阅读说明：本技术 一种野生滇龙胆libs识别方法及系统 (LIBS (laser-induced breakdown Spectroscopy) identification method and system for wild gentiana rigescens ) 是由 孙明华 孔汶汶 孙永祺 于 2020-06-04 设计创作，主要内容包括：本发明涉及一种野生滇龙胆LIBS识别方法及系统。所述方法包括获取待检测的滇龙胆植株的压片样本；利用激光诱导击穿光谱仪确定所述压片样本的激光诱导击穿光谱；根据所述激光诱导击穿光谱,利用matlab中的findpeaks算法和有效激发峰阈值,确定所述激光诱导击穿光谱的有效激发峰；根据所述激光诱导击穿光谱中的有效激发峰确定重构后的激光诱导击穿光谱；根据所述重构后的激光诱导击穿光谱和识别模型确定识别结果。本发明能够快速、准确地识别野生滇龙胆。(The invention relates to a wild gentiana rigescens LIBS identification method and system. The method comprises the steps of obtaining a tabletting sample of a Gentiana rigescens plant to be detected; determining a laser-induced breakdown spectrum of the tabletting sample by using a laser-induced breakdown spectrometer; determining an effective excitation peak of the laser-induced breakdown spectrum by using a findpeaks algorithm and an effective excitation peak threshold in matlab according to the laser-induced breakdown spectrum; determining a reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum; and determining a recognition result according to the reconstructed laser-induced breakdown spectrum and the recognition model. The method can quickly and accurately identify the wild gentiana rigescens.)

1. A wild Gentiana rigescens LIBS recognition method is characterized by comprising the following steps:

obtaining a tabletting sample of a gentiana rigescens plant to be detected;

determining a laser-induced breakdown spectrum of the tabletting sample by using a laser-induced breakdown spectrometer;

determining an effective excitation peak of the laser-induced breakdown spectrum by using a findpeaks algorithm and an effective excitation peak threshold in matlab according to the laser-induced breakdown spectrum; the effective excitation peak is a spectrum value of the current wavelength which is greater than a spectrum value of the previous wavelength and a spectrum value of the next wavelength in the laser-induced breakdown spectrum, and the spectrum value of the current wavelength is greater than the threshold value of the effective excitation peak;

determining a reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum;

determining an identification result according to the reconstructed laser-induced breakdown spectrum and an identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

2. The method for LIBS identification of Gentiana rigescens Linn as claimed in claim 1, wherein the determining the laser-induced breakdown spectrum of the pressed sample by using a laser-induced breakdown spectrometer further comprises:

performing wavelet denoising on the laser-induced breakdown spectrum to obtain a denoised laser-induced breakdown spectrum;

and carrying out baseline correction and area normalization on the denoised laser-induced breakdown spectrum.

3. The method for identifying the LIBS of the wild gentiana rigescens according to claim 1, wherein the step of determining the reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum specifically comprises the following steps:

determining the peak position of the effective excitation peak according to the effective excitation peak;

determining the intensity of an effective excitation peak according to the peak position;

determining a reconstructed spectrum variable according to the intensity of the effective excitation peak;

and determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

4. The method for identifying the wild gentiana rigescens LIBS according to claim 1, wherein the process for determining the identification model specifically comprises the following steps:

obtaining a tabletting sample of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens;

determining a laser-induced breakdown spectrum of the sample set from the pressed samples of the sample set;

determining an effective excitation peak of the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold in matlab according to the laser-induced breakdown spectrum of the sample set;

determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set;

determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data;

sequencing the spectrum angles to obtain a sequencing result;

acquiring a maximum spectrum angle and a minimum spectrum angle in the sequencing result, and determining a spectrum corresponding to the maximum spectrum angle and a spectrum corresponding to the minimum spectrum angle to construct a modeling set;

dividing other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals, and acquiring a spectrum corresponding to a middle spectrum angle in each interval to construct a verification set; placing spectra corresponding to other spectral angles in the plurality of intervals except for an intermediate spectral angle into the modeling set;

constructing an identification model by taking the modeling set as input and the identification result as output;

and verifying the identification model by using the verification set.

5. A wild gentiana rigescens LIBS recognition system is characterized by comprising:

the first tabletting sample acquisition module is used for acquiring a tabletting sample of the Gentiana rigescens plant to be detected;

the first laser-induced breakdown spectroscopy determining module is used for determining the laser-induced breakdown spectroscopy of the tabletting sample by using a laser-induced breakdown spectrometer;

the first effective excitation peak determining module is used for determining an effective excitation peak of the laser-induced breakdown spectrum by utilizing a findpeaks algorithm and an effective excitation peak threshold value in matlab according to the laser-induced breakdown spectrum; the effective excitation peak is a spectrum value of the current wavelength which is greater than a spectrum value of the previous wavelength and a spectrum value of the next wavelength in the laser-induced breakdown spectrum, and the spectrum value of the current wavelength is greater than the threshold value of the effective excitation peak;

the first reconstructed laser-induced breakdown spectrum determining module is used for determining a reconstructed laser-induced breakdown spectrum according to an effective excitation peak in the laser-induced breakdown spectrum;

the identification result determining module is used for determining an identification result according to the reconstructed laser-induced breakdown spectrum and the identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

6. The wild gentiana rigescens LIBS recognition system of claim 5, further comprising:

the de-noised laser-induced breakdown spectrum determining module is used for performing wavelet de-noising on the laser-induced breakdown spectrum to obtain a de-noised laser-induced breakdown spectrum;

and the baseline correction and area normalization processing module is used for performing baseline correction and area normalization on the denoised laser-induced breakdown spectrum.

7. The wild gentiana rigescens LIBS recognition system of claim 5, wherein the first reconstructed laser-induced breakdown spectroscopy determination module specifically comprises:

a peak position determining unit for determining a peak position of the effective excitation peak according to the effective excitation peak;

an intensity determination unit of the effective excitation peak, which is used for determining the intensity of the effective excitation peak according to the peak position;

the reconstructed spectral variable determining unit is used for determining a reconstructed spectral variable according to the intensity of the effective excitation peak;

and the reconstructed laser-induced breakdown spectrum determining unit is used for determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

8. The wild gentiana rigescens LIBS recognition system of claim 5, wherein the identification model determination process specifically comprises:

the second tabletting sample acquisition module is used for acquiring tabletting samples of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens;

a second laser-induced breakdown spectroscopy determination module for determining a laser-induced breakdown spectroscopy of the sample set from the pressed samples of the sample set;

the second effective excitation peak determining module is used for determining an effective excitation peak of the laser-induced breakdown spectrum of the sample set according to the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold in matlab;

the second reconstructed laser-induced breakdown spectrum determining module is used for determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set;

the spectral angle data determining module is used for determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data;

the sequencing result determining module is used for sequencing the spectrum angles to obtain a sequencing result;

the modeling set determining module is used for acquiring the maximum spectral angle and the minimum spectral angle in the sequencing result, and determining the spectrum corresponding to the maximum spectral angle and the spectrum corresponding to the minimum spectral angle so as to construct a modeling set;

a verification set determining module, configured to divide other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals, and obtain a spectrum corresponding to an intermediate spectrum angle in each interval to construct a verification set; placing spectra corresponding to other spectral angles in the plurality of intervals except for an intermediate spectral angle into the modeling set;

the identification model construction module is used for constructing an identification model by taking the modeling set as input and the identification result as output;

and the identification model verification module is used for verifying the identification model by utilizing the verification set.

Technical Field

The invention relates to the field of spectrum detection, in particular to a wild gentiana rigescens LIBS identification method and system.

Background

The Gentiana (Gentiana) plants mostly grow in mountains or plateaus, the climatic conditions of the growing environment are extreme and changeable, because of the difference between the growing environment and the growing period of the Gentiana rigescens, the family products and the wild products have obvious difference in some aspects, and the Gentiana rigescens is rich and loose in appearance, thin and compact in wild, long in wild growing period, balanced and rich in components, good in curative effect, and nonstandard in planting technology and abuse of hormones in the medicinal materials of the family, so the wild identification of the Gentiana rigescens serving as the traditional Chinese medicinal material is particularly important.

In the aspect of traditional technical means, the quality of the Chinese herbal medicine is determined by measuring the composition proportion and content value of various nutritional ingredients in the Chinese herbal medicine mainly through a chemical method, and the method has the advantages of good repeatability, high accuracy, high labor cost and complex operation, has a long time from sampling to result output, and cannot meet the requirement of real-time and rapid detection of the market.

Disclosure of Invention

The invention aims to provide a method and a system for identifying LIBS of wild gentiana rigescens, which can quickly and accurately identify the wild gentiana rigescens.

In order to achieve the purpose, the invention provides the following scheme:

a wild Gentiana rigescens LIBS recognition method comprises the following steps:

obtaining a tabletting sample of a gentiana rigescens plant to be detected;

determining a laser-induced breakdown spectrum of the tabletting sample by using a laser-induced breakdown spectrometer;

determining an effective excitation peak of the laser-induced breakdown spectrum by using a findpeaks algorithm and an effective excitation peak threshold in matlab according to the laser-induced breakdown spectrum; the effective excitation peak is a spectrum value of the current wavelength which is greater than a spectrum value of the previous wavelength and a spectrum value of the next wavelength in the laser-induced breakdown spectrum, and the spectrum value of the current wavelength is greater than the threshold value of the effective excitation peak;

determining a reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum;

determining an identification result according to the reconstructed laser-induced breakdown spectrum and an identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

Optionally, the determining the laser-induced breakdown spectrum of the pressed sheet sample by using the laser-induced breakdown spectrometer further includes:

performing wavelet denoising on the laser-induced breakdown spectrum to obtain a denoised laser-induced breakdown spectrum;

and carrying out baseline correction and area normalization on the denoised laser-induced breakdown spectrum.

Optionally, the determining the reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum specifically includes:

determining the peak position of the effective excitation peak according to the effective excitation peak;

determining the intensity of an effective excitation peak according to the peak position;

determining a reconstructed spectrum variable according to the intensity of the effective excitation peak;

and determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

Optionally, the determining process of the recognition model specifically includes:

obtaining a tabletting sample of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens;

determining a laser-induced breakdown spectrum of the sample set from the pressed samples of the sample set;

determining an effective excitation peak of the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold in matlab according to the laser-induced breakdown spectrum of the sample set;

determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set;

determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data;

sequencing the spectrum angles to obtain a sequencing result;

acquiring a maximum spectrum angle and a minimum spectrum angle in the sequencing result, and determining a spectrum corresponding to the maximum spectrum angle and a spectrum corresponding to the minimum spectrum angle to construct a modeling set;

dividing other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals, and acquiring a spectrum corresponding to a middle spectrum angle in each interval to construct a verification set; placing spectra corresponding to other spectral angles in the plurality of intervals except for an intermediate spectral angle into the modeling set;

constructing an identification model by taking the modeling set as input and the identification result as output;

and verifying the identification model by using the verification set.

A wild gentiana rigescens LIBS recognition system comprises:

the first tabletting sample acquisition module is used for acquiring a tabletting sample of the Gentiana rigescens plant to be detected;

the first laser-induced breakdown spectroscopy determining module is used for determining the laser-induced breakdown spectroscopy of the tabletting sample by using a laser-induced breakdown spectrometer;

the first effective excitation peak determining module is used for determining an effective excitation peak of the laser-induced breakdown spectrum by utilizing a findpeaks algorithm and an effective excitation peak threshold value in matlab according to the laser-induced breakdown spectrum; the effective excitation peak is a spectrum value of the current wavelength which is greater than a spectrum value of the previous wavelength and a spectrum value of the next wavelength in the laser-induced breakdown spectrum, and the spectrum value of the current wavelength is greater than the threshold value of the effective excitation peak;

the first reconstructed laser-induced breakdown spectrum determining module is used for determining a reconstructed laser-induced breakdown spectrum according to an effective excitation peak in the laser-induced breakdown spectrum;

the identification result determining module is used for determining an identification result according to the reconstructed laser-induced breakdown spectrum and the identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

Optionally, the method further includes:

the de-noised laser-induced breakdown spectrum determining module is used for performing wavelet de-noising on the laser-induced breakdown spectrum to obtain a de-noised laser-induced breakdown spectrum;

and the baseline correction and area normalization processing module is used for performing baseline correction and area normalization on the denoised laser-induced breakdown spectrum.

Optionally, the module for determining a laser-induced breakdown spectroscopy after the first reconstruction specifically includes:

a peak position determining unit for determining a peak position of the effective excitation peak according to the effective excitation peak;

an intensity determination unit of the effective excitation peak, which is used for determining the intensity of the effective excitation peak according to the peak position;

the reconstructed spectral variable determining unit is used for determining a reconstructed spectral variable according to the intensity of the effective excitation peak;

and the reconstructed laser-induced breakdown spectrum determining unit is used for determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

Optionally, the determining process of the recognition model specifically includes:

the second tabletting sample acquisition module is used for acquiring tabletting samples of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens;

a second laser-induced breakdown spectroscopy determination module for determining a laser-induced breakdown spectroscopy of the sample set from the pressed samples of the sample set;

the second effective excitation peak determining module is used for determining an effective excitation peak of the laser-induced breakdown spectrum of the sample set according to the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold in matlab;

the second reconstructed laser-induced breakdown spectrum determining module is used for determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set;

the spectral angle data determining module is used for determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data;

the sequencing result determining module is used for sequencing the spectrum angles to obtain a sequencing result;

the modeling set determining module is used for acquiring the maximum spectral angle and the minimum spectral angle in the sequencing result, and determining the spectrum corresponding to the maximum spectral angle and the spectrum corresponding to the minimum spectral angle so as to construct a modeling set;

a verification set determining module, configured to divide other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals, and obtain a spectrum corresponding to an intermediate spectrum angle in each interval to construct a verification set; placing spectra corresponding to other spectral angles in the plurality of intervals except for an intermediate spectral angle into the modeling set;

the identification model construction module is used for constructing an identification model by taking the modeling set as input and the identification result as output;

and the identification model verification module is used for verifying the identification model by utilizing the verification set.

According to the specific embodiment provided by the invention, the invention discloses the following technical effects:

according to the method and the system for identifying the LIBS of the wild gentiana rigescens, provided by the invention, the background noise is filtered by screening effective excitation peak signals, the information of element types and relative contents in the gentiana rigescens can be basically extracted by screening the effective excitation peaks, the number of spectrum variables is reduced from 22015 to 342, and the model training time is greatly shortened. The identification model can quickly identify the category of the gentiana rigescens by screening the effective excitation peak signal to obtain an identification result, and the method has the characteristics of high detection speed, few required samples, no pollution, high detection precision and high reliability. The invention provides a technical means for screening the wild identification of the gentiana rigescens. And the requirements of rapid market detection, simple sample treatment, high efficiency and the like are met.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.

FIG. 1 is a schematic flow chart of a wild Gentiana rigescens LIBS identification method provided by the invention;

FIG. 2 is a graph of the original laser-induced breakdown spectrum of Gentiana rigescens;

FIG. 3 is a schematic diagram of efficient excitation peak screening;

FIG. 4 is a graph of laser induced breakdown spectra after screening for effective excitation peaks;

FIG. 5 is a graph of laser-induced breakdown spectroscopy elemental information for Gentiana rigescens;

fig. 6 is a schematic structural diagram of a wild gentiana rigescens LIBS recognition system provided by the invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The invention aims to provide a method and a system for identifying LIBS of wild gentiana rigescens, which can quickly and accurately identify the wild gentiana rigescens.

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.

Laser Induced Breakdown Spectroscopy (LIBS) is a novel spectroscopic measurement technique that uses laser to excite a plasma and collect the emission spectra of atoms or ions in the cooling process of the plasma to obtain the elemental composition and content information of the measured object.

Fig. 1 is a schematic flow chart of a wild gentiana rigescens LIBS identification method provided by the present invention, and as shown in fig. 1, the wild gentiana rigescens LIBS identification method provided by the present invention comprises:

s101, obtaining a tabletting sample of the Gentiana rigescens plant to be detected.

And S102, determining the laser-induced breakdown spectrum of the tabletting sample by using a laser-induced breakdown spectrometer. I.e., the original laser induced breakdown spectrum, and as shown in fig. 2, there are 22015 laser induced breakdown spectrum variables, each representing element type and relative content information.

And carrying out wavelet denoising on the laser-induced breakdown spectrum to obtain a denoised laser-induced breakdown spectrum. And searching the optimal wavelet parameter with the highest signal-to-noise ratio in different wavelet transformation parameters with the wavelet basis functions of db 3-db 10 and the number of layers of 3-10 as db5, and performing wavelet denoising on the laser induced breakdown spectrum on the basis that the optimal wavelet parameter with the highest signal-to-noise ratio is found out and the number of layers is 3.

And in order to eliminate instrument parameter errors and fluctuation caused by experimental environment, baseline correction and area normalization are carried out on the denoised laser-induced breakdown spectrum.

S103, determining an effective excitation peak of the laser-induced breakdown spectrum according to the laser-induced breakdown spectrum by using a findpeaks algorithm and an effective excitation peak threshold in matlab; the effective excitation peak is a spectrum value of the current wavelength in the laser-induced breakdown spectrum which is larger than a spectrum value of the previous wavelength and a spectrum value of the next wavelength, and the spectrum value of the current wavelength is larger than the threshold value of the effective excitation peak. Is effectiveExcitation peak screening scheme and is shown in figure 3. Effective excitation peak threshold of 5.75x10^-5. Because the threshold value of the effective excitation peak is fixed, the positions and the number of the effective excitation peaks of different laser-induced breakdown spectrums are different, and the representativeness of the positions of the effective excitation peaks is ensured.

The findpeaks algorithm is that if the spectrum value at a certain wavelength is larger than the spectrum values at the previous wavelength and the next wavelength.

And S104, determining the reconstructed laser-induced breakdown spectrum according to the effective excitation peak in the laser-induced breakdown spectrum, as shown in FIG. 4.

And determining the peak position of the effective excitation peak according to the effective excitation peak. And taking the average value of the number of the effective excitation peaks in all the laser-induced breakdown spectrums after being rounded downwards as a reference value, finding the laser-induced breakdown spectrums with the same number of the effective excitation peaks as the reference value as the reference spectrum, and recording the peak positions of the spectrums.

And determining the intensity of the effective excitation peak according to the peak position. Because the plasma excited at the sample surface has a complex environment of high temperature and high electron density in the spectra collected by the LIBS system, there may be a slight shift in the peak position. In the screening process, two points before and after the peak position are taken, and the maximum value of the spectrum intensity of the five points represents the intensity of the effective excitation peak at the peak position.

And determining the reconstructed spectral variable according to the intensity of the effective excitation peak.

And determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

The wavelength of the original laser-induced breakdown spectrum recording spectral line is lambda₁、λ₂、……、λ_p(p 22015) a spectrum having 22015 wavelengths corresponding to a line intensity I_λ1,I_λ2,……I_λpThe spectral vector can be represented as X_i，

(I-1, 2 … … 2048), I representing the number of spectral slices, for a total of 2048 spectra, combined with the findpeaks algorithm in matlab, I_λGreater than threshold 5.75x10^-5The effective excitation peak is marked, and the wavelength corresponding to the effective excitation peak is recorded as₁,₂,……_n(n-342), a total of 342 wavelengths are marked, and a new spectral variation is obtained(i is 1,2 … … 2048), i represents the number of spectra, and there are 2048 spectra in total, but the number of variables in the spectrum is reduced from 22015 to 342. Greatly shortens the training time of the model and improves the verification effect of the model

S105, determining an identification result according to the reconstructed laser-induced breakdown spectrum and an identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

The process for determining the identification model specifically includes:

obtaining a tabletting sample of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens. The sample set is Gentiana rigescens plants of different producing areas in Yunnan province, the roots of the Gentiana rigescens are cleaned, the Gentiana rigescens are dried and ground, 0.15g of each sample is taken and pressed into tablets, the thickness of the tablets is constant, the errors caused by the fact that the laser focusing point is not consistent with the surface height of the sample can be reduced, the wild samples are taken from the rear rock mountain heads, the total number of the wild samples is ten, and the rest are the family. Sample types and amounts are shown in table 1:

TABLE 1

Determining a laser induced breakdown spectrum of the sample set from the pressed sample of the sample set. Spectra of 16 different positions of one tablet are collected, and each spectrum is the average result of 5 times of accumulation of laser impact, namely 16 spectra are collected by one tablet, and 2048 spectra are collected in total. Each spectrum has 22015 wavelengths, the spectral range is 229.1-877.48nm, namely each spectrum has 22015 spectral variables, and the LIBS system parameters for collecting the spectrum are as follows: laser energy 60 mJ; the laser wavelength is 532 nm; the depth of focus is 2 mm; the delay time is 2 mus; the integration time was 20. mu.s.

And determining the effective excitation peak of the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold value in matlab according to the laser-induced breakdown spectrum of the sample set.

And determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set.

Determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; and the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data.

And sequencing the spectral angles to obtain a sequencing result.

And acquiring a maximum spectrum angle and a minimum spectrum angle in the sequencing result, and determining a spectrum corresponding to the maximum spectrum angle and a spectrum corresponding to the minimum spectrum angle to construct a modeling set.

Dividing other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals, and acquiring a spectrum corresponding to a middle spectrum angle in each interval to construct a verification set; placing spectra corresponding to spectral angles other than the intermediate spectral angle in a plurality of said intervals into said modeling set.

The ratio of the modeling set to the validation set is approximately 2: 1. And (3) dividing the 2048 spectrums of the gentiana rigescens into a modeling set and a verification set according to spectrum angles, and as a result, 1445 spectrums exist in the modeling set and 603 spectrums exist in the verification set.

The essence of the spectrum angle is the included angle of two spectrum vectors, the deviation degree of the spectrum can be effectively represented, the larger the included angle is, the larger the difference of the two spectra is, the spectrum angle of each spectrum and the average spectrum of the spectrum is solved for the Gentiana rigescens spectrum of each production place, the spectrum angle is divided into a plurality of intervals, the spectrum corresponding to the spectrum angle in the middle of each interval is put into a verification set, the uniformity of the modeling set and the verification set can be effectively ensured, the model representation performance is improved, and the modeling effect is greatly improved.

And constructing an identification model by taking the modeling set as input and the identification result as output.

And verifying the identification model by using the verification set.

The family sample label is 1, the wild sample label is 2, an extreme learning machine (ELM, extreme learning machine), a Support Vector Machine (SVM), a random forest (RF, random forest), a linear recognition Analysis (PLS-DA, Partial Least square prediction Analysis) and a radial basis function neural network (RBFNN, radial basis function neural network) model are established, the input of each model is a modeling set and a verification set, the output is the accuracy of the modeling set and the verification set under the model, and the effect is as shown in table 2:

TABLE 2

Model	Accuracy of Calibration set	Accuracy of Prediction set
			ELM	99.58	97.24
SVM	99.37	97.56
			RF	100	95.13
PLS-DA	97.16	95.62
			RBFNN	98.75	95.94

If no effective excitation peak signal is screened out, the full spectrum modeling result is adopted as follows (blank represents that the program has too long running time and no result is ended), as shown in table 3:

TABLE 3

Model	Accuracy of Calibration set	Accuracy of Prediction set
			ELM	97.73	94.94
SVM
			RF	99.20	91.47
PLS-DA
			RBFNN

As can be seen from the table data, after the effective excitation peak is screened out, the program operation time is reduced, the model accuracy is greatly improved, the background information seriously interferes with the modeling, the background noise is very necessary to be filtered out, and the method for screening the effective excitation peak can effectively reduce the input information and improve the model efficiency.

In the aspect of LIBS spectral data processing of traditional Chinese medicine, another commonly used method for reducing input information is to reduce the number of spectral wavelengths according to corresponding wavelengths of nutritious elements, trace elements and the like of gentiana rigescens in LIBS spectrum, and also to reduce input information, and then an identification model is established by using this method, referring to the spectral wavelengths corresponding to elements provided by the National Institute of Standards and Technology (NIST), the nutritious elements and trace elements of gentiana rigescens obtained in this experiment are labeled, and 42 spectral wavelengths correspond to the spectral wavelengths, as shown in fig. 5 (all element information is labeled, but the same element corresponds to a plurality of excitation lines, and only a part of the excitation lines are selected).

So that LIBS data is reduced from 22015 spectral variations to spectral variations corresponding to 42 relevant elements. The correspondence between elements and wavelengths is shown in table 4:

TABLE 4

According to the wavelengths corresponding to the 42 related elements, screening out the spectral data corresponding to the wavelengths from the original spectrum, and establishing an identification model according to the data, the results are shown in the following table 5:

TABLE 5

Model	Accuracy of Calibration set	Accuracy of Prediction set
			ELM	98.92	96.29
SVM	100	96.58
			RF	99.71	97.18
PLS-DA	92.51	91.38
			RBFNN	97.19	95.69

Compared with the identification model under the method for screening the effective excitation peak, the method for screening the effective excitation peak is shown to have lower accuracy of the verification set except for the RF model, thereby fully showing that the method for screening the effective excitation peak is superior to the common method, and fully ensuring the accuracy of the discrimination model while reducing the input variable.

Fig. 6 is a schematic structural diagram of a wild gentiana rigescens LIBS recognition system provided by the present invention, and as shown in fig. 6, the wild gentiana rigescens LIBS recognition system provided by the present invention includes: the device comprises a first tabletting sample acquisition module 601, a first laser-induced breakdown spectrum determination module 602, a first effective excitation peak determination module 603, a first reconstructed laser-induced breakdown spectrum determination module 604 and an identification result determination module 605.

The first tabletting sample obtaining module 601 is used for obtaining a tabletting sample of a gentiana rigescens plant to be detected;

the first laser induced breakdown spectroscopy determination module 602 is configured to determine a laser induced breakdown spectroscopy of the preform sample using a laser induced breakdown spectrometer.

The first effective excitation peak determining module 603 is configured to determine an effective excitation peak of the laser-induced breakdown spectrum according to the laser-induced breakdown spectrum by using a findpeaks algorithm and an effective excitation peak threshold in the matlab; the effective excitation peak is a spectrum value of the current wavelength in the laser-induced breakdown spectrum which is larger than a spectrum value of the previous wavelength and a spectrum value of the next wavelength, and the spectrum value of the current wavelength is larger than the threshold value of the effective excitation peak.

The first reconstructed laser-induced breakdown spectrum determining module 604 is configured to determine a reconstructed laser-induced breakdown spectrum according to an effective excitation peak in the laser-induced breakdown spectrum.

The identification result determining module 605 is configured to determine an identification result according to the reconstructed laser-induced breakdown spectrum and the identification model; the identification model takes the reconstructed laser-induced breakdown spectrum as input and takes the identification result as output; the identification result comprises wild gentiana rigescens and domestic gentiana rigescens; the recognition model comprises an extreme learning machine model, a support vector machine model, a random forest model, a linear recognition analysis model and a radial basis function neural network model.

The invention provides a wild gentiana rigescens LIBS recognition system, which further comprises: the device comprises a denoised laser-induced breakdown spectrum determining module and a baseline correction and area normalization processing module.

The de-noised laser-induced breakdown spectrum determining module is used for performing wavelet de-noising on the laser-induced breakdown spectrum to obtain the de-noised laser-induced breakdown spectrum.

And the baseline correction and area normalization processing module is used for performing baseline correction and area normalization on the denoised laser-induced breakdown spectrum.

The laser-induced breakdown spectroscopy determination module after the first reconstruction specifically includes: the device comprises a peak position determining unit, an effective excitation peak intensity determining unit, a reconstructed spectral variable determining unit and a reconstructed laser-induced breakdown spectrum determining unit.

And the peak position determining unit is used for determining the peak position of the effective excitation peak according to the effective excitation peak.

The intensity determination unit of the effective excitation peak is used for determining the intensity of the effective excitation peak according to the peak position.

And the reconstructed spectral variable determining unit is used for determining the reconstructed spectral variable according to the intensity of the effective excitation peak.

And the reconstructed laser-induced breakdown spectrum determining unit is used for determining the reconstructed laser-induced breakdown spectrum according to the reconstructed spectrum variable.

The process for determining the identification model specifically includes: the device comprises a second tabletting sample acquisition module, a second laser-induced breakdown spectrum determination module, a second effective excitation peak determination module, a second reconstructed laser-induced breakdown spectrum determination module, a spectrum angle data determination module, a sequencing result determination module, a modeling set determination module, a verification set determination module, an identification model construction module and an identification model verification module.

The second tabletting sample acquisition module is used for acquiring tabletting samples of the sample set; the sample set comprises wild gentiana rigescens and domestic gentiana rigescens.

And the second laser-induced breakdown spectrum determination module is used for determining the laser-induced breakdown spectrum of the sample set according to the pressed sample of the sample set.

And the second effective excitation peak determining module is used for determining the effective excitation peak of the laser-induced breakdown spectrum of the sample set by using a findpeaks algorithm and an effective excitation peak threshold value in matlab according to the laser-induced breakdown spectrum of the sample set.

And the second reconstructed laser-induced breakdown spectrum determination module is used for determining the reconstructed laser-induced breakdown spectrum of the sample set according to the effective excitation peak in the laser-induced breakdown spectrum of the sample set.

The spectral angle data determining module is used for determining spectral angle data according to the reconstructed laser-induced breakdown spectrum of the sample set; the spectral angle data comprises a plurality of spectral angles; and the spectral angle is an included angle between the reconstructed laser-induced breakdown spectrum and the average spectrum of all sample sets in the spectral data.

And the sequencing result determining module is used for sequencing the spectrum angles to obtain a sequencing result.

And the modeling set determining module is used for acquiring the maximum spectral angle and the minimum spectral angle in the sequencing result, and determining the spectrum corresponding to the maximum spectral angle and the spectrum corresponding to the minimum spectral angle so as to construct a modeling set.

The verification set determining module is used for dividing other spectrum angles except the maximum spectrum angle and the minimum spectrum angle into a plurality of intervals and acquiring a spectrum corresponding to a middle spectrum angle in each interval to construct a verification set; placing spectra corresponding to spectral angles other than the intermediate spectral angle in a plurality of said intervals into said modeling set.

And the identification model construction module is used for constructing an identification model by taking the modeling set as input and the identification result as output.

And the identification model verification module is used for verifying the identification model by utilizing the verification set.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.

The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.