Near infrared spectrum analysis method and device based on characteristic spectrum region reorganization data enhancement

Abstract

The application discloses a near-infrared spectrum analysis method and device based on feature spectrum region reorganization data enhancement, and relates to the field of near-infrared spectrum analysis.The method comprises the following steps: acquiring near-infrared spectrum data of a sample, and performing pretreatment to obtain pretreated spectrum data; according to the signal response of the near-infrared spectrum, the pretreated spectrum data is divided into three feature spectrum regions, namely, spectrum region I, spectrum region II and spectrum region III; the three feature spectrum regions are subjected to normalization treatment to obtain three normalized feature spectrum regions; the three normalized feature spectrum regions are subjected to interpolation treatment to obtain three interpolated feature spectrum regions; the three interpolated feature spectrum regions are used for spectrum reorganization to obtain a reorganized spectrum matrix; the reorganized spectrum matrix is subjected to feature information extraction to obtain effective feature information; and a quantitative or qualitative analysis model is established based on the effective feature information.The application can significantly improve the accuracy and reliability of near-infrared spectrum analysis.

Description

Technical Field

[0001] This application relates to the field of near-infrared spectroscopy analysis technology, and in particular to a near-infrared spectroscopy analysis method and apparatus based on feature spectral region reconstruction data enhancement. Background Technology

[0002] Near-infrared spectroscopy (NIR) technology has been widely used in agriculture, food, petrochemicals, pharmaceuticals and other fields due to its advantages such as speed, non-destructive nature and environmental friendliness. However, NIR spectral information is complex, containing a large amount of overlapping overtones and combination frequencies, and traditional full-spectrum modeling methods often fail to fully extract the effective information in the spectrum. The main shortcomings are as follows: (1) treating the entire spectrum as a whole, ignoring the characteristics of the signal response relationship between different bands; (2) failing to effectively handle the intensity differences and baseline drift between different bands; (3) lacking in-depth mining of spectral information, especially the interaction information between different bands; and (4) a large amount of redundant information affects the model performance.

[0003] There is an urgent need to develop a general spectral analysis method that can fully utilize signals from different characteristic spectral regions of the near-infrared spectrum, effectively mine inter-band interactive information, and eliminate redundant information. Summary of the Invention

[0004] The purpose of this application is to provide a near-infrared spectral analysis method and apparatus based on feature spectral region reconstruction data enhancement. The method can divide the spectrum into different bands according to the signal response of the near-infrared spectrum, eliminate the differences between bands by normalization, unify the data dimension by interpolation, and then realize the reconstruction data enhancement by linear combination. Finally, redundant information is removed by feature signal extraction, which significantly improves the accuracy and reliability of near-infrared spectral analysis.

[0005] To achieve the above objectives, this application provides the following solution.

[0006] In a first aspect, this application provides a near-infrared spectral analysis method based on characteristic spectral region reconstructed data enhancement, the near-infrared spectral analysis method based on characteristic spectral region reconstructed data enhancement includes the following steps.

[0007] Obtain near-infrared spectral data of the sample.

[0008] The near-infrared spectral data is preprocessed to obtain preprocessed spectral data.

[0009] Based on the signal response of the near-infrared spectrum, the preprocessed spectral data is divided into three characteristic spectral regions: spectral region I, spectral region II, and spectral region III. Spectral region I includes the second and third overtones of the stretching vibrations of hydrogen-containing groups. Spectral region II includes the first overtone of the stretching vibrations of hydrogen-containing groups. Spectral region III includes the combination frequency of the stretching vibrations of hydrogen-containing groups and the second overtone of the stretching vibrations of carbonyl groups.

[0010] The three characteristic spectral regions are normalized to obtain the three normalized characteristic spectral regions.

[0011] Interpolation is performed on the three normalized feature spectral regions to obtain the three interpolated feature spectral regions.

[0012] Based on the three interpolated characteristic spectral regions, spectral reconstruction is performed to obtain the reconstructed spectral matrix.

[0013] Feature information is extracted from the recombined spectral matrix to obtain effective feature information.

[0014] Quantitative or qualitative analysis models are established based on the aforementioned effective feature information.

[0015] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and capable of running on the processor, wherein the processor executes the computer program to implement the aforementioned near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement.

[0016] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement.

[0017] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement.

[0018] Based on the specific embodiments provided in this application, the following technical effects are disclosed.

[0019] This application provides a near-infrared spectral analysis method and apparatus based on feature spectral region reconstruction data enhancement. The method includes: acquiring near-infrared spectral data of a sample; preprocessing the near-infrared spectral data to obtain preprocessed spectral data; this step removes noise, baseline drift, and other interferences, improving the quality of the spectral data and providing more reliable data for subsequent spectral region division and other processing. Based on the signal response of the near-infrared spectrum, the preprocessed spectral data is divided into three feature spectral regions: spectral region I, spectral region II, and spectral region III. Spectral region I includes the second and third overtones of the stretching vibrations of hydrogen-containing groups; spectral region II includes the first overtone of the stretching vibrations of hydrogen-containing groups; and spectral region III includes the combination frequency of the stretching vibrations of hydrogen-containing groups and the second overtone of the stretching vibrations of carbonyl groups. This step allows for targeted focusing of the vibrational information of different hydrogen-containing groups and carbonyl groups, making the characteristics of each spectral region clearer and facilitating subsequent processing. The three characteristic spectral regions are normalized to obtain three normalized characteristic spectral regions. This step eliminates the influence of intensity differences between different spectral regions, making the data from each spectral region comparable and facilitating subsequent interpolation and recombination. The three normalized characteristic spectral regions are then interpolated to obtain three interpolated characteristic spectral regions. This step unifies the number of data points in each spectral region, facilitating spectral recombination and ensuring the consistency of the recombined spectral matrix. Based on the interpolated three characteristic spectral regions, spectral recombination is performed to obtain a recombined spectral matrix. This step integrates the effective information from each spectral region, forming a more comprehensive spectral matrix and laying the foundation for feature information extraction. Feature information is extracted from the recombined spectral matrix to obtain effective feature information. This step filters out key information useful for analysis from complex spectral data, reducing redundancy and improving the efficiency and accuracy of model building. A quantitative or qualitative analysis model is established based on the effective feature information. This step improves the predictive ability and stability of the model, making it more suitable for the analysis of actual samples. This method is highly versatile and interpretable, significantly improving prediction accuracy compared to traditional methods. Through a series of ordered processing steps, this application progressively optimizes spectral data, extracts effective features, and ultimately establishes a model with higher accuracy and reliability, better meeting practical analytical needs. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1This is an application environment diagram of a near-infrared spectroscopy analysis method based on feature spectral region reconstruction data enhancement according to an embodiment of this application.

[0022] Figure 2 This is a flowchart illustrating a near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement, provided as an embodiment of this application.

[0023] Figure 3 This is a schematic diagram of the PLS modeling results after extracting feature information of straw moisture content from near-infrared spectroscopy using the CARS method.

[0024] Figure 4 This is a schematic diagram of the PLS modeling results after extracting feature information of straw moisture content from near-infrared spectroscopy using the method described in this application.

[0025] Figure 5 This is a schematic diagram of the PLS modeling results after extracting feature information from the near-infrared spectroscopy analysis of wheat straw cellulose content using the CARS method.

[0026] Figure 6 This is a schematic diagram of the PLS modeling results after extracting feature information from the near-infrared spectroscopy analysis of wheat straw cellulose content using the method described in this application.

[0027] Figure 7 This is a schematic diagram of the PLS modeling results after extracting feature information from near-infrared spectroscopy analysis of crude protein content in wheat straw using the CARS method.

[0028] Figure 8 This is a schematic diagram of the PLS modeling results after extracting feature information from the crude protein content of wheat straw using the method described in this application.

[0029] Figure 9 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0030] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0031] This application relates to a near-infrared spectral analysis method based on a fusion strategy of linear reconstruction of characteristic spectral region information for data enhancement and feature information extraction. Specifically, it is a general method that enhances spectral data by dividing the near-infrared spectrum into different characteristic spectral regions according to the signal response, normalizing, interpolating, and reconstructing the different spectral regions, and extracting feature information for calibration. The near-infrared spectral analysis fusion strategy provided in this application has broad application prospects and can be used for near-infrared spectroscopy-based agricultural product quality detection, soil nutrient analysis, crop breeding screening, etc. It is also applicable to near-infrared spectroscopy-based food component analysis, adulteration detection, process monitoring, as well as drug component analysis, quality control, and raw material identification.

[0032] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0033] The near-infrared spectroscopy analysis method based on feature spectral region reconstruction data enhancement provided in this application can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be set up independently, integrated into server 104, or placed in the cloud or on another server. Terminal 102 can send the acquired near-infrared spectral data of the sample to server 104. After receiving the near-infrared spectral data, server 104 preprocesses the near-infrared spectral data to obtain preprocessed spectral data. Based on the signal response of the near-infrared spectrum, the preprocessed spectral data is divided into three characteristic spectral regions: spectral region I, spectral region II, and spectral region III. Spectral region I includes the second and third overtones of the stretching vibrations of hydrogen-containing groups; spectral region II includes the stretching vibrations of hydrogen-containing groups... The three characteristic spectral regions are: the first harmonic of the stretching vibration of hydrogen-containing groups and the second harmonic of the stretching vibration of carbonyl groups; the three characteristic spectral regions are normalized to obtain three normalized characteristic spectral regions; the three normalized characteristic spectral regions are interpolated to obtain three interpolated characteristic spectral regions; the spectra are reconstructed based on the three interpolated characteristic spectral regions to obtain a reconstructed spectral matrix; feature information is extracted from the reconstructed spectral matrix to obtain effective feature information; a quantitative or qualitative analysis model is established based on the effective feature information. The server 104 can feed back the obtained quantitative or qualitative analysis model to the terminal 102. Furthermore, in some embodiments, the near-infrared spectral analysis method based on feature spectral region reconstructed data enhancement can also be implemented separately by the server 104 or the terminal 102. For example, the terminal 102 can directly perform near-infrared spectral analysis based on feature spectral region reconstructed data enhancement on the near-infrared spectral data of the sample, or the server 104 can obtain the near-infrared spectral data of the sample from the data storage system and perform near-infrared spectral analysis based on feature spectral region reconstructed data enhancement on the near-infrared spectral data of the sample.

[0034] The terminal 102 can be, but is not limited to, various desktop computers, laptops, smartphones, and tablets. The server 104 can be implemented using a standalone server or a server cluster consisting of multiple servers, or it can be a cloud server.

[0035] In one exemplary embodiment, such as Figure 2 As shown, a near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement is provided. This method is executed by a computer device, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is applied to... Figure 1 Taking server 104 as an example, the following steps are included.

[0036] S1: Obtain the near-infrared spectral data of the sample.

[0037] S2: Preprocess the near-infrared spectral data to obtain preprocessed spectral data.

[0038] S3: Based on the signal response of the near-infrared spectrum, the preprocessed spectral data is divided into three characteristic spectral regions, namely spectral region I, spectral region II, and spectral region III; spectral region I includes: second and third overtones of the stretching vibration of hydrogen-containing groups; spectral region II includes: first overtone of the stretching vibration of hydrogen-containing groups; spectral region III includes: combination frequency of the stretching vibration of hydrogen-containing groups and second overtone of the stretching vibration of carbonyl groups.

[0039] S4: Normalize the three characteristic spectral regions to obtain the normalized three characteristic spectral regions.

[0040] S5: Perform interpolation on the three feature spectral regions after normalization to obtain the three interpolated feature spectral regions.

[0041] S6: Perform spectral reconstruction based on the three interpolated characteristic spectral regions to obtain the reconstructed spectral matrix.

[0042] S7: Extract feature information from the recombined spectral matrix to obtain effective feature information.

[0043] S8: Establish a quantitative or qualitative analysis model based on the effective feature information.

[0044] By implementing steps S1 to S8 above, this method divides the spectrum into different bands based on the signal response of the near-infrared spectrum, eliminates the differences between bands through normalization, unifies the data dimension through interpolation, enhances the recombined data through linear combination, and finally removes redundant information through feature signal extraction, which can significantly improve the accuracy and reliability of near-infrared spectral analysis.

[0045] In practical applications, the near-infrared spectroscopy analysis method based on the reconstruction of characteristic spectral regions includes the following steps.

[0046] (1) Collect near-infrared spectral data of the sample.

[0047] (2) The acquired spectra are preprocessed, including at least one of the following: standard normal variable transformation (SNV), multiple scattering correction (MSC), adaptive iterative reweighted penalized least squares (airPLS) baseline correction and Savitzky-Golay (SG) derivative smoothing.

[0048] (3) Based on the principle that the same information has multiple responses in the near-infrared spectral region in different forms such as first-order, second-order and third-order harmonics and combination frequencies, the preprocessed spectrum is divided into three characteristic spectral regions.

[0049] Spectral region I: 800–1200 nm ), mainly including the second and third harmonics of the stretching vibrations of hydrogen-containing groups.

[0050] Spectral region II: 1200–1800 nm ( ), mainly including the first harmonic of the stretching vibration of hydrogen-containing groups.

[0051] Spectral Region III: 1800–2500 nm ), mainly including the combination frequency of the stretching vibration of hydrogen-containing groups and the second harmonic of the stretching vibration of carbonyl groups.

[0052] (4) Normalize the three characteristic spectral regions respectively to eliminate the differences between different bands caused by baseline drift and spectral intensity.

[0053] (5) Use cubic spline interpolation to interpolate the three normalized characteristic spectral regions so that each characteristic spectral region has the same number of wavelength points.

[0054] (6) Stack the interpolated spectra of the three characteristic spectral regions in three dimensions—sample, wavelength point, and characteristic spectral region—to form a multidimensional spectral matrix, and then expand it into a two-dimensional spectral matrix to achieve linear recombination of information from different characteristic spectral regions and achieve the purpose of data enhancement.

[0055] (7) Use the competitive adaptive reweighted sampling (CARS) method to extract feature information from the expanded two-dimensional spectral matrix, eliminate redundant information introduced by data augmentation, and screen out effective feature information;

[0056] (8) Establish quantitative or qualitative analysis models based on the selected feature information.

[0057] As an optional implementation, the instrument used to acquire near-infrared spectra in step (1) is a near-infrared spectral acquisition device with wavelengths covering three characteristic band ranges.

[0058] As an optional implementation, the preprocessing method in step (2) can be selected from different methods or combinations of methods depending on the specific object and application.

[0059] As an optional implementation, the normalization process in step (4) maps the spectral values ​​of each spectral region to the [0, 1] interval, preferably using the maximum-minimum normalization method.

[0060] As an optional implementation, cubic spline interpolation is used in step (5) to maintain the continuity and smoothness of the spectrum while achieving data alignment.

[0061] As an optional implementation, in step (7), the parameters of the CARS method are set to retain about 90% of the variables in each iteration, and 10-fold cross-validation is used to determine the optimal subset of variables.

[0062] As an optional implementation, the establishment of quantitative and qualitative models in step (8) is applicable to various calibration algorithms.

[0063] As an optional implementation method, all spectral acquisition temperatures are 20~30℃.

[0064] It should be noted that this method is applicable to near-infrared spectral analysis of various substances, including agricultural products, food, pharmaceuticals, and petrochemical products.

[0065] This application has the following beneficial effects.

[0066] 1. High versatility: Based on the signal response of near-infrared spectroscopy, it is segmented and applicable to near-infrared spectral analysis of various substances, without being limited by sample type or instrument model.

[0067] 2. Good interpretability: Each spectral region corresponds to a clear signal response mechanism, which is easy to understand and apply.

[0068] 3. Sufficient information extraction: Information exchange between different bands is achieved through spectral reconstruction, and potential relevant information is mined.

[0069] 4. High robustness: Normalization and interpolation processing ensure that the method can adapt to spectral data under different instrument and measurement conditions.

[0070] 5. High prediction accuracy: Compared with traditional near-infrared spectroscopy analysis methods, this method has significantly improved prediction accuracy.

[0071] This application provides a near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement. Its core features are: (1) scientifically segmenting the spectrum according to the signal response of the near-infrared spectrum; (2) eliminating systematic differences between bands through normalization; (3) aligning data using interpolation methods; (4) enhancing data reconstruction through linear combination to achieve information exchange between bands; and (5) eliminating redundant information and extracting feature information using the CARS method. The detailed steps of the method are shown below.

[0072] Step 1: Spectral acquisition and preprocessing.

[0073] The near-infrared spectrum of the sample was collected, with the spectral range preferably being... (1000~2500nm), with the preferred resolution. Choose an appropriate pretreatment method based on sample characteristics and analytical requirements, such as SNV, MSC, SG derivative, or airPLS baseline correction.

[0074] Step 2: Segment the near-infrared spectrum based on the signal response.

[0075] The preprocessed near-infrared spectrum was divided into three characteristic spectral regions.

[0076] Spectral region I ( This region mainly contains the second and third harmonic frequencies of the stretching vibrations of hydrogen-containing groups such as CH, NH, and OH;

[0077] Spectral Region II ( This region mainly contains the first harmonic of the stretching vibrations of hydrogen-containing groups;

[0078] Spectral region III ( This region mainly contains combination frequency information and overtone information of groups such as C=O.

[0079] Step 3: Normalization.

[0080] The three characteristic spectral regions were normalized separately using the maximum-minimum normalization method.

[0081] .

[0082] in, These are the three characteristic spectral regions after normalization. The original spectral intensity; This represents the maximum value in the spectral region. This represents the minimum value in the spectral region. Normalization can effectively eliminate intensity differences between different spectral regions caused by factors such as baseline drift, wavelength, and optical path differences, enabling information from different bands to be fused on the same scale.

[0083] Step 4: Interpolation processing.

[0084] Cubic spline interpolation was used to interpolate the three spectral regions, ensuring that each region had the same number of wavelength points. The target number of wavelength points was set to the maximum value among the original three spectral regions. Cubic spline interpolation guarantees the second-order continuity of the interpolation function while preserving the morphological characteristics of the original spectrum to the greatest extent possible.

[0085] Step 5: Spectral reconstruction.

[0086] The interpolated spectra of the three spectral regions are recombined as follows: First, a three-dimensional spectral matrix X (I×J×K) is constructed, where I is the number of samples, J is the number of wavelength points after unification, and K is the number of spectral regions.

[0087] The three-dimensional matrix is ​​then expanded into a two-dimensional matrix X (I×3J), achieving a linear combination of information from different spectral regions. This recombination method allows spectral information from different spectral regions to interact and form new feature combinations.

[0088] Step 6: Feature information extraction.

[0089] The CARS method is used to extract feature information from the reconstructed spectral matrix. CARS uses Monte Carlo sampling to generate multiple sub-models, assesses variable importance based on the absolute value of regression coefficients, gradually reduces the number of retained variables using an exponential decay function, and selects the optimal subset of variables using cross-validation. Through these methods, the CARS method can effectively identify and retain wavelength variables that contribute significantly to the prediction target, eliminating redundant information generated by spectral reconstruction.

[0090] Step 7: Model building.

[0091] Based on the characteristic wavelengths selected by CARS, an analysis model is established using partial least squares (PLS) or other suitable methods.

[0092] The method of this application will be further described below with reference to the embodiments.

[0093] Example 1: Near-infrared spectroscopy analysis of moisture content in wheat straw.

[0094] 110 wheat straw samples were collected and divided into a calibration set (77 samples) and a validation set (33 samples) at a 7:3 ratio. Feature selection and model building were performed on the calibration set, while prediction and evaluation were conducted on the validation set. Spectra were acquired using a Fourier transform near-infrared spectrometer, covering a range of... resolution .

[0095] The following processing shall be carried out in accordance with the method described in this application.

[0096] 1. Perform SNV preprocessing on the original spectrum.

[0097] 2. Divide the spectrum into three characteristic spectral regions and normalize them respectively.

[0098] 3. Use cubic spline interpolation to unify wavelength points.

[0099] 4. Construct and expand the three-dimensional spectral matrix.

[0100] 5. Use the CARS method to select 245 characteristic wavelengths.

[0101] 6. Establish the PLS model.

[0102] In contrast, the CARS method is used to directly extract feature information from the entire spectrum and then PLS modeling is employed.

[0103] The results show that: Figure 4 As shown, the analysis results of the method in this application are as follows: RMSECV=0.25%DM RMSEP = 0.25% DM; For example Figure 3 As shown, the analysis results are obtained after directly extracting features from the full spectrum using CARS: RMSECV=0.89%DM RMSEP = 0.29%DM. The prediction accuracy of the method in this application is significantly better than that of traditional methods.

[0104] Example 2: Near-infrared spectroscopy analysis of cellulose content in wheat straw.

[0105] The same sample set and instrument conditions as in Example 1 were used.

[0106] The preprocessing was performed using the MSC + first derivative combination, and CARS selected 312 characteristic wavelengths.

[0107] The results show that: Figure 6 As shown, the analysis results of the method in this application are as follows: RMSECV = 1.05% DM RMSEP = 0.96% DM; For example Figure 5 As shown, the analysis results are obtained after directly extracting features from the full spectrum using CARS: RMSECV = 2.08%DM The RMSEP value was 2.44% DM, indicating that the method has good predictive power for different components.

[0108] Example 3: Near-infrared spectroscopy analysis of crude protein content in wheat straw.

[0109] The same sample set and instrument conditions as in Example 1 were used.

[0110] The preprocessing was performed according to the method described in this application, using SNV and selecting 550 characteristic wavelengths for CARS.

[0111] The results show that: Figure 8 As shown, the analysis results of the method in this application are as follows: RMSECV=0.12%, RMSEP = 0.10%; For example Figure 7 As shown), the analysis results were obtained after directly extracting features from the full spectrum using CARS: RMSECV=0.89%, The RMSEP was 0.77%, indicating that the method has good predictive power for different components.

[0112] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 9 As shown, the computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores the near-infrared spectral data of the samples. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement.

[0113] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0114] In one exemplary embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the above-described method embodiments.

[0115] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the above-described method embodiments.

[0116] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the above-described method embodiments.

[0117] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0118] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0119] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0120] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0121] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement, characterized in that, The near-infrared spectroscopy analysis method based on feature spectral region reconstruction data enhancement includes: Obtain near-infrared spectral data of the sample; The near-infrared spectral data is preprocessed to obtain preprocessed spectral data; Based on the signal response of the near-infrared spectrum, the preprocessed spectral data is divided into three characteristic spectral regions: spectral region I, spectral region II, and spectral region III. Spectral region I includes the second and third overtones of the stretching vibrations of hydrogen-containing groups; spectral region II includes the first overtone of the stretching vibrations of hydrogen-containing groups; and spectral region III includes the combination frequency of the stretching vibrations of hydrogen-containing groups and the second overtone of the stretching vibrations of carbonyl groups. The three characteristic spectral regions are normalized to obtain the three normalized characteristic spectral regions. Interpolation is performed on the three normalized feature spectral regions to obtain the three interpolated feature spectral regions. Based on the three interpolated characteristic spectral regions, spectral reconstruction is performed to obtain the reconstructed spectral matrix; Feature information is extracted from the recombined spectral matrix to obtain effective feature information; Quantitative or qualitative analysis models are established based on the aforementioned effective feature information; Based on the three interpolated characteristic spectral regions, spectral reconstruction is performed to obtain the reconstructed spectral matrix, specifically including: The three interpolated characteristic spectral regions are stacked in three dimensions—sample, wavelength point, and characteristic spectral region—to form a multidimensional spectral matrix. The multidimensional spectral matrix is ​​expanded into a two-dimensional spectral moment.

2. The near-infrared spectral analysis method based on characteristic spectral region reconstruction data enhancement according to claim 1, characterized in that, The expression for the normalization process is: ； in, These are the three characteristic spectral regions after normalization. The original spectral intensity; This represents the maximum value in the spectral region. This is the minimum value in the spectral region.

3. The near-infrared spectral analysis method based on characteristic spectral region reconstruction data enhancement according to claim 1, characterized in that, The preprocessing includes at least one of the following: standard normal variable transformation, multiple scattering correction, adaptive iterative reweighted penalized least squares baseline correction, and Savitzky-Golay derivative smoothing.

4. The near-infrared spectral analysis method based on characteristic spectral region reconstruction data enhancement according to claim 1, characterized in that, The three characteristic spectral regions after normalization are interpolated using cubic spline interpolation.

5. The near-infrared spectral analysis method based on characteristic spectral region reconstruction data enhancement according to claim 1, characterized in that, The feature signals of the recombined spectral matrix are extracted using a competitive adaptive reweighted sampling method.

6. The near-infrared spectral analysis method based on characteristic spectral region reconstruction data enhancement according to claim 1, characterized in that, All spectral acquisition temperatures were between 20 and 30°C.

7. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that the processor executes the computer program to implement the near-infrared spectral analysis method based on feature spectral region reconstruction data enhancement as described in any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the near-infrared spectral analysis method based on the reconstruction data of characteristic spectral regions as described in any one of claims 1-6.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the near-infrared spectral analysis method based on the reconstruction data of characteristic spectral regions as described in any one of claims 1-6.

Images