Spectral analysis method based on Gaussian information enhancement

Abstract

The invention discloses a spectral analysis method based on Gaussian information enhancement, and the method comprises the following steps: S1, collecting sample spectral data and physicochemical attribute data, and constructing a sample analysis data set; and S2, carrying out data cleaning on the sample spectrum data to obtain a sample one-dimensional spectrogram. And S3, performing Gaussian information enhancement on the sample one-dimensional spectrogram to generate a sample two-dimensional spectrogram database. S4, randomly dividing the sample analysis data set into a training set, a verification set and a test set, constructing a convolutional neural network model by taking the two-dimensional spectrogram of the sample as input and taking the materialized attribute data as output, and training and verifying the constructed convolutional neural network model by utilizing the divided training set and verification set to obtain a convolutional neural network model; and obtaining a convolutional neural network model for predicting spectral materialization attribute data. And S5, predicting the two-dimensional spectrograms of the samples in the test set by using the trained convolutional neural network model to obtain materialization attribute data of the samples in the test set.

Description

Technical Field

[0001] The present invention relates to the field of spectral data analysis, and in particular to a spectral analysis method based on Gaussian information enhancement. Background Art

[0002] Compared with conventional laboratory analysis methods, spectral analysis techniques such as UV-visible spectroscopy, near-infrared spectroscopy, mid-infrared spectroscopy and Raman spectroscopy have the advantages of fast, accurate and non-destructive detection. At present, spectral analysis technology has been widely used in many industries such as petrochemicals, agriculture and forestry, medicine, and food. As an indirect detection method, the accuracy of measuring spectral data and the reliability of qualitative or quantitative models are the basis for achieving accurate spectral analysis.

[0003] In the actual measurement and analysis process, factors such as fluctuations in raw material properties, changes in the measurement environment, and instrument vibration lead to measurement errors in the collected spectral data. For spectral analysis, the existence of measurement errors leads to the loss of some spectral information and causes input uncertainty problems in spectral modeling and analysis. Summary of the invention

[0004] A brief summary of one or more aspects is given below to provide a basic understanding of these aspects. This summary is not an exhaustive overview of all conceived aspects, and is neither intended to identify the key or critical elements of all aspects nor to define the scope of any or all aspects. Its only purpose is to give some concepts of one or more aspects in a simplified form as a prelude to a more detailed description that will be given later.

[0005] The purpose of the present invention is to solve the above-mentioned problems and provide a spectral analysis method based on Gaussian information enhancement. Gaussian information enhancement is used to convert the sample one-dimensional spectrum into a sample two-dimensional spectrum, and then deep feature extraction is performed through a trained convolutional neural network, thereby improving the accuracy and reliability of spectral analysis.

[0006] The technical solution of the present invention is:

[0007] The present invention provides a spectrum analysis method based on Gaussian information enhancement, comprising the following steps:

[0008] Step S1: Collect sample spectral data and physicochemical property data to construct a sample analysis data set;

[0009] Step S2: performing data cleaning on the sample spectrum data to obtain a one-dimensional spectrum of the sample;

[0010] Step S3: Perform Gaussian information enhancement on the sample one-dimensional spectrum to generate a sample two-dimensional spectrum database;

[0011] Step S4: randomly divide the sample analysis data set into a training set, a validation set and a test set, and construct a convolutional neural network model with the sample two-dimensional spectrum as input and the physicochemical property data as output, and use the divided training set and validation set to train and verify the constructed convolutional neural network model to obtain a convolutional neural network model for predicting spectral physicochemical property data;

[0012] Step S5: Use the trained convolutional neural network model to predict the two-dimensional spectra of the samples in the test set, obtain the physicochemical property data of the samples in the test set, and complete the spectral analysis of the collected samples.

[0013] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, in step S2, the sample spectral data includes wavelengths of different frequencies and corresponding absorbance values; wherein, when the spectral analysis method based on Gaussian information enhancement performs data cleaning on the sample spectral data, abnormal samples and wavelengths containing strong noise are eliminated, thereby obtaining a one-dimensional spectrum after cleaning.

[0014] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, in step S3, when the spectral analysis method based on Gaussian information enhancement performs Gaussian information enhancement on a one-dimensional spectrum, the average value and standard deviation of the absorbance of each wavelength in the sample spectral data are first calculated, and then the spectral data is enhanced based on the Gaussian error law to obtain the corresponding sample two-dimensional spectrum.

[0015] According to an embodiment of the spectrum analysis method based on Gaussian information enhancement of the present invention, the spectrum analysis method based on Gaussian information enhancement calculates the spectrum data matrix X∈R by the following formula: N×M The average absorbance at each wavelength is:

[0016]

[0017] in, represents the average absorbance at the jth wavelength, X i,j Represents the absorbance value of the i-th sample at the j-th wavelength.

[0018] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, after the spectral analysis method based on Gaussian information enhancement calculates the average value of absorbance at each wavelength, the spectral data matrix X∈R is calculated by the following formula: N×M The standard deviation of absorbance at each wavelength in:

[0019]

[0020] Among them, σ j represents the standard deviation of absorbance at the jth wavelength, Xi,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the average value of absorbance at the jth wavelength.

[0021] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, after calculating the standard deviation of absorbance at each wavelength, the spectral analysis method based on Gaussian information enhancement performs Gaussian information enhancement on the spectral data using the following formula:

[0022]

[0023] Among them, k l ∈[min(X),max(X)],l=1,2,…,L represents the lth Gaussian information enhancement variable, x i,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the enhanced two-dimensional spectrum of the i-th sample, σ j represents the standard deviation of absorbance at the jth wavelength.

[0024] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, when the spectral analysis method based on Gaussian information enhancement uses a convolutional neural network to analyze the sample two-dimensional spectrogram in the prediction set, the convolution layer, pooling layer and connection layer in the convolutional neural network are used to extract the input sample two-dimensional spectrogram for deep feature extraction, thereby completing the prediction of the physicochemical property data of the sample two-dimensional spectrogram; wherein the convolution layer uses the following formula to extract the sample two-dimensional spectrum Z∈R L×M Perform feature mapping:

[0025]

[0026] Among them, C k represents the kth convolution kernel output of the convolution layer, f(·) represents the activation function of the convolution layer, represents the weight parameter of the convolution kernel, represents convolution calculation, b k Represents the bias coefficient of the convolutional layer.

[0027] According to an embodiment of the spectrum analysis method based on Gaussian information enhancement of the present invention, after the convolutional neural network completes the feature mapping of the sample two-dimensional spectrum through the convolutional layer, the output of the convolutional layer is downsampled through the pooling layer, thereby reducing the feature dimension; wherein, the pooling window size is set to h*w, the pooling step size is s, and the convolutional neural network is performed through the maximum pooling as shown below and average pooling Perform a pooling operation on the output of the convolution:

[0028]

[0029] in, Represents the region R m,n The maximum pooling layer output within, Indicates region R m,n The maximum average pooling output within, C i,j Represents the activation value in the pooling area, and m and n represent the row and column index representation in the pooling area respectively.

[0030] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, after the convolutional neural network performs multiple convolution-pooling operations on the spectral data matrix through the convolution layer and the pooling layer, the feature matrix obtained by the multiple convolution-pooling operations is fully connected and expanded through the connection layer, so as to calculate the output of the convolutional neural network. The formula is as follows:

[0031] y out =f(Wβ+η)

[0032] Among them, y out is the output of the convolutional neural network, β is the fully connected vector, W is the fully connected weight matrix, and η is the bias vector.

[0033] According to an embodiment of the spectral analysis method based on Gaussian information enhancement of the present invention, after the spectral analysis method based on Gaussian information enhancement completes the prediction of the test set samples through the convolutional neural network model, the network structure of the convolutional neural network model is adjusted according to the prediction results of the test set, and the error back propagation algorithm is used to optimize the model parameters, so as to obtain the optimal convolutional neural network model.

[0034] 14. According to an embodiment of the spectrum analysis method based on Gaussian information enhancement of the present invention, the spectrum analysis method based on Gaussian information enhancement adopts the determination coefficient R 2 And the root mean square error RMSE are used to verify the prediction results of the test set, so as to adjust the network structure of the convolutional neural network model; among them, the determination coefficient R 2 The calculation formula for the root mean square error RMSE is as follows:

[0035]

[0036] Among them, y represents the true value, represents the predicted value, represents the average value of the true value, and T represents the number of prediction samples.

[0037] The present invention also provides a computer readable medium storing a computer program code, wherein the computer program code implements the method as described above when executed by a processor.

[0038] The present invention also provides a spectrum analysis device based on Gaussian information enhancement, comprising:

[0039] a memory for storing instructions executable by a processor; and

[0040] A processor is used to execute the instructions to implement the method as described above.

[0041] Compared with the prior art, the present invention has the following beneficial effects: for spectral data with measurement errors, after preliminary cleaning, the present invention uses Gaussian information enhancement to convert the sample one-dimensional spectrum into a sample two-dimensional spectrum, and then performs deep feature extraction on the trained convolutional neural network, thereby completing the predictive analysis of the spectral data. Compared with the prior art, the present invention realizes spectral enhancement according to the statistical distribution of the absorbance values ​​of the sample at each wavelength. On the basis of converting the one-dimensional spectrum into a two-dimensional spectrum through Gaussian information enhancement, a trained convolutional neural network is combined to perform deep feature extraction on the enhanced two-dimensional spectrum, which solves the uncertainty of the input of spectral modeling analysis caused by the measurement errors in the collected spectral data due to factors such as fluctuations in the properties of raw materials, changes in the measurement environment, and instrument jitter, and improves the accuracy and reliability of spectral analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The above features and advantages of the present invention can be better understood after reading the detailed description of the embodiments of the present disclosure in conjunction with the following drawings. In the drawings, the components are not necessarily drawn to scale, and components with similar related properties or features may have the same or similar reference numerals.

[0043] Figure 1 It is a flow chart showing an embodiment of a spectrum analysis method based on Gaussian information enhancement of the present invention.

[0044] Figure 2 It is a sample one-dimensional spectrum diagram showing an embodiment of data cleaning of gasoline sample spectral data according to the present invention.

[0045] Figure 3 1 is a sample two-dimensional spectrum diagram showing an embodiment of the present invention in which Gaussian enhancement is performed on a sample one-dimensional spectrum diagram. DETAILED DESCRIPTION

[0046] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of the present application. For ordinary technicians in this field, the present application can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0047] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0048] Unless otherwise specifically stated, the relative arrangement, numerical expressions and numerical values ​​of the parts and steps set forth in these embodiments do not limit the scope of the present application. At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the accompanying drawings are not drawn according to the actual proportional relationship. The technology, method and equipment known to those of ordinary skill in the relevant field may not be discussed in detail, but in appropriate cases, the technology, method and equipment should be considered as a part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as being merely exemplary, rather than as a limitation. Therefore, other examples of exemplary embodiments may have different values. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so that once a certain item is defined in an accompanying drawing, it does not need to be further discussed in subsequent drawings.

[0049] When describing the embodiments of the present invention in detail, for the sake of convenience, the cross-sectional views showing the device structure will not be partially enlarged according to the general scale, and the schematic views are only examples, which should not limit the scope of protection of the present invention. In addition, in actual production, the three-dimensional dimensions of length, width and depth should be included.

[0050] In the description of the present application, it should be understood that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "lateral, vertical, perpendicular, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the drawings, and are only for the convenience of describing the present application and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the device or element referred to must have a specific direction or be constructed and operated in a specific direction, and therefore cannot be understood as limiting the scope of protection of the present application; the directional words "inside and outside" refer to the inside and outside relative to the contours of each component itself.

[0051] Disclosed herein is an embodiment of a spectrum analysis method based on Gaussian information enhancement. Figure 1 FIG. 1 is a flow chart showing an embodiment of a spectrum analysis method based on Gaussian information enhancement according to the present invention. Figure 1 ,The following is a detailed description of each step of the spectral analysis method based on Gaussian ,information enhancement.

[0052] Step S1: Collect sample spectral data and physicochemical property data to construct a sample analysis data set.

[0053] Step S2: Perform data cleaning on the sample spectrum data to obtain a one-dimensional spectrum of the sample.

[0054] In this embodiment, the sample spectrum data includes wavelengths (or wave numbers) of different frequencies and corresponding absorbance values, and the physicochemical property data is one or more physicochemical indicators related to the analysis purpose. After obtaining the sample spectrum data and physicochemical property data, it is necessary to remove abnormal samples in the sample spectrum data and wavelengths containing detailed noise at both ends, so as to obtain a cleaned one-dimensional spectrum.

[0055] Taking the prediction of gasoline octane number as an example, 466 gasoline samples are collected, and the octane number is set as the physicochemical property data of the gasoline samples. The sample spectrum data and octane number of these 466 gasoline samples are collected to construct a sample analysis data set. The abnormal samples in the sample analysis data set and the wavelengths with detailed noise at both ends are removed to obtain the following: Figure 2 The one-dimensional spectrum of the sample is shown.

[0056] Step S3: Perform Gaussian information enhancement on the sample one-dimensional spectrum to generate a sample two-dimensional spectrum database.

[0057] In this embodiment, after obtaining the cleaned one-dimensional spectrum through the above steps, Gaussian information enhancement is performed on the one-dimensional spectrum. First, the average value of the absorbance of each wavelength in the sample spectrum data is calculated, and then the standard deviation of the absorbance of each wavelength is calculated using the calculated average value. Finally, the spectrum data is enhanced based on the Gaussian error law and the standard deviation of the absorbance of each wavelength, thereby obtaining the corresponding sample two-dimensional spectrum. Among them, the two-dimensional spectrum of the first sample is as follows Figure 3 shown.

[0058] Specifically, in this embodiment, for the spectral data matrix X∈R N×M , the spectral data matrix X∈R is calculated by the following formula N×M The average absorbance at each wavelength is:

[0059]

[0060] in, represents the average absorbance at the jth wavelength, x i,j Represents the absorbance value of the i-th sample at the j-th wavelength. This formula is used to complete the spectral data matrix X∈R N×M After calculating the average value of absorbance at each wavelength, the standard deviation of absorbance at each wavelength is calculated using the following formula:

[0061]

[0062] Among them, σ j represents the standard deviation of absorbance at the jth wavelength, x i,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the average value of absorbance at the jth wavelength.

[0063] In this embodiment, the spectral data matrix X∈R is calculated through the above steps. N×M After obtaining the standard deviation of absorbance at each wavelength, the spectral data is enhanced by Gaussian information using the formula shown below to obtain a two-dimensional spectrum of the sample:

[0064]

[0065] Among them, k l ∈[min(X),max(X)],l=1,2,…,L represents the lth Gaussian information enhancement variable, X i,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the enhanced two-dimensional spectrum of the i-th sample, σ j represents the standard deviation of absorbance at the jth wavelength.

[0066] Step S4: Randomly divide the sample analysis data set into a training set, a validation set and a test set, and construct a convolutional neural network model with the sample two-dimensional spectrum as input and the physicochemical property data as output. Use the divided training set and validation set to train and verify the constructed convolutional neural network model to obtain a convolutional neural network model for predicting spectral physicochemical property data.

[0067] In this embodiment, after the Gaussian information enhancement of the sample data is completed through the above steps, the sample analysis data set is randomly divided into a training set, a validation set and a test set, and then a convolutional neural network model is constructed with the sample two-dimensional spectrum as input and the physicochemical property data as output, and the constructed convolutional neural network model is trained and verified using the divided training set and validation set, thereby obtaining a convolutional neural network model for predicting spectral physicochemical property data. Among them, when the convolutional neural network model analyzes the sample two-dimensional spectrum in the prediction set, the convolutional layer, pooling layer and connection layer in the convolutional neural network are used to extract the input sample two-dimensional spectrum for deep feature extraction, thereby completing the prediction of the physicochemical property data of the sample two-dimensional spectrum.

[0068] Specifically, in this embodiment, when the convolutional neural network makes a prediction, the convolutional layer is first discussed to analyze the sample two-dimensional spectrogram Z∈R L×M For feature mapping, the formula is as follows:

[0069]

[0070] Among them, C k represents the kth convolution kernel output of the convolution layer, f(·) represents the activation function of the convolution layer, represents the weight parameter of the convolution kernel, represents convolution calculation, b k Represents the bias coefficient of the convolution layer. After the convolution layer completes the feature mapping of the sample two-dimensional spectrogram, the pooling layer is then used to downsample the output of the convolution layer to reduce the feature dimension; where the pooling window size is set to h*w and the pooling step size is s. The convolutional neural network is constructed by the maximum pooling as shown below and average pooling Perform a pooling operation on the output of the convolution:

[0071]

[0072] in, Represents the region R m,n The maximum pooling layer output within, Indicates region R m,n The maximum average pooling output within, C i,j Represents the activation value in the pooling area, and m and n represent the row and column index representation in the pooling area respectively. In this way, the convolution layer and the pooling layer are used to implement multiple convolution-pooling operations on the spectral data matrix. Finally, the feature matrix obtained by multiple convolution-pooling operations is fully connected and expanded through the connection layer to calculate the output of the convolutional neural network. The formula is as follows:

[0073] y out =f(Wβ+η)

[0074] Among them, y out is the output of the convolutional neural network, β is the fully connected vector, W is the fully connected weight matrix, and η is the bias vector.

[0075] In one embodiment, the analysis data set is randomly divided into a training set, a validation set, and a test set in a ratio of 8:1:1, and then the training set and the validation set are used to train the convolutional neural network. The network structure of the convolutional neural network is set to include 4 convolutional layers-pooling layers, wherein the sizes of the 4 convolutional kernels are 7×7, 3×3, 3×3, and 3×3, respectively, and the numbers of convolutional kernels are 64, 64, 128, and 128, respectively, and the pooling layer window size is set to 3×3, the step size is set to 2, the convolution activation function uses ReLU, the initial learning rate is set to 0.0002, and the network parameter optimization method uses the Adam algorithm. Then, the divided training set and validation set are used to train and verify the constructed convolutional neural network model, thereby obtaining a convolutional neural network model for predicting spectral physicochemical property data.

[0076] Step S5: Use the trained convolutional neural network model to predict the two-dimensional spectra of the samples in the test set, obtain the physicochemical property data of the samples in the test set, and complete the spectral analysis of the collected samples.

[0077] In this embodiment, after the trained convolutional neural network model completes the prediction of the test set, the network structure of the convolutional neural network model can be adjusted according to the prediction results of the test set, and the error back propagation algorithm is used to optimize the model parameters to obtain the best convolutional neural network model. When verifying the convolutional neural network model, the determination coefficient R 2 And root mean square error RMSE are used to verify the convolutional neural network model. Among them, the determination coefficient R 2 The calculation formula for the root mean square error RMSE is as follows:

[0078]

[0079] Among them, y represents the true value, represents the predicted value, represents the average value of the true value, and T represents the number of predicted samples. In one embodiment, the coefficient of determination R of the model 2 The root mean square error (RMSE) is 0.917 and 0.2203 respectively. The determination coefficient R 2 The root mean square error (RMSE) is used to verify the effectiveness of the proposed method.

[0080] The present specification also provides a computer-readable medium storing a computer program code, which, when executed by a processor, implements the spectral analysis method based on Gaussian information enhancement as described above.

[0081] This specification also provides a spectral analysis device based on Gaussian information enhancement, including an instruction memory for storing instructions executable by a processor, and a processor for executing instructions in the instruction memory to implement the spectral analysis method based on Gaussian information enhancement as described above.

[0082] The previous description of the disclosure is provided to enable any person skilled in the art to make or use the disclosure. Various modifications to the disclosure will be apparent to those skilled in the art, and the general principles defined herein may be applied to other variations without departing from the spirit or scope of the disclosure. Thus, the disclosure is not intended to be limited to the examples and designs described herein, but should be granted the widest scope consistent with the principles and novel features disclosed herein.

[0083] Those skilled in the art will further appreciate that the various illustrative logic blocks, modules, circuits, and algorithm steps described in conjunction with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, circuits, and steps are generally described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the specific application and the design constraints imposed on the overall system. The technician may implement the described functionality in different ways for each specific application, but such implementation decisions should not be interpreted as resulting in a departure from the scope of the present invention.

[0084] The various illustrative logic blocks, modules, and circuits described in conjunction with the embodiments disclosed herein may be implemented or performed with a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in cooperation with a DSP core, or any other such configuration.

[0085] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in a RAM memory, a flash memory, a ROM memory, an EPROM memory, an EEPROM memory, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor so that the processor can read and write information from / to the storage medium. In an alternative, a storage medium may be integrated into a processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside in a user terminal as discrete components.

[0086] In one or more exemplary embodiments, the functions described may be implemented in hardware, software, firmware, or any combination thereof. If implemented as a computer program product in software, each function may be stored on or transmitted by a computer-readable medium as one or more instructions or codes. Computer-readable media include both computer storage media and communication media, including any medium that facilitates the transfer of a computer program from one place to another. Storage media may be any available medium that can be accessed by a computer. As an example and not limitation, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer. Any connection is also properly referred to as a computer-readable medium. For example, if the software is transmitted from a website, a server, or other remote source using a coaxial cable, a fiber optic cable, a twisted pair, a digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwaves, the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of the medium. Disk and disc as used herein include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk and Blu-ray disc, wherein disk often reproduces data magnetically, while disc reproduces data optically with lasers. Combinations of the above should also be included within the scope of computer-readable media.

Claims

1. A spectrum analysis method based on Gaussian information enhancement, characterized in that: The following steps are involved: Step S1: Collect sample spectral data and physicochemical property data to construct a sample analysis data set; Step S2: performing data cleaning on the sample spectrum data to obtain a one-dimensional spectrum of the sample; Step S3: Perform Gaussian information enhancement on the sample one-dimensional spectrum to generate a sample two-dimensional spectrum database; Step S4: randomly divide the sample analysis data set into a training set, a validation set and a test set, and construct a convolutional neural network model with the sample two-dimensional spectrum as input and the physicochemical property data as output, and use the divided training set and validation set to train and verify the constructed convolutional neural network model to obtain a convolutional neural network model for predicting spectral physicochemical property data; Step S5: Use the trained convolutional neural network model to predict the two-dimensional spectra of the samples in the test set, obtain the physicochemical property data of the samples in the test set, and complete the spectral analysis of the collected samples.

2. The spectral analysis method based on Gaussian information enhancement according to claim 1, characterized in that: In step S2, the sample spectral data includes wavelengths of different frequencies and corresponding absorbance values; wherein, when the spectral analysis method based on Gaussian information enhancement performs data cleaning on the sample spectral data, abnormal samples and wavelengths containing strong noise are removed, thereby obtaining a cleaned one-dimensional spectrum.

3. The spectral analysis method based on Gaussian information enhancement according to claim 2, characterized in that: In step S3, when the spectral analysis method based on Gaussian information enhancement performs Gaussian information enhancement on the one-dimensional spectrum, the average value and standard deviation of the absorbance of each wavelength in the sample spectral data are first calculated, and then the spectral data is enhanced based on the Gaussian error law to obtain the corresponding sample two-dimensional spectrum.

4. The spectral analysis method based on Gaussian information enhancement according to claim 3, characterized in that: The spectral analysis method based on Gaussian information enhancement calculates the spectral data matrix X∈R by the following formula: N×M The average absorbance at each wavelength is: in, represents the average absorbance at the jth wavelength, x i,j Represents the absorbance value of the i-th sample at the j-th wavelength.

5. The spectrum analysis method based on Gaussian information enhancement according to claim 4, characterized in that: After the spectral analysis method based on Gaussian information enhancement calculates the average value of absorbance at each wavelength, the spectral data matrix X∈R is calculated by the following formula: N×M The standard deviation of absorbance at each wavelength in: Among them, σ j represents the standard deviation of absorbance at the jth wavelength, x i,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the average value of absorbance at the jth wavelength.

6. The spectrum analysis method based on Gaussian information enhancement according to claim 5, characterized in that: After the spectral analysis method based on Gaussian information enhancement calculates the standard deviation of absorbance at each wavelength, Gaussian information enhancement is performed on the spectral data using the following formula: Among them, k l ∈[min(X),max(X)],l=1,2,…,L represents the lth Gaussian information enhancement variable, x i,j represents the absorbance value of the i-th sample at the j-th wavelength, represents the enhanced two-dimensional spectrum of the i-th sample, σ j represents the standard deviation of absorbance at the jth wavelength.

7. The spectrum analysis method based on Gaussian information enhancement according to claim 4, characterized in that: When the spectral analysis method based on Gaussian information enhancement uses a convolutional neural network to analyze the sample two-dimensional spectra in the prediction set, the convolutional layer, pooling layer and connection layer in the convolutional neural network are used to extract the input sample two-dimensional spectra for deep feature extraction, thereby completing the prediction of the physicochemical property data of the sample two-dimensional spectra; wherein the convolutional layer uses the following formula to extract the sample two-dimensional spectrum Z∈R L×M Perform feature mapping: Among them, C k represents the kth convolution kernel output of the convolution layer, f(·) represents the activation function of the convolution layer, represents the weight parameter of the convolution kernel, represents convolution calculation, b k Represents the bias coefficient of the convolutional layer.

8. The spectrum analysis method based on Gaussian information enhancement according to claim 7, characterized in that: After the convolutional neural network completes the feature mapping of the sample two-dimensional spectrogram through the convolution layer, it downsamples the output of the convolution layer through the pooling layer to reduce the feature dimension; among them, the pooling window size is set to h*w and the pooling step size is s. The convolutional neural network uses the maximum pooling as shown below and average pooling Perform a pooling operation on the output of the convolution: in, Represents the region R m,n The maximum pooling layer output within, Indicates region R m,n The maximum average pooling output within, C i,j Represents the activation value in the pooling area, and m and n represent the row and column index representation in the pooling area respectively.

9. The spectrum analysis method based on Gaussian information enhancement according to claim 8, characterized in that: After the convolutional neural network performs multiple convolution-pooling operations on the spectral data matrix through the convolution layer and the pooling layer, the feature matrix obtained by the multiple convolution-pooling operations is fully connected and expanded through the connection layer to calculate the output of the convolutional neural network. The formula is as follows: y out =f(Wβ+η) Among them, y out is the output of the convolutional neural network, β is the fully connected vector, W is the fully connected weight matrix, and η is the bias vector.

10. The spectrum analysis method based on Gaussian information enhancement according to claim 1, characterized in that: After the spectral analysis method based on Gaussian information enhancement completes the prediction of the test set samples through the convolutional neural network model, the network structure of the convolutional neural network model is adjusted according to the prediction results of the test set, and the model parameters are optimized by using the error back propagation algorithm, so as to obtain the optimal convolutional neural network model.

11. The spectrum analysis method based on Gaussian information enhancement according to claim 1, characterized in that: The spectral analysis method based on Gaussian information enhancement adopts the determination coefficient R 2 And the root mean square error RMSE are used to verify the prediction results of the test set, so as to adjust the network structure of the convolutional neural network model; among them, the determination coefficient R 2 The calculation formula for the root mean square error RMSE is as follows: Among them, y represents the true value, represents the predicted value, represents the average value of the true value, and T represents the number of prediction samples.

12. A computer readable medium storing computer program code, characterized in that: The computer program code implements the method according to any one of claims 1 to 11 when executed by a processor.

13. A spectrum analysis device based on Gaussian information enhancement, characterized in that: include: a memory for storing instructions executable by a processor; as well as A processor, configured to execute the instructions to implement the method according to any one of claims 1 to 11.

Images