CN116754497B - High-efficiency spectrum sensing method and system based on sparse statistics - Google Patents
High-efficiency spectrum sensing method and system based on sparse statistics Download PDFInfo
- Publication number
- CN116754497B CN116754497B CN202311052634.7A CN202311052634A CN116754497B CN 116754497 B CN116754497 B CN 116754497B CN 202311052634 A CN202311052634 A CN 202311052634A CN 116754497 B CN116754497 B CN 116754497B
- Authority
- CN
- China
- Prior art keywords
- spectrum
- dictionary
- spectral
- negative
- target scene
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 73
- 238000000034 method Methods 0.000 title claims abstract description 64
- 230000003595 spectral effect Effects 0.000 claims abstract description 50
- 230000008447 perception Effects 0.000 claims abstract description 19
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 15
- 230000008878 coupling Effects 0.000 claims abstract description 12
- 238000010168 coupling process Methods 0.000 claims abstract description 12
- 238000005859 coupling reaction Methods 0.000 claims abstract description 12
- 238000005316 response function Methods 0.000 claims abstract description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 9
- 239000011159 matrix material Substances 0.000 claims description 11
- 238000005286 illumination Methods 0.000 claims description 10
- 238000000354 decomposition reaction Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 230000011218 segmentation Effects 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 5
- 230000006870 function Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 238000007619 statistical method Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 abstract description 6
- 238000000701 chemical imaging Methods 0.000 abstract 1
- 238000003384 imaging method Methods 0.000 description 6
- 238000005070 sampling Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 238000001514 detection method Methods 0.000 description 2
- 239000005262 ferroelectric liquid crystals (FLCs) Substances 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000004891 communication Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
Classifications
-
- 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
-
- 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
-
- 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
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
- G06F18/2133—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on naturality criteria, e.g. with non-negative factorisation or negative correlation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/28—Determining representative reference patterns, e.g. by averaging or distorting; Generating dictionaries
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Pathology (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Immunology (AREA)
- Chemical & Material Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to the technical field of photography, and discloses a high-efficiency spectrum sensing method and a system based on sparse statistics, wherein the method comprises the steps of constructing a non-negative spectrum-space feature dictionary according to a spectrum feature curve and space features of a spectrum image dataset; illuminating the target scene with spectral feature curves in a non-negative spectral-spatial feature dictionary characterized by response functions inherent to the coded light source and the sensor to produce target scene reflected light; coupling and modulating the reflected light of the target scene by using a preset modulation method based on the spatial characteristics in the non-negative spectrum-spatial characteristic dictionary to obtain coupled multispectral information; and calculating the coupled multispectral information by using a preset perception algorithm so as to output spectrum perception information according to a calculation result. According to the invention, spectral imaging is not needed, and the adjustable light source and the structural light modulator are used for directly performing the dimension reduction and intelligent perception on the high-dimensional information.
Description
Technical Field
The invention relates to the technical field of computational photography, in particular to a high-efficiency spectrum sensing method and system based on sparse statistics.
Background
The spectrum has the unique characteristic and is widely applied to the aspects of space remote sensing, biomedicine, detection, industrial quality inspection and the like. Compared with the traditional RGB image and gray scale image, the spectrum image has more dimensional information and has large data volume. Spectral images are often used for recognition, classification, etc. in the following, whereas traditional methods require imaging before sensing. The sensing precision depends on the imaging precision, and huge hardware and software resources are consumed, so that the cost and the operation complexity of the existing sensing system are high, and meanwhile, the sensing speed is low and the communication load is heavy. In order to solve the above problems, a method for spectral imaging-free sensing is proposed, and non-imaging sensing is realized, but double sensors are needed, and the structure is complex.
The spectrum sensing has huge redundant information, and more hardware and software resources are directly acquired. In practice, multispectral images have redundant information in the spectral and spatial dimensions, and highly redundant information can be reduced and compressed using sparse prior characteristics. In the spectrum dimension, the spectrum is subjected to dimension reduction by using a principal component analysis-based method and dictionary learning algorithm, and a single-pixel spectrometer which realizes principal component characteristic illumination by using a coded light source and a response function design appears in the aspect of hardware. In the space dimension, based on the image sparse characteristic, a single-pixel sparse sampling method is realized by using the data statistical characteristic and a designed sampling function.
Disclosure of Invention
The present invention aims to solve at least one of the technical problems in the related art to some extent.
Therefore, the invention provides a high-efficiency spectrum sensing method based on sparse statistics, which does not need spectrum imaging, and uses an adjustable light source and a structural light modulator to directly conduct dimensionality reduction and intelligent sensing on high-dimensional information. The optical system is used for realizing the preliminary perception of multispectral target information before acquisition, and classification, segmentation and target detection can be realized by combining a subsequent algorithm.
Another object of the present invention is to provide a high-efficiency spectrum sensing system based on sparse statistics.
In order to achieve the above objective, in one aspect, the present invention provides a high-efficiency spectrum sensing method based on sparse statistics, including:
constructing a non-negative spectrum-space characteristic dictionary according to the spectrum characteristic curve and the space characteristic of the spectrum image data set;
illuminating the target scene with spectral feature curves in the non-negative spectral-spatial feature dictionary characterized by response functions inherent to the coded light source and the sensor to produce target scene reflected light;
coupling and modulating the reflected light of the target scene by using a preset modulation method based on the spatial features in the non-negative spectrum-spatial feature dictionary to obtain coupled multispectral information;
and calculating the coupled multispectral information by using a preset perception algorithm so as to output spectrum perception information according to a calculation result.
In addition, the sparse statistics-based high-efficiency spectrum sensing method according to the above embodiment of the present invention may further have the following additional technical features:
further, in one embodiment of the present invention, a spectral feature curve and a spatial feature of the spectral image dataset are acquired using a preset learning method; the learning method comprises one of dictionary learning, principal component analysis and statistical analysis-based methods.
Further, in one embodiment of the present invention, the coded light source comprises one of an LED, an LCD, and a laser.
Further, in an embodiment of the present invention, the modulation method includes one of fourier-based modulation, hadamard-based modulation, and optimized mask modulation.
Further, in one embodiment of the present invention, the coupled multispectral information is obtained by modulating the reflected light of the target scene with a structured light modulator; the structure light modulator comprises one of a digital micro-mirror array, a ferroelectric liquid crystal array and a spatial light modulation array.
Further, in one embodiment of the invention, the perceptual algorithm comprises one of an iterative optimization method and a network-based method.
Further, in an embodiment of the present invention, the spectrum sensing information includes a category, a position, and a segmentation result of the object in the target scene.
Further, in one embodiment of the present invention, an objective function is obtained that the spectral characteristic meets:
wherein,for a dataset comprising p spectral curves, < +.>Is a non-negative dictionary corresponding to +.>Spectral characteristics,/>Is a coefficient matrix.
To achieve the above object, another aspect of the present invention provides a high-efficiency spectrum sensing system based on sparse statistics, including:
the light source module is used for representing a spectrum characteristic curve in the non-negative spectrum-space characteristic dictionary by utilizing the inherent response functions of the coded light source and the sensor so as to set the operation time sequence of the light source array, and transmitting illumination with the spectrum characteristic curve to a target scene according to the operation time sequence;
the modulating module is used for modulating the reflected light of the target scene through the structural light modulating device so as to acquire spectral information in the space dimension;
the sensing module is used for collecting coupling measured values of the spectrum information at different moments;
and the calculating module is used for calculating the coupling measured value and outputting spectrum sensing information.
The high-efficiency spectrum sensing method and the system based on sparse statistics overcome the defect that the traditional spectrum sensing is performed by sampling and then dimension reducing, and based on the ideas of sparsity and statistics, the sensor and the adjustable light source are used for modulating a scene, so that spectrum-space joint dimension reduction is performed on the spectrum information, and then intelligent sensing without reconstruction is performed according to measured values.
Additional aspects and advantages of the invention will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the invention.
Drawings
The foregoing and/or additional aspects and advantages of the invention will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings, in which:
FIG. 1 is a flow chart of a sparse statistics-based high efficiency spectrum sensing method in accordance with an embodiment of the present invention;
FIG. 2 is a network frame diagram of a sparse statistics-based high-efficiency spectrum sensing method in accordance with an embodiment of the present invention;
fig. 3 is a schematic structural diagram of a sparse statistics-based high-efficiency spectrum sensing system according to an embodiment of the invention.
Detailed Description
It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other. The invention will be described in detail below with reference to the drawings in connection with embodiments.
In order that those skilled in the art will better understand the present invention, a technical solution in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in which it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, shall fall within the scope of the present invention.
The following describes a high-efficiency spectrum sensing method and system based on sparse statistics according to an embodiment of the present invention with reference to the accompanying drawings.
Fig. 1 is a flowchart of a sparse statistics-based high-efficiency spectrum sensing method according to an embodiment of the present invention.
As shown in fig. 1, the method includes, but is not limited to, the steps of:
s1, constructing a non-negative spectrum-space feature dictionary according to a spectrum feature curve and space features of a spectrum image dataset;
s2, utilizing a spectrum characteristic curve in a non-negative spectrum-space characteristic dictionary which is characterized by an inherent response function of the coded light source and the sensor to irradiate a target scene to generate target scene reflected light;
s3, coupling and modulating the reflected light of the target scene by using a preset modulation method based on the spatial features in the non-negative spectrum-spatial feature dictionary to obtain coupled multispectral information;
s4, calculating the coupled multispectral information by using a preset sensing algorithm so as to output spectrum sensing information according to a calculation result.
Illustratively, the spectral characteristic curves and spatial features of the spectral image dataset are learned and counted, wherein the manner of learning and counting is divided into two steps: spectral feature learning and spatial feature learning, including but not limited to dictionary learning, principal component analysis, and statistical analysis-based methods.
Illustratively, learning the non-negative spectral features, mainly satisfies the following objective functions:
wherein the method comprises the steps ofFor a dataset comprising p spectral curves, < +.>Is a non-negative dictionary corresponding to +.>Spectral characteristics,/>Is a coefficient matrix.
Illustratively, the spectral signature in the dictionary is characterized using coded light sources and sensor response functions, with tunable light sources including, but not limited to, LEDs, LCDs, lasers, such that the light sources emit multi-band illumination consistent with the learned spectral dimension signature.
Illustratively, the modulating the target scene reflected light is coupled in accordance with spatial features in a dictionary using a structured light modulator, including but not limited to fourier-based modulation, hadamard-based modulation, and optimized mask modulation.
Illustratively, the target scene reflected light is modulated in accordance with the spatial feature coupling in the dictionary using a structured light modulator, including but not limited to a digital micromirror array, a ferroelectric liquid crystal array, a spatial light modulation array.
Illustratively, the acquired data is input into a sensing algorithm, the spectrum multispectral sensing information is output, and the sensing result can be directly output without reconstructing a spectrum image, and the method comprises, but is not limited to, a traditional iterative optimization method and a network-based method.
Illustratively, the output spectral awareness information includes, but is not limited to, the class, location, and segmentation results of objects in the output scene.
Fig. 2 is a network frame diagram of the sparse statistics-based high-efficiency spectrum sensing method of the present invention. As shown in fig. 2, in the embodiment of the present invention, the spectral characteristic curve and the spatial characteristic of the hyperspectral dataset (Indian fine, pavia University, kennedy Space Center, and Pavia Center) are first learned to form a non-negative spectral-spatial characteristic dictionary. The spectral dimension features are learned by using an NN-K-SVD dictionary learning method, and a spectral dictionary is obtained and comprises n spectral feature curves. The method comprises the following specific steps:
first vectorizing spectral data toTo generate a non-negative dictionary->(sum coefficient matrix)) It is necessary to force non-negative coefficients during sparse spectral coding phase and force dictionary matrices to remain positive during dictionary updating phase. Specific stepsThe method comprises initializing non-negative random normalized dictionary matrix ++>. n is the spectral feature number, K isThe number of rows of vectors.
In one embodiment of the invention, the spectrum is sparsely encoded: fixed dictionaryTracking algorithm using non-negative decomposition for each +.>Calculate->:
Dictionary updating stage: internal circulation
Define a set of usesIs>
Calculating an error matrix：
Select onlyMiddle and->Corresponding columns, composition->。
Further, decomposition using KSVDTaking the first left->Right singular vector->And the first singular value +.>Respectively->And->Truncated negative term is null and normalizes +.>。
Repeating the dictionary updating stage and the spectrum sparse coding process until convergence.
In one embodiment of the invention, spatial features are learned by a method based on Fourier domain sparse statistics, and the hyperspectral dataset is randomly clipped to 256256-pixel image block, which is 20000 sheets. The fourier coefficients of the image blocks are then modulo summed and then sorted from large to small. Due to the nature of conjugate symmetry, only the upper half of the fourier coefficients need be ordered. This sequence is recorded as a spatial dimension dictionary. And combining non-uniform sparse sampling optimization to obtain a Fourier frequency chart to be sampled, and taking the Fourier frequency chart as a space dimension dictionary.
Further, a plurality of high-brightness LEDs with different wave bands are used for combination and programming to obtain the multispectral LED array. Different light sources in the array are lightened according to the time sequence, and illumination consistent with the obtained spectrum characteristic curve is emitted at different moments to irradiate the target microlens array. Duration t, for a total of n different sets of illumination. And projecting a sine mask corresponding to a certain frequency point on the space dimension dictionary by using the digital micromirror array, and projecting the sine mask to a target scene. In this example a 3-step fourier mask is used. The light source is shifted one round during the time of each mask projection (t=nt).
Further, the modulated target information is converged to a single pixel sensor measurement using a lens. According to the time sequence, a plurality of illuminations and masks are transformed to illuminate the target object, and the scene is modulated for a plurality of times.
Further, the data after the multiple modulations are input into a computer, and multispectral classification without an image reconstruction process is output by using full-connection layer+Unet network calculation.
Thus, the invention learns and counts the spectrum characteristic curve and the space characteristic of the multispectral image dataset through the implementation process of FIG. 2 to form a non-negative spectrum-space characteristic dictionary; characterizing spectral signature curves in a dictionary using response functions inherent to coded light sources and sensors, projecting them to a target scene; modulating the reflected light of the target scene according to the spatial feature coupling in the dictionary using a structured light modulator; collecting the coupled multispectral information by using a sensor; the acquired data is input into a perception algorithm, and spectrum multispectral perception information is output in a mode of no need of reconstructing an image, wherein the spectrum multispectral perception information comprises the types, the positions and the segmentation results of objects in an output scene.
According to the sparse statistics-based high-efficiency spectrum sensing method, spectrum imaging is not needed, and the adjustable light source and the structural light modulator are used for directly performing dimensionality reduction and intelligent sensing on high-dimensional information. The optical system is used for realizing the primary perception of the multispectral target information, and the spectrum perception information can be output by combining a subsequent perception algorithm. The method reduces the calculation complexity, and has the advantages of high response speed and high perception precision.
In order to implement the above embodiment, as shown in fig. 3, there is further provided a high-efficiency spectrum sensing system 10 based on sparse statistics, where the system 10 includes:
a light source module 100 for characterizing a spectral characteristic curve in a non-negative spectral-spatial characteristic dictionary using a response function inherent to the coded light source and the sensor to set an operation timing of the light source array, and transmitting illumination having the spectral characteristic curve to a target scene according to the operation timing;
the modulating module 200 is configured to modulate the reflected light of the target scene through the structural light modulating device, so as to collect spectral information in a spatial dimension;
the sensing module 300 is used for collecting coupling measured values of the spectrum information at different moments;
the calculation module 400 is configured to calculate the coupling measurement value and output spectrum sensing information.
According to the sparse statistics-based high-efficiency spectrum sensing system provided by the embodiment of the invention, spectrum imaging is not needed, and the adjustable light source and the structural light modulator are used for directly performing the dimension reduction and intelligent sensing on the high-dimensional information. The optical system is used for realizing the primary perception of the multispectral target information, and the spectrum perception information can be output by combining a subsequent perception algorithm. The method reduces the calculation complexity, and has the advantages of high response speed and high perception precision.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, the different embodiments or examples described in this specification and the features of the different embodiments or examples may be combined and combined by those skilled in the art without contradiction.
Furthermore, the terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying a relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include at least one such feature. In the description of the present invention, the meaning of "plurality" means at least two, for example, two, three, etc., unless specifically defined otherwise.
Claims (8)
1. The high-efficiency spectrum sensing method based on sparse statistics is characterized by comprising the following steps of:
constructing a non-negative spectrum-space characteristic dictionary according to the spectrum characteristic curve and the space characteristic of the spectrum image data set;
illuminating the target scene with spectral feature curves in the non-negative spectral-spatial feature dictionary characterized by response functions inherent to the coded light source and the sensor to produce target scene reflected light; the method comprises the steps that a target scene reflected light is obtained by combining and programming a plurality of high-brightness LEDs with different wave bands to obtain a multispectral LED array, different light sources in the multispectral LED array are lightened according to a current time sequence, illumination consistent with a spectrum characteristic curve is emitted at different moments and irradiates a target, and a digital micromirror array is used for projecting a sine mask corresponding to a certain frequency point on a spatial characteristic dictionary so as to project the sine mask to the target scene;
coupling and modulating the reflected light of the target scene by using a preset modulation method based on the spatial features in the non-negative spectrum-spatial feature dictionary to obtain coupled multispectral information; calculating the coupled multispectral information by using a preset perception algorithm so as to output spectrum perception information according to a calculation result;
the method further comprises the steps of:
learning by using an NN-K-SVD dictionary learning method to obtain spectral features in the non-negative spectrum-space feature dictionary, wherein the spectral features comprise n spectral feature curves; comprising the following steps:
vectorizing spectral image datasets intoTo generate a non-negative dictionary->Sum coefficient matrixForcing non-negative coefficients in sparse spectrum coding, and forcing dictionary matrixes to keep positive values in a dictionary updating stage; initializing a non-negative random normalized dictionary matrix +.>N is the spectral feature number, K is +.>The number of rows of vectors;
spectrum sparse coding: fixed dictionaryTracking algorithm using non-negative decomposition for each +.>Calculate->:
Dictionary updating stage: internal circulation；
Define a set of usesIs>
Calculating an error matrix：
Select onlyMiddle and->Corresponding columns, composition->；
Decomposition by KSVDTaking the first left singular vector +.>Right singular vector->And the first singular value +.>Respectively isAnd->Truncated negative term is null and normalizes +.>；
Repeating the dictionary updating stage and the spectrum sparse coding process until convergence;
for a dataset comprising p spectral curves, < +.>Is a non-negative dictionary corresponding to +.>A spectral signature.
2. The method of claim 1, wherein the spatial features of the spectral image dataset are acquired using a pre-set learning method; wherein the learning method comprises a method based on statistical analysis.
3. The method of claim 1, wherein the modulation method comprises one of fourier-based modulation, hadamard-based modulation, and optimized mask modulation.
4. The method of claim 1, wherein the coupled multispectral information is obtained by modulating the reflected light of the target scene with a structured light modulator; wherein the structural light modulator comprises a spatial light modulation array.
5. The method of claim 1, wherein the perceptual algorithm comprises one of an iterative optimization method and a network-based method.
6. The method of claim 1, wherein the spectral awareness information includes a class, a location, and a segmentation result of objects in the target scene.
7. The method according to claim 1, wherein an objective function is obtained that the spectral signature satisfies:
8. a sparse statistics-based high efficiency spectrum sensing system, comprising:
the light source module is used for representing a spectrum characteristic curve in the non-negative spectrum-space characteristic dictionary by utilizing the inherent response functions of the coded light source and the sensor so as to set the operation time sequence of the light source array, and transmitting illumination with the spectrum characteristic curve to a target scene according to the operation time sequence; the modulating module is used for modulating the reflected light of the target scene through the structural light modulating device so as to acquire spectral information in the space dimension; the method comprises the steps that a target scene reflected light is obtained by combining and programming a plurality of high-brightness LEDs with different wave bands to obtain a multispectral LED array, different light sources in the multispectral LED array are lightened according to a current time sequence, illumination consistent with a spectrum characteristic curve is emitted at different moments and irradiates a target, and a digital micromirror array is used for projecting a sine mask corresponding to a certain frequency point on a spatial characteristic dictionary so as to project the sine mask to the target scene;
the sensing module is used for collecting coupling measured values of the spectrum information at different moments;
the calculation module is used for calculating the coupling measured value and outputting spectrum sensing information;
the light source module is further configured to:
learning by using an NN-K-SVD dictionary learning method to obtain the spectral characteristics of the non-negative spectrum-space characteristic dictionary, wherein the spectral characteristics comprise n spectral characteristic curves; comprising the following steps:
vectorizing spectral image datasets intoTo generate a non-negative dictionary->Sum coefficient matrixForcing non-negative coefficients in sparse spectrum coding, and forcing dictionary matrixes to keep positive values in a dictionary updating stage; initializing a non-negative random normalized dictionary matrix +.>N is the spectral feature number, K is +.>The number of rows of vectors;
spectrum sparse coding: fixed dictionaryTracking algorithm using non-negative decomposition for each +.>Calculate->:
Dictionary updating stage: internal circulation；
Define a set of usesIs>
Calculating an error matrix：
Select onlyMiddle and->Corresponding columns, composition->；
Decomposition by KSVDTaking the first left singular vector +.>Right singular vector->And the first singular value +.>Respectively isAnd->Truncated negative term is null and normalizes +.>；
Repeating the dictionary updating stage and the spectrum sparse coding process until convergence;
wherein,for a dataset comprising p spectral curves, < +.>Is a non-negative dictionary corresponding to +.>A spectral signature.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311052634.7A CN116754497B (en) | 2023-08-21 | 2023-08-21 | High-efficiency spectrum sensing method and system based on sparse statistics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311052634.7A CN116754497B (en) | 2023-08-21 | 2023-08-21 | High-efficiency spectrum sensing method and system based on sparse statistics |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116754497A CN116754497A (en) | 2023-09-15 |
| CN116754497B true CN116754497B (en) | 2023-11-07 |
Family
ID=87948271
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311052634.7A Active CN116754497B (en) | 2023-08-21 | 2023-08-21 | High-efficiency spectrum sensing method and system based on sparse statistics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116754497B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104063857A (en) * | 2014-06-30 | 2014-09-24 | 清华大学 | Hyperspectral image generating method and system |
| CN106952317A (en) * | 2017-03-23 | 2017-07-14 | 西安电子科技大学 | Based on the high spectrum image method for reconstructing that structure is sparse |
| CN109238463A (en) * | 2018-08-22 | 2019-01-18 | 天津大学 | A kind of active EO-1 hyperion detection system of LED based low cost |
| CN114441035A (en) * | 2021-12-31 | 2022-05-06 | 复旦大学 | Multispectral imaging method and device based on high-speed adjustable multicolor LED light source |
| CN115144075A (en) * | 2022-06-30 | 2022-10-04 | 北京理工大学 | High-speed spectral imaging method and device |
| CN115272861A (en) * | 2022-08-05 | 2022-11-01 | 西安交通大学 | Subspace sparse representation hyperspectral target detection method based on spectral correlation |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| AU2015309700A1 (en) * | 2014-08-29 | 2017-03-30 | Commonwealth Scientific And Industrial Research Organisation | Imaging method and apparatus |
| CN110717354B (en) * | 2018-07-11 | 2023-05-12 | 哈尔滨工业大学 | Super-pixel classification method based on semi-supervised K-SVD and multi-scale sparse representation |
-
2023
- 2023-08-21 CN CN202311052634.7A patent/CN116754497B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104063857A (en) * | 2014-06-30 | 2014-09-24 | 清华大学 | Hyperspectral image generating method and system |
| CN106952317A (en) * | 2017-03-23 | 2017-07-14 | 西安电子科技大学 | Based on the high spectrum image method for reconstructing that structure is sparse |
| CN109238463A (en) * | 2018-08-22 | 2019-01-18 | 天津大学 | A kind of active EO-1 hyperion detection system of LED based low cost |
| CN114441035A (en) * | 2021-12-31 | 2022-05-06 | 复旦大学 | Multispectral imaging method and device based on high-speed adjustable multicolor LED light source |
| CN115144075A (en) * | 2022-06-30 | 2022-10-04 | 北京理工大学 | High-speed spectral imaging method and device |
| CN115272861A (en) * | 2022-08-05 | 2022-11-01 | 西安交通大学 | Subspace sparse representation hyperspectral target detection method based on spectral correlation |
Non-Patent Citations (2)
| Title |
|---|
| "基于空间-光谱字典的不完备高光谱图像重构";练秋生 等;《仪器仪表学报》;第34卷(第01期);第112-118页 * |
| "空间光谱联合稀疏表示的高光谱图像超分辨率方法";许蒙恩 等;《激光与光电子学进展》;第55卷(第07期);第231-239页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116754497A (en) | 2023-09-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20230401447A1 (en) | Devices and methods employing optical-based machine learning using diffractive deep neural networks | |
| US20190096049A1 (en) | Method and Apparatus for Reconstructing Hyperspectral Image Using Artificial Intelligence | |
| CN110441271B (en) | Light field high-resolution deconvolution method and system based on convolutional neural network | |
| Liu et al. | End-to-end computational optics with a singlet lens for large depth-of-field imaging | |
| US20210233316A1 (en) | Method for Generating an Augmented Set of Images | |
| CN112130308A (en) | High-resolution microscopic imaging system with multi-angle illumination | |
| Marquez et al. | Snapshot compressive spectral depth imaging from coded aberrations | |
| CN112465922A (en) | Hyperspectral imaging system combining chromatic aberration fuzzy imaging and image reconstruction technology | |
| Zhang et al. | Ten-mega-pixel snapshot compressive imaging with a hybrid coded aperture | |
| Ma et al. | Multi-scale ghost imaging LiDAR via sparsity constraints using push-broom scanning | |
| Li et al. | Generative adversarial network for superresolution imaging through a fiber | |
| CN116754497B (en) | High-efficiency spectrum sensing method and system based on sparse statistics | |
| Lee et al. | Design and single-shot fabrication of lensless cameras with arbitrary point spread functions | |
| WO2022123047A1 (en) | Optical method | |
| Dumas et al. | From modeling to hardware: an experimental evaluation of image plane and Fourier plane coded compressive optical imaging | |
| CN115144075B (en) | High-speed spectrum imaging method and device | |
| Lau et al. | Single-pixel image reconstruction based on block compressive sensing and convolutional neural network | |
| CN116091640A (en) | Remote sensing hyperspectral reconstruction method and system based on spectrum self-attention mechanism | |
| Sancho et al. | An embedded GPU accelerated hyperspectral video classification system in real-time | |
| CN210427972U (en) | Gray level image diaphragm diffraction imaging device | |
| Hemsley et al. | Optimized coded aperture for frugal hyperspectral image recovery using a dual-disperser system | |
| US5926296A (en) | Vector normalizing apparatus | |
| Kou et al. | Large depth-of-field computational imaging with multi-spectral and dual-aperture optics | |
| CN114234846B (en) | Rapid nonlinear compensation method based on double-response curve fitting | |
| Li et al. | MWDNs: reconstruction in multi-scale feature spaces for lensless imaging |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |