Adaptive characteristic spectral line screening method and system based on atomic emission spectroscopy

Abstract

The application provides an adaptive characteristic spectral line screening method and system based on atomic emission spectroscopy, which comprises the following steps: based on a spectral data set, a set of characteristic screening optimization methods is used for characteristic screening in several optimization rounds to obtain an initial spectral data set of each round of characteristic screening and initial characteristic population genes corresponding to the initial spectral data set; based on the initial spectral data set and the initial characteristic population genes, the optimal characteristic population genes of each round are obtained through a set of analysis methods, a fitness function and the iteration of a genetic algorithm; when the several optimization rounds reach a set of optimization rounds, an optimized spectral characteristic information set is obtained; and the optimized spectral characteristic information set is subjected to combination statistics and discriminant analysis to complete the screening of adaptive characteristic spectral lines. The application can efficiently and automatically screen out characteristic spectral lines that are adapted to the analysis requirements in complex atomic emission spectroscopy, thereby ensuring the effectiveness and accuracy of the screened characteristic spectral lines.

Description

Technical Field

[0001] This invention relates to the field of atomic emission spectroscopy analysis technology, and in particular to an adaptive characteristic spectral line screening method and system based on atomic emission spectroscopy. Background Technology

[0002] Atomic emission spectrometry (AES) possesses advantages such as speed, sensitivity, and high selectivity, playing a crucial role in the qualitative, semi-quantitative, and quantitative analysis of various inorganic materials. With the emergence of new light sources and technologies, AES has become recognized as one of the most prestigious modern analytical techniques.

[0003] Atomic emission spectroscopy (AES) analyzes materials qualitatively or quantitatively based on the wavelength and intensity of the characteristic spectra emitted by excited-state atoms. Therefore, the effective identification and selection of characteristic spectral lines is a crucial factor affecting the analytical capability of emission spectroscopy. With the diversification of analytical testing technologies, instruments with various hardware configurations, such as portable, benchtop, and online devices, have emerged. The resolution and detection capabilities of the corresponding spectroscopic systems also vary. Given the vast number of atomic emission lines, selection is primarily based on empirical verification using standard spectral databases combined with actual analytical results. This process requires significant manpower and time for spectral calibration and verification, resulting in high time consumption and low efficiency. While increasing research utilizes algorithms for spectral line selection, the richness and complexity of atomic spectra limit their efficiency, effectiveness, and automation in practical applications. Therefore, efficiently and accurately selecting characteristic spectral lines for different analytical needs is the primary task and a crucial link in the entire analytical process. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive characteristic spectral line screening method and system based on atomic emission spectroscopy, which can efficiently and automatically screen characteristic spectral lines in complex atomic emission spectra to meet analytical requirements, ensuring the effectiveness and accuracy of the screening of characteristic spectral lines, and providing a reliable guarantee for the accurate and stable execution of analytical tasks.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] An adaptive feature spectral line screening method based on atomic emission spectroscopy includes the following steps:

[0007] Based on the original spectral signal of the sample to be analyzed, the spectral dataset is determined;

[0008] performing feature screening on the spectral dataset for several optimization rounds by using a set of feature screening optimization methods, obtaining an initialized spectral dataset for each round of feature screening and an initialized feature population gene corresponding to the initialized spectral dataset;

[0009] Based on the initialized spectral dataset and the initialized feature population gene, the optimal feature population gene for each round is obtained by a set of analysis methods, fitness function and iteration of genetic algorithm;

[0010] When the several optimization rounds reach a set of optimization rounds, an optimal spectral feature information set corresponding to an optimal feature population gene set composed of the optimal feature population gene for each round in the set of optimization rounds is obtained;

[0011] Combination statistics and discriminant analysis are performed on the optimal spectral feature information set to complete the screening of adaptive characteristic spectral lines.

[0012] Further, the spectral dataset is determined based on the original spectral signal of the sample to be analyzed, comprising:

[0013] According to the analysis requirements, samples meeting the element and content coverage range are selected as the samples to be analyzed;

[0014] The original spectral signal of the sample to be analyzed is obtained by performing spectral analysis on the sample to be analyzed;

[0015] The original spectral signal is preprocessed to obtain spectral information;

[0016] All or part of the spectral information preprocessed by the spectrum is selected to construct the spectral dataset.

[0017] Further, the spectral preprocessing includes background correction and filtering and noise reduction.

[0018] Further, the spectral dataset is determined based on the original spectral signal of the sample to be analyzed, comprising:

[0019] The spectral dataset is screened for several optimization rounds, and part or all of the spectral information in the spectral dataset is selected for each round of feature screening to obtain the initialized spectral dataset;

[0020] Based on the spectral characteristics, the spectral information in the initialized spectral dataset is encoded to obtain an uninitialized feature population gene, and the initialized feature population gene is determined based on the uninitialized feature population gene.

[0021] Further, the optimal feature population gene of each round is obtained by setting an analysis method, a fitness function and iteration of a genetic algorithm based on the initial spectral data set and the initial feature population gene, comprising:

[0022] According to the analysis requirement, a corresponding analysis method is selected as the set analysis method, and the set analysis method determines the model parameters and evaluation indexes of the analysis method based on the initial spectral data set and the initial feature population gene;

[0023] The fitness function is based on the initial feature population gene and the model parameters and evaluation indexes of the analysis method to obtain the population fitness;

[0024] The genetic algorithm is based on the population fitness and the initial feature population gene to obtain a new feature population gene until the iteration number of the genetic algorithm reaches a set maximum value or the population fitness reaches a set fitness threshold, and the finally obtained new feature population gene is the optimal feature population gene.

[0025] Further, the original spectral signal is an atomic emission spectral signal, which includes a wide-band continuous spectral signal obtained by an array detection device or a narrow-band spectral signal obtained by a photoelectric detection device.

[0026] Further, the set analysis method is a quantitative analysis, a semi-quantitative analysis, a discriminant analysis or an analysis method that can be characterized by modeling.

[0027] Further, the combination statistics and discriminant analysis of the optimal spectral feature information set are performed to complete the screening of the adaptive feature spectral lines, comprising:

[0028] The optimal spectral feature information set is subjected to one or more of probability analysis, frequency analysis and arbitrary combination result analysis to obtain statistical information of each feature spectral line;

[0029] Through discriminant analysis, if the evaluation value corresponding to the statistical information of each feature spectral line is greater than a set screening threshold, the screening of the adaptive feature spectral lines of the optimal spectral feature information set is completed.

[0030] The application also provides an adaptive feature spectral line screening system based on atomic emission spectroscopy, which is applied to any one of the adaptive feature spectral line screening methods based on atomic emission spectroscopy.

[0031] The first determination module is configured to determine a spectral data set based on the original spectral signal of the sample to be analyzed;

[0032] The first acquisition module is configured to perform feature screening of the spectral data set for a plurality of optimization rounds by using a set feature screening optimization method, and to acquire an initialized spectral data set and an initialized feature population gene corresponding to the initialized spectral data set for each round of feature screening.

[0033] The second acquisition module is configured to acquire an optimal feature population gene for each round of feature screening by using a set analysis method, a fitness function, and an iteration of a genetic algorithm based on the initialized spectral data set and the initialized feature population gene.

[0034] The third acquisition module is configured to acquire an optimal feature population gene set corresponding to an optimal spectral feature information set when the plurality of optimization rounds reaches a set optimization round, and the optimal feature population gene set is composed of the optimal feature population gene for each round of the set optimization round.

[0035] The first data processing module is configured to perform combination statistics and discriminant analysis on the optimal spectral feature information set, and complete the screening of the adaptive characteristic spectral line.

[0036] The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of the atomic emission spectrum-based adaptive characteristic spectral line screening method according to any one of the above embodiments when executing the program.

[0037] According to the atomic emission spectrum-based adaptive characteristic spectral line screening method provided by the present application, the following technical effects are achieved: for different analysis requirements of a sample to be analyzed, a set feature screening optimization method is used to initialize a data set and obtain a corresponding initialized feature population gene; based on the initialized feature population gene, an optimal feature spectral information set is obtained by using a set analysis method, a fitness function, and an iteration of a genetic algorithm; and the characteristic spectral line that is adaptive to the analysis requirement is efficiently and automatically screened from the complex atomic emission spectrum, thereby ensuring the effectiveness and accuracy of the screened characteristic spectral line and providing a reliable guarantee for the accurate and stable implementation of the analysis task. The method is applicable to atomic emission spectrometry of spark light sources, arc light sources, plasma light sources, and laser light sources; and is used to realize adaptive screening of characteristic spectral lines of analysis elements in atomic emission spectrometry. BRIEF DESCRIPTION OF DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0039] Figure 1A flowchart of an atomic emission spectrum-based adaptive characteristic spectral line screening method of the present application is shown in the figure.

[0040] Figure 2 A structural diagram of an atomic emission spectrum-based adaptive characteristic spectral line screening system of the present application is shown in the figure. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0042] The present application aims to provide an atomic emission spectrum-based adaptive characteristic spectral line screening method and system, which can efficiently and automatically screen out characteristic spectral lines that meet the analysis requirements from complex atomic emission spectra, ensuring the effectiveness and accuracy of the screened characteristic spectral lines and providing reliable guarantee for the accurate and stable analysis.

[0043] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, the present application will be further described in detail below with reference to the drawings and specific embodiments.

[0044] As shown in the figure, the atomic emission spectrum-based adaptive characteristic spectral line screening method provided by the present application comprises the following steps: Figure 1 S1, determining a spectrum data set based on an original spectrum signal of a sample to be analyzed;

[0045] The sample to be analyzed can be a standard substance or an experimental sample to be analyzed. The original spectrum signal can be an atomic emission spectrum signal, including a wide-spectrum continuous spectrum signal acquired by an array detector or a narrow-band spectrum signal acquired by a photoelectric detector, such as a photomultiplier tube.

[0046] S2, performing several optimization rounds of characteristic screening on the spectrum data set by using a set characteristic screening optimization method, to obtain an initialization spectrum data set of each round of characteristic screening and an initialization characteristic population gene corresponding to the initialization spectrum data set. The set characteristic screening optimization method can be expressed as: optMtd(gav,dataset,t), wherein gav represents the current genetic algorithm basic parameter and the characteristic screening result, dataset represents the spectrum data set, and t represents the optimization round.

[0047]

[0048] ​S3, obtaining optimal feature population genes of each round based on the initialization spectral dataset and the initialization feature population genes by a set of analysis methods, fitness function and iteration of genetic algorithm;

[0049] S4, obtaining an optimal spectral feature information set corresponding to an optimal feature population gene set composed of optimal feature population genes of a set of optimal rounds when the set of optimal rounds reaches a set of optimization rounds; wherein, decoding the optimal feature population genes in the optimal feature population gene set obtains corresponding optimal spectral feature information, thereby composing the optimal spectral feature information set;

[0050] S5, completing the screening of adaptive characteristic spectral lines by combining and analyzing the optimal spectral feature information set.

[0051] For example, the spectral dataset is determined based on the original spectral signal of the sample to be analyzed, comprising:

[0052] According to the analysis requirements, select samples that meet the element and content coverage range as the sample to be analyzed;

[0053] Perform spectral analysis on the sample to be analyzed to obtain the original spectral signal of the sample to be analyzed; wherein, specifically, a spectral analyzer can be used to perform spectral analysis on the sample to be analyzed;

[0054] Perform spectral preprocessing on the original spectral signal to obtain spectral information; wherein, the spectral preprocessing includes background correction and filtering and denoising;

[0055] Select all or part of the spectral information after spectral preprocessing to construct the spectral dataset.

[0056] For example, the spectral dataset is determined based on the original spectral signal of the sample to be analyzed, comprising:

[0057] The spectral dataset is subjected to a set of optimization rounds of feature screening, and each round of feature screening can select part or all of the spectral information from the spectral dataset to obtain the initialization spectral dataset; wherein, the selection method includes random selection and use of specific selection, and each round of feature screening selects different spectral information from the spectral dataset, which can improve the effectiveness of feature screening;

[0058] Based on the spectral characteristics, the spectral information in the initialization spectral dataset is encoded to obtain un-initialized feature population genes, and the initialization feature population genes are determined based on the un-initialized feature population genes.

[0059] wherein, gene is a concept of genetic algorithm, encoding is to encode the spectral information into gene which can be used and optimized by genetic algorithm, to complete the mapping from phenotype to genotype.

[0060] Further, the determining the initialized feature population gene based on the un-initialized feature population gene specifically comprises:

[0061] In the first round of optimization screening, the un-initialized feature population gene is initialized by using an initialization algorithm to obtain the initialized feature population gene.

[0062] In the second and subsequent rounds of optimization screening, the un-initialized feature population gene is initialized by referring to the optimal feature population gene set and the initialization algorithm to obtain the initialized feature population gene.

[0063] The feature screening optimization method optMtd can obtain the initialized feature population gene according to the optimal feature population gene set.

[0064] For example, the optimal feature population gene of each round is obtained by setting an analysis method, a fitness function and iteration of a genetic algorithm based on the initialized spectral data set and the initialized feature population gene, which comprises:

[0065] According to the analysis requirement, a corresponding analysis method is selected as the set analysis method, which determines the model parameters and evaluation indexes of the analysis method based on the initialized spectral data set and the initialized feature population gene; wherein, the set analysis method is quantitative analysis, semi-quantitative analysis, discriminant analysis or an analysis method that can be represented by modeling.

[0066] The fitness function obtains the population fitness based on the initialized feature population gene and the model parameters and evaluation indexes of the analysis method.

[0067] Wherein, the fitness function can be expressed as fnMtd(fg, k, Mv), wherein fg is the initialized feature population gene, k is the evaluation index of the analysis method, and Mv is the model parameter of the analysis method, so as to obtain the population fitness fitness.

[0068] The genetic algorithm iterates based on the population fitness and the initialized feature population gene to obtain new feature population genes until the iteration number of the genetic algorithm reaches a set maximum value or the population fitness reaches a set fitness threshold, and the last obtained new feature population gene is the optimal feature population gene.

[0069] The adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0070] The adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0071] The adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0072] The adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0073] In summary, the adaptive characteristic spectral line screening method based on atomic emission spectrum provided by the present application can initialize a data set by using a set of characteristic screening optimization methods according to different analysis requirements of a sample to be analyzed, and obtain corresponding initial characteristic population genes. Based on the initial characteristic population genes, an optimized characteristic spectral information set is obtained by using a set of analysis methods, a fitness function and an iterative genetic algorithm. The method can efficiently and automatically screen out characteristic spectral lines that are adaptive to the analysis requirements from complex atomic emission spectrum, thereby ensuring the effectiveness and accuracy of the screened characteristic spectral lines and providing reliable guarantee for accurate and stable analysis. The method is suitable for atomic emission spectrum of spark light sources, arc light sources, plasma light sources and laser light sources, and is used for adaptive screening of characteristic spectral lines of analysis elements in atomic emission spectrum analysis technology.

[0074] The adaptive characteristic spectral line screening method based on atomic emission spectrum provided by the present application can be described in detail below by taking three series of stainless steel waste as examples.

[0075] The three series of stainless steel waste include waste stainless steel 200 series, 300 series and 400 series samples. In this embodiment, the adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0076] The adaptive characteristic spectral line screening method based on atomic emission spectrum comprises the following steps.

[0077] In air environment, laser-induced breakdown spectrometer is used to test the above sample, and after signal acquisition, background correction and normalization by laser spectrum software, pre-processed spectral signal value is obtained, and all spectral information is selected as spectral data set;

[0078] According to the classification analysis requirement, SVM (Support Vector Machine) algorithm is selected as the analysis method to verify the accuracy rate as the analysis method evaluation index parameter ksvm;

[0079] The fitness function fnMtd (fgsvm, ksvm, Msvm) of the multi-dimensional feature spectral line quality evaluation of the genetic algorithm is set and optimized, wherein fgsvm represents the feature population gene, and Msvm represents the analysis method model parameter; thus, the population fitness is fitness;

[0080] The feature screening optimization method optMtd (gav, dataset, t) is established, wherein gav represents the current genetic algorithm basic parameter, including mutation rate, crossover rate, population size, evolution generation number and feature screening result, and t represents the optimization round.

[0081] The feature spectral line optimization screening process includes:

[0082] (1) The feature population gene and the spectral data set are initialized by the feature screening optimization method optMtd;

[0083] (2) The new feature population gene is obtained by the genetic algorithm; the method model parameter and the evaluation index are obtained by the analysis method; and the population fitness fitness is obtained by the fitness function fnMtd;

[0084] The fitness fitness is the evaluation value of the gene quality, and the quality of the feature spectral line corresponding to the gene is determined by the value;

[0085] (3) Until the iteration generation number of the genetic algorithm reaches the maximum or the population fitness reaches the corresponding threshold value, the optimal feature population gene I i of the current round is obtained;

[0086] (4) Step (1) is repeated until the optimization number reaches the corresponding threshold value, and the optimal feature population gene set I = (I1, I2, …, I n ) is obtained;

[0087] (5) The optimal feature population gene set is decoded to obtain the optimized feature spectral information set.

[0088] The obtained optimized characteristic spectrum set is combined and counted to obtain statistical information Q(i), and discriminant analysis is performed to complete adaptive feature screening of the optimal characteristic spectrum set, and thus the characteristic spectrum line of the input spectrum signal is obtained.

[0089] For the 200 series, 300 series and 400 series stainless steel actual samples, the method of the application screens out the spectral lines with relatively strong energy in the spectral atlas database, such as Ni 341.476 nm, Cr 425.435 nm, Si 251.612 nm, Mo 281.615 nm and Cu 327.396 nm, which are also the classic spectral lines commonly screened by artificial experience, and also screens out other characteristic spectral lines, such as 229.749 nm, 338.057 nm and 349.296 nm for the main element Ni, and 357.869 nm, 359.349 nm and 427.48 nm for the main element Cr.

[0090] The embodiment of the application adopts SVM to perform classification modeling and test actual samples, tests 3673 actual stainless steel samples, analyzes classification analysis of different series of 200, 300 and 400 series, and analyzes classification analysis of the grade of 300 series stainless steel, and the results are shown in Tables 1 and 2, the classification accuracy of the characteristic spectral line screened by the method of the application is better than that of the characteristic spectral line screened by artificial experience, the advantage is more obvious in the application of grade subdivision classification, and the effectiveness of the method of the application and the superiority of feature screening for analysis tasks are illustrated.

[0091] Table 1 Classification analysis results of different feature screening methods for 200, 300 and 400 series stainless steel

[0092]

[0093] Table 2 Classification analysis results of different feature screening methods for 300 series stainless steel grade

[0094]

[0095]

[0096] Due to the rich atomic emission spectrum lines of the stainless steel material and the serious interference caused by the matrix, meanwhile, the characteristic element content difference is small and the overlap is serious for the classification analysis of the grade, therefore, the quality and quantity of the characteristic spectrum line selected by artificial experience cannot meet the detection requirements of the stainless steel classification analysis, and the insufficient classification analysis ability of the grade is particularly obvious, as shown in Table 2, the classification result accuracy is obviously weaker than that of the spectrum line selected by the adaptive characteristic screening, and in terms of the effectiveness of the characteristic spectrum line, the verification and screening process of the artificial experience characteristic spectrum line is also very tedious and inefficient, and it is difficult to converge to the optimal, which is not conducive to providing reliable and effective spectrum line information for the analysis method.

[0097] The embodiment of the present application selects an analysis method model for different analysis requirements, introduces a characteristic analysis evaluation, establishes a screening optimization method and an optimal spectrum screening method, performs spectrum analysis on the selected sample, obtains an atomic emission spectrum signal and performs adaptive characteristic spectrum line screening, so as to obtain the characteristic spectrum information required for analysis.

[0098] As shown in Figure 2 The present application also provides an adaptive characteristic spectrum line screening system based on atomic emission spectrum, which is applied to the adaptive characteristic spectrum line screening method based on atomic emission spectrum of any one of the above, and comprises:

[0099] A first determination module 201 is configured to determine a spectrum data set based on an original spectrum signal of a sample to be analyzed;

[0100] A first acquisition module 202 is configured to perform characteristic screening on the spectrum data set for a plurality of optimization rounds by using a set characteristic screening optimization method, and acquire an initialization spectrum data set of each round of characteristic screening and an initialization characteristic population gene corresponding to the initialization spectrum data set;

[0101] A second acquisition module 203 is configured to acquire an optimal characteristic population gene of each round by using a set analysis method, an adaptability function and an iteration of a genetic algorithm based on the initialization spectrum data set and the initialization characteristic population gene;

[0102] A third acquisition module 204 is configured to acquire an optimal characteristic population gene set corresponding to an optimal spectrum characteristic information set when the plurality of optimization rounds reaches a set optimization round, the optimal characteristic population gene set being composed of the optimal characteristic population gene of each round in the set optimization round;

[0103] A first data processing module 205 is configured to perform combination statistics and discriminant analysis on the optimal spectrum characteristic information set, and complete the screening of the adaptive characteristic spectrum line.

[0104] The application further provides an electronic device, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the atomic emission spectrum-based adaptive characteristic spectral line screening method according to any one of the above.

[0105] The application further provides a non-transitory computer-readable storage medium, which stores a computer program, wherein the computer program is executable by a processor to implement the steps of the atomic emission spectrum-based adaptive characteristic spectral line screening method according to any one of the above.

[0106] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, i.e., may be located in one place, or may be distributed on multiple network units. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0107] From the above description of the embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and necessary universal hardware platforms, and of course can also be realized by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0108] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.​​​​

Claims

1. An adaptive characteristic spectral line screening method based on atomic emission spectroscopy, characterized in that, The method comprises the following steps: determining a spectral data set based on original spectral signals of a sample to be analyzed; performing feature screening on the spectral data set for several optimization rounds using a set feature screening optimization method to obtain an initialized spectral data set and an initialized feature population gene corresponding to the initialized spectral data set for each round of feature screening, including: performing feature screening on the spectral data set for several optimization rounds, and selecting part or all of the spectral information from the spectral data set for each round of feature screening to obtain the initialized spectral data set; encoding the spectral information in the initialized spectral data set based on spectral characteristics to obtain an uninitialized feature population gene, and determining the initialized feature population gene based on the uninitialized feature population gene; obtaining the optimal feature population gene for each round through a set analysis method, an adaptive function, and iteration of a genetic algorithm based on the initialized spectral data set and the initialized feature population gene, including: selecting a corresponding analysis method as the set analysis method according to the analysis requirement, and determining the model parameters and evaluation indexes of the analysis method based on the initialized spectral data set and the initialized feature population gene; obtaining the population fitness based on the initialized feature population gene and the model parameters and evaluation indexes of the analysis method, and the adaptive function is expressed as fnMtd(fg, k, Mv), wherein fg is the initialized feature population gene, k is the evaluation index of the analysis method, Mv is the model parameter of the analysis method, and the population fitness is fitness; iterating the genetic algorithm based on the population fitness and the initialized feature population gene to obtain a new feature population gene until the iteration number of the genetic algorithm reaches a set maximum value or the population fitness reaches a set fitness threshold, and the new feature population gene obtained finally is the optimal feature population gene; when the several optimization rounds reach a set optimization round, an optimal spectral feature information set corresponding to a set optimal feature population gene set composed of the optimal feature population genes for each round in the set optimization round is obtained; performing combination statistics and discriminant analysis on the optimal spectral feature information set to complete the screening of adaptive feature spectral lines, including: performing one or more of probability analysis and frequency analysis on the optimal spectral feature information set to obtain statistical information of each feature spectral line; through discriminant analysis, if the evaluation value corresponding to the statistical information of each feature spectral line is greater than a set screening threshold, the screening of the adaptive feature spectral lines of the optimal spectral feature information set is completed.

2. The method of claim 1, wherein, The method comprises the following steps: selecting a sample meeting the element and content coverage range as the sample to be analyzed according to the analysis requirement; performing spectral analysis on the sample to be analyzed to obtain the original spectral signals of the sample to be analyzed; performing spectral pretreatment on the original spectral signals to obtain spectral information; selecting all or part of the spectral information after spectral pretreatment to construct the spectral data set.

3. The method of claim 2, wherein, The spectral preprocessing includes background correction and filter denoising.

4. The method of claim 1, wherein, The original spectral signal is an atomic emission spectral signal, including a wide-band continuous spectral signal acquired by an array detection device or a narrow-band spectral signal acquired by a photoelectric detection device.

5. The method of claim 1, wherein, The set analysis method is quantitative analysis, semi-quantitative analysis, discriminant analysis or an analysis method that can be characterized by modeling.

6. An adaptive characteristic spectral line screening system based on atomic emission spectrum, applied to the adaptive characteristic spectral line screening method based on atomic emission spectrum in any one of claims 1 to 5, characterized in that, The method comprises the following steps: The first determining module is configured to determine a spectral data set based on an original spectral signal of a sample to be analyzed; The first obtaining module is configured to perform feature screening of the spectral data set for a plurality of optimization rounds by using a set feature screening optimization method, to obtain an initialization spectral data set of each round of feature screening and an initialization feature population gene corresponding to the initialization spectral data set; The second obtaining module is configured to obtain an optimal feature population gene of each round by using a set analysis method, a fitness function and an iteration of a genetic algorithm based on the initialization spectral data set and the initialization feature population gene; The third obtaining module is configured to obtain an optimal feature population gene set corresponding to an optimal spectral feature information set when the plurality of optimization rounds reaches a set optimization round, the optimal feature population gene set being composed of the optimal feature population gene of each round in the set optimization round; The first data processing module is configured to perform combination statistics and discriminant analysis on the optimal spectral feature information set, to complete the screening of adaptive characteristic spectral lines.

7. An electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, The processor executes the program to implement the steps of the atomic emission spectral-based adaptive characteristic spectral line screening method according to any one of claims 1 to 5.

Images