Differential spectrum reconstruction method, system and terminal

Abstract

The embodiment of the invention relates to the field of optical engineering, and provides a differential spectrum reconstruction method, system and terminal. The method comprises the following steps: generating a broadband filter array imitating soil random shape pores; processing transmission characteristics of filtering units in the broadband filter array simulating the soil random-shaped pores by adopting a differential calculation structure to obtain a reconstructed measurement matrix; and reconstructing a spectrum by using a compressed sensing algorithm based on the reconstructed measurement matrix. According to the invention, the spectrum recovery fidelity as high as 98% is realized, and the resolution of 2 nanometers is realized.

Description

Technical Field

[0001] The present invention relates to the field of optical engineering technology, and in particular to a method, system and terminal for differential spectrum reconstruction. Background Art

[0002] Miniaturized spectrometers have attracted widespread attention from academia and industry to meet the demand for compact, economical spectral analysis equipment in consumer electronics, industrial production, scientific research, and other fields. Traditional laboratory benchtop spectrometers, consisting of bulky optical components, moving parts, and long optical paths, can provide unparalleled ultra-fine resolution and a wide spectral range, making them suitable for scientific research, teaching, testing, and other fields that require high spectral analysis but do not require portability. However, the application space for spectral analysis is rapidly growing, with the need to reduce physical size, cost, and power consumption outweighing the demand for high performance.

[0003] Based on their operating principles, miniaturized spectrometers can be broadly categorized into four types: dispersion-based, Fourier transform-based, narrowband filtering-based, and computational reconstruction-based. Computational reconstruction-based spectrometers have garnered increasing attention due to their small size, large bandwidth, and compatibility with spectral imaging. As spectral reconstruction algorithms mature, they will be able to offset the performance compromises associated with miniaturization, enabling ultra-compact and high-performance spectrometers. Compared to deep learning, compressed sensing reconstruction algorithms do not require large training datasets, but they do require that the transmission characteristics of the broadband filter array be sufficiently non-correlated, meaning that the measurement matrix approximates a random Gaussian matrix to achieve sufficient spectral reconstruction accuracy.

[0004] Current designs are limited by regular geometric structures, resulting in low non-correlation between transmission properties, which significantly affects the resolution and fidelity of spectral reconstruction. Summary of the Invention

[0005] In view of the deficiencies in the prior art, the present invention proposes a method, system and terminal for differential spectrum reconstruction to solve the technical problems in the prior art.

[0006] The technical solution adopted in the present invention is:

[0007] A first aspect of an embodiment of the present application provides a method for differential spectrum reconstruction, comprising:

[0008] Generate a broadband filter array that mimics randomly shaped pores in soil;

[0009] A differential calculation structure is used to process the transmission characteristics of the filter units in the broadband filter array simulating random-shaped pores in soil to obtain a reconstructed measurement matrix;

[0010] Based on the reconstructed measurement matrix, a spectrum is reconstructed using a compressed sensing algorithm.

[0011] Furthermore, the broadband filter array for generating randomly shaped pores simulating soil comprises:

[0012] SiO2 is used as the substrate and TiO2 as the membrane layer to simulate and generate soil microscopic pores, which are then etched on the TiO2 membrane layer to generate a broadband filter array that mimics the randomly shaped pores of the soil.

[0013] Furthermore, the simulation and generation of soil microscopic pores, and etching of the microscopic pores on the TiO2 film layer to generate a broadband filter array imitating random-shaped pores in the soil, comprises:

[0014] The four-parameter random growth QSGS algorithm is used to simulate soil micropores and obtain the three-dimensional soil micropore structure.

[0015] extracting a two-dimensional random pore pattern from the three-dimensional soil microscopic pore structure using a slicing technique;

[0016] The two-dimensional random pore pattern is preprocessed to model a broadband filter array that simulates random-shaped pores in soil.

[0017] Furthermore, the two-dimensional random pore pattern is preprocessed to model a broadband filter array that simulates random-shaped pores in soil, including:

[0018] Blurring and sharpening the two-dimensional random pore pattern to obtain a two-dimensional pore pattern with sharp edges;

[0019] Independent holes in the two-dimensional pore pattern with sharp edges are removed, and the two-dimensional pore pattern with sharp edges without independent holes is divided according to a preset size to obtain the basic shape of the filtering unit of the broadband filter array, and a broadband filter array that simulates random-shaped pores in soil is modeled.

[0020] Furthermore, the differential calculation structure is used to process the transmission characteristics of the filter units in the broadband filter array simulating random-shaped pores in soil to obtain a reconstructed measurement matrix, including:

[0021] Obtaining the transmission characteristics of each filter unit of the broadband filter array that simulates soil random-shaped pores;

[0022] The difference between the projection characteristics of two adjacent filter units in the broadband filter array is used as a row vector of a reconstructed measurement matrix, and a reconstructed measurement matrix is ​​constructed based on the row vector.

[0023] Furthermore, the soil-simulating random-shaped pores in the broadband filter array of soil-simulating random-shaped pores are nanopores.

[0024] Furthermore, the compressed sensing algorithm includes: an orthogonal matching pursuit algorithm OMP and a basis pursuit algorithm BP.

[0025] A second aspect of the embodiments of the present application provides a differential computational reconstruction spectroscopy system, comprising:

[0026] A soil-simulated random pore generation module generates the basic pattern of filter units in a broadband filter array;

[0027] Filter unit batch modeling module, batch build filter units;

[0028] Filter unit transmission characteristics calculation module, simulates and calculates the transmission characteristics of different filter units;

[0029] The differential processing module performs differential processing on the transmission characteristics and constructs a measurement matrix;

[0030] Non-correlation verification module, verifies the non-correlation of the measurement matrix;

[0031] The compressed sensing spectrum reconstruction module reconstructs the spectrum using a compressed sensing algorithm based on the measurement matrix.

[0032] A third aspect of an embodiment of the present application provides a terminal, comprising a processor, an input device, an output device, and a memory, wherein the processor, input device, output device, and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions and execute the step instructions in the first aspect of the embodiment of the present application.

[0033] It can be seen from the above technical solution that the beneficial technical effects of the present invention are as follows:

[0034] 1. This paper proposes a random metasurface pattern generation scheme by simulating soil micropores and explores the factors that affect the correlation of the transmission characteristics of random metasurface patterns, solving the problem of high correlation between filtering units due to the limitations of traditional regular shapes.

[0035] 2. This invention proposes a differential calculation structure that mitigates the effects of shared features between different filter units and the influence of the filter pattern's through-hole ratio on correlation, further reducing correlation and improving the resolution and fidelity of the reconstructed spectrum. Using a compressed sensing algorithm, the measured spectrum can be restored with a fidelity of 98% and a spectral resolution of 2 nm.

[0036] 3. The spectrum reconstruction scheme of the present invention does not need to generate a large number of random structures, but can select a combination with high non-correlation to construct a measurement matrix, which greatly improves the modeling efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0038] Figure 1 This is a working principle diagram of a reconstruction spectrometer according to an embodiment of the present invention; Figure 1 (a) is the incident spectrum; Figure 1 (b) is a diagram of a broadband filter array; Figure 1 (c) is a diagram of the detector array; Figure 1 (d) is the difference result diagram of the transmission characteristics of the filter unit; Figure 1 (e) is the result diagram of the difference in transmitted light intensity; Figure 1 (f) is the flowchart of compressed sensing algorithm reconstruction;

[0039] Figure 2 This is a flow chart of an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the process of generating random-shaped nanopores according to an embodiment of the present invention; Figure 3 (a) A two-dimensional random-shaped pore structure image generated using the QSGS algorithm; Figure 3 (b) is the first blurred image; Figure 3 (c) is the first sharpened image; Figure 3 (d) is the second blurred image; Figure 3 (e) is the second sharpened image; Figure 3 (f) is the image with isolated holes removed; Figure 3 (g) is the basic shape image of the filter unit formed by segmentation;

[0041] Figure 4 A schematic diagram of modeling for an embodiment of the present invention;

[0042] Figure 5 Schematic diagram of original transmission characteristics and differential results of different filter units according to an embodiment of the present invention; Figure 5 (a) Transmission characteristics of random hole 1 and random hole 2 and their difference results; Figure 5 (b) Transmission characteristics of random hole 3 and random hole 4 and their difference results;

[0043] Figure 6 Schematic diagram of the variation trend of the average correlation coefficient of the non-differential method and the differential method with the average hole ratio of the filter unit according to an embodiment of the present invention;

[0044] Figure 7 This is a spectrum type diagram used to test the spectrum reconstruction effect in an embodiment of the present invention; Figure 7 (a) is a broadband spectrum; Figure 7 (b) is a narrow-band double-peak spectrum with an interval of 8 nm; Figure 7 (c) is a narrow-band double-peak spectrum with an interval of 3 nm; Figure 7 (d) Figure 7 (c) A partial enlarged view;

[0045] Figure 8 This is a system framework diagram of an embodiment of the present invention;

[0046] Figure 9 This is a terminal framework diagram of an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following embodiments of the technical solution of the present invention will be described in detail with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention and are therefore only examples and are not intended to limit the scope of protection of the present invention.

[0048] It should be noted that, unless otherwise specified, the technical or scientific terms used in this application should have the common meanings understood by those skilled in the art to which the present invention belongs.

[0049] Example

[0050] In one embodiment, a method for differential spectrum reconstruction is provided.

[0051] The working principle of the embodiment is described in detail below:

[0052] Working principle of reconstruction spectrometer:

[0053] The computational reconstruction spectrometer modulates the unknown spectrum through a broadband filter array. The detector array obtains the transmitted light intensity of different filter units and uses a compressed sensing algorithm to recover the measured spectrum. Generally, when the influence of noise is ignored, the modulation effect of the filter array on the incident spectrum can be described as:

[0054]

[0055] Where: x(λ) is the unknown spectrum to be measured, λ represents the wavelength; Φ i (1≤i≤M) is the transmission characteristic of the i-th filter unit, M is the total number of filter units; I i (λ)(1≤i≤M) is the measurement result. Discretize equation (1) into:

[0056]

[0057] Where: n (1≤n≤N) is the nth discrete wavelength, and N is the total number of discrete wavelengths. Rewrite equation (2) into the form of a linear equation system:

[0058] Φ M×N x N×1 = I M×1 (3)

[0059] In the formula: Φ M×N = (Φ1, Φ2, Φ3…Φ M ) T , I M×N = (I1, I2, I3…I M ) T .

[0060] Equation (3) can represent a conventional spectral sampling model, and the measured signal x(λ n ) can be recovered using a compressive sensing reconstruction algorithm. However, it is found in the research that the non-correlation between filtering units is a factor that must be considered in the spectral reconstruction process. It not only affects the effect of spectral recovery, but often requires generating a large number of filtering structures and selecting a combination with relatively high non-correlation from them to construct a measurement matrix, and the modeling and design process is rather cumbersome. Based on this, the present invention proposes a differential calculation structure, which uses the subtraction of the transmission characteristics of adjacent filtering units to construct a measurement matrix, so as to improve the non-correlation of the measurement matrix and effectively solve the above technical problems.

[0061] The principle of differential spectral reconstruction is as Figure 1 shown. For Equation (3), considering the transmission characteristics Φ i and Φ i+1 of the i-th and the (i + 1)-th adjacent filtering units, the corresponding transmitted light intensities are I i and I i+1 , and their relationship is:

[0062]

[0063] Subtract the two equations in Equation (4):

[0064] (Φ i - Φ i+1 )x = (I i - I i+1 ) (5)

[0065] As Figure 1 (f) shown.

[0066] Since M << N, the solution of this equation is a NP (Nondeterministic Polynominal) problem. However, the spectral signals in nature are usually sparse in some transform domains. Use the dictionary matrix Ψ to sparsely represent the measured spectrum x:

[0067] x N×1 = Ψ N×N sN×1 (6)

[0068] Adding the sparsity constraint of equation (7) to the NP problem of equation (6), the problem can be transformed into an optimization problem of solving the minimum l0 norm of a vector:

[0069]

[0070] This matches the theory of compressive sensing. For spectral reconstruction, low-rate compressive sampling consists of two parts: sparse representation and observation. It compresses an N-dimensional signal into an M-dimensional (M << N) signal, and the reconstruction process recovers the original N-dimensional signal from the M-dimensional sampled signal. The reconstruction process is as Figure 1 (f) shown. After reconstructing the sparse signal s using the greedy algorithm, the inverse operation of the sparse representation can be used to recover the spectral x to be measured.

[0071] In this embodiment, a method for differential spectral reconstruction is provided, as Figure 2 shown: including:

[0072] Step S01: Generate a broadband filter array with randomly shaped pores imitating soil;

[0073] [[ID= twenty-three ]] Step S02: Process the transmission characteristics of the filtering units in the broadband filter array with randomly shaped pores imitating soil using a differential calculation structure to obtain a reconstructed measurement matrix;

[0074] Step S03: Based on the reconstructed measurement matrix, use the compressive sensing algorithm to reconstruct the spectrum.

[0075] Specifically: In step S01, generating a broadband filter array with randomly shaped pores imitating soil includes:

[0076] Using SiO2 as the substrate and TiO2 as the film layer, simulate and generate soil micro-pores, and etch them on the TiO2 film layer to generate a broadband filter array with randomly shaped pores imitating soil.

[0077] The simulation and generation of soil micro-pores, and etching them on the TiO2 film layer to generate a broadband filter array with randomly shaped pores imitating soil includes:

[0078] Use the QSGS (Four-Parameter Stochastic Growth Method) algorithm to simulate soil micro-pores to obtain a three-dimensional soil micro-pore structure;

[0079] Use slicing technology to extract a two-dimensional random pore pattern from the three-dimensional soil micro-pore structure;

[0080] Preprocess the two-dimensional random pore structure and model it to obtain a broadband filter array with randomly shaped pores imitating soil.

[0081] The two-dimensional random-shaped pores are pre-processed to model a broadband filter array that simulates random-shaped pores in soil, including:

[0082] Blurring and sharpening the two-dimensional random pore pattern to obtain a two-dimensional pore pattern with sharp edges;

[0083] Independent holes in the two-dimensional pore pattern with sharp edges are removed, and the two-dimensional pore pattern with sharp edges without independent holes is divided according to a preset size to obtain the basic shape of the filtering unit of the broadband filter array, and a broadband filter array that simulates random-shaped pores in soil is modeled.

[0084] The soil-simulating random-shaped pores in the broadband filter array of soil-simulating random-shaped pores are nanopores.

[0085] In this embodiment, a broadband filter array employs a nanopore structure for ease of fabrication. To avoid the high correlation in the transmission characteristics of each filter element caused by conventional regular geometric shapes, and to construct a sufficiently non-correlated measurement matrix, the use of randomly shaped nanopores was considered. For non-metallic nanopores, their shape modulates incident light through various effects, including refraction, scattering, and diffraction. When the pore shape is sufficiently random, the modulation of the incident light is also nearly random.

[0086] In this embodiment, the QSGS (four-parameter random growth method) algorithm is used to simulate soil micropores.

[0087] like Figure 3 As shown in the figure, the two-dimensional random pore structure is obtained by simulating the soil microscopic pores, as shown in the figure. Figure 3 As shown in (a). Since there are many tiny burrs on the edge of the pore, it is not conducive to simulation modeling. Figure 3 As shown in (b) to 3(d), two blurring and sharpening operations are performed to obtain a two-dimensional pore pattern with sharp edges, as shown in Figure 3 As shown in (e), the pore edges become sharper. However, there are some isolated holes, such as Figure 3 As shown in the box area in (e), the size is too small to be processed, so it is detected and eliminated. Figure 3 (f) Figure 3 (f) Divide according to a certain size to form the basic shape of the broadband filter unit, such as Figure 3 As shown in (g), by repeating each shape into a periodic structure, different filter units can be obtained.

[0088] In this embodiment, the filtering unit is modeled as follows Figure 4As shown, SiO2 is used as the substrate and TiO2 as the film. The film thickness t is 300 nm, and periodic boundary conditions are applied in both the x and y directions, with a period p of 1.25 μm. A plane wave is incident along the negative z-axis, polarized along the x-direction, and the transmission characteristics of the filter unit are calculated using the FDTD (finite-difference time-domain) method. Factors influencing the transmission characteristics mainly fall into two categories. First, the film's inherent properties, including material absorption and the Fabry-Pérot (Fabry-Pérot) cavity effect, determine the overall trend of the transmission characteristics. This trend is significant and reduces the correlation between the transmission characteristics of different filter units. Second, the spectral modulation of the random pores, including scattering and diffraction, determines the randomness of the transmission characteristics of different filter units. This randomness increases the correlation between the transmission characteristics, but it is not significant and is overwhelmed by the overall trend. Therefore, to obtain transmission characteristics with strong correlation, factor one should be suppressed and factor two should be enhanced. To this end, the differential calculation structure effectively addresses this issue.

[0089] Specifically: in step S02, a differential calculation structure is used to process the transmission characteristics of the filter units in the broadband filter array simulating random-shaped pores in soil to obtain a reconstructed measurement matrix;

[0090] The differential calculation structure is used to process the transmission characteristics of the filter units in the broadband filter array simulating random-shaped pores in soil to obtain a reconstructed measurement matrix, including:

[0091] Obtaining the transmission characteristics of each filter unit of the broadband filter array that simulates soil random-shaped pores;

[0092] The difference between the projection characteristics of two adjacent filter units in the broadband filter array is used as a row vector of a reconstructed measurement matrix, and a reconstructed measurement matrix is ​​constructed based on the row vector.

[0093] In this embodiment, the differential calculation structure is shown in the following formula. The form of the measurement matrix is ​​constructed to obtain the transmission characteristics of each filter unit of the broadband filter array that simulates the random-shaped pores in the soil. The difference in the projection characteristics of two adjacent filter units in the broadband filter array is used as the row vector of the reconstructed measurement matrix. Based on this row vector, the reconstructed measurement matrix Φ′ is constructed. The difference in the light intensity collected by two adjacent detector array elements is used as the row vector of I′. The specific construction form of Φ′ and I′ is:

[0094] I' M×1 =Φ' M×N x N×1 (8)

[0095]

[0096] In this embodiment, the differential operation is as follows Figure 5 As shown, the transmission characteristics of random hole 1 and random hole 2 have roughly the same trend characteristics. Subtracting these two transmission characteristics yields the first differential result. Then, subtracting the transmission characteristics of random hole 2 from random hole 3 yields the second differential result. Similarly, by subtracting the transmission characteristics of two adjacent filter units, only M+1 filter units are needed to obtain M differential results. The measurement matrix Φ′ constructed from these M differential results has a high degree of non-correlation.

[0097] In this embodiment, a linear correlation coefficient is used to measure the non-correlation between transmission characteristics. The correlation coefficient between filter units i and j is defined as:

[0098]

[0099] Where: Φ i and Φ j (i, j = 1, 2, ..., M + 1 and i ≠ j) represents the transmission characteristics of the i-th and j-th filtering units, E represents the mean, and D represents the variance.

[0100] In this example, 200 random pore structures were generated to verify the effect of the differential method on non-correlation. When the average pore ratio of the 200 random pore structures varied in the range of 0.2-0.7, the average correlation coefficients of the differential method and the non-differential method were as follows: Figure 6 As shown in the figure, the average correlation coefficient after the differential operation is stable at around 0.21. The advantage of the differential operation is that a low and stable average correlation coefficient can be achieved under different pore ratios.

[0101] Specifically: in step S03, based on the reconstructed measurement matrix, a spectrum is reconstructed using a compressed sensing algorithm; the compressed sensing algorithm includes: an orthogonal matching pursuit algorithm (OMP), a basis pursuit algorithm (BP), etc.

[0102] In this embodiment, since the spectrum signal in nature can be represented by Gaussian signal and its linear combination, Gaussian signal and its linear combination are used to simulate the spectrum signal to verify the spectrum reconstruction effect. Figure 7 As shown, broadband spectrum, narrowband double peaks with an interval of 8 nm, and narrowband double peaks with an interval of 3 nm are used for spectral reconstruction to verify the recovery effects of broadband spectrum and narrowband spectrum respectively, and its resolution is determined by reducing the interval between narrowband double peaks.

[0103] In this embodiment, since spectral signals in nature are naturally sparse, a discrete cosine transform (DCT) is used to create a sparse representation of the measured spectrum. A DCT basis is obtained, and a dictionary matrix is ​​constructed. The measurement matrix is ​​multiplied by the dictionary matrix to obtain the sensing matrix. Since the dictionary matrix is ​​orthogonal, it does not alter the correlation of the measurement matrix, resulting in a highly non-correlated sensing matrix. The sparse signal can be solved using the orthogonal matching pursuit algorithm, and the measured spectrum can then be reconstructed using the inverse DCT.

[0104] In this embodiment, the following formula is used to determine its fidelity:

[0105]

[0106] Where: x(λ) is the spectrum to be measured; To reconstruct the signal.

[0107] In this embodiment, the greater the number of filtering units, the higher the resolution and fidelity of spectrum recovery, but the higher the device size and price, and a compromise should be made based on actual conditions.

[0108] In this embodiment, a differential calculation and reconstruction spectral system is also provided, such as Figure 8 As shown, the system includes:

[0109] Simulated soil random pore generation module 501 generates a basic pattern of filter units in a broadband filter array;

[0110] Filter unit batch modeling module 502, batch building filter units;

[0111] The filter unit transmission characteristic calculation module 503 simulates and calculates the transmission characteristics of different filter units;

[0112] The differential processing module 504 performs differential processing on the transmission characteristics to construct a measurement matrix;

[0113] A non-correlation verification module 505 verifies the non-correlation of the measurement matrix;

[0114] The compressed sensing spectrum reconstruction module 506 reconstructs the spectrum using a compressed sensing algorithm based on the measurement matrix.

[0115] The embodiment of the present application also provides a terminal, such as Figure 9 As shown, it includes a processor, an input device, an output device and a memory, and the processor, input device, output device and memory are interconnected, wherein the memory is used to store a computer program, and the computer program includes program instructions, and the processor is configured to call the program instructions and execute the step instructions in the aforementioned embodiment.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A method for differential spectrum reconstruction, characterized in that: include: Generate broadband filter arrays that mimic randomly shaped pores in soil; The transmission characteristics of the filter units in the broadband filter array simulating random-shaped pores in soil are processed by using a differential calculation structure to obtain a reconstructed measurement matrix; Based on the reconstructed measurement matrix, a spectrum is reconstructed using a compressed sensing algorithm.

2. The method for differential spectrum reconstruction according to claim 1, characterized in that: The broadband filter array for generating random-shaped pores simulating soil comprises: SiO2 is used as the substrate and TiO2 as the membrane layer to simulate and generate soil microscopic pores, which are then etched on the TiO2 membrane layer to generate a broadband filter array that imitates the random-shaped pores of the soil.

3. The method for differential spectrum reconstruction according to claim 2, characterized in that: The method simulates and generates soil microscopic pores, and etches them on the TiO2 film layer to generate a broadband filter array that simulates random-shaped pores in the soil, including: The four-parameter random growth QSGS algorithm is used to simulate soil micropores and obtain the three-dimensional soil micropore structure. extracting a two-dimensional random pore pattern from the three-dimensional soil microscopic pore structure using a slicing technique; The two-dimensional random pore pattern is preprocessed to model a broadband filter array that simulates random-shaped pores in soil.

4. The method for differential spectrum reconstruction according to claim 3, characterized in that: The two-dimensional random pore pattern is preprocessed to model a broadband filter array that simulates random-shaped pores in soil, including: The two-dimensional random pore pattern is blurred and sharpened to obtain a two-dimensional pore pattern with sharp edges; Independent holes in the two-dimensional pore pattern with sharp edges are eliminated, and the two-dimensional pore pattern with sharp edges without independent holes is divided according to a preset size to obtain the basic shape of the filtering unit of the broadband filter array, and a broadband filter array that simulates random-shaped pores in soil is modeled.

5. A method for differential spectrum reconstruction according to any one of claims 1 to 4, characterized in that: The differential calculation structure is used to process the transmission characteristics of the filter unit in the broadband filter array simulating the random-shaped pores of the soil to obtain a reconstructed measurement matrix, including: Acquiring the transmission characteristics of each filter unit of the broadband filter array simulating random-shaped pores of soil; The difference between the projection characteristics of two adjacent filter units in the broadband filter array is used as a row vector of a reconstructed measurement matrix, and a reconstructed measurement matrix is ​​constructed based on the row vector.

6. The method for differential spectrum reconstruction according to claim 1, characterized in that: The soil-simulating random-shaped pores in the broadband filter array of soil-simulating random-shaped pores are nanopores.

7. A method for differential spectrum reconstruction according to any one of claims 1 or 6, characterized in that: The compressed sensing algorithm includes: an orthogonal matching pursuit algorithm OMP and a basis pursuit algorithm BP.

8. A differential calculation and reconstruction spectral system, characterized in that: include: A soil-simulated random pore generation module generates the basic pattern of filter units in a broadband filter array; Filter unit batch modeling module, batch build filter units; Filter unit transmission characteristic calculation module, which simulates and calculates the transmission characteristics of different filter units; The differential processing module performs differential processing on the transmission characteristics and constructs a measurement matrix; A non-correlation verification module verifies the non-correlation of the measurement matrix; The compressed sensing spectrum reconstruction module reconstructs the spectrum using a compressed sensing algorithm based on the measurement matrix.

9. A terminal, characterized in that: The method comprises a processor, an input device, an output device and a memory, wherein the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program comprises program instructions, and the processor is configured to call the program instructions to execute the method according to any one of claims 1 to 7.

Images