KR102276065B1 - Method and apparatus for robust raman spectrum discrimination against measurement environmental conditions - Google Patents

Abstract

An embodiment of the present invention provides a method and apparatus for increasing chemical substance discrimination performance through efficient analysis of a Raman spectrum. According to an embodiment of the present invention, a method for determining a Raman spectrum robust to a measurement environment condition includes: i) measuring a Raman spectrum for each chemical sample in advance, collecting a library spectrum, and forming a library database; ii) collecting an input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one chemical sample; iii) adjusting the resolution of the input spectrum and removing noise; iv) determining a peak cluster region of the input spectrum by scanning the input spectrum using a moving window and measuring a peak value of the input spectrum; v) dividing the peak cluster region of the input spectrum into at least one subspectrum; vi) analyzing a degree of similarity between each of the partial spectra and the library spectrum; and vii) deriving a chemical substance included in the measurement target.

Description

A method and apparatus for determining a Raman spectrum that is robust to measurement environmental conditions {METHOD AND APPARATUS FOR ROBUST RAMAN SPECTRUM DISCRIMINATION AGAINST MEASUREMENT ENVIRONMENTAL CONDITIONS}

The present invention relates to a method and apparatus for determining a Raman spectrum that is robust to measurement environmental conditions, and more particularly, to a method and apparatus for increasing chemical substance identification performance through efficient analysis of a Raman spectrum.

When a material is irradiated with short-wavelength visible light or ultraviolet light, it causes a change in polarization during molecular vibration, so incident light is scattered by changing the wavelength, and a Raman spectrum is obtained by displaying the scattering intensity against the change in wavelength. Spectra are used in the analysis of chemicals because they are material-specific.

As a calibration method of the Raman spectrum, there is a spectrum calibration method using SRM. This method assumes that a transfer function that causes spectral differences between Raman instruments exists, and after obtaining it, performs spectrum intensity correction, and once the transfer function is obtained, it can be determined using a simple algorithm such as a correlation coefficient. . However, there is a problem in that an additional process to obtain the transfer function is required, and it is difficult to respond to other environmental variables such as the measured temperature.

As a method for discriminating the Raman spectrum, there is the SHQI method. Since this method divides the entire spectral region and determines the degree of correlation in each partial spectral region, it is a discrimination algorithm based on the intensity difference characteristic, and has the advantage of high classification efficiency of the Raman spectrum. However, since the peak distribution characteristic is not reflected, there is a disadvantage that the correlation value is uneven, and there is a problem that a value lower than 0.95, which is a general critical correlation, often occurs.

And, in the case of the existing algorithm, when selecting a partial region from the entire spectral region, a fixed window parameter is used, so there is a problem in that the peak distribution characteristic of the spectrum cannot be modeled.

In Korean Patent Laid-Open No. 10-2018-0127939 (Title of the Invention: Quantitative Molecular Sensing Apparatus and Method Using Raman Peak Point Variation), an illumination optical system including a light source irradiating excitation light to a subject; a detection optical system including a photodetector for detecting light scattered from the subject; and a signal processor for analyzing physical properties of an object by using an output signal from a detection optical system. A quantitative molecular sensing device using a Raman peak point variation is disclosed.

Republic of Korea Patent Publication No. 10-2018-0127939

An object of the present invention for solving the above problems is to increase the chemical substance discrimination performance through efficient analysis of the Raman spectrum.

In addition, it is an object of the present invention to maintain chemical substance discrimination performance even when a difference in peak intensity of a Raman spectrum occurs depending on ambient temperature conditions and the like.

The technical problems to be achieved by the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned can be clearly understood by those of ordinary skill in the art to which the present invention belongs from the description below. There will be.

The configuration of the present invention for achieving the above object includes the steps of: i) collecting a library spectrum by measuring a Raman spectrum for each chemical sample in advance, and forming a library database; ii) collecting an input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one chemical sample; iii) adjusting the resolution of the input spectrum and removing noise; iv) determining a peak cluster region of the input spectrum by scanning the input spectrum using a moving window and measuring a peak value of the input spectrum; v) dividing the peak cluster region of the input spectrum into at least one subspectrum; vi) analyzing a degree of similarity between each of the partial spectra and the library spectrum; and vii) deriving a chemical substance included in the measurement target.

In an embodiment of the present invention, in step v), the partial spectrum may be a spectrum obtained by dividing a peak cluster region of the input spectrum by a preset threshold value.

In an embodiment of the present invention, in step vi), a degree of similarity between the partial spectrum and the library spectrum is derived by performing a cross-correlation diagram to obtain a correlation using peak area information of each of the partial spectrum and the library spectrum. can do.

In an embodiment of the present invention, in step vi), the partial correlation is derived after normalizing the i-th partial spectrum, which is any one of the plurality of (n) partial spectra, and the partial correlation is calculated can be used to derive the overall correlation.

In an embodiment of the present invention, the total correlations for the plurality of partial spectra may be an average value of the partial correlations by the total number of the partial spectra.

In an embodiment of the present invention, the overall correlation for a part of the library spectrum may be an average value of the partial correlation by the number of divisions of the part of the library spectrum.

In an embodiment of the present invention, the final correlation may be the sum of the total correlations for the plurality of partial spectra and the total correlations with respect to some of the library spectra.

In an embodiment of the present invention, in step vii), the library spectrum obtained with the greatest final correlation with the input spectrum may be determined as a Raman spectrum for the chemical substance.

According to an aspect of the present invention, there is provided a signal acquisition unit configured to collect an input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one or more chemical samples; It is connected to the signal acquisition unit, adjusts the resolution of the input spectrum and removes noise, scans the input spectrum using a moving window, and measures the peak value of the input spectrum to determine the peak clustering area of the input spectrum signal processing unit; a signal discrimination unit connected to the signal processing unit, dividing a peak cluster region of the input spectrum into at least one partial spectrum, and analyzing a degree of similarity between each partial spectrum and the library spectrum; and an output unit connected to the signal discrimination unit and displaying the chemical substance included in the measurement target on a display screen.

The effect of the present invention according to the above configuration is that discrimination performance can be strengthened through more detailed and efficient correlation evaluation through adaptive window modeling for modeling the peak distribution of the Raman spectrum.

In addition, an effect of the present invention is that by discriminating a chemical substance through correlation analysis between a partial spectrum of the measured Raman spectrum and a library spectrum, chemical substance discrimination efficiency can be improved even under conditions such as temperature change in the surrounding environment.

It should be understood that the effects of the present invention are not limited to the above-described effects, and include all effects that can be inferred from the configuration of the invention described in the detailed description or claims of the present invention.

1 is a graph of an input spectrum according to an embodiment of the present invention.
2 is a graph illustrating a process of forming a partial spectrum according to an embodiment of the present invention.
3 is a graph for comparing a library spectrum and an input spectrum according to an embodiment of the present invention.
4 is a graph for comparing a library spectrum and a partial spectrum according to an embodiment of the present invention.
5 is a block diagram of an apparatus for determining a Raman spectrum according to an embodiment of the present invention.
6 is a graph showing correlation distribution when a plurality of chemical substance discrimination methods including the Raman spectrum discrimination method of the present invention are performed.

Hereinafter, the present invention will be described with reference to the accompanying drawings. However, the present invention may be embodied in several different forms, and thus is not limited to the embodiments described herein. And in order to clearly explain the present invention in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout the specification, when a part is said to be “connected (connected, contacted, coupled)” with another part, it is not only “directly connected” but also “indirectly connected” with another member interposed therebetween. "Including cases where In addition, when a part "includes" a certain component, this means that other components may be further provided without excluding other components unless otherwise stated.

The terms used herein are used only to describe specific embodiments, and are not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In this specification, terms such as "comprises" or "have" are intended to designate that the features, numbers, steps, operations, components, parts, or combinations thereof described in the specification exist, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Hereinafter, the Raman spectrum discrimination method of the present invention will be described. 1 is a graph of an input spectrum according to an embodiment of the present invention, and FIG. 2 is a graph illustrating a process of forming a partial spectrum according to an embodiment of the present invention. A moving box in FIG. 1 may represent a moving window.

In the first step, by measuring the Raman spectrum for each chemical sample in advance, the library spectrum can be collected and a library database can be formed. Here, first, after adjusting the resolution of the Raman spectrum of each chemical sample and removing noise, the spectrum of the chemical sample is divided by scanning the Raman spectrum of the chemical sample using a moving window and measuring the peak value of the spectrum of the chemical sample After selection, the library spectrum can be obtained. This matter is also used for processing the input spectrum, and specific details will be described in detail in the following related matters. Then, the library spectrum can be converted into data to form a library database.

In the second step, Raman spectroscopy may be performed on a measurement target including at least one chemical sample to collect an input spectrum including at least one Raman spectrum. The second step can be performed by conventional Raman spectroscopy.

In the third step, we can adjust the resolution of the input spectrum and remove the noise. Here, it is possible to adjust the resolution by performing resampling on the input spectrum, and to remove the addition noise and background noise for the input spectrum. As this is a prior art, a detailed description thereof will be omitted.

In the fourth step, as shown in FIG. 1 , a peak cluster region of the input spectrum may be determined by scanning the input spectrum using a moving window and measuring the peak value of the input spectrum.

First, a threshold value is set, and a meaningful threshold value setting may be necessary because the distribution of the peak value cluster of the input spectrum must be known. Specifically, the largest peak in the input spectrum may be divided into equal parts, and the threshold value may be determined as the largest value among noises. However, the present invention is not limited thereto. In addition, the length of the moving window and the moving length of the moving window may be arbitrarily determined.

When searching using the moving window, if the largest peak value within the moving window is greater than the set threshold value, the corresponding region is regarded as a peak region, and if it is smaller than the threshold value, the region may be determined as an area without a peak.

By performing the same process as the fourth step, it is possible to determine a region in which peaks are clustered in the input spectrum.

In the fifth step, as shown in FIG. 2 , the peak cluster region of the input spectrum may be divided into at least one partial spectrum. Here, the partial spectrum may be a spectrum obtained by dividing the peak cluster region of the input spectrum by a preset threshold value.

In the fourth step, when the peak clustering region in the input spectrum is determined, partial spectra may be collected by dividing the peak clustering region in the input spectrum by a critical dimension that is an arbitrarily preset region size. Here, division may not be performed on an area having a size less than or equal to the critical dimension, but division may be performed on an area having a size greater than the critical dimension. Specifically, as shown in the division by the boxes in (A) and (B) of FIG. 2, in the region of (A) of FIG. 2, the length of the peak cluster region in the input spectrum exceeds the length of the critical dimension, so that the critical dimension is set. Segmentation is performed based on the reference, and in the region (B) of FIG. 2 , the length of the peak cluster region in the input spectrum is formed to be less than or equal to the length of the critical dimension, so the division based on the critical dimension may not be performed.

In the sixth step, the similarity between each partial spectrum and the library spectrum may be analyzed. Hereinafter, the similarity analysis process of both spectra will be described.

In the sixth step, cross-correlation may be performed to obtain a correlation using peak area information of each of the partial spectrum and the library spectrum.

First, a partial correlation may be derived after normalizing an i-th partial spectrum, which is a partial spectrum of any one of a plurality of (n) partial spectra, and an overall correlation may be derived using the partial correlation diagram.

The partial correlation of the i-th partial spectrum may be derived by [Equation 1] below.

[Formula 1]

Here, S _i is the i-th partial spectral partial correlation value, and x _i and y _i are data obtained by normalizing the i-th partial spectral region.

), 동일한 2개의 데이터끼리의 내적을 구하면 1이되므로, 상관도의 최대값은 1이고 최소값은 0일 수 있다. 즉, 상관도가 1이면 완전 상관을 의미할 수 있다.)(In [Equation 1], '·' means dot product, so in the case of a normalized vector (for example, if vector a=(a1, a2) is normalized, normr(a)=

), the dot product between two identical data becomes 1, so the maximum value of the correlation may be 1 and the minimum value may be 0. That is, if the correlation is 1, it may mean perfect correlation.)

In addition, the total degree of correlation for a plurality of partial spectra may be an average value of partial correlations by the total number of partial spectra. Accordingly, the overall correlation for a plurality of partial spectra can be derived by [Equation 2].

[Formula 2]

Here, S _i is the i-th subspectral partial correlation value, and n is the number of partial spectra.

In addition, the overall correlation for a part of the library spectrum may be an average value of the partial correlation by the number of divisions of the part of the library spectrum. Accordingly, the overall correlation for a plurality of library spectra can be derived by [Equation 3].

[Equation 3]

Here, S _i is the i-th subspectral partial correlation value, and m is the number of library spectra.

The final correlation may be the sum of the total correlations for a plurality of partial spectra and the total correlations with respect to some of the library spectra, as shown in Equation 4 below.

[Equation 4]

Hereinafter, for a visual understanding of the similarity analysis, the library spectrum of acetonitrile and the process of determining acetonitrile measured by inVia-785.0 nm equipment will be described.

3 is a graph for comparing a library spectrum and an input spectrum according to an embodiment of the present invention, and FIG. 4 is a graph for comparing a library spectrum and a partial spectrum according to an embodiment of the present invention.

In graph (1) of FIG. 3 , graph a may represent a library spectrum, and graph b may represent an input spectrum. And, in the (2) graph of FIG. 3, the (a) to (d) graphs are graphs comparing the library spectrum and the input spectrum in each partial region of the (a) to (d) region in the (1) graph of FIG. can be

In the graph (1) of FIG. 4 , graph a may represent a library spectrum, and graph b may represent a partial spectrum. And, in the (2) graph of FIG. 4, the (a) to (d) graphs are graphs comparing the library spectrum and the partial spectrum in each partial region of the (a) to (d) region in the (1) graph of FIG. can be

As shown in FIG. 3 , it can be confirmed that the library spectrum and the input spectrum have a low correlation value because the size of the peak area is significantly different from each other. On the other hand, there is a region where there is some discrepancy between the library spectrum and the partial spectrum, but it can be seen that the library spectrum and the partial spectrum have similar peak areas to obtain a partial correlation of 0.9 or more, and the other partial spectrum is also the same Since a relatively high correlation is obtained through the process, it can be confirmed that the overall correlation between the library spectrum and the partial spectrum is higher than the overall correlation between the library spectrum and the input spectrum.

In the seventh step, a chemical substance included in the measurement target can be derived. Here, the library spectrum obtained with the greatest final correlation with the input spectrum may be determined as a Raman spectrum for a chemical substance. And, such information may be displayed on the display screen.

Hereinafter, the Raman spectrum discriminating apparatus of the present invention will be described. 5 is a block diagram of an apparatus for determining a Raman spectrum according to an embodiment of the present invention.

As shown in FIG. 5, the Raman spectrum discriminating device of the present invention includes a signal acquisition unit that collects an input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one or more chemical samples ( 100); A signal that is connected to the signal acquisition unit 100, adjusts the resolution of the input spectrum and removes noise, scans the input spectrum using a moving window, and measures the peak value of the input spectrum to determine the peak clustering area of the input spectrum processing unit 200; a signal discrimination unit 300 connected to the signal processing unit 200, dividing a peak cluster region of an input spectrum into at least one partial spectrum, and analyzing a similarity between each partial spectrum and a library spectrum; and an output unit 400 connected to the signal discrimination unit 300 and deriving a chemical substance included in the measurement target and displaying it on the display screen. In addition, the signal discriminating unit 300 may be connected to the data unit 500 storing the library database, so that data on the library Raman spectrum may be transmitted from the data unit 500 to the signal discriminating unit 300 .

Here, the signal acquisition unit 100 may be a Raman spectrometer. And, the detailed description of each configuration is the same as the description of the corresponding matter in the Raman spectrum discrimination method of the present invention described above.

Hereinafter, an experiment using the Raman spectrum discrimination method of the present invention will be described.

The experiment consisted of two parts. First, a test data set was prepared that measured 10 chemical substances at three different instruments (Renishaw 2000 514.5 nm, inVia 632.8 nm, and inVia 785.0 nm), and the main experiment for recognizing them in the library spectrum and the present invention In order to validate the Raman spectrum discrimination method, a pre-experiment to recognize each other in the test data set was performed, [Table 1] below is the mechanical specification of the experimental equipment.

Method Rank Results for Test data set from Instrument 1 and Instrument 2 Mean score Segmental HQI One 0.977 0.946 0.931 0.986 0.960 0.951 0.977 0.974 0.965 0.963 0.963 2 0.524 0.597 0.431 0.593 0.625 0.806 0.777 0.726 0.312 0.589 0.598 Adaptive HQI One 0.978 0.973 0.963 0.988 0.977 0.955 0.968 0.978 0.981 0.970 0.970 2 0.458 0.502 0.496 0.553 0.576 0.634 0.595 0.583 0.519 0.585 0.549

Here, the 10 chemical substances are Acetone, Acetonitrile, Benzene, Cyclohexane, Ethyl Alcohol, Ethylene Glycol, Hexane, Nitromethane, 2-Nitrotoluene and Toluene. And, in order to perform the Raman spectrum discrimination method (Adaptive HQI, AHQI) of the present invention, the threshold value is determined to be the largest value among noises, the moving window length value is set to 100, and the moving length value of the moving window is The experiment was performed by setting it to 50 (half the value of the moving window length).

(1) Validity test between experimental data

First, validation was performed within the set based on experimental data measuring 10 types of substances in 3 types of instruments, and the final correlation results of the substances that ranked 1st and 2nd were shown in [Table 2] to [Table 4]. Hereinafter, Adaptive HQI may refer to the Raman spectrum discrimination method of the present invention, and Segmental HQI may refer to a conventional chemical substance discrimination method.

Here, [Table 2] is the experimental data for the correlation experiment with respect to the spectrum measured in Instruments 1 and 2, and [Table 3] is the experimental data for the correlation experiment for the spectrum measured in Instruments 1 and 3, [Table 4] is the experimental data for the correlation experiment for the spectrum measured in Instruments 2 and 3.

Method Rank Results for Test data set from Instrument 1 and Instrument 2 Mean score Segmental HQI One 0.977 0.946 0.931 0.986 0.960 0.951 0.977 0.974 0.965 0.963 0.963 2 0.524 0.597 0.431 0.593 0.625 0.806 0.777 0.726 0.312 0.589 0.598 Adaptive HQI One 0.978 0.973 0.963 0.988 0.977 0.955 0.968 0.978 0.981 0.970 0.970 2 0.458 0.502 0.496 0.553 0.576 0.634 0.595 0.583 0.519 0.585 0.549

Method Rank Results for Test data set from Instrument 1 and Instrument 3 Mean score Segmental HQI One 0.830 0.901 0.853 0.887 0.924 0.937 0.964 0.954 0.927 0.880 0.906 2 0.594 0.518 0.447 0.443 0.665 0.824 0.763 0.730 0.353 0.570 0.591 Adaptive HQI One 0.957 0.970 0.977 0.989 0.976 0.953 0.976 0.967 0.984 0.952 0.973 2 0.492 0.501 0.505 0.433 0.564 0.614 0.616 0.600 0.464 0.706 0.550

Method Rank Results for Test data set from Instrument 2 and Instrument 3 Mean score Segmental HQI One 0.861 0.920 0.882 0.911 0.957 0.970 0.986 0.979 0.940 0.908 0.931 2 0.610 0.535 0.333 0.431 0.659 0.792 0.735 0.696 0.332 0.574 0.570 Adaptive HQI One 0.973 0.962 0.955 0.994 0.991 0.988 0.987 0.987 0.979 0.980 0.980 2 0.516 0.494 0.453 0.434 0.522 0.599 0.609 0.604 0.458 0.651 0.534

Experimental results showed 100% discrimination results in both the conventional chemical substance discrimination method (Segmental HQI, SHQI) and the Raman spectrum discrimination method of the present invention (Adaptive HQI, AHQI). However, in the case of the conventional chemical substance identification method (SHQI), it can be seen that the correlation of a low value is shown. In particular, looking at [Table 3], a correlation of 0.834 is shown, and such a low correlation can be directly related to the problem of reliability of the discrimination system.

In contrast, the Raman spectrum discrimination method (AHQI) of the present invention shows a high correlation result of 0.95 or more in three experimental conditions. On the other hand, when looking at the correlation of the second-ranked material rather than the first-ranked material, the correlation was determined to be lower than that of the conventional chemical substance identification method (SHQI) in 22 out of 30 spectra. It can be confirmed that the Raman spectrum discrimination method (AHQI) of the present invention does not simply increase the correlation score, but provides a strong discrimination power that increases the correlation score of the target material and decreases the score of the remaining substances.

(2) Experimental result of correlation method for experimental data and library target material

In [Table 5], the results of the experiment for discriminating input spectrum data using 14033 library spectra in the library database are presented. Looking at the experimental results, it can be seen that the correlation with the target material has increased overall. In particular, in the case of the input spectrum measured at inVia 785.0nm, which showed a clear difference in the library spectrum and the peak intensity, the existing chemical substance discrimination In the case of the method (SHQI), the average correlation was 0.9295, while the Raman spectrum discrimination method (ASHQI) of the present invention showed a correlation of 0.9720.

measuring equipment test substance HQI SHQI(HW) ASHQI (TW) Hit# Score Hit# Score Hit# Score inVia
785.0nm Acetone 59 0.5413 One 0.9210 One 0.9669 Acetonitrile 4 0.6387 One 0.8849 One 0.9763 Benzene 13 0.7468 One 0.9188 One 0.9933 Cyclohexane 17 0.5964 One 0.9595 One 0.9530 Ethyl Alcohol 2 0.7958 One 0.9576 One 0.9624 Ethylene Glycol One 0.9532 One 0.9822 One 0.9861 Hexane 222 0.8030 One 0.9726 One 0.9711 Nitromethane One 0.7832 One 0.9426 One 0.9724 2-Nitrotoluene One 0.9088 One 0.8544 One 0.9584 Toluene 118 0.6034 One 0.9013 One 0.9797 average correlation 0.7371 0.9295 0.9720 Standard Deviation 0.1311 0.0389 0.0118 Renishaw 2000 514.5nm Acetone One 0.9738 One 0.9925 One 0.9973 Acetonitrile One 0.9726 One 0.9844 One 0.9932 Benzene One 0.8840 One 0.9906 One 0.9974 Cyclohexane 4 0.9754 One 0.9803 One 0.9812 Ethyl Alcohol 1585 0.8622 One 0.9909 One 0.9970 Ethylene Glycol 4763 0.5660 One 0.9781 One 0.9828 Hexane One 0.9916 One 0.9765 One 0.9930 Nitromethane One 0.8605 One 0.9865 One 0.9820 2-Nitrotoluene One 0.9492 One 0.9781 One 0.9883 Toluene One 0.9278 One 0.9843 One 0.9741 average correlation 0.8963 0.9842 0.9886 Standard Deviation 0.1194 0.0056 0.0078 inVia
632.8nm Acetone One 0.9812 One 0.9607 One 0.9680 Acetonitrile One 0.9783 One 0.9410 One 0.9515 Benzene One 0.9677 One 0.9898 One 0.9923 Cyclohexane One 0.9905 One 0.9835 One 0.9565 Ethyl Alcohol One 0.9804 One 0.9533 One 0.9611 Ethylene Glycol One 0.9580 One 0.9786 One 0.9646 Hexane One 0.9933 One 0.9815 One 0.9806 Nitromethane One 0.9748 One 0.9755 One 0.9786 2-Nitrotoluene One 0.9673 One 0.9604 One 0.9762 Toluene One 0.9257 One 0.9495 One 0.9483 mean correlation 0.9717 0.9674 0.9678 Standard Deviation 0.0184 0.0157 0.0133

As described above, when the Raman spectrum discrimination method (ASHQI) of the present invention is used, the discrimination power between the target material and the rest of the material is further strengthened, and in the case of verifying the authenticity of the spectrum, it is generally considered a reliable correlation. It shows a consistent correlation result close to a correlation of 0.95.

6 is a graph showing correlation distribution when a plurality of chemical substance discrimination methods including the Raman spectrum discrimination method of the present invention are performed. Figure 6 (a) is a graph showing the correlation distribution of each chemical substance discrimination method in each equipment, Figure 6 (b) is a graph showing the overall correlation of each chemical substance discrimination method in each equipment to be.

6 shows the distribution of correlation values according to each method. As shown in Figure 6, compared with the existing chemical substance discrimination methods (HQI, SHQI), the distribution of correlation values in the chemical discrimination using each equipment in the Raman spectrum discrimination method (ASHQI) of the present invention is significantly higher, It can be seen that the degree of distribution of scores is not large.

The above description of the present invention is for illustration, and those of ordinary skill in the art to which the present invention pertains can understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a dispersed form, and likewise components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present invention.

100: signal acquisition unit
200: signal processing unit
300: signal discrimination unit
400: output unit
500: data part

Claims (10)

i) measuring Raman spectra for each chemical sample in advance to collect library spectra and form a library database;
ii) collecting an input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one chemical sample;
iii) adjusting the resolution of the input spectrum and removing noise;
iv) scanning the input spectrum using a moving window and measuring a peak value of the input spectrum to determine a peak clustering region, which is a region in which peaks are clustered in the input spectrum;
v) dividing the peak cluster region of the input spectrum into at least one subspectrum;
vi) analyzing a degree of similarity between each of the partial spectra and the library spectrum; and
vii) deriving a chemical substance included in the measurement target;
In step iv), the threshold value and the length of the moving window are set, and when searching using the moving window, if the largest peak value within the moving window is greater than the set threshold value, the corresponding area is regarded as the peak area. and if it is less than the threshold value, the peak cluster area is determined by judging it as an area without a peak,
In step vi), the library spectrum and the partial spectrum have an increased similarity in peak areas to each other, so that the i-th partial spectrum, which is any one of the plurality (n) of the partial spectrum, is normalized and then derived. A Raman spectrum discrimination method robust to measurement environmental conditions, characterized in that correlation is increased.
The method according to claim 1,
In step v), the partial spectrum is a spectrum obtained by dividing a peak cluster region of the input spectrum by a preset threshold value.
The method according to claim 1,
In step vi), the partial spectrum and the library spectrum are measured environment conditions, characterized in that the degree of similarity between the partial spectrum and the library spectrum is derived by performing a cross-correlation diagram to obtain a correlation using the peak area information of each. A robust Raman spectrum discrimination method.
4. The method according to claim 3,
In step vi), a method for determining a Raman spectrum robust to a measurement environment condition, characterized in that deriving an overall correlation using the partial correlation.
5. The method according to claim 4,
A method for determining a Raman spectrum robust to a measurement environment condition, characterized in that the partial correlation of the i-th partial spectrum is derived by the following polynomial.

Here, S _i is the i-th partial spectral partial correlation value, and x _i and y _i are data obtained by normalizing the i-th partial spectral region.
6. The method of claim 5,
The Raman spectrum discrimination method robust to a measurement environment condition, characterized in that the total degree of correlation with respect to the plurality of partial spectra is an average value of the partial correlation degree by the total number of the partial spectrum.
7. The method of claim 6,
A Raman spectrum discrimination method robust to measurement environmental conditions, characterized in that the overall correlation for a part of the library spectrum is an average value of the partial correlation by the number of divisions of the part of the library spectrum.
8. The method of claim 7,
The final correlation degree is a Raman spectrum determination method robust to measurement environmental conditions, characterized in that it is the sum of the total correlation degrees for the plurality of partial spectra and the total correlation degrees for some of the library spectra.
9. The method of claim 8,
In step vii), the method for determining a Raman spectrum robust to a measurement environment condition, characterized in that the library spectrum, which has obtained the highest final correlation with the input spectrum, is determined as a Raman spectrum for the chemical substance.
In the Raman spectrum discrimination apparatus used in the Raman spectrum discrimination method robust to the measurement environment conditions of claim 1,
a signal acquisition unit configured to collect the input spectrum including at least one Raman spectrum by performing Raman spectroscopy on a measurement target including at least one or more chemical samples;
It is connected to the signal acquisition unit, adjusts the resolution of the input spectrum and removes noise, scans the input spectrum using a moving window, and measures the peak value of the input spectrum to determine a peak cluster area of the input spectrum signal processing unit;
a signal discrimination unit connected to the signal processing unit, dividing a peak cluster region of the input spectrum into at least one partial spectrum, and analyzing a similarity between each partial spectrum and the library spectrum; and
and an output unit connected to the signal discrimination unit and displaying the chemical substance included in the measurement target on a display screen.

Images