
It should be noted that the embodiments and features of the embodiments may be combined with each other without conflict.
In the description of the present invention, it is to be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", and the like, indicate orientations or positional relationships based on those shown in the drawings, and are used only for convenience in describing the present invention and for simplicity in description, and do not indicate or imply that the referenced devices or elements must have a particular orientation, be constructed and operated in a particular orientation, and thus, are not to be construed as limiting the present invention. Furthermore, the terms "first", "second", etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first," "second," etc. may explicitly or implicitly include one or more of that feature. In the description of the present invention, "a plurality" means two or more unless otherwise specified.
In the description of the present invention, it should be noted that, unless otherwise explicitly specified or limited, the terms "mounted," "connected," and "connected" are to be construed broadly, e.g., as meaning either a fixed connection, a removable connection, or an integral connection; can be mechanically or electrically connected; they may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art through specific situations.
The invention provides an end-to-end spectrum reconstruction method, as shown in fig. 1, comprising the following steps:
s1 reconstruction of hyperspectral image
S1-1, constructing a sensing model of the adaptive coding aperture snapshot system through the adaptive coding template:
y=Hf+n
wherein y is an original image acquired by the sensor, H represents a system observation matrix, f represents an original hyperspectral image to be restored, and n is sensor noise. In particular, the sensor collects raw imagesThe system observation matrixThe original hyperspectral image f to be restored belongs to RMNLThe sensing noiseWherein R represents dimensionDegree, M denotes the length of the original image in space, N denotes the width of the original image in space, and L denotes the number of spectral bands.
The system observation matrix H is obtained by:
H=TPD
wherein T is equal to {0,1} to represent the state of the encoding template, 0 represents closing, and 1 represents opening; p represents a discretization model of the dispersive element; d represents the spatial extraction related to the size of the sensor pixel, and the discretization model of the dispersive element is specifically expressed as P epsilon RM(N+L-1)×MNL。
S1-2, according to the model, preliminarily reconstructing a hyperspectral image
Specifically, the method for preliminarily reconstructing the hyperspectral image comprises the steps of adopting image prior as regularization to constrain solution space, solving a minimization problem to obtain the hyperspectral image:
where τ is a balance parameter, μ is a penalty parameter, e ∈ RMNLAs an auxiliary variable, the number of variables,the middle subscript 2 represents the L2 norm and the upper right 2 represents the square, and the solution to the equation can be split into two subproblems to find the minimized f and e.
S2, obtaining the hyperspectral imageAs the input of the optimized neural network, training neural network parameters and calculating a loss function according to an end-to-end spectrum reconstruction model;
specifically, the optimized neural network is a CA-uet network, which may be denoted as p (·), and the minimized e may be denoted as:
it can be understood that the optimized neural network CA-uet is mainly composed of two parts, a uet module and a channel attention module. In the Unet network, the first convolutional layer generates a tensor with a characteristic size of 64 using a 3 × 3 × 31 filter to enhance sparsity of spectral gradients. The network then generates a multi-scale feature, a contraction path with a largest pool and an expansion path with an upcurl layer. For each level, two convolutional layers encode spatial spectral features. By skipping the connection, the scaling feature will be connected to the upscaling feature. In the channel attention module, extracting 1 × 1 × 64 global information from M × N feature maps of 64 channels, then performing full connection to obtain a 64/r-dimensional vector, performing Relu activation, performing full connection again, converting the 64/r-dimensional vector into a 64-dimensional vector, and performing sigmoid activation to enable the numerical value to be between 0 and 1 to obtain a weight matrix. The weight matrix is multiplied by the feature map. Finally, 31 convolution layers of 3 × 3 × 64 are used to generate a tensor of the original hyperspectral cube size, namely, the restored hyperspectral image.
Further, the end-to-end spectral reconstruction model is as follows:
the method comprises the steps of obtaining a spectrum reconstruction network, determining a system observation matrix, a code template, a system observation matrix, a whole spectrum reconstruction network, and alpha R (t), wherein theta represents optimization parameters of the spectrum reconstruction network and comprises design parameters of the code template, H is the system observation matrix and is used for optimizing the spectrum reconstruction network, M (-) represents the whole spectrum reconstruction network, and alpha R (t) represents regularization to realize binarization of the code template. And (4) according to the end-to-end spectrum reconstruction model, the coded aperture template can be optimally designed in a combined mode.
Said R (t) is represented by:
wherein, ti,jE {0,1} represents the binarization of the encoding template.
The loss function is expressed as:
wherein the content of the first and second substances,for the high spectral value of each reconstructed pixel point, the spectral loss minimizes the angle between the reconstructed image and the true spectral feature, and the loss function value can be calculated by the above formula.
And S3, reversely transmitting the error to the coding template to be designed, changing the design parameters of the coding template, and continuously training the model until a high-quality hyperspectral image is reconstructed.
Specifically, the design parameter of the coding template is a finite equidistant property constant δ of a system sensing matrix, wherein δ is expressed as a function of a random coding aperture pattern structure: t ═ T (δ), where T (·) denotes a mapping function of finite equidistant property constants δ to the system observation matrix, which ensures maximum separation of the coding patterns to be designed in the horizontal direction and maximum incoherence in the vertical direction. The initial delta setting is 0.5.
The error back propagation adopts a gradient descent method:
where λ, ε is the gradient descent step, λ is set to 0.001, and ε is set to 0.002.
In the present embodiment, the number of iterations is set to 20. And obtaining the final delta and a designed self-adaptive coding template pattern, and reconstructing a high-quality hyperspectral image.
The invention also discloses an end-to-end spectrum reconstruction system, which comprises an objective lens, an adaptive coding template, a relay lens, a prism and a gray-scale camera, wherein the adaptive coding template comprises a memory, a processor and a computer program which is stored in the memory and can execute the end-to-end spectrum reconstruction method on the processor, as shown in fig. 2.
The working principle of the technical scheme is as follows:
the target scene is transmitted to the self-adaptive coding template through the objective lens, the hyperspectral image is reconstructed through the end-to-end spectrum reconstruction method, the reconstructed hyperspectral image is obtained through the relay lens and the prism and the grayscale camera, the reconstructed hyperspectral image is optimized through the neural network and is sent to the self-adaptive coding template for iteration, and finally the high-quality hyperspectral image is output.
The embodiments of the present invention have been described in detail with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. It will be apparent to those skilled in the art that various changes, modifications, substitutions and alterations can be made in these embodiments, including the components, without departing from the principles and spirit of the invention, and still fall within the scope of the invention.
Claims (10)
1. A method of end-to-end spectral reconstruction, comprising the steps of:
s1 reconstruction of hyperspectral image
S1-1, constructing a sensing model of the adaptive coding aperture snapshot system through an adaptive coding template:
y=Hf+n
y is an original image acquired by a sensor, H represents a system observation matrix, f represents an original hyperspectral image to be restored, and n is sensor noise;
s1-2, according to the model, preliminarily reconstructing a hyperspectral image;
s2, taking the obtained hyperspectral image as the input of an optimized neural network, training neural network parameters according to an end-to-end spectral reconstruction model, and calculating a loss function;
and S3, reversely transmitting the error to the coding template to be designed, changing the design parameters of the coding template, and continuously training the model until a high-quality hyperspectral image is reconstructed.
2. The end-to-end spectral reconstruction method of claim 1, wherein in step S1, the system observation matrix H is obtained by the following formula: h ═ TPD
Wherein T is equal to {0,1} to represent the state of the encoding template, 0 represents closing, and 1 represents opening; p represents a discretization model of the dispersive element; d represents the spatial extraction related to the size of the sensor pixel, and the discretization model of the dispersive element is specifically expressed as P epsilon RM(N +L-1)×MNL。
3. The end-to-end spectral reconstruction method of claim 2, wherein the raw image acquired by the sensor isThe system observation matrixThe original hyperspectral image f to be restored belongs to RMNLThe sensing noiseWhere R represents the dimension, M represents the length of the original image in space, N represents the width of the original image in space, and L represents the number of spectral bands.
4. An end-to-end spectrum reconstruction method according to claim 3, wherein in step S1-2, the primary reconstruction method of the hyperspectral image is to use image priors as regularization to constrain solution space and solve a minimization problem to obtain the hyperspectral image:
where τ is the balance parameter, μ is the penalty parameter, e ∈ RMNLAs an auxiliary variable, the number of variables,the middle subscript 2 represents the L2 norm and the upper right 2 represents the square, and the solution to the equation can be split into two subproblems to find the minimized f and e.
5. The end-to-end spectral reconstruction method of claim 1, wherein said optimized neural network in step S2 is a CA-uet network.
6. The end-to-end spectral reconstruction method of claim 5, wherein the end-to-end spectral reconstruction model in step S2 is:
the method comprises the steps of obtaining a spectrum reconstruction network, determining a system observation matrix, a code template, a system observation matrix, a whole spectrum reconstruction network, and alpha R (t), wherein theta represents optimization parameters of the spectrum reconstruction network and comprises design parameters of the code template, H is the system observation matrix and is used for optimizing the spectrum reconstruction network, M (-) represents the whole spectrum reconstruction network, and alpha R (t) represents regularization to realize binarization of the code template.
8. The end-to-end spectral reconstruction method of claim 7, wherein said loss function is expressed as:
9. The end-to-end spectral reconstruction method of claim 7, wherein in step S3, the encoding template design parameter is a finite equidistant property constant δ of the system sensing matrix, δ being expressed as a function of the random encoding aperture pattern structure: t ═ T (δ), where T (·) represents the mapping function of the finite equidistant property constant δ to the system observation matrix.
10. An end-to-end spectral reconstruction system comprising an objective lens, an adaptive encoding template, a relay lens, a prism, a grayscale camera, the adaptive encoding template comprising a memory, a processor, and a computer program stored in the memory and operable on the processor to perform the end-to-end spectral reconstruction method of any of claims 1-9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210343128.2A CN114719980A (en) | 2022-04-02 | 2022-04-02 | End-to-end spectrum reconstruction method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210343128.2A CN114719980A (en) | 2022-04-02 | 2022-04-02 | End-to-end spectrum reconstruction method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114719980A true CN114719980A (en) | 2022-07-08 |
Family
ID=82241088
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210343128.2A Pending CN114719980A (en) | 2022-04-02 | 2022-04-02 | End-to-end spectrum reconstruction method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114719980A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115187870A (en) * | 2022-09-13 | 2022-10-14 | 浙江蓝景科技有限公司杭州分公司 | Marine plastic waste material identification method and system, electronic equipment and storage medium |
-
2022
- 2022-04-02 CN CN202210343128.2A patent/CN114719980A/en active Pending
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115187870A (en) * | 2022-09-13 | 2022-10-14 | 浙江蓝景科技有限公司杭州分公司 | Marine plastic waste material identification method and system, electronic equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110555446B (en) | Remote sensing image scene classification method based on multi-scale depth feature fusion and migration learning | |
Deng et al. | Machine learning in pansharpening: A benchmark, from shallow to deep networks | |
Yuan et al. | Plug-and-play algorithms for video snapshot compressive imaging | |
US10274420B2 (en) | Compact multifunctional system for imaging spectroscopy | |
CN109697697B (en) | Reconstruction method of spectral imaging system based on optimization heuristic neural network | |
CN109883548B (en) | Optimization heuristic-based coding optimization method for spectral imaging system of neural network | |
CN111145131A (en) | Infrared and visible light image fusion method based on multi-scale generation type countermeasure network | |
CN114119444B (en) | Multi-source remote sensing image fusion method based on deep neural network | |
CN109886898B (en) | Imaging method of spectral imaging system based on optimization heuristic neural network | |
Han et al. | Tensor FISTA-Net for real-time snapshot compressive imaging | |
CN107730479A (en) | High dynamic range images based on compressed sensing go artifact fusion method | |
AU2020281143B1 (en) | Creating super-resolution images | |
CN111598962B (en) | Single-pixel imaging method and device based on matrix sketch analysis | |
CN114719980A (en) | End-to-end spectrum reconstruction method and system | |
Karwowska et al. | Using super-resolution algorithms for small satellite imagery: A systematic review | |
CN110717947A (en) | High-quality spectrum reconstruction method based on external and internal training | |
CN113008370A (en) | Three-dimensional self-adaptive compression reconstruction method based on liquid crystal hyperspectral calculation imaging system | |
CN110335197B (en) | Demosaicing method based on non-local statistical eigen | |
CN113008371B (en) | Hyperspectral imaging method for deep learning dispersion-based fuzzy solution | |
CN113724308A (en) | Cross-waveband stereo matching algorithm based on mutual attention of luminosity and contrast | |
CN109978802A (en) | High dynamic range images fusion method in compressed sensing domain based on NSCT and PCNN | |
Yang et al. | Learning Degradation-Aware Deep Prior for Hyperspectral Image Reconstruction | |
Chen et al. | Prior image guided snapshot compressive spectral imaging | |
CN114140359B (en) | Remote sensing image fusion sharpening method based on progressive cross-scale neural network | |
CN113670440B (en) | Compression spectrum imaging method based on self-adaptive dictionary |
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 |