High-throughput Raman hyperspectral imaging method and system based on spectral distance modeling

Abstract

A high-throughput Raman hyperspectral imaging method and system based on spectral distance modeling. The method includes: S1. Represent the initial Raman hyperspectral image as the superposition of the optimized Raman hyperspectral image and noise; then construct an optimization problem with the goal of solving the optimized Raman hyperspectral image; S2. Introduce auxiliary variables to transform the optimization problem to obtain the transformed optimization problem, and then construct a Lagrangian function based on the transformed optimization problem; S3. Construct multiple sub-problems based on the Lagrangian function, and solve the multiple sub-problems through iterative optimization until the preset conditions are met, ending the iterative optimization, so as to obtain the optimized Raman hyperspectral image. The present invention realizes efficient low signal-to-noise ratio image reconstruction. At the same time, an unsupervised neural network is designed by using the long-range correlation between spectra and space, significantly improving the imaging quality, not relying on large-scale training data, having strong adaptability, and reducing the dependence on equipment.

Description

Technical Field

[0001] The present invention relates to the technical fields of image processing and spectral information processing, and particularly relates to a high-throughput Raman hyperspectral imaging method and system based on spectral distance modeling. Background Art

[0002] Raman hyperspectral imaging (HSI) is an analytical tool that combines Raman spectroscopy and imaging technology. It can provide molecular vibration information and its spatial distribution, and is widely used in fields such as biomedicine, drug research and development, materials science, and pharmaceuticals. Compared with other chemical imaging methods, such as fluorescence imaging and Fourier transform infrared spectroscopy imaging, Raman hyperspectral imaging technology has the advantages of non-destructive analysis, simple sample preparation, high spatial and spectral resolution; however, there are still major challenges in how to improve the quality of Raman hyperspectral imaging;

[0003] With the research of computational imaging algorithms, a method based on computational imaging algorithms provides a way to reconstruct Raman hyperspectral images with high signal-to-noise ratio (SNR) from low-quality or incomplete measurements, for example:

[0004] Traditional noise reduction algorithms (such as singular value decomposition SVD): Perform SVD decomposition on low-quality Raman spectral images and save their main component signals. However, such methods have limited effects in separating weak Raman signals and complex background noise;

[0005] Deep learning methods: In recent years, research has shown that deep learning methods can significantly improve imaging quality, but they rely on large-scale paired datasets of low-signal-to-noise ratio and high-signal-to-noise ratio Raman spectral images for training. The collection cost of such data is high or even impossible to achieve. In addition, such methods generally adopt a "training-testing" framework. First, the training process requires a large amount of high-quality data, and the training process is complex and time-consuming. In addition, the instrument dependence is strong, and the data characteristics collected by different devices vary greatly. The model needs to be pre-trained for specific devices or data before it can be applied. And the generalization ability of existing algorithms is limited and cannot process spectral data collected by different Raman instruments. Summary of the Invention

[0006] The present invention provides a high-throughput Raman hyperspectral imaging method and system based on spectral distance modeling to solve the technical problems mentioned in the background art.

[0007] To achieve the above object, the technical solution of the present invention is realized as follows:

[0008] The present invention provides a high-throughput Raman hyperspectral imaging method based on spectral distance modeling, including the following steps:

[0009] S1. Represent the obtained initial Raman hyperspectral image as an optimized Raman hyperspectral image superimposed with noise to obtain a Raman hyperspectral imaging mathematical model, and then construct an optimization problem based on the Raman hyperspectral imaging mathematical model;

[0010] S2. Introduce auxiliary variables to transform the optimization problem to obtain a transformed optimization problem, and then construct a Lagrangian function based on the transformed optimization problem;

[0011] S3. Construct multiple sub-problems based on the Lagrangian function and solve the multiple sub-problems through iterative optimization until a preset condition is met, ending the iterative optimization, thereby obtaining the optimized Raman hyperspectral image .

[0012] Furthermore, the specific steps of S1 are as follows:

[0013] S11. Represent the obtained initial Raman hyperspectral image as an optimized Raman hyperspectral image superimposed with noise to obtain a Raman hyperspectral imaging mathematical model. The Raman hyperspectral imaging mathematical model is specifically:

[0014] ;

[0015] S12. Then, with the goal of solving the optimized Raman hyperspectral image and based on the Raman hyperspectral imaging mathematical model, construct an optimization problem. The optimization problem is specifically as follows:

[0016] ;

[0017] where respectively represent the optimized Raman hyperspectral image to be solved and the noise to be solved , represents the Frobenius norm; represents the L1 norm; is a regularization term, that is, the prior information constraint on the data; , are two different balance coefficients respectively.

[0018] Furthermore, the specific steps of S2 are as follows:

[0019] S21. Introduce an auxiliary variable to transform the optimization problem in S1, where , to obtain a transformed optimization problem. The transformed optimization problem is specifically as follows:

[0020] ;

[0021] Among them, represents the auxiliary variable to be solved ;

[0022] S22. Construct the Lagrangian function according to the transformed optimization problem. The specific construction of the Lagrangian function is as follows:

[0023] ;

[0024] Among them, represents the Lagrangian function with respect to , , , , is the Lagrange multiplier, is the balance factor.

[0025] Furthermore, the specific steps of S3 are as follows:

[0026] S31. Construct multiple sub-problems according to the Lagrangian function. The multiple sub-problems are respectively sub-problem, sub-problem, sub-problem, and sub-problem;

[0027] S32. Iteratively solve the sub-problem, sub-problem, sub-problem, and sub-problem multiple times;

[0028] S33. Check whether the following convergence conditions are satisfied, or whether the preset maximum number of iterations is reached. If so, stop the iteration to obtain the optimized Raman hyperspectral image ;

[0029] Among them, the convergence conditions are specifically as follows:

[0030] ;

[0031] Among them, the auxiliary variable obtained in the th iteration; represents the norm; represents a given extremely small constant.

[0032] Furthermore, the specific mathematical expression of the sub-problem is as follows:

[0033] ;

[0034] Among them, represents the Raman hyperspectral image obtained in the th iteration; represents the noise obtained in the th iteration; represents the auxiliary variable obtained in the th iteration; represents the Lagrange multiplier obtained in the th iteration;

[0035] The sub - problem is specifically as follows:

[0036] ;

[0037] Among them, represents the auxiliary variable obtained in the th iteration;

[0038] The sub - problem is specifically as follows:

[0039] ;

[0040] Among them, represents the noise obtained in the th iteration;

[0041] The sub - problem is specifically as follows:

[0042] ;

[0043] Among them, represents the Lagrange multiplier obtained in the th iteration.

[0044] Furthermore, the specific steps of S32 are as follows:

[0045] S321. Solve and update the sub - problem, the sub - problem, the sub - problem, and the sub - problem multiple times in sequence;

[0046] S322. After the th iteration, obtain , and ;

[0047] S323. For The sub - problem performs the th iteration, and inputs , , and into the sub - problem, and obtains the solution of the th iteration of the sub - problem as: ; wherein,

[0048] ;

[0049] where is the identity matrix, and is the scaled identity matrix;

[0050] S324. Perform the th iteration on the sub - problem, take and as inputs, and finally obtain the solution of the sub - problem at the th iteration, that is ; ;

[0051] S325. Take as an input, and update the noise variable through the objective function two; The objective function two is specifically as follows:

[0052] ;

[0053] S326. Take , and as inputs, and update the Lagrange multiplier through the dual ascent method in the augmented Lagrangian method, and obtain the Lagrange multiplier after the th iteration; The solution formula of the Lagrange multiplier after the th iteration is specifically as follows:

[0054] .

[0055] Furthermore, the S324 specifically includes the following steps:

[0056] S3241. In the th iteration, decompose into the prior image PI and the optimal spectral distance OSD;

[0057] S3242. Initialize the prior image PI to obtain the initialized prior image , and then use the prior image Initialize the optimal spectral distance (OSD) respectively to obtain the initialized spectral distance ;

[0058] The initialized prior image Expressed by the formula as follows:

[0059] ;

[0060] where B represents the number of spectral bands, represents the initialized variable i.e., the variable in the first round of training in the th iteration , represents all the values of the bth band, represents the bth band;

[0061] The initialized spectral distance Expressed by the formula as follows:

[0062] ;

[0063] S3243. Use the predefined objective function two, regularization term , the initialized spectral distance to perform multiple rounds of training and update on the weight parameters of the unsupervised neural network UNN to obtain the updated unsupervised neural network UNN; where the objective function two is specifically as follows:

[0064] ;

[0065] where, is the network parameter in the tth round of training in the th iteration; is the spectral distance in the tth round of training in the th iteration; is the prior image in the tth round of training in the th iteration;

[0066] The regularization term is specifically as follows:

[0067] ;

[0068] where, is the observation value consistency constraint, i.e., minimizing the L1 distance between the generated image and the initial Raman hyperspectral image ; is the spectral-spatial smoothness constraint; is the balance factor of the spectral-spatial smoothness

[0069] S3244. Input the spectral distance of the t-th round of training in the th iteration into the updated unsupervised neural network UNN to obtain the optimized spectral distance . It is specifically expressed by the formula as follows: .

[0070] ;

[0071] S3245. Calculate the mean value of all bands of the current reconstructed image as the optimized prior image . It is specifically expressed by the formula as follows:

[0072] ;

[0073] S3246. Then add the optimized spectral distance and the optimized prior image to obtain the reconstructed image , or the auxiliary variable of the th iteration.

[0074] Further, the unsupervised neural network UNN includes an encoder and a decoder connected to the encoder;

[0075] Among them, the encoder includes three identical sub-modules one connected in sequence. The sub-module one includes a convolutional layer one, an activation function layer one, an SDTrans module, and an activation function layer two connected in sequence;

[0076] The decoder includes three identical sub-modules two connected in sequence. The sub-module two includes an upsampling layer, a convolutional layer two, an activation function layer three, an SDTrans module, and an activation function layer four connected in sequence. The inputs of the three sub-modules one of the encoder are respectively connected to the outputs of the upsampling layers in the three sub-modules two of the decoder through a splicing operation and a reverse connection method.

[0077] Further, the SDTrans module includes a Transformer module and an attention mechanism connected to the Transformer module;

[0078] Among them, the attention mechanism is used to perform pooling operations, matrix multiplications, and row-wise averaging on the output of the Transformer module respectively to obtain the weights of each band.

[0079] On the other hand, the present invention also provides a high-throughput Raman hyperspectral imaging system, including a computer device, and the computer device is programmed or configured to execute the above high-throughput Raman hyperspectral imaging method.

[0080] Advantages of the present invention:

[0081] 1. The present invention discloses a high-throughput Raman hyperspectral imaging method based on spectral distance modeling, which combines an iterative optimization framework to achieve efficient low signal-to-noise ratio image reconstruction. At the same time, an unsupervised neural network is designed using the long-range correlation between spectra and space, significantly improving the imaging quality, being independent of large-scale training data, having strong adaptability, and reducing the dependence on equipment.

[0082] 2. By predicting the spectral distance instead of directly reconstructing the complete spectrum, the present invention significantly reduces the computational complexity. The present invention reduces the amount of calculation during the optimization process, making the image reconstruction process more efficient, especially outstanding in the case of limited computing resources.

[0083] 3. The present invention does not rely on a large-scale training data set, so it can effectively perform high-quality image reconstruction even under low signal-to-noise ratio conditions. This makes the present invention particularly important in practical application scenarios where data acquisition is difficult and can avoid the dependence on a large amount of labeled data.

[0084] 4. The present invention utilizes the long-range correlation between spectra and space to optimize the image reconstruction process, thereby significantly improving the signal-to-noise ratio and imaging quality of the image. This correlation performance helps the model better capture potential image information, reduces noise interference, and can provide a clearer reconstructed image.

[0085] 5. The present invention has high generality and adaptability and can be applied to a variety of complex imaging scenarios, such as high-precision medical imaging, industrial inspection, etc., and has broader application potential.

[0086] 6. The present invention can not only accelerate the imaging process, significantly improve work efficiency, but also expand the applicability of Raman hyperspectral imaging (HSI) in dynamic sample detection and resource-constrained environments. Especially in scenarios where rapid imaging is required, the present invention can effectively provide real-time simulation results and test data, having important academic value and broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 is a flowchart of the high-throughput Raman hyperspectral imaging method in the present invention;

[0088] Figure 2 is a structural block diagram of the unsupervised neural network UNN in the present invention;

[0089] Figure 3 is a structural block diagram of the SDTrans module in the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0090] For the convenience of understanding the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Preferred embodiments of the present invention are shown in the drawings. However, the present invention can be implemented in many other different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided to make the understanding of the disclosure of the present invention more thorough and comprehensive.

[0091] Referring to Figure 1 , an embodiment of the present application provides a high-throughput Raman hyperspectral imaging method based on spectral distance modeling, including the following steps:

[0092] S1. Represent the acquired initial Raman hyperspectral image as the superimposition of the optimized Raman hyperspectral image and the noise to obtain a Raman hyperspectral imaging mathematical model, and then construct an optimization problem according to the Raman hyperspectral imaging mathematical model; wherein, , , H, W, and B respectively represent the height, width, and spectral band number of the Raman spectral image; represents a vector space;

[0093] Preferably, the initial Raman hyperspectral image is low signal-to-noise ratio data, that is, a low-quality Raman hyperspectral image ; the optimized Raman hyperspectral image is a high-quality Raman hyperspectral image ;

[0094] S2. Introduce an auxiliary variable to transform the optimization problem to obtain a transformed optimization problem, and then construct a Lagrangian function according to the transformed optimization problem;

[0095] S3. Construct multiple sub-problems according to the Lagrangian function, and solve the multiple sub-problems through iterative optimization until a preset condition is met, and end the iterative optimization, so as to obtain a high-quality Raman hyperspectral image .

[0096] The present invention discloses a high-throughput Raman hyperspectral imaging method based on spectral distance modeling, which combines an iterative optimization framework to achieve efficient low signal-to-noise ratio image reconstruction. At the same time, an unsupervised neural network is designed by using the long-range correlation between spectra and space, which significantly improves the imaging quality, does not rely on large-scale training data, has strong adaptability, and reduces the dependence on equipment.

[0097] In some embodiments, step S1 specifically includes the following steps:

[0098] S11. Represent the acquired low-quality Raman hyperspectral image as a high-quality Raman hyperspectral image Superposed with noise to obtain a Raman hyperspectral imaging mathematical model, and the Raman hyperspectral imaging mathematical model is specifically:

[0099] ;

[0100] S12. Then, taking the solution of a high-quality Raman hyperspectral image as the goal and constructing an optimization problem based on the Raman hyperspectral imaging mathematical model, the optimization problem is specifically as follows:

[0101] ;

[0102] wherein, respectively represent the high-quality Raman hyperspectral image to be solved and the noise to be solved , represents the Frobenius norm; represents the L1 norm; is a regularization term, that is, the prior information constraint on the data; , are two different balancing coefficients respectively. The term helps to reduce the influence of sparse noise interference such as cosmic rays by constraining the sparsity of the noise.

[0103] In some embodiments, the S2 specifically includes the following steps:

[0104] S21. Introduce an auxiliary variable to transform the optimization problem in S1, where , and obtain the transformed optimization problem, and the transformed optimization problem is specifically as follows:

[0105] ;

[0106] wherein, represents the auxiliary variable to be solved ;

[0107] S22. Construct a Lagrangian function based on the transformed optimization problem, and the constructed Lagrangian function is specifically:

[0108] ;

[0109] wherein, represents the Lagrangian function with respect to , , , , is the Lagrange multiplier, is the balancing factor.

[0110] In some embodiments, S3 specifically includes the following steps:

[0111] S31. Construct multiple sub-problems based on the Lagrangian function. The sub-problems are Sub-questions, Sub-questions, Sub-questions and Sub-questions;

[0112] S32, in turn Sub-questions, Sub-questions, Sub-questions and The sub-problems are solved iteratively multiple times;

[0113] S33, check whether the following convergence conditions are met, or whether the preset maximum number of iterations is reached, if yes, stop the iteration, so as to obtain the optimized Raman hyperspectral image ;

[0114] The convergence conditions are as follows:

[0115] ;

[0116] in, No. Auxiliary variables obtained by iteration; express norm; Represents a given very small constant.

[0117] In some embodiments, the The sub-problems are expressed mathematically as follows:

[0118] ;

[0119] in, Indicates The Raman hyperspectral image obtained by the iteration; Indicates The noise obtained by the iteration; Indicates Auxiliary variables obtained by iteration; Indicates The Lagrange multiplier obtained by the iteration;

[0120] Said The sub-questions are as follows:

[0121] ;

[0122] in, denote the auxiliary variable obtained in the

[0123] sub - problem is as follows:

[0124] ;

[0125] wherein, denote the noise obtained in the

[0126] sub - problem is as follows:

[0127] ;

[0128] wherein, denote the Lagrange multiplier obtained in the

[0129] In some embodiments, S32 specifically includes the following steps:

[0130] S321. Solve and update the sub - problem, sub - problem, sub - problem, and sub - problem multiple times;

[0131] Specifically, in the first iteration, i.e., when and are initialized to 0, is a low - quality Raman spectrum image ;

[0132] S322. After the th iteration, obtain , and ;

[0133] S323. Perform the th iteration on the sub - problem, input , , and into the sub - problem, and obtain the solution of the th iteration sub - problem as:

[0134] ;

[0135] wherein, is the identity matrix,​​ is a scaled identity matrix;

[0136] S324. Take the sub - problem for the th iteration, take and as inputs, and finally obtain the solution of the sub - problem for the th iteration, that is ;

[0137] S325. Take as an input, and update the noise variable through the second objective function; The second objective function is as follows:

[0138] ;

[0139] S326. Take , and as inputs, and update the Lagrange multiplier through the dual ascent method in the augmented Lagrangian method to obtain the Lagrange multiplier after the th iteration; The solution formula of the Lagrange multiplier after the th iteration is as follows:

[0140] .

[0141] In some embodiments, S324 specifically includes the following steps:

[0142] S3241. In the th iteration, decompose into a prior image PI and an optimal spectral distance OSD; Expressed by a formula, it is specifically as follows:

[0143] ;

[0144] wherein, the prior image (PI) is obtained by taking the average value of all bands of the low - signal - to - noise Raman spectral image, and is used to provide a preliminary high - signal - to - noise image. The optimal spectral distance (OSD) is dynamically optimized and generated by a designed lightweight neural network (UNN) and is used to supplement detailed information;

[0145] S3242. Initialize the prior image PI to obtain the initialized prior image , and then use the prior image to initialize the optimal spectral distance OSD respectively to obtain the initialized spectral distance ;

[0146] Prior image after initialization Expressed by a formula, specifically as follows:

[0147] ;

[0148] Where B represents the number of spectral bands, represents the initialized variable That is, the variable in the first round of training in the th iteration , represents all the values of the b-th band, represents the b-th band;

[0149] Spectral distance after initialization Expressed by a formula, specifically as follows:

[0150] ;

[0151] S3243. Utilize the predefined objective function two and regularization term , the spectral distance after initialization to update the weight parameters of the unsupervised neural network UNN for multiple rounds of training to obtain the updated unsupervised neural network UNN; where the objective function two is specifically as follows:

[0152] ;

[0153] Where, is the network parameter in the t-th round of training in the th iteration; is the spectral distance in the t-th round of training in the th iteration; is the prior image in the t-th round of training in the th iteration;

[0154] Regularization term is specifically as follows:

[0155] ;

[0156] Where, is the observation value consistency constraint, that is, to minimize the L1 distance between the generated image (i.e., the auxiliary variable obtained in the (t + 1)-th iteration) and the initial Raman hyperspectral image ; is the spectral-spatial smoothness constraint; is the balance factor of spectral-spatial smoothness; SSTV is the spectral-spatial total variation (Spectral-Spatial Total Variation); Adopt spectral-spatial total variation to enhance smoothness;

[0157] S3244. Input the spectral distance of the t-th round of training in the k-th iteration into the updated unsupervised neural network UNN to obtain the optimized spectral distance , which is specifically expressed by the formula as follows: , specifically as follows:

[0158] ;

[0159] S3245. Calculate the mean value of all bands of the current reconstructed image (i.e., the auxiliary variable obtained in the t-th iteration) as the optimized prior image ; specifically expressed by the formula as follows:

[0160] ;

[0161] S3246. Then add the optimized spectral distance and the optimized prior image to obtain the reconstructed image , or the auxiliary variable of the k-th iteration.

[0162] By predicting the spectral distance instead of directly reconstructing the complete spectrum, the present invention significantly reduces the computational complexity. The present invention reduces the amount of calculation during the optimization process, making the image reconstruction process more efficient, especially outstanding in the case of limited computing resources.

[0163] In some embodiments, referring to Figure 2 , the unsupervised neural network UNN includes an encoder and a decoder connected to the encoder;

[0164] The encoder includes three identical sub-modules one connected in sequence, and the sub-module one includes a convolutional layer one, an activation function layer one, an SDTrans module, and an activation function layer two connected in sequence; the convolutional kernel size of the convolutional layer one is 3 and the stride is 2, which is used to extract local features and reduce the resolution of the feature map; the SDTrans module is used to capture spectral correlation and non-local spatial dependence and is the core component for spectral distance modeling.

[0165] The decoder includes three identical sub-modules two connected in sequence, and the sub-module two includes an upsampling layer, a convolutional layer two, an activation function layer three, an SDTrans module, and an activation function layer four connected in sequence. The inputs of the three sub-modules one of the encoder are respectively connected to the outputs of the upsampling layers in the three sub-modules two of the decoder through a splicing operation and a reverse connection method.

[0166] The upsampling layer is used to restore the spatial dimension of the feature map;

[0167] The convolutional kernel size of the second convolutional layer is 3 and the stride is 1, which is used to extract local features;

[0168] In some embodiments, referring to Figure 3 , the SDTrans module includes a Transformer module and an attention mechanism connected to the Transformer module (i.e., Figure 3 the spatial perception attention in

[0169] Among them, the attention mechanism is used to perform pooling operations, matrix multiplication, and row-wise averaging on the output of the Transformer module respectively to obtain the weights of each band.

[0170] The SDTrans module captures long-range dependencies by introducing a Transformer module to make up for the limitations of traditional convolutional layers. It mainly includes (multi-head) self-attention mechanism and convolutional projection layer, and stabilizes the training process through residual connection and normalization. This design fully considers the correlation between spectrum and space, enabling the unsupervised neural network UNN to have efficient and stable modeling capabilities in complex imaging scenarios.

[0171] On the other hand, the present invention also provides a high-throughput Raman hyperspectral imaging system, including a computer device, which is programmed or configured to execute the above high-throughput Raman hyperspectral imaging method.

[0172] As mentioned above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, and all of them should be covered by the protection scope of the present invention. Moreover, the technical solutions between various embodiments of the present invention can be combined with each other, but it must be based on the premise that those skilled in the art can implement them. When the combination of technical solutions appears to be contradictory or unable to be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. A high-throughput Raman hyperspectral imaging method based on spectral distance modeling, characterized in that It includes the following steps: S1. Represent the obtained initial Raman hyperspectral image as an optimized Raman hyperspectral image superimposed with noise to obtain a Raman hyperspectral imaging mathematical model, and then construct an optimization problem based on the Raman hyperspectral imaging mathematical model; S2. Introduce auxiliary variables to transform the optimization problem, obtain the transformed optimization problem, and then construct a Lagrangian function based on the transformed optimization problem; The specific transformed optimization problem is as follows: ； Among them, represents the auxiliary variable to be solved , ; respectively represent the optimized Raman hyperspectral image to be solved and the noise to be solved , represents the Frobenius norm; represents the L1 norm; is the regularization term, , are two different balance coefficients respectively; S3. Construct multiple sub-problems based on the Lagrangian function. The multiple sub-problems are respectively sub-problem sub-problem sub-problem, and sub-problem; and solve the multiple sub-problems through iterative optimization until the preset conditions are met, end the iterative optimization, and thus obtain the optimized Raman hyperspectral image ; The said The sub-questions are as follows: ； Among them, represents the auxiliary variable obtained in the k-th iteration; represents the Raman hyperspectral image obtained in the k-th iteration; represents the Lagrange multiplier obtained in the (k - 1)-th iteration; is the balancing factor.

2. The high-throughput Raman hyperspectral imaging method according to claim 1, wherein The specific steps of S1 are as follows: S11. Represent the acquired initial Raman hyperspectral image as an optimized Raman hyperspectral image superimposed with noise to obtain a Raman hyperspectral imaging mathematical model, and the Raman hyperspectral imaging mathematical model is specifically: ； S12. Then, with the goal of solving the optimized Raman hyperspectral image as the objective and based on the Raman hyperspectral imaging mathematical model, an optimization problem is constructed as follows: ； Among them, respectively represent the optimized Raman hyperspectral image to be solved and the noise to be solved , represents the Frobenius norm; represents the L1 norm; is the regularization term, that is, the prior information constraint on the data; , are two different balance coefficients respectively.

3. The high-throughput Raman hyperspectral imaging method according to claim 2, characterized in that, The specific steps of S2 are as follows: S21. Introduce an auxiliary variable Transform the optimization problem in S1, where , to obtain the transformed optimization problem; S22. Construct a Lagrangian function based on the transformed optimization problem. The specific construction of the Lagrangian function is: ； Among them, represents the Lagrangian function with respect to , and is the Lagrange multiplier.

4. The high-throughput Raman hyperspectral imaging method according to claim 3, wherein The specific steps of S3 are as follows: S31. Construct multiple sub-problems based on the Lagrangian function. The multiple sub-problems are respectively sub-problem, sub-problem, sub-problem, and sub-problem; S32. Iteratively solve the sub-problem, sub-problem, sub-problem, and sub-problem multiple times; S33. Check whether the following convergence conditions are met, or whether the preset maximum number of iterations is reached. If so, stop the iteration to obtain the optimized Raman hyperspectral image ; Among them, the convergence condition is specifically as follows: ； Among them, The auxiliary variable obtained in the (k + 1)-th iteration; Denotes the L2 norm; Denotes a given extremely small constant.

5. The high-throughput Raman hyperspectral imaging method according to claim 4, wherein, The sub-problem is represented by a mathematical formula as follows: ； Among them, represents the noise obtained in the (k - 1)-th iteration; represents the auxiliary variable obtained in the (k - 1)-th iteration; The sub-problems are as follows: ； Among them, represents the noise obtained in the k-th iteration; The said The sub-problems are as follows: ； Among them, represents the Lagrange multiplier obtained in the k-th iteration.

6. The high-throughput Raman hyperspectral imaging method according to claim 5, characterized in that The specific steps of S32 are as follows: S321. Sequentially perform multiple solution updates on sub-question, sub-question, sub-question, and sub-question; After the (k - 1)-th iteration, obtain , and ; S323. Perform the k-th iteration on the sub-problem, and input , , and into the sub-problem. The solution of the k-th iteration sub-problem is: ； Among them, is the identity matrix, is the scaled identity matrix; S324. Perform the k-th iteration on the sub-problem, and use and as inputs. Finally, obtain the solution of the k-th iteration of the sub-problem, which is ; S325. Take as the input and update the noise variable through the second objective function ; The second objective function is as follows: ； S326. Take , and as inputs, and update the Lagrange multiplier through the dual ascent method in the augmented Lagrangian method to obtain the Lagrange multiplier after the k-th iteration; the solution formula for the Lagrange multiplier after the k-th iteration is specifically as follows: 。 7. The high-throughput Raman hyperspectral imaging method according to claim 6, characterized in that The specific steps of S324 are as follows: S3241. In the k-th iteration, decompose into a priori image PI and optimal spectral distance OSD; S3242. Initialize the prior image PI to obtain the initialized prior image , and then use the prior image to initialize the optimal spectral distance OSD respectively to obtain the initialized spectral distance ; Prior image after initialization Expressed by a formula, it is as follows: ； Among them, B represents the number of spectral bands, represents the initialized variable , that is, the variable for the first round of training in the k-th iteration , represents all the values of the b-th band, represents the b-th band; Spectral distance after initialization Expressed by a formula, it is as follows: ； S3243. Using the predefined objective function two and regularization term . The spectral distance after initialization . For the weight parameters of the unsupervised neural network UNN . Perform multiple rounds of training and update to obtain the updated unsupervised neural network UNN; where the objective function two is specifically as follows: ； Among them, are the network parameters for the t-th round of training in the k-th iteration; is the spectral distance for the t-th round of training in the k-th iteration; is the prior image for the t-th round of training in the k-th iteration; Regularization term The details are as follows: ； Among them, is the observed value consistency constraint, that is, minimizing the generated image and the initial Raman hyperspectral image 's L1 distance; is the spectral-spatial smoothness constraint; is the balance factor of spectral-spatial smoothness; S3244. Input the spectral distance of the t-th round of training in the k-th iteration into the updated unsupervised neural network UNN to obtain the optimized spectral distance , which is specifically expressed by the formula as follows: ； S3245. Calculate the mean value of all bands of the current reconstructed image as the optimized prior image ; It is specifically expressed by the formula as follows: ; It is expressed by the formula as follows: ； S3246. Then, add the optimized spectral distance and the optimized prior image to obtain the reconstructed image , which is also called the auxiliary variable at the k-th iteration.

8. The high-throughput Raman hyperspectral imaging method according to claim 7, wherein The unsupervised neural network UNN includes an encoder and a decoder connected to the encoder; Among them, the encoder includes three identical sub-modules one connected in sequence. The sub-module one includes a convolutional layer one, an activation function layer one, an SDTrans module, and an activation function layer two connected in sequence; The decoder includes three identical sub-modules two connected in sequence. The sub-module two includes an upsampling layer, a convolutional layer two, an activation function layer three, an SDTrans module, and an activation function layer four connected in sequence. The inputs of the three sub-modules one of the encoder are respectively connected to the outputs of the upsampling layers in the three sub-modules two of the decoder through a splicing operation and a reverse connection method.

9. The high-throughput Raman hyperspectral imaging method according to claim 8, characterized in that, The SDTrans module includes a Transformer module and an attention mechanism connected to the Transformer module; Among them, the attention mechanism is used to perform pooling operations, matrix multiplications, and row-wise averaging on the outputs of the Transformer module respectively to obtain the weights of each band.

10. A high-throughput Raman hyperspectral imaging system, including a computer device, characterized in that, The computer device is programmed or configured to execute the high-throughput Raman hyperspectral imaging method according to any one of claims 1 to 9.

Images