Single-channel Raman spectrum unmixing method

Abstract

The invention relates to the technical field of spectrum unmixing, in particular to a single-channel Raman spectrum unmixing method. Comprising the following steps: acquiring a mixed Raman spectrum, and performing data coding on the mixed Raman spectrum to obtain high-dimensional features; cutting the high-dimensional feature into a plurality of mutually overlapped three-dimensional tensors; inputting the three-dimensional tensor into an RSSNet separation module to obtain a separation feature tensor; overlapping and adding the separated feature tensors to obtain a high-dimensional feature sequence; inputting the high-dimensional feature sequence into a mask network to obtain a mask; inputting the high-dimensional features into a feature separation module to obtain masks; and inputting the high-dimensional features and the masks into a decoding and spectrum reconstruction module to obtain a pure spectrum. According to the method, the dependence on multispectral input is broken through through a specially designed coding-separation-decoding deep neural network architecture, single-channel unmixing is realized, and meanwhile, the method has high robustness and anti-noise capability and has a practical value of application in a real environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of spectral unmixing technology, and particularly relates to a single-channel Raman spectral unmixing method. BACKGROUND

[0002] Raman spectroscopy is a scattering spectroscopy technique based on molecular vibration and rotation, which can provide detailed chemical structure information of a substance, like a "molecular fingerprint" of the substance. Due to its advantages of rapidity, non-destructiveness, non-contact, etc., Raman spectroscopy has been widely applied in many fields such as material science, geological exploration, life science, industrial process control, etc.

[0003] In many practical application scenarios, the sample to be measured is often not a single pure substance, but a mixture composed of multiple chemical components. At this time, the collected Raman spectrum is the result of linear or nonlinear superposition of the pure spectrum signals of each component and background noise, resulting in the characteristic peaks of different substances overlapping and masking each other, which is difficult to directly identify. In order to identify the original components and their contents from such a complex mixed spectrum, a technique called "spectral unmixing" must be used. The goal of spectral unmixing is to decompose a mixed spectrum into pure spectra and corresponding abundances.

[0004] At present, for Raman spectral unmixing, the existing technologies can be mainly divided into the following categories, but all have significant limitations: 1. Methods based on geometry and statistics: This kind of method is the mainstream of traditional methods, including geometric methods such as VCA (Vertex Component Analysis); statistical methods such as ICA (Independent Component Analysis), NMF (Non-negative Matrix Factorization) or some methods based on Bayesian estimation. The design basis of these methods is an overdetermined system, that is, it is assumed to have multiple mixed spectrum data collected from different spatial points or under different conditions, for example, a Raman imaging image, each pixel of which corresponds to a spectrum. By analyzing the overall characteristics of this group of spectral data, the pure spectrum of the substance and its distribution are inversely calculated.

[0005] The disadvantage is that this kind of method is heavily dependent on "multi-channel" or "multi-spectral" input. However, in many key applications such as on-site rapid detection, portable device analysis or non-cooperative target detection, only a single mixed spectrum can usually be obtained. In these "single-channel" scenarios, the above methods are not applicable due to not meeting their basic data input requirements, greatly limiting the application range of Raman spectroscopy technology.

[0006] 2. Deep learning methods based on autoencoders: In recent years, with the development of deep learning techniques, autoencoder networks have also been applied to hyperspectral image unmixing, and some research has explored their application in Raman imaging unmixing. This type of method compresses the input spectral data to a low-dimensional latent space through an encoder, and then reconstructs it through a decoder. During the training process, the weights of the decoder are optimized to fit the characteristics of the pure spectrum, thereby achieving unmixing.

[0007] However, this type of autoencoder-based method has similar fundamental flaws as traditional methods: First, it also relies on multispectral input. This type of method is essentially still designed for image data. Like traditional geometric or statistical methods, they require a Raman image containing hundreds or even thousands of pixel points of spectra as input. Therefore, when faced with a "single-channel" detection task with only a single mixed spectrum, this type of method is also not applicable. Second, it is difficult to deal with large-scale spectral libraries. Existing autoencoder unmixing models can usually only handle a limited number of pure components. In scenarios where unmixing needs to be performed from a large-scale spectral library containing hundreds or even thousands of substances, the performance of these models will be severely degraded, or even completely ineffective, failing to meet the complex demands of practical applications.

[0008] 3. Method based on sparse regression: There are also unmixing algorithms based on sparse regression in the prior art, such as SUnSAL (Sparse Unmixing by Variable Splitting and Augmented Lagrange). This method assumes that the mixed spectrum can be sparsely linearly combined from a few pure spectra in a large pre-defined spectral library, and the unmixing process is to find this optimal sparse subset. In theory, this method can handle a single mixed spectrum.

[0009] The fatal flaw is that this type of algorithm is very sensitive to noise. The actual collected Raman spectrum will inevitably be disturbed by noise from multiple sources such as sample fluorescence, detector dark current, stray light, etc. In the presence of noise, the performance of the sparse regression algorithm will decrease sharply, often leading to unmixing failure or producing seriously erroneous results, making it lack of practicality and reliability in real and complex application environments.

[0010] Therefore, there is an urgent need in the art for a completely new technical solution that can achieve accurate and robust unmixing based only on a single, noisy mixed Raman spectrum, especially when faced with a large-scale spectral library containing hundreds or even thousands of substances. The prior art has failed to effectively solve this difficult problem. SUMMARY

[0011] The present application aims to at least solve one of the technical problems existing in the related art. To this end, the present application provides a single-channel Raman spectrum unmixing method, which breaks through the dependence on multi-spectral input, realizes single-channel unmixing, has high robustness and noise resistance, and has practical value in real environment applications.

[0012] The present application provides a single-channel Raman spectrum unmixing method, comprising: S1: obtaining a mixed Raman spectrum, and performing data encoding on the mixed Raman spectrum to obtain a high-dimensional feature; S2: inputting the high-dimensional feature into a feature separation module to obtain a mask, comprising: S21: cutting the high-dimensional feature into a plurality of three-dimensional tensors that overlap with each other; S22: inputting the three-dimensional tensors into an RSSNet separation module to obtain a separation feature tensor; S23: overlapping and adding the separation feature tensors to obtain a high-dimensional feature sequence; S24: inputting the high-dimensional feature sequence into a mask network to obtain a mask; S3: inputting the high-dimensional feature and the mask into a decoding and spectrum reconstruction module to obtain a pure spectrum.

[0013] According to the single-channel Raman spectrum unmixing method provided by the present application, step S1 comprises: S11: performing one-dimensional convolution on the mixed Raman spectrum to obtain a local feature; S12: performing global layer normalization on the local feature to obtain a normalized local feature; S13: performing parameterized linear correction on the normalized local feature to obtain a high-dimensional feature.

[0014] According to the single-channel Raman spectrum unmixing method provided by the present application, the RSSNet separation module is based on a U-Net module and is composed of a global attention mechanism and a local attention mechanism.

[0015] According to the single-channel Raman spectrum unmixing method provided by the present application, the process of the RSSNet separation module is: S221: inputting the three-dimensional tensor into an intra-block processing module to obtain an intra-block processing result; The process of the intra-block processing module is: S2211: performing depth separable convolution on the three-dimensional tensor to obtain an intra-block convolution result; S2212: transposing the three-dimensional tensor to obtain a first transposed three-dimensional tensor; S2213: inputting the first transposed three-dimensional tensor into a TDA module to obtain an intra-block TDA result; S2214: The intra-block convolution result and the intra-block TDA result are added element-wise to obtain the intra-block processing result; S222: The intra-block processing result is input to the inter-block processing module to obtain the inter-block processing result. The flow of the inter-block processing module is as follows: S2221: Perform depthwise separable convolution on the inter-block processing results to obtain inter-block convolution results; S2222: Transpose the inter-block processing result to obtain a second transposed three-dimensional tensor; S2223: Input the second transposed 3D tensor into the TDA module to obtain the inter-block TDA result; S2224: The inter-block convolution result and the inter-block TDA result are added element-wise to obtain the inter-block processing result. S223: The inter-block processing result and the three-dimensional tensor are added element-wise to obtain the separated feature tensor.

[0016] According to the single-channel Raman spectroscopy unmixing method provided by the present invention, the transpose directions in the intra-block processing module and the inter-block processing module are different, and the dimension of the first transpose three-dimensional tensor is... The dimension of the second transposed three-dimensional tensor is ,in, The number of blocks to be divided. For data block length, For feature dimensions.

[0017] According to a single-channel Raman spectroscopy unmixing method provided by the present invention, the separation feature tensor is subjected to depthwise separable convolution to obtain a cyclic input; the cyclic input is used as the input of the RSSNet separation module, and one or more cycles are performed to improve the separation effect of the RSSNet separation module.

[0018] According to the single-channel Raman spectroscopy demixing method provided by the present invention, the flow of the TDA module is as follows: S100: Convolve the input of the TDA module to obtain the TDA convolution result; S200: The TDA convolution result is downsampled multiple times to obtain multiple downsampled results; S300: Perform average pooling on each downsampling result to obtain multiple corresponding average pooling results, and combine the average pooling results with the last downsampling result. The overall information and global features are obtained by adding elements together. : in, For the first The second downsampling result For the number of samples, , is the total number of samples, is the TDA convolution result, is the average pooling function, is the element-wise addition; S400: input the overall information global feature into a global attention module to obtain a top-layer attention signal ; S500: up-sample the top-layer attention signal to obtain an up-sampling result, and element-wise add the up-sampling result and a down-sampling result to obtain an LA input value: wherein, is the nth up-sampling result, is the nth LA input value; S600: input the LA input value into an LA module to obtain an LA output result: wherein, is the nth LA output result, is the function of the LA module; S700: the last LA output result is the output result of the TDA module. According to the single-channel Raman spectrum unmixing method provided by the application, the flow of the global attention module is:

[0019] S110: input into a multi-head self-attention layer to obtain a multi-head self-attention result; S120: element-wise add the multi-head self-attention result and to obtain a feedforward network layer input value; S130: input the feedforward network layer input value into a feedforward network layer to obtain a feedforward network layer output; S140: element-wise add the feedforward network layer input value and the feedforward network layer output to obtain .

[0020] According to the single-channel Raman spectrum unmixing method provided by the application, the flow of the LA module is: S210: perform convolution operation on to obtain an LA first convolution result, perform Sigmoid activation on the LA first convolution result, and then perform interpolation to obtain an LA first intermediate value; S230: perform convolution operation on ​​​The convolution operation is performed to obtain an LA second convolution result, and interpolation is performed on the LA second convolution result to obtain an LA second intermediate value. S240: obtaining an LA third intermediate value by performing a convolution operation on the LA first intermediate value. The convolution operation is performed to obtain an LA third intermediate value. S250: obtaining an LA fourth intermediate value by multiplying the LA first intermediate value and the LA third intermediate value element by element. S260: obtaining an LA fifth intermediate value by adding the LA fourth intermediate value and the LA second intermediate value element by element. .

[0021] According to the single-channel Raman spectrum demixing method provided by the application, step S3 comprises: S31: obtaining a shielded feature by multiplying the high-dimensional feature and the mask element by element. S32: obtaining a pure spectrum by performing one-dimensional transpose convolution on the shielded feature.

[0022] According to the single-channel Raman spectrum demixing method provided by the application, when the component of the mixed Raman spectrum is greater than 2, the mixed Raman spectrum is demixed, and two demixed spectra are output, and the demixed spectra are continuously demixed until a pure spectrum is output.

[0023] According to the single-channel Raman spectrum demixing method provided by the application, when the component of the mixed Raman spectrum is greater than 2, the mixed Raman spectrum is demixed, and a plurality of pure spectra corresponding to the number of components are directly output.

[0024] The one or more technical solutions in the embodiments of the application have at least one of the following technical effects: The single-channel Raman spectrum demixing method provided by the application overcomes many defects in the prior art and has the following remarkable beneficial effects: 1. The dependence on multi-spectrum input is broken, and single-channel demixing is realized: the existing mainstream geometric, statistical and auto-encoder demixing methods all need to rely on “multi-channel” or “multi-spectrum” data input, and cannot be applied to the fundamental problem of more general single-channel detection scenarios. The application proposes a new end-to-end demixing paradigm. The application only needs a single mixed Raman spectrum obtained in a single measurement as input, and can complete the high-precision demixing task. This greatly expands the application scenarios of Raman spectrum demixing.

[0025] 2. High robustness and anti-noise capability, practical value in real environment: In view of the fatal defect that the sparse regression method in the prior art is the only method available for single channel, which is very sensitive to noise and leads to sharp performance decline and unreliable results in practical application, the application can deeply learn the internal structure and dependency of spectral features through the specially designed coding-separating-decoding deep neural network architecture, especially the unique double-path separation mechanism and attention module inside it. This enables the application to effectively suppress noise interference and obtain accurate and reliable unmixing results even under low signal-to-noise ratio conditions. The application fundamentally improves the stability and practicality of single-channel Raman unmixing technology in complex and real environments.

[0026] 3. Support for open-domain unmixing of large-scale spectral library, significantly improve the application range of the method: In view of the limitation that some existing deep learning methods (such as autoencoders) and geometry-based methods can only handle a few known components and are difficult to cope with complex mixture systems, the network structure designed by the application can efficiently unmix from a large-scale predefined spectral library containing thousands of candidate substances. This enables the application to perform material identification tasks in open-domain and non-cooperative scenarios, such as quickly identifying the components of unknown mixtures from a large chemical library in security checks, which has important practical application significance.

[0027] 4. A training method based on synthetic data is proposed, realizing low-cost model training: The application trains the model by constructing a large-scale synthetic mixed spectral dataset, avoiding the huge cost and time required for collecting and labeling a large number of diverse mixture spectral samples in the real world. Experiments show that a model trained only on synthetic data can successfully generalize to real measurement data and achieve excellent results, providing an economical and efficient technical path for the widespread application of the technology.

[0028] Additional aspects and advantages of the application will be in part apparent and in part pointed out hereinafter in the description. BRIEF DESCRIPTION OF DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0030] Figure 1 is a flowchart of the single-channel Raman spectral unmixing method provided by the application.

[0031] Figure 2 is a flow chart of the single-channel Raman spectrum unmixing method provided by the present application.

[0032] Figure 3 is a flow chart of step S2 of the present application.

[0033] Figure 4 is a flow chart of the TDA module of the present application.

[0034] Figure 5 is a flow chart of the LA module of the present application.

[0035] Figure 6 is a comparison chart of decoding results of the present application and other methods.

[0036] Figure 7 is a decoding result chart of the present application, embodiment 2.

[0037] Figure 8 is a decoding result chart of the present application, embodiment 3, when three components are involved.

[0038] Figure 9 is a decoding result chart of the present application, embodiment 3, when four components are involved.

[0039] Figure 10 is a decoding result chart of the present application, embodiment 3, when five components are involved. DETAILED DESCRIPTION

[0040] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below. Obviously, the described embodiments are some embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application. The following embodiments are used to illustrate the present application, but cannot be used to limit the scope of the present application.

[0041] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0042] The following will be described in combination withFigures 1 to 10 The application is described.

[0043] Embodiment 1 As shown in Figure 1 and Figure 2 The application provides a single-channel Raman spectrum unmixing method, comprising: S1: obtaining a mixed Raman spectrum, and performing data coding on the mixed Raman spectrum to obtain high-dimensional features; S2: inputting the high-dimensional features into a feature separation module to obtain a mask; S3: inputting the high-dimensional features and the mask into a decoding and spectrum reconstruction module to obtain pure spectra.

[0044] Specifically, step S1 comprises: S11: performing one-dimensional convolution on the mixed Raman spectrum to obtain local features.

[0045] Specifically, the application performs one-dimensional convolution operation on the input spectrum to extract local features. The convolution layer has a specific convolution kernel size and step.

[0046] S12: performing global layer normalization on the local features to obtain normalized local features.

[0047] Specifically, the application performs normalization processing on the features after convolution to stabilize the training process.

[0048] S13: performing parameterized linear correction on the normalized local features to obtain high-dimensional features.

[0049] Specifically, the application uses a PReLU function as an activation function on the normalized local features, and performs nonlinear transformation on the normalized features. After processing by the coding module, high-dimensional features are obtained, where is a tensor with a dimension of , wherein is a feature dimension, is a sequence length after transformation.

[0050] Therefore, the formula of step S1 is: wherein, is a parameterized linear correction unit activation function, is a global layer normalization operation, is one-dimensional convolution.

[0051] Specifically, the RSSNet separation module is based on a U-Net module, and is composed of a global attention mechanism and a local attention mechanism. As shown in Figure 3As shown, step S2 includes: S2: inputting the high-dimensional features into a feature separation module to obtain a mask, including: S21: cutting the high-dimensional features into several three-dimensional tensors that overlap with each other.

[0052] Specifically, in order to model the local and global dependencies of the spectrum at the same time, first, the spectrum is cut into several overlapping blocks. For example, a sequence with a length of 128 is divided into 16 data blocks with a length of 8, and there is 50% overlap between adjacent blocks. After all the blocks are stacked, a three-dimensional tensor with dimensions of 16x8x128 is formed. S22: inputting the three-dimensional tensor into an RSSNet separation module to obtain a separation feature tensor. Specifically, the process of the RSSNet separation module is as follows: S221: inputting the three-dimensional tensor into an intra-block processing module to obtain an intra-block processing result, wherein the process of the intra-block processing module is as follows:

[0053] S2211: performing deep separable convolution (DWConv) on the three-dimensional tensor to obtain an intra-block convolution result.

[0054] S2212: transposing the three-dimensional tensor to obtain a first transposed three-dimensional tensor; wherein the three-dimensional tensor after transposition obtains a first transposed three-dimensional tensor with dimensions of 16x8x128.

[0055] S2213: inputting the first transposed three-dimensional tensor into a TDA module to obtain an intra-block TDA result. S2214: element-wise adding the intra-block convolution result and the intra-block TDA result to obtain the intra-block processing result. S222: inputting the intra-block processing result into an inter-block processing module to obtain an inter-block processing result, wherein the process of the inter-block processing module is as follows:

[0056] S2221: performing deep separable convolution on the intra-block processing result to obtain an inter-block convolution result; S2222: transposing the intra-block processing result to obtain a second transposed three-dimensional tensor; wherein the intra-block processing result

[0057] after transposition obtains a second transposed three-dimensional tensor with dimensions of 16x8x128.

[0058]

[0059] ​​​​​​S2223: inputting the second transposed three-dimensional tensor into a TDA module to obtain an inter-block TDA result; S2224: element-wise adding the inter-block convolution result and the inter-block TDA result to obtain an inter-block processing result.

[0060] S223: element-wise adding the inter-block processing result and the three-dimensional tensor to obtain a separated feature tensor .

[0061] S23: overlap-adding the separated feature tensor to obtain a high-dimensional feature sequence. Specifically, the separated feature tensor is seamlessly spliced back into a continuous high-dimensional feature sequence (dimension ) This operation is the inverse process of step S21.

[0062] S24: inputting the high-dimensional feature sequence into a mask network to obtain a mask. Specifically, the high-dimensional feature sequence is input into a mask network, which is usually composed of a PReLU activation layer and a one-dimensional convolution layer with a kernel size of 1. The output of the mask network is a mask , where is the mask sequence number, is the preset number of pure spectra to be separated, and in this embodiment . The dimension of each mask is the same as that of the high-dimensional feature .

[0063] Specifically, as shown in Figure 4 , the TDA module (Top-Down Attention Module) is a processing unit for deep feature extraction and modulation, and its internal architecture is an enhanced structure based on a U-Net network. The process of the TDA module described in the present application is as follows: S100: convolving the input of the TDA module to obtain a TDA convolution result; S200: down-sampling the TDA convolution result to obtain a down-sampled result: wherein, is the down-sampled result, is a down-sampling function, is the sampling number, , is the total number of samples, is the TDA convolution result; in the embodiment of the present application .

[0064] S300: Perform average pooling on each downsampling result to obtain multiple corresponding average pooling results, and combine the average pooling results with the last downsampling result. The overall information and global features are obtained by adding elements together. : in, For average pooling function, This is element-wise addition; S400: Input the overall information global features into the global attention module to obtain the top-level attention signal. ; S500: Upsample the top-level attention signal to obtain an upsampling result, and add the upsampling result and the downsampling result element-wise to obtain the LA input value. in, For the first Second upsampling result, For upsampling operation, For the first The second LA input value, specifically, ; S600: Input the LA input value to the LA module to obtain the LA output result: in, For the first The next LA output result, For the first The next LA output result, Functions for the LA module, especially , S700: The last LA output result is the output result of the TDA module.

[0065] The process of the global attention module is as follows: S110: Will Input to a multi-head self-attention layer (MHA) to obtain multi-head self-attention results; S120: Combine the multi-head self-attention results with... The input values ​​of the feedforward network layer are obtained by adding elements together. S130: Input the input value of the feedforward network layer to the feedforward network layer (FFN) to obtain the output of the feedforward network layer; S140: The input and output values ​​of the feedforward network layer are added element-wise to obtain... .

[0066] As Figure 5 shown, the application introduces a LA module (Up-sampling and Local Attention Module), which dynamically generates parameters such as gain and bias for fine adjustment of the current scale feature according to the information of adjacent and coarser scale feature maps. This local and adaptive modulation process can gradually and finely reconstruct a feature representation containing rich details.

[0067] The process of the LA module is as follows: S210: Perform convolution operation on the first LA convolution result to obtain the LA first intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. S230: Perform convolution operation on the second LA convolution result to obtain the LA second intermediate value. .

[0068] Specifically, step S3 includes: S31: Multiply the high-dimensional feature and the mask by element to obtain the screened feature.

[0069] Specifically, the high-dimensional feature is multiplied by each mask by element to obtain C screened features . This operation is equivalent to filtering out the exclusive feature information of each pure component.

[0070] S32: Perform one-dimensional transpose convolution on the screened feature to obtain the pure spectrum.

[0071] Each screened feature performs one-dimensional transpose convolution operation. This operation is the inverse operation of one-dimensional convolution in the encoding module, which is responsible for mapping the high-dimensional feature back to the original spectral dimension, thereby obtaining the final estimated pure spectrum .

[0072] Specifically, the formula of step S3 is: wherein, It is a one-dimensional transpose convolution function. This is element-wise multiplication.

[0073] In particular, for this system to work accurately, it needs to be trained. This invention also includes a unique training method, the process of which is as follows: 1. Construct a synthetic training dataset: First, obtain a large-scale, open-source library of pure Raman spectra of substances, such as the RRUFF database, which contains pure spectra of thousands of known substances.

[0074] D different pure spectra are randomly selected from the library.

[0075] Random abundance coefficients are assigned to these D pure spectra based on a linear mixture model. The abundance coefficients sum to 1 and are linearly added together.

[0076] By adding superimposed random red and white noise within a specific signal-to-noise ratio (SNR) range to the synthesized mixed spectrum, a noisy single mixed spectrum simulating real-world conditions is generated and used as a training data point. Tens of thousands of mixed spectra are generated using this method as the training dataset.

[0077] This process can be represented by the following formula: in, It is a mixed spectrum. For the first A pure spectrum, This is for superimposed noise.

[0078] 2. Define the loss function: The negative value of the scale-invariant signal-to-noise ratio (SI-SNR) is used as the loss function. The SI-SNR measures the difference between the estimated pure spectrum output by the system and the true pure spectrum. A higher SI-SNR value indicates that the estimation result is closer to the true value.

[0079] 3. Model optimization: The synthesized mixed spectrum is input into the system to obtain the estimated spectrum. The SI-SNR loss between the estimated spectrum and the pure spectrum is calculated. Using optimization algorithms such as gradient descent and the Adam optimizer, the weight parameters of all neural network modules in the system are continuously adjusted through backpropagation to maximize the SI-SNR value, i.e., minimize the loss. This process is repeated until the model converges.

[0080] like Figure 6 As shown, Figure 6 This is a comparison chart of the decoding results of Embodiment 1 of the present invention and other methods. Figure 6The Raman spectrum in the middle (a) is formed by mixing realgar and tianhe stone 1:1, the spectrum diagram separated by the method (RSSNet prediction) adopted in the application is basically the same as the pure spectrum, and the spectrum diagram separated by other methods (SUnSAL prediction and NNOMP prediction) is completely different from the pure spectrum; Figure 6 The Raman spectrum in the middle (b) is formed by mixing realgar and tianhe stone 1:1, the spectrum diagram separated by the method (RSSNet prediction) adopted in the application is basically the same as the pure spectrum, and the spectrum diagram separated by other methods (SUnSAL prediction and NNOMP prediction) is completely different from the pure spectrum; Figure 6 The Raman spectrum in the middle (c) is formed by mixing phlogopite and ocher 1:2, the spectrum diagram separated by the method (RSSNet prediction) adopted in the application is basically the same as the pure spectrum, and the spectrum diagram separated by other methods (SUnSAL prediction and NNOMP prediction) is completely different from the pure spectrum; Figure 6 The Raman spectrum in the middle (d) is formed by mixing realgar and calcite 1:2, the spectrum diagram separated by the method (RSSNet prediction) adopted in the application is basically the same as the pure spectrum, and the spectrum diagram separated by other methods (SUnSAL prediction and NNOMP prediction) is completely different from the pure spectrum.

[0081] Example 2 When the number of components of the Raman spectrum is greater than 2, the Raman spectrum is unblended, and by modifying the output number of the mask in the mask network, a plurality of pure spectra corresponding to the number of components can be directly output.

[0082] As Figure 7 shown, Figure 7 is the decoding result diagram of Example 2 of the application.

[0083] Among them, Figure 7 The Raman spectrum in the middle (a) is the Raman spectrum of the spectrum mixed by three components of rose calcium arsenate, barium silicate and triagonal barium.

[0084] Figure 7 The Raman spectrum in the middle (b) is the Raman spectrum of the spectrum mixed by three components of calcium vanadium uraninite, hard silicon calcium and column phosphorus strontium lithium.

[0085] Figure 7 The Raman spectrum in the middle (c) is the Raman spectrum of the spectrum mixed by three components of phosphorus calcium vanadium, lead sulfocarbonate and cesium.

[0086] Example 3 When the Raman spectrum has more than two components, it can be first demixed to output two demixed spectra. These demixed spectra can then be further demixed until a pure spectrum is output. For example, a three-component mixed spectrum. like Figure 8 As shown, Figure 8 This is a decoding result diagram of the three components in Embodiment 3 of the present invention.

[0087] in, Figure 8 In the middle (a), the Raman spectrum of the mixture of three components, Busselite, chrysoprase, and tetrahydrate, is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0088] Figure 8 In the middle (b), the Raman spectrum of the mixture of sodalite, argentite and potassium gypsum is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0089] Figure 8 In the middle (c), the Raman spectrum of the mixture of three components, potassium fluoride borate, calcium oxalate, and barite, is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0090] like Figure 9 As shown, Figure 9 This is a decoding result diagram of four components in Embodiment 3 of the present invention.

[0091] in, Figure 9 In the middle (a), the Raman spectrum of the mixture of four components, dolomite, arsenopyrite, tetragonal hydroxymanganese and zabuyeite, is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0092] Figure 9 In the middle (b), the Raman spectrum of the mixture of four components, namely phosphogypsum, hydrazine, serpentine and barite, is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0093] Figure 9 In the middle (c), the Raman spectrum of the mixture of four components, boehmite, vanadium lead ore, lithium iron phosphate and hydrous magnesium calcium ore, is shown. The spectrum separated by the method (RSSNet prediction) used in this invention is basically the same as the pure spectrum.

[0094] like Figure 10 As shown, Figure 10 This is a diagram showing the decoding results of the five components in Embodiment 3 of the present invention.

[0095] wherein, Figure 10 In the (a), the Raman spectrum of the spectrum mixed by five components of wickenbergite, barytoid, red lead chromate, aluminofluorogypsum and pollucite is basically the same as the spectrum separated by the method (RSSNet prediction) adopted in the application.

[0096] Figure 10 In the (b), the Raman spectrum of the spectrum mixed by five components of xanthiosite, kurchatovite, strontium spar, pollucite and thenardite is basically the same as the spectrum separated by the method (RSSNet prediction) adopted in the application.

[0097] Figure 10 In the (c), the Raman spectrum of the spectrum mixed by five components of dolomite, natroncrossite, adapt, hydrozincbertrandite and barite is basically the same as the spectrum separated by the method (RSSNet prediction) adopted in the application.

[0098] In summary, the application proposes a new end-to-end unmixing paradigm to solve the fundamental problem that the existing mainstream geometric, statistical and autoencoder unmixing methods all need to rely on “multi-channel” or “multi-spectral” data input and cannot be applied to more general single-channel detection scenarios. The application only needs a single mixed Raman spectrum obtained in a single measurement as input to complete the high-precision unmixing task. This greatly expands the application scenarios of Raman spectrum unmixing. At the same time, the method proposed in the application can also be used to unmix the mixed Raman spectrum containing noise. Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the application, and not to limit them. Although the application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that they can modify the technical solutions recorded in the foregoing examples, or make equivalent substitutions for some of the technical features. These modifications or substitutions do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application.

[0099] It should be noted that the embodiments of the present disclosure can be realized by hardware, software or a combination of software and hardware. The hardware part can be realized by special logic: the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a special designed hardware. Those skilled in the art can understand that the above devices and methods can be realized by computer executable instructions and / or included in processor control code, for example, such code is provided on a programmable memory or a data carrier such as an optical or electronic signal carrier.

[0100] Further, although operations of the methods of the present disclosure are described in a particular order in the figures, this is not required or implied in any particular order to perform the operations, or that all operations be performed, to achieve desirable results. Rather, the steps depicted in the flowcharts can change order. Additionally or alternatively, certain steps can be omitted, combined into a single step, and / or separated into multiple steps. It should also be noted that features and functions of two or more devices according to the present disclosure can be embodied in a single device. Conversely, features and functions of one device described above can be further partitioned into multiple devices.

[0101] While the present disclosure has been described with reference to several particular embodiments, it is to be understood that the disclosure is not limited to the particular embodiments disclosed. Rather, the disclosure is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

Claims

1. A single channel Raman spectral unmixing method, characterized in that, The method comprises the following steps: S1: obtaining a mixed Raman spectrum, and performing data coding on the mixed Raman spectrum to obtain a high-dimensional feature; S2: inputting the high-dimensional feature into a feature separation module to obtain a mask, comprising: S21: cutting the high-dimensional feature into a plurality of three-dimensional tensors that overlap with each other; S22: inputting the three-dimensional tensors into an RSSNet separation module to obtain a separation feature tensor; S23: overlapping and adding the separation feature tensors to obtain a high-dimensional feature sequence; S24: inputting the high-dimensional feature sequence into a mask network to obtain a mask; S3: inputting the high-dimensional feature and the mask into a decoding and spectrum reconstruction module to obtain a pure spectrum.

2. The single channel Raman spectral unmixing method of claim 1, wherein, Step S1 comprises: S11: performing one-dimensional convolution on the mixed Raman spectrum to obtain a local feature; S12: performing global layer normalization on the local feature to obtain a normalized local feature; S13: performing parameterized linear correction on the normalized local feature to obtain a high-dimensional feature.

3. The single channel Raman spectral unmixing method of claim 1, wherein, The RSSNet separation module is based on a double-path structure and comprises a global feature processing module and a local feature processing module.

4. The single channel Raman spectral unmixing method of claim 1, wherein, The process of the RSSNet separation module is as follows: S221: inputting the three-dimensional tensor into an intra-block processing module to obtain an intra-block processing result; The process of the intra-block processing module is as follows: S2211: performing depth separable convolution on the three-dimensional tensor to obtain an intra-block convolution result; S2212: transposing the three-dimensional tensor to obtain a first transposed three-dimensional tensor; S2213: inputting the first transposed three-dimensional tensor into a TDA module to obtain an intra-block TDA result; S2214: elementarily adding the intra-block convolution result and the intra-block TDA result to obtain the intra-block processing result; S222: inputting the intra-block processing result into an inter-block processing module to obtain an inter-block processing result, and the process of the inter-block processing module is as follows: S2221: performing depth separable convolution on the inter-block processing result to obtain an inter-block convolution result; S2222: transposing the inter-block processing result to obtain a second transposed three-dimensional tensor; S2223: inputting the second transposed three-dimensional tensor into a TDA module to obtain an inter-block TDA result; S2224: elementarily adding the inter-block convolution result and the inter-block TDA result to obtain the inter-block processing result, S223: elementarily adding the inter-block processing result and the three-dimensional tensor to obtain a separation feature tensor.

5. The single channel Raman spectral unmixing method of claim 4, wherein, The transposition directions in the intra-block processing module and the inter-block processing module are different, and the dimensions of the first transposed three-dimensional tensor are ; The dimensions of the second transposed three-dimensional tensor are wherein, is the number of cut blocks, is the data block length, is the feature dimension.

6. The single channel Raman spectral unmixing method of claim 3, wherein, The separation feature tensor is subjected to depth separable convolution to obtain a cyclic input; the cyclic input is taken as the input of the RSSNet separation module and is subjected to one or more cycles to improve the separation effect of the RSSNet separation module.

7. The single channel Raman spectral unmixing method of claim 4, wherein, The process of the TDA module is as follows: S100: performing convolution on the input of the TDA module to obtain a TDA convolution result; S200: performing multiple times of down-sampling on the TDA convolution result to obtain a plurality of down-sampling results; S300: average pooling is respectively performed on each down-sampling result to obtain a plurality of corresponding average pooling results, and the average pooling results and the last down-sampling result are added element by element to obtain the global feature of the whole image The global feature of the whole image is obtained by adding elements : wherein, is the first down-sampling result, is the number of sampling, , is the total number of sampling, is the TDA convolution result, is the average pooling function, is the element-wise addition; S400: input the whole information global feature into a global attention module to obtain a top-level attention signal ; S500: performing up-sampling on the top-layer attention signal to obtain an up-sampling result, and elementarily adding the up-sampling result and the down-sampling result to obtain an LA input value: wherein, is the first upsampled result, is the first LA input value; S600: inputting the LA input value into an LA module to obtain an LA output result: wherein, is the LA output result, is a function of the LA module; S700: The last LA output result is the output result of the TDA module.

8. The single channel Raman spectral unmixing method of claim 7, wherein, The flow of the global attention module is as follows: S110: input to the multi-head self-attention layer to obtain a multi-head self-attention result; input to the multi-head self-attention layer to obtain a multi-head self-attention result; S120: add the multi-head self-attention result and element-wise addition to obtain the feedforward network layer input value; S130: Input the feedforward network layer input value into the feedforward network layer to obtain a feedforward network layer output; S140: element-wise adding the feedforward network layer input values and the feedforward network layer output values to obtain .

9. The single channel Raman spectral unmixing method of claim 7, wherein, The flow of the LA module is as follows: S210: obtaining performing a convolution operation to obtain a LA first convolution result, performing a Sigmoid activation on the LA first convolution result, and performing interpolation to obtain a LA first intermediate value; S230: obtaining a second convolution result of LA by performing a convolution operation, and obtaining a second intermediate value of LA by performing interpolation on the second convolution result of LA. performing a convolution operation to obtain a second convolution result of LA, and performing interpolation on the second convolution result of LA to obtain a second intermediate value of LA. S240: obtaining the third intermediate value LA by performing a convolution operation on the first intermediate value LB and the second intermediate value LC. obtaining the third intermediate value LA by performing a convolution S250: Multiply the LA first intermediate value and the LA third intermediate value element by element to obtain an LA fourth intermediate value; S260: element-wise adding the LA fourth intermediate value and the LA second intermediate value to obtain .

10. The single channel Raman spectral unmixing method of claim 1, wherein, Step S3 comprises: S31: Multiply the high-dimensional feature and the mask element by element to obtain a screened feature; S32: Perform one-dimensional transpose convolution on the screened feature to obtain a pure spectrum.

11. The single channel Raman spectral unmixing method of claim 1, wherein, When the number of components of the mixed Raman spectrum is greater than 2, the mixed Raman spectrum is unblended, and two unblended spectra are output. The unblended spectra are continuously unblended until a pure spectrum is output.

12. The single channel Raman spectral unmixing method of claim 1, wherein, When the number of components of the mixed Raman spectrum is greater than 2, the mixed Raman spectrum is unblended, and a plurality of pure spectra corresponding to the number of components are directly output.

Images