A raman spectroscopy discrimination method and system

Abstract

The application discloses a Raman spectrum discrimination method and system, and relates to the field of Raman spectrum analysis. The method comprises the following steps: pre-processing a to-be-detected sample signal to obtain a first Raman scattering signal of the to-be-detected sample signal; calculating a correlation coefficient according to the first Raman scattering signal and a preset spectrum database; and obtaining a matching spectrum type of the to-be-detected sample according to the correlation coefficient. According to the method, the correlation coefficient is calculated according to the first Raman scattering signal and the preset spectrum database, and the matching spectrum type of the to-be-detected sample is obtained according to the correlation coefficient. Compared with the traditional method, the calculation amount is greatly reduced, the algorithm running time is shortened, the algorithm time consumption is greatly reduced, and the operation is simple.

Description

Technical Field

[0001] This invention relates to the field of Raman spectroscopy analysis, and more particularly to a Raman spectroscopy discrimination method and system. Background Technology

[0002] After acquiring the Raman scattering signal of the sample using a Raman spectrometer, the acquired spectral signal needs to be matched. Since Raman signals are fingerprint-like and unique, a match can be found by comparing them one-to-one with database files. Assuming there are 100 database files, simply calculating the similarity between the acquired spectral signal and each of these 100 files will find the database file that best matches the sample, thus completing the Raman spectral result matching.

[0003] However, generally speaking, a Raman spectrum data dimension is 2048 data points, and the size of commonly used database files is more than 2000. Some large databases have more than 10,000 spectral data points. If the similarity between the sampled signal and the database spectrum is calculated one by one, the amount of computation is very large and the computation time is very long, which is very unfavorable for online detection. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a Raman spectroscopy discrimination method and system to address the shortcomings of the prior art.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] A Raman spectroscopy discrimination method includes:

[0007] The sample signal to be tested is preprocessed to obtain the first Raman scattering signal of the sample signal to be tested.

[0008] The correlation coefficient is calculated based on the first Raman scattering signal and a preset spectral database;

[0009] The matching spectral type of the sample to be tested is obtained based on the correlation coefficient.

[0010] The beneficial effects of this invention are: This solution calculates the correlation coefficient based on the first Raman scattering signal and a preset spectral database, and obtains the matching spectral type of the sample to be tested based on the correlation coefficient. Compared with traditional methods, the amount of calculation is greatly reduced, the algorithm running time is shortened, the algorithm consumption time is greatly reduced, and the operation is simple.

[0011] Furthermore, it also includes:

[0012] Extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases;

[0013] The step of calculating the correlation coefficient based on the first Raman scattering signal and a preset spectral database specifically includes:

[0014] Based on the first Raman scattering signal, a matching first sub-spectral database is selected from multiple sub-spectral databases;

[0015] Based on the first Raman scattering signal and the first sub-spectral database, the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database is calculated.

[0016] The beneficial effects of adopting the above-mentioned further scheme are: by classifying the database, the algorithm can adaptively select the database type, reduce the amount of computation, and facilitate online detection.

[0017] Furthermore, the step of selecting a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal specifically includes:

[0018] Calculate the signal-to-noise ratio of the first Raman scattering signal;

[0019] Extract the number of spectral peaks in the first Raman scattering signal;

[0020] The first sub-spectral database is selected based on the signal-to-noise ratio and the number of spectral peaks.

[0021] Furthermore, the preprocessing of the sample signal to be tested specifically includes:

[0022] The spectral baseline removal and spectral noise reduction processes are performed on the signal of the sample to be tested.

[0023] Furthermore, the extraction of spectral peaks from each spectrum in the preset spectral database specifically includes:

[0024] The spectral peaks of each spectrum in the preset spectral database are extracted using a second-order difference algorithm.

[0025] Another technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0026] A Raman spectroscopy discrimination system includes: a preprocessing module, a correlation parameter calculation module, and a type discrimination module;

[0027] The preprocessing module is used to preprocess the sample signal to be tested and obtain the first Raman scattering signal of the sample signal to be tested.

[0028] The correlation parameter calculation module is used to calculate the correlation coefficient based on the first Raman scattering signal and a preset spectral database;

[0029] The type discrimination module is used to obtain the matching spectral type of the sample to be tested based on the correlation coefficient.

[0030] The beneficial effects of this invention are: This solution calculates the correlation coefficient based on the first Raman scattering signal and a preset spectral database, and obtains the matching spectral type of the sample to be tested based on the correlation coefficient. Compared with traditional methods, the amount of calculation is greatly reduced, the algorithm running time is shortened, the algorithm consumption time is greatly reduced, and the operation is simple.

[0031] Furthermore, it also includes: a data block grouping module, used to extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases;

[0032] The correlation parameter calculation module is specifically used to select a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal.

[0033] Based on the first Raman scattering signal and the first sub-spectral database, the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database is calculated.

[0034] The beneficial effects of adopting the above-mentioned further scheme are: by classifying the database, the algorithm can adaptively select the database type, reduce the amount of computation, and facilitate online detection.

[0035] Furthermore, the correlation parameter calculation module is specifically used to calculate the signal-to-noise ratio of the first Raman scattering signal;

[0036] Extract the number of spectral peaks in the first Raman scattering signal;

[0037] The first sub-spectral database is selected based on the signal-to-noise ratio and the number of spectral peaks.

[0038] Furthermore, the preprocessing module is used to perform spectral baseline removal and spectral noise reduction on the sample signal to be tested.

[0039] Furthermore, the data block grouping module is specifically used to extract the spectral peaks of each spectrum in the preset spectral database using a second-order difference algorithm.

[0040] The advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0041] Figure 1 A schematic flowchart of a Raman spectroscopy discrimination method provided for an embodiment of the present invention;

[0042] Figure 2 A structural framework diagram of a Raman spectroscopy discrimination system provided for embodiments of the present invention;

[0043] Figure 3 Original spectra of losartan potassium provided for other embodiments of the present invention;

[0044] Figure 4 Spectral baseline removal schematic diagram provided for other embodiments of the present invention;

[0045] Figure 5 Schematic diagram of spectral denoising provided for other embodiments of the present invention;

[0046] Figure 6 A schematic diagram of database processing logic provided for other embodiments of the present invention;

[0047] Figure 7 Database design logic diagrams provided for other embodiments of the present invention;

[0048] Figure 8 A schematic diagram of database selection logic provided for other embodiments of the present invention;

[0049] Figure 9 A schematic diagram illustrating the principle of spectral peak finding for other embodiments of the present invention;

[0050] Figure 10 A schematic diagram illustrating the peak identification effect of ammonium nitrate spectrum in other embodiments of the present invention;

[0051] Figure 11 A schematic diagram of the measured spectrum of ammonium nitrate provided for other embodiments of the present invention;

[0052] Figure 12 A schematic diagram of a Raman spectrometer structure provided for other embodiments of the present invention. Detailed Implementation

[0053] The principles and features of the present invention are described below with reference to the accompanying drawings. The embodiments described are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0054] like Figure 1 As shown, an embodiment of the present invention provides a Raman spectroscopy discrimination method, comprising:

[0055] S1, preprocess the sample signal to be tested to obtain the first Raman scattering signal of the sample signal to be tested;

[0056] It should be noted that, in one embodiment, preprocessing may include spectral baseline removal and spectral denoising. There are many methods for removing baseline interference. This patent uses the Automated Baseline Estimation (ABE) method proposed by H. Georg Schulze as an example. This method uses a cyclic sliding window averaging method to smooth the spectrum, removing spectral peaks in the original spectrum that are larger than the smoothed spectrum. The spectral denoising algorithm uses wavelet denoising, employing the Daubechies wavelet function. The Daubechies wavelet was constructed by the world-renowned wavelet analyst Ingrid Daubechies and is generally abbreviated as dbN, where N is the order of the wavelet. In this patent, N is set to 4, i.e., db4.

[0057] S2, calculate the correlation coefficient based on the first Raman scattering signal and a preset spectral database;

[0058] In one embodiment, calculating the correlation coefficient may include:

[0059] The correlation coefficient used is the Pearson correlation coefficient, also known as the Pearson product-moment correlation coefficient. It is commonly used to measure the correlation between two variables X and Y. Its value ranges between -1 and 1. The closer it is to 1, the more linearly related X and Y are. It is usually represented by r, and the calculation method is as follows:

[0060]

[0061] In the formula, X and Y represent two variables, which in this patent can represent measured spectra and database spectra, respectively. i Represents the i-th element of spectrum X. Let X represent the average value of X, and n represent the dimension of the spectrum.

[0062] S3, obtain the matching spectral type of the sample to be tested based on the correlation coefficient.

[0063] This solution calculates the correlation coefficient based on the first Raman scattering signal and a preset spectral database, and obtains the matching spectral type of the sample to be tested based on the correlation coefficient. Compared with traditional methods, the computational load is greatly reduced, the algorithm running time is shortened, the algorithm consumption is significantly reduced, and the operation is simple.

[0064] Optionally, in some embodiments, it further includes:

[0065] Extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases;

[0066] The step of calculating the correlation coefficient based on the first Raman scattering signal and a preset spectral database specifically includes:

[0067] Based on the first Raman scattering signal, a matching first sub-spectral database is selected from multiple sub-spectral databases;

[0068] Based on the first Raman scattering signal and the first sub-spectral database, the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database is calculated.

[0069] In one embodiment, such as Figure 6 As shown, the main peaks are grouped at intervals d. The main peak is the spectral peak with the highest spectral intensity, and the secondary peaks are those whose spectral intensities are arranged in descending order, excluding the main peak. The value of d depends on the spectral resolution of the equipment used. Taking the CR2000 handheld Raman spectrometer from Beijing Huatai Nuoan Detection Technology Co., Ltd. as an example, the typical spectral resolution is 10 cm⁻¹. -1 Then d can be taken as 10. Taking ammonium nitrate and potassium nitrate as examples, their main peaks are located at 1048 cm⁻¹. -1 and 1050cm -1 ,according to Figure 6 The logic shown should be grouped into one group, i.e., database 1 is grouped at 10cm intervals. -1 If the sample to be tested is ammonium nitrate, the database obtained by grouping the main peaks will only calculate the correlation coefficient between the spectrum to be tested and the data spectrum in one group in Database 1, which will greatly reduce the amount of computation.

[0070] This solution categorizes databases, allowing the algorithm to adaptively select database types, reducing computational load and facilitating online detection.

[0071] Optionally, in some embodiments, the step of selecting a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal specifically includes:

[0072] Calculate the signal-to-noise ratio of the first Raman scattering signal;

[0073] Extract the number of spectral peaks in the first Raman scattering signal;

[0074] The first sub-spectral database is selected based on the signal-to-noise ratio and the number of spectral peaks.

[0075] In one embodiment, after the Raman spectrometer acquires the spectrum of the sample to be tested, the signal-to-noise ratio (SNR) of the measured spectrum is first determined, because spectral preprocessing and subsequent database selection are highly correlated with the SNR. Since there is no unified method for calculating the SNR performance of the spectrum, this patent adopts a more commonly used calculation method in the industry, as shown below:

[0076]

[0077] I peak This indicates the intensity of the main peak, such as the main peak of ammonium nitrate at 1046 cm⁻¹. -1 Strength, σ N The standard deviation of noise.

[0078] A low signal-to-noise ratio will lead to inaccurate peak identification and the presence of "false peaks". Therefore, the database conditions for peak identification need to be relaxed. If the measured spectrum has a high signal-to-noise ratio, it can be assumed that the spectrum is less affected by noise and the identified peaks are accurate. In this case, the matching database conditions can be made stricter.

[0079] The signal-to-noise ratio threshold is determined by the performance of the equipment used, and can generally be obtained by collecting a batch of data and statistically analyzing it. It can be continuously optimized during the algorithm iteration process.

[0080] This patent also allows users to actively select the database type. Users can actively choose to use database 1, database 2, etc., based on their professional knowledge or experience, or the algorithm can adaptively select the database type. Analysis of the database design principles clearly shows that while unclassified matching (full database matching) is computationally intensive and time-consuming, its accuracy is unaffected by database classification, resulting in the highest accuracy. Matching with database 1, database 2, and database 3 results in progressively decreasing computational intensity and time consumption, but the accuracy also decreases accordingly with each level.

[0081] Optionally, in some embodiments, the preprocessing of the sample signal to be tested specifically includes:

[0082] The spectral baseline removal and spectral noise reduction processes are performed on the signal of the sample to be tested.

[0083] Optionally, in some embodiments, extracting the spectral peaks of each spectrum in the preset spectral database specifically includes:

[0084] The spectral peaks of each spectrum in the preset spectral database are extracted using a second-order difference algorithm.

[0085] In one embodiment, after acquiring the spectral signal of the sample, the Raman spectrometer, taking losartan potassium as an example, uses a CR2000 handheld Raman spectrometer from Beijing Huatai Nuoan Detection Technology Co., Ltd. to acquire the Raman spectrum of the sample. The integration time is 1 second, and the laser power is set to 120mW. The obtained raw spectrum of losartan potassium is as follows: Figure 3 As shown:

[0086] A true Raman spectrum should consist of three parts: the useful Raman signal, baseline interference in the measured spectrum, and baseline interference including fluorescence and phosphorescence background from the sample itself and its container, blackbody radiation from the sample and its surrounding environment, as well as various noises, including shot noise, dark current noise and readout noise from the CCD detector, emission noise introduced by the excitation source, and cosmic rays, as shown in the following formula:

[0087] S1 = S0 + B + N,

[0088] Where S1 represents the measured Raman signal, S0 represents the true Raman signal, B represents baseline interference, and N represents the sum of various noises. Since mature algorithms already exist for spectral baseline removal and spectral denoising, and this patent focuses on the establishment of Raman spectral data, algorithms for spectral baseline removal and spectral noise removal are not analyzed in detail. This patent uses the Automated Basedline Estimation (ABE) method proposed by H. Georg Schulze to remove the spectral baseline. This method uses a cyclic sliding window averaging method to smooth the spectrum, and positions in the original spectrum larger than the smoothed spectrum are taken as spectral peaks and removed. The spectral denoising algorithm uses wavelet denoising, employing the Daubechies wavelet function. The Daubechies wavelet was constructed by the world-renowned wavelet analyst Ingrid Daubechies and is generally abbreviated as dbN, where N is the order of the wavelet. In this patent, N is set to 4, i.e., db4. A schematic diagram of using ABE to remove the spectral baseline is shown below. Figure 4 As shown in the figure. The dashed line in the figure represents the spectrum after baseline removal, and the solid line represents the spectral baseline. After baseline removal, spectral denoising is performed, and the denoised spectrum is shown below. Figure 5As shown. After excluding the baseline and noise of the original spectrum, the retained signal is the Raman scattering signal S0 of the sample to be tested. The correlation coefficient is calculated by comparing S0 with the spectra in the database (the spectra in the database consist of x and y coordinates, where x represents the wavenumber and y represents the spectral intensity; the spectra in the database are those acquired by large-scale equipment or those acquired and compared multiple times by equipment of the same level, containing the Raman scattering signal of the substance). This yields the substance in the database that is closest to the sample to be tested. The correlation coefficient used in this patent is the Pearson correlation coefficient, also known as the Pearson product-moment correlation coefficient, which is commonly used to measure the correlation between two variables X and Y. Its value ranges between -1 and 1; the closer to 1, the more linearly correlated X and Y are. It is commonly represented by r, and the calculation method is as follows:

[0089]

[0090] In the formula, X and Y represent two variables, which in this patent can represent the measured spectrum after preprocessing (S0) and the database spectrum, respectively. i Represents the i-th element of spectrum X. Let X represent the average value of X, and n represent the dimension of the spectrum.

[0091] In one embodiment, to reduce computational load and shorten algorithm execution time, the database is processed as follows:

[0092] like Figure 6 As shown, the main peaks are grouped at intervals d. The main peak is the spectral peak with the highest spectral intensity, and the secondary peaks are those whose spectral intensities are arranged in descending order, excluding the main peak. The value of d depends on the spectral resolution of the equipment used. Taking the CR2000 handheld Raman spectrometer from Beijing Huatai Nuoan Detection Technology Co., Ltd. as an example, the typical spectral resolution is 10 cm⁻¹. -1 Then d can be taken as 10. Taking ammonium nitrate and potassium nitrate as examples, their main peaks are located at 1048 cm⁻¹. -1 and 1050cm -1 ,according to Figure 6 The logic shown should be grouped into one group, i.e., database 1 is grouped at 10cm intervals. -1 If the sample to be tested is ammonium nitrate, the database obtained by grouping the main peaks will only calculate the correlation coefficient between the spectrum to be tested and the data spectrum in one group in Database 1, which will greatly reduce the amount of computation.

[0093] Similarly, sorting the peak intensities in descending order yields Database 2 and Database 3. Database 2 and Database 3 are based on the main peak grouping, further reducing the computational load. The database setup principle is as follows:

[0094] Figure 7 This patent uses a primary peak as an example to illustrate the database design principle. A is part of database 1, B is part of database 2, and C is part of database 3. It should be noted that primary peak 1, secondary peak 1_1, and tertiary peak 1_1_1 do not represent a specific location, but rather a range with a width of 10cm. -1 .

[0095] However, some substances have fewer spectral peaks, or the testing environment is poor, resulting in too few peaks being obtained. For example, potassium nitrate has very few peaks at 200 cm⁻¹. -1 ~3000cm -1 The range is only 1056cm -1 and 1353cm- 1 The two strong spectral peaks are obviously not suitable for matching calculations with data 2 and data 3. Therefore, it is necessary to comprehensively consider the signal-to-noise ratio of the measured spectrum and then determine the database category.

[0096] In one embodiment, the database selection logic is as follows: Figure 8 As shown: After the Raman spectrometer acquires the spectrum of the sample to be tested, the signal-to-noise ratio (SNR) of the measured spectrum is first determined, because spectral preprocessing and subsequent database selection are highly correlated with the SNR. Since there is no unified method for calculating the SNR performance of the spectrum, this patent adopts a more commonly used calculation method in the industry, as shown below:

[0097]

[0098] I peak This indicates the intensity of the main peak, such as the main peak of ammonium nitrate at 1046 cm⁻¹. -1 Strength, σ N The standard deviation of noise.

[0099] Spectral preprocessing includes baseline removal and spectral denoising. Many methods exist for removing baseline interference; this patent uses the Automated Basedline Estimation (ABE) method proposed by H. Georg Schulze as an example. This method uses a cyclic sliding window averaging method to smooth the spectrum, removing spectral peaks in the original spectrum that are larger than those in the smoothed spectrum. The spectral denoising algorithm employs wavelet denoising, using the Daubechies wavelet function. The Daubechies wavelet was constructed by the world-renowned wavelet analyst Ingrid Daubechies and is generally abbreviated as dbN, where N is the order of the wavelet. In this patent, N is set to 4, i.e., db4. The algorithm results are as follows: Figure 4 and Figure 5 As shown.

[0100] In one embodiment, after preprocessing the measured spectrum, the spectral peaks of the preprocessed spectrum can be extracted. Many algorithms exist for peak extraction; this patent employs a second-order difference algorithm. The principle is to find the boundary points, i.e., extrema, where the monotonicity of the spectral intensity changes. The intensity monotonicity of discrete data can be determined using the difference method, as detailed below. Figure 9 As shown, the peak identification effect of ammonium nitrate spectrum is as follows: Figure 10 As shown.

[0101] After extracting the spectral peaks, it's necessary to compare the overall signal-to-noise ratio (SNR) of the spectrum. If the SNR is too low, peak searching will identify many "pseudo-peaks." These pseudo-peaks are not Raman signals from the sample but rather spectral noise. Figure 11 The image shows the measured spectrum of ammonium nitrate. Clearly, the signal-to-noise ratio (SNR) is very low, which is caused by probe defocusing during sampling. There are two main reasons for a low SNR: first, interference from the sampling environment, including probe defocusing and strong background light interference; and second, the sample itself has low Raman activity. A low SNR will lead to inaccurate peak identification and the presence of "false peaks." Therefore, the database conditions for peak identification need to be relaxed. If the measured SNR is high, the spectrum can be considered less affected by noise, and the identified peaks are accurate. In this case, the matching database conditions can be stricter.

[0102] The signal-to-noise ratio threshold is determined by the performance of the equipment used, and can generally be obtained by collecting a batch of data and statistically analyzing it. It can be continuously optimized during the algorithm iteration process.

[0103] This patent also allows users to actively select the database type. Users can actively choose to use database 1, database 2, etc., based on their professional knowledge or experience, or the algorithm can adaptively select the database type. Analysis of the database design principles clearly shows that while unclassified matching (full database matching) is computationally intensive and time-consuming, its accuracy is unaffected by database classification, resulting in the highest accuracy. Matching with database 1, database 2, and database 3 results in progressively decreasing computational intensity and time consumption, but the accuracy also decreases accordingly with each level.

[0104] In one embodiment, such as Figure 2 As shown, a Raman spectroscopy discrimination system includes: a preprocessing module 1101, a correlation parameter calculation module 1102, and a type discrimination module 1103;

[0105] The preprocessing module 1101 is used to preprocess the sample signal to be tested to obtain the first Raman scattering signal of the sample signal to be tested.

[0106] The correlation parameter calculation module 1102 is used to calculate the correlation coefficient based on the first Raman scattering signal and a preset spectral database;

[0107] The type discrimination module 1103 is used to obtain the matching spectral type of the sample to be tested based on the correlation coefficient.

[0108] This solution calculates the correlation coefficient based on the first Raman scattering signal and a preset spectral database, and obtains the matching spectral type of the sample to be tested based on the correlation coefficient. Compared with traditional methods, the computational load is greatly reduced, the algorithm running time is shortened, the algorithm consumption is significantly reduced, and the operation is simple.

[0109] Optionally, in some embodiments, it further includes: a data block grouping module, used to extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases;

[0110] The correlation parameter calculation module is specifically used to select a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal.

[0111] Based on the first Raman scattering signal and the first sub-spectral database, the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database is calculated.

[0112] This solution categorizes databases, allowing the algorithm to adaptively select database types, reducing computational load and facilitating online detection.

[0113] Optionally, in some embodiments, the correlation parameter calculation module 1102 is specifically used to calculate the signal-to-noise ratio of the first Raman scattering signal;

[0114] Extract the number of spectral peaks in the first Raman scattering signal;

[0115] The first sub-spectral database is selected based on the signal-to-noise ratio and the number of spectral peaks.

[0116] Optionally, in some embodiments, the preprocessing module 1101 is used to perform spectral baseline removal and spectral noise reduction on the sample signal to be tested.

[0117] Optionally, in some embodiments, the data block grouping module is specifically used to extract the spectral peaks of each spectrum in the preset spectral database using a second-order difference algorithm.

[0118] It is understood that in some embodiments, some or all of the optional implementations as described in the above embodiments may be included.

[0119] It should be noted that the above embodiments are product embodiments corresponding to the prior method embodiments. For the description of each optional implementation in the product embodiments, please refer to the corresponding description in the above method embodiments, which will not be repeated here.

[0120] In some embodiments, light irradiates a material, resulting in elastic and inelastic scattering. Elastic scattering produces light with the same wavelength as the excitation light, while inelastic scattering produces light with components longer and shorter than the excitation light wavelength; this is collectively known as the Raman effect. Because the Raman scattered light is modulated by the material and carries structural information, the Raman spectrum of a material can be used to study its structure. Among numerous spectroscopic analysis techniques, Raman spectroscopy has unique advantages:

[0121] The excitation light selection offers a high degree of freedom. As is known from the Raman scattering mechanism, the frequency shift of the Raman spectrum is not limited by the frequency of the light source, allowing for the selection of different wavelengths of excitation light sources based on the characteristics of different samples.

[0122] It has a wide detection range. Since both polar and nonpolar molecules can produce Raman spectra, Raman spectroscopy can detect a wide range of substances, including inorganic substances, organic substances, polymers, minerals, animal and plant tissues, catalysts, etc.

[0123] Almost no sample preparation is required. Because Raman spectroscopy is a scattering spectroscopy, Raman spectroscopy measurements do not require sample preparation as is the case with infrared absorption spectroscopy.

[0124] It enables non-contact, non-destructive measurement. Since Raman spectroscopy does not require sample preparation, we can directly measure the analyte without damaging it. Therefore, Raman spectroscopy analysis technology can achieve completely non-contact, non-destructive testing.

[0125] It allows for direct measurement of substances in aqueous solutions and glassware. Because water and glass exhibit very low Raman scattering, it enables direct detection of solutes within glass containers or aqueous solutions, which is crucial in industrial production.

[0126] Due to its numerous advantages, Raman spectroscopy has been widely applied in all aspects of people's production and daily life. The use of Raman spectroscopy can be simply summarized in two points:

[0127] Raman spectroscopy acquisition;

[0128] Raman spectroscopy results matched;

[0129] Raman spectroscopy acquisition involves using a Raman spectrometer to collect the Raman spectral signal of a sample. A Raman spectrometer typically consists of an excitation source, a Raman probe, a spectral scattering system, and a data processing system. Figure 12 As shown.

[0130] L0 is the light source section of the Raman spectrometer, and R1 represents the laser. The laser emitted by the laser is reflected by the mirror M1 into the D0 Raman probe, then split and reflected by the high-reflection lens L1, and finally focused by the focusing lens onto the sample S0. The Raman scattering signal generated by the excitation light of S0 is collimated by L1, passes through the high-pass filter F1 to the Raman filter F2, filters out the excitation light, and allows the Raman scattering signal to pass through. The Raman scattering signal is converged by L2 to the D2 spectral scattering and data processing system, and then coupled to the concave mirror M2 through the spatial filter of the slit S. The spectral signal is collimated by M2 and illuminated by the grating G. The function of the grating G is to disperse the spectral signal according to different wavelengths. After dispersion by G, the spectral signals of different wavelengths are finally converged by the imaging mirror M3 onto the sensor D. The sensor D is essentially a photoelectric conversion element such as a charge-coupled device (CCD) or a complementary metal-oxide-semiconductor (CMOS), which is responsible for converting the optical signal into an electrical signal. The data processing system then performs data analysis on the electrical signal.

[0131] After acquiring the Raman scattering signal of the sample using a Raman spectrometer, the acquired spectral signal needs to be matched. Since Raman signals are fingerprint-like and unique, a match can be found by comparing them one-to-one with database files. Assuming there are 100 database files, simply calculating the similarity between the acquired spectral signal and each of these 100 files will find the database file that best matches the sample, thus completing the Raman spectral result matching.

[0132] However, generally speaking, a Raman spectrum data dimension is 2048 data points, and the size of commonly used database files is more than 2000. Some large databases have more than 10,000 spectral data points. If the similarity between the sampled signal and the database spectrum is calculated one by one, the amount of computation is very large and the computation time is very long, which is very unfavorable for online detection.

[0133] Therefore, this invention provides a fast database matching method with low computational load, short computation time, and simple operation.

[0134] Readers should understand that in the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0135] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the method embodiments described above are merely illustrative. For instance, the division of steps is only a logical functional division, and there may be other division methods in actual implementation. For example, multiple steps may be combined or integrated into another step, or some features may be ignored or not executed.

[0136] If the above methods are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0137] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A Raman spectroscopy discrimination method, characterized in that, include: The sample signal to be tested is preprocessed to obtain the first Raman scattering signal of the sample signal to be tested. The correlation coefficient is calculated based on the first Raman scattering signal and a preset spectral database; The matching spectral type of the sample to be tested is obtained based on the correlation coefficient. Also includes: Extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases; The step of calculating the correlation coefficient based on the first Raman scattering signal and a preset spectral database specifically includes: Based on the first Raman scattering signal, a matching first sub-spectral database is selected from multiple sub-spectral databases; Based on the first Raman scattering signal and the first sub-spectral database, calculate the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database; The step of selecting a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal specifically includes: Calculate the signal-to-noise ratio of the first Raman scattering signal; Extract the number of spectral peaks in the first Raman scattering signal; The first sub-spectral database is selected based on the signal-to-noise ratio and the number of spectral peaks.

2. The Raman spectroscopy discrimination method according to claim 1, characterized in that, The preprocessing of the sample signal to be tested specifically includes: The spectral baseline removal and spectral noise reduction processes are performed on the signal of the sample to be tested.

3. The Raman spectroscopy discrimination method according to claim 1, characterized in that, The extraction of spectral peaks for each spectrum from the preset spectral database specifically includes: The spectral peaks of each spectrum in the preset spectral database are extracted using a second-order difference algorithm.

4. A Raman spectroscopy discrimination system, characterized in that, include: The module consists of a preprocessing module, a correlation parameter calculation module, and a type discrimination module. The preprocessing module is used to preprocess the sample signal to be tested and obtain the first Raman scattering signal of the sample signal to be tested. The correlation parameter calculation module is used to calculate the correlation coefficient based on the first Raman scattering signal and a preset spectral database; The type discrimination module is used to obtain the matching spectral type of the sample to be tested based on the correlation coefficient; It also includes: a data block grouping module, used to extract the spectral peaks of each spectrum in the preset spectral database, and group the preset spectral database according to the spectral peaks of each spectrum to obtain multiple sub-spectral databases; The correlation parameter calculation module is specifically used to select a matching first sub-spectral database from multiple sub-spectral databases based on the first Raman scattering signal; and to calculate the correlation coefficient between the first Raman scattering signal and each spectrum in the first sub-spectral database based on the first Raman scattering signal and the first sub-spectral database. The correlation parameter calculation module is specifically used to calculate the signal-to-noise ratio of the first Raman scattering signal; extract the number of spectral peaks of the first Raman scattering signal; and select the first sub-spectral database based on the signal-to-noise ratio and the number of spectral peaks.

5. The Raman spectroscopy discrimination system according to claim 4, characterized in that, The preprocessing module is used to perform spectral baseline removal and spectral noise reduction on the sample signal to be tested.

6. The Raman spectroscopy discrimination system according to claim 4, characterized in that, The data block grouping module is specifically used to extract the spectral peaks of each spectrum in the preset spectral database using a second-order difference algorithm.

Images