Laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration

Abstract

The present invention proposes a laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration, including step 1, adjusting the laser intensity until the intensity value of the water Raman peak of the spectrum curve of the tested liquor is consistent with the set value; step 2, translating the spectrum curve along the horizontal axis, so that the wavelength position corresponding to the currently measured water Raman peak is moved to the theoretical wavelength position corresponding to the Raman peak generated by laser-excited water; step 3, calculating the correlation between the spectrum curve obtained in step 2 and the spectrum curve stored in the database, and taking the spectrum curve with the highest correlation with the spectrum curve obtained in step 2 in the database as a backup match; step 4, judging whether the highest correlation is higher than the set matching threshold; if so, displaying it and giving the liquor name corresponding to the backup match, and determining that the tested liquor is true; otherwise, outputting no match. The above method can alleviate the problem of detection error caused by differences in different instruments, external environmental interference and control errors.

Description

Technical Field

[0001] The invention relates to the technical field of wine authenticity detection, and in particular to a laser spectrum liquor detection method based on water Raman peak intensity and wavelength calibration. Background Art

[0002] With the development of the wine industry, wine testing research has always been an important research direction, from traditional physical and chemical testing, to the combination of physical and chemical testing and modern analytical technology, to spectral analysis and combined technology such as spectroscopy and chromatography. The instruments used in spectral technology testing are easy to operate, fast in detection speed, and will not pollute the environment during the detection process. The detection accuracy is high, and it is widely used in food safety testing, especially suitable for the detection and analysis of liquids such as wine.

[0003] In the actual application scenarios of liquor detection based on spectral technology, differences between different instruments, external environmental interference and control errors will cause a certain degree of fluctuation in the obtained spectral curve, which will cause errors in the test results; and because the stability of portable equipment is inferior to that of non-portable high-precision instruments in the laboratory, the above problems are particularly serious in the portable liquor detection scenario. Summary of the invention

[0004] In order to solve the problems existing in the prior art, the present application proposes a laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration, so as to greatly alleviate the detection error problems caused by differences between different instruments, external environmental interference and control errors, and to improve the spectral matching accuracy in big data scenarios.

[0005] In order to achieve the above purpose, the present application proposes a laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration, comprising the following steps:

[0006] Step 1, water Raman peak intensity calibration: within the measurement range of the spectrometer used, set an intensity value for the water Raman peak, record it as the set value, and adjust the laser intensity until the intensity value of the water Raman peak of the spectrum curve of the measured liquor is consistent with the set value;

[0007] Step 2, water Raman peak wavelength calibration: the spectrum curve obtained in step 1 is translated along the horizontal axis, so that the wavelength position corresponding to the currently measured water Raman peak is moved to the theoretical wavelength position corresponding to the Raman peak generated by laser excitation water;

[0008] Step 3, calculating the correlation between the spectral curve obtained in step 2 and the spectral curve stored in the database using a curve matching algorithm, and taking the spectral curve in the database with the highest correlation with the spectral curve obtained in step 2 as a backup matching item;

[0009] Step 4: determine whether the highest correlation calculated in step 3 is higher than the set matching threshold;

[0010] Step 5: When the correlation is higher than the set matching threshold, it is displayed in percentage form and the name of the liquor corresponding to the backup matching item is given. At this time, the tested liquor is considered to be true;

[0011] Step 6: When the correlation is not higher than the set matching threshold, output no matching item.

[0012] In some embodiments, in step 3, the curve matching algorithm adopts a relative offset distance algorithm.

[0013] In some embodiments, the process of calculating the correlation between the spectral curve obtained in step 2 and the spectral curve in the database using the relative offset distance algorithm is as follows:

[0014] The relative distance dis(B, A) between the spectral curve obtained in step 2 and the spectral curve in the database is calculated by formula (1):

[0015]

[0016] Where A represents the spectral curve in the database, B represents the spectral curve obtained in step 2, l represents the horizontal pixel number of the spectrometer detector, and A i represents the intensity value of the spectral curve in the database at pixel i, B i represents the intensity value of the spectral curve obtained in step 2 at pixel i;

[0017] Formula (2) is used to calculate the correlation score (B, A) between the spectral curve obtained in step 2 and the spectral curve in the database:

[0018]

[0019] The beneficial effect of the scheme of the present application is that the above-mentioned laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration can greatly alleviate the problem of detection errors caused by differences between different instruments, external environmental interference and control errors, and can improve the accuracy of spectral matching in big data scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 A typical laser spectrum curve of liquor is shown.

[0021] Figure 2 A flow chart of a laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration in an embodiment is shown.

[0022] Figure 3The sample matching error bar graphs in the embodiments are shown, wherein (a) is a sample matching error bar graph obtained by using the Pearson correlation coefficient matching algorithm, and (b) is a sample matching error bar graph obtained by using the method involved in the present application. DETAILED DESCRIPTION

[0023] The specific implementation of the present application is further described below in conjunction with the accompanying drawings.

[0024] Laser has the advantages of high intensity, high brightness, good monochromaticity, etc. When incident on liquor, it can stimulate water, ethanol and soluble molecules in liquor to produce detectable laser spectrum signals such as Rayleigh scattering, Mie scattering, Raman scattering, fluorescence scattering, etc. Since the spectrum of the laser excitation is different for different liquor ingredients, if the composition or content of the liquor changes, it will generally produce obvious differences in the spectrum, which provides a relatively objective and reliable basis for the authenticity identification and quality judgment of liquor. Figure 1 This is a typical spectral curve produced by the excitation of laser incident on liquor, which includes the laser scattering peak, the Raman scattering peak of ethanol, the Raman scattering peak of water, and the fluorescence spectrum of liquor.

[0025] like Figure 2 As shown, the laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration involved in this application includes the following steps:

[0026] Step 1, water Raman peak intensity calibration: within the measurement range of the spectrometer used, set an intensity value for the water Raman peak, record it as the set value, and adjust the laser intensity until the intensity value of the water Raman peak of the spectral curve of the measured liquor is consistent with the set value.

[0027] Step 2, water Raman peak wavelength calibration: the spectrum curve obtained in step 1 is translated along the horizontal axis to move the wavelength position corresponding to the currently measured water Raman peak to the theoretical wavelength position corresponding to the Raman peak generated by laser-excited water.

[0028] Step 3: Calculate the correlation between the spectral curve obtained in step 2 and the spectral curves stored in the database using a curve matching algorithm, and take the spectral curve in the database with the highest correlation with the spectral curve obtained in step 2 as a backup matching item.

[0029] Step 4: Determine whether the highest correlation calculated in step 3 is higher than the set matching threshold.

[0030] Step 5: When the correlation is higher than the set matching threshold, it is displayed in percentage form and the name of the liquor corresponding to the alternative matching item is given. At this time, the tested liquor is considered to be true.

[0031] Step 6: When the correlation is not higher than the set matching threshold, output no matching item.

[0032] The above-mentioned laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration can greatly alleviate the problem of detection errors caused by differences in different instruments, external environmental interference and control errors, and can improve the accuracy of spectral matching in big data scenarios.

[0033] In step 3, the curve matching algorithm may be a relative offset distance algorithm, a Euclidean distance algorithm, a pattern distance algorithm, a shape distance algorithm, or a DTW (dynamic time normalization) algorithm. Since the spectral curve of the measured liquor has been calibrated for water Raman peak intensity and wavelength, there is no deviation in the horizontal coordinate direction (wavelength) between the calibrated spectral curve and the spectral curve in the database, and the spectral intensity is also adjusted to the same horizontal range. Therefore, the relative offset distance between the curves can most intuitively reflect the difference between the curves, so in this embodiment, the relative offset distance algorithm is preferably used as the curve matching algorithm, which has the advantages of low computational complexity and fast matching speed, and can improve the overall matching efficiency of the system.

[0034] Specifically, the process of calculating the correlation between the spectral curve obtained in step 2 and the spectral curve in the database using the relative offset distance algorithm is as follows:

[0035] First, the relative distance dis(B, A) between the spectral curve obtained in step 2 and the spectral curve in the database is calculated by formula (1):

[0036]

[0037] Where A represents the spectral curve in the database, B represents the spectral curve obtained in step 2, l represents the horizontal pixel number of the spectrometer detector, and A i represents the intensity value of the spectral curve in the database at pixel i, B i represents the intensity value of the spectral curve obtained in step 2 at pixel i;

[0038] Then, the correlation score (B, A) between the spectral curve obtained in step 2 and the spectral curve in the database is calculated using formula (2):

[0039]

[0040] The laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration involved in the present application is compared with the Pearson correlation coefficient matching algorithm to verify the superiority of the method involved in the present application.

[0041] The names and numbers of the tested liquors (samples) are: Beijing Erguotou (#1), Weihaiwei Laoshao (#2), Dingxiangqing (#3), Yuanjiang (#4), Beida Cang (#5), Kujie (#6), Jingzhi Baigan (#7), Hengshui Laobaigan (#8), Laojiuhu (#9), Niulanshan (#10), Duerjiu (#11) and Xiaolangjiu (#12). Table 1 shows the matching results of the Pearson correlation coefficient matching algorithm, and Table 2 shows the matching results of the method involved in this application.

[0042] Table 1 Matching results of Pearson correlation coefficient matching algorithm (%)

[0043]

[0044] Table 2 Matching results of the methods involved in this application (%)

[0045]

[0046] By comparing the two tables, it can be seen that although the method involved in this application obtains a lower matching degree of samples with the same substance content than the Pearson correlation coefficient matching algorithm, the method involved in this application can ensure that the matching degree of samples with the same substance content is always ranked first, and the matching degree is more clearly distinguished from the matching degree of different samples. Therefore, the method involved in this application has significantly improved the ability to distinguish different substances, so that it can better distinguish true and false samples, which is more in line with actual application scenarios.

[0047] The most significant advantage of the method involved in this application is that it can greatly alleviate the matching instability problem caused by external environmental interference and instrument measurement errors. In order to further verify this advantage of the method proposed in this application, a matching stability experiment was conducted. Specifically, for each wine sample, ten test samples were set, and each sample was measured ten times. The matching error bar graphs were as follows: Figure 3 (a) and Figure 3 (b) is shown. The error bar can reflect the uncertainty of the measured data. The error bar is a line segment drawn in the direction of the measured value with the arithmetic mean of the measured value as the midpoint. Half of the length of the line segment is equal to the (standard or extended) uncertainty. By comparison, it can be found that the method proposed in this application has been significantly improved in stability, so that it can be better applied to actual detection scenarios.

[0048] The laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration involved in the present application can greatly alleviate the problem of detection errors caused by differences in different instruments, external environmental interference and control errors, and can improve the accuracy of spectral matching in big data scenarios; it can also greatly improve the stability of matching, so that different samples have better distinction.

[0049] The above is only a preferred specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent substitutions or changes according to the technical solution and concept of the present application within the technical scope disclosed in the present application, which should be covered by the protection scope of the present application.

Claims

1. A laser spectroscopy liquor detection method based on water Raman peak intensity and wavelength calibration, characterized in that: The following steps are involved: Step 1, water Raman peak intensity calibration: within the measurement range of the spectrometer used, set an intensity value for the water Raman peak, record it as the set value, and adjust the laser intensity until the intensity value of the water Raman peak of the spectrum curve of the measured liquor is consistent with the set value; Step 2, water Raman peak wavelength calibration: the spectrum curve obtained in step 1 is translated along the horizontal axis, so that the wavelength position corresponding to the currently measured water Raman peak is moved to the theoretical wavelength position corresponding to the Raman peak generated by laser excitation water; Step 3: Calculate the correlation between the spectral curve obtained in step 2 and the spectral curve stored in the database using a curve matching algorithm, and take the spectral curve in the database with the highest correlation with the spectral curve obtained in step 2 as a backup matching item; wherein the curve matching algorithm adopts a relative offset distance algorithm; the process of calculating the correlation between the spectral curve obtained in step 2 and the spectral curve in the database using the relative offset distance algorithm is as follows: The relative distance dis(B, A) between the spectral curve obtained in step 2 and the spectral curve in the database is calculated by formula (1): Where A represents the spectral curve in the database, B represents the spectral curve obtained in step 2, l represents the number of horizontal pixels of the spectrometer detector, and A i represents the intensity value of the spectral curve in the database at pixel i, B i represents the intensity value of the spectral curve obtained in step 2 at pixel i; Formula (2) is used to calculate the correlation score (B, A) between the spectral curve obtained in step 2 and the spectral curve in the database: Step 4: determine whether the highest correlation calculated in step 3 is higher than the set matching threshold; Step 5: When the correlation is higher than the set matching threshold, it is displayed in percentage form and the name of the liquor corresponding to the backup matching item is given. At this time, the tested liquor is considered to be true; Step 6: When the correlation is not higher than the set matching threshold, output no matching item.

Images