Plant beverage flavor component detection system and method based on artificial intelligence

Abstract

The invention provides a plant beverage flavor component detection system and method based on artificial intelligence, and the method comprises the steps: analyzing the sensory contribution spectrum of a target plant beverage sample through a multi-dimensional original data matrix of the flavor substance content in the target plant beverage sample; extracting principal component vectors of flavor substances in the target plant beverage sample from the multi-dimensional original data matrix, and determining key flavor markers of flavor components in the target plant beverage sample in combination with a sensory contribution spectrum; extracting key spectrum fingerprint characteristics of flavor components in the target plant beverage sample from a spectrum detection result of the target beverage sample based on the key flavor marker; and constructing a flavor component prediction model of the target plant beverage sample, and further determining a flavor score of the flavor component in the target plant beverage sample in combination with the key spectrum fingerprint characteristics. By adopting the scheme of the invention, the specific contribution of each substance to the overall sensory attribute can be analyzed so as to identify the key flavor substance in the flavor evaluation of the traditional plant beverage.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of flavor component detection, and more particularly to a plant beverage flavor component detection system and method based on artificial intelligence. BACKGROUND

[0002] Flavor component detection is a key technology in the cross field of food science, sensory analysis, chemistry and artificial intelligence, aiming to identify, quantify and analyze the chemical components in food and beverage that affect the sensory experience (aroma, taste, mouthfeel), and the core goal is to establish an objective link between chemical components and human sensory perception, to realize the scientific, digital and intelligent evaluation of flavor.

[0003] With the rapid development of the plant beverage market, consumers' demand for product flavor diversification and personalization is increasing, and flavor has become a core indicator of plant beverage quality and market competitiveness. The flavor of plant beverages is determined by a variety of volatile substances (such as aroma components such as esters, alcohols, aldehydes) and non-volatile substances (such as taste components such as sugars, organic acids, amino acids). However, plant beverages contain hundreds or even thousands of volatile (such as aroma) and non-volatile (such as taste, mouthfeel) compounds, and the contributions of these substances to the overall flavor are very different, some are key, some are background or interference. Traditional chemical analysis methods (such as chromatography-mass spectrometry) can detect a large number of substances, but it is difficult to efficiently and accurately screen out the core flavor substances that really play a decisive role in sensory attributes (such as "rich floral aroma", "mellow mouthfeel", "balanced bitterness") from them. Therefore, how to analyze the specific contribution of each substance to the overall sensory attribute to identify the key flavor substances in traditional plant beverage flavor evaluation has become a problem faced by the industry. SUMMARY

[0004] The present application provides a plant beverage flavor component detection system and method based on artificial intelligence, which can analyze the specific contribution of each substance to the overall sensory attribute to identify the key flavor substances in traditional plant beverage flavor evaluation.

[0005] In a first aspect, the present application provides a plant beverage flavor component detection method based on artificial intelligence, comprising the following steps: detecting the content of each flavor substance in the target plant beverage sample to obtain a multi-dimensional original data matrix of the flavor substance content in the target plant beverage sample; Based on the sensory evaluation data of the target plant beverage sample, the multi-dimensional original data matrix is analyzed to obtain a sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to overall sensory attributes of the target plant beverage sample, and principal component analysis is performed on the multi-dimensional original data matrix to obtain a principal component vector of the flavor substances in the target plant beverage sample, which is used to reflect a synergistic change characteristic among the flavor substances. By the sensory contribution spectrum and the synergistic change characteristic among the flavor substances reflected by the principal component vector, contribution degree analysis is performed on core ingredients affecting the flavor of the plant beverage in the target plant beverage sample to obtain key flavor markers of the flavor ingredients in the target plant beverage sample. Spectrum detection is performed on the target plant beverage sample, and key spectral fingerprint characteristics of the flavor ingredients in the target plant beverage sample are extracted from the spectrum detection result based on the key flavor markers. A flavor ingredient prediction model of the target plant beverage sample is constructed according to historical spectrum data of the plant beverage, and the key spectral fingerprint characteristics are input into the flavor ingredient prediction model to output a flavor score of the flavor ingredients in the target plant beverage sample.

[0006] In some embodiments, based on the sensory evaluation data of the target plant beverage sample, the multi-dimensional original data matrix is analyzed to obtain a sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to overall sensory attributes of the target plant beverage sample, specifically including: Obtaining sensory evaluation data of the target plant beverage sample; According to the sensory evaluation data and the multi-dimensional original data matrix, correlation coefficients between volatile and non-volatile flavor substances and each sensory attribute are calculated; According to all the correlation coefficients, a sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to overall sensory attributes of the target plant beverage sample is constructed.

[0007] In some embodiments, the multi-dimensional original data matrix is processed by principal component analysis to obtain a principal component vector of the flavor substances in the target plant beverage sample, specifically including: The multi-dimensional original data matrix is standardized to obtain a standardized multi-dimensional original data matrix; Eigenvalue calculation is performed on the standardized multi-dimensional original data matrix by a principal component analysis algorithm to obtain eigenvalue data of the flavor substances in the target plant beverage sample; A characteristic threshold of the flavor substances in the target plant beverage sample is determined; The eigenvalue data is judged according to the characteristic threshold, and a principal component vector of the flavor substances in the target plant beverage sample is obtained.

[0008] In some embodiments, the key flavor markers of the flavor components in the target plant beverage sample are obtained by analyzing the contribution of the core components affecting the flavor of the plant beverage in the target plant beverage sample based on the synergistic variation characteristics between the flavor substances reflected by the sensory contribution spectrum and the principal component vector, specifically including: extracting a plurality of high-contribution flavor substances affecting the flavor of the plant beverage from the target plant beverage sample according to the sensory contribution spectrum; a plurality of core flavor substances screened from the principal component load of the principal component vector; performing intersection analysis on each high-contribution flavor substance and each core flavor substance to obtain a plurality of candidate core components of the target plant beverage sample; calculating the contribution score of each candidate core component affecting the flavor of the plant beverage, and further obtaining the key flavor markers of the flavor components in the target plant beverage sample.

[0009] In some embodiments, the key spectral fingerprint characteristics of the flavor components in the target plant beverage sample are extracted from the results of spectral detection based on the key flavor markers, specifically including: performing spectral detection on the target plant beverage sample to obtain the results of spectral detection; performing denoising processing on the results of spectral detection to obtain the denoised spectral detection results; extracting the characteristic spectral interval, absorption peak intensity and peak shape change highly related to the content change of the key flavor markers from the denoised spectral detection results based on the key flavor markers; taking the characteristic spectral interval, absorption peak intensity and peak shape change as the key spectral fingerprint characteristics of the flavor components in the target plant beverage sample.

[0010] In some embodiments, the flavor component prediction model of the target plant beverage sample is constructed according to the historical spectral data of the plant beverage, specifically including: obtaining the historical spectral data of the plant beverage; dividing the historical spectral data into a training set and a validation set; constructing the flavor component prediction model of the target plant beverage sample based on an artificial intelligence algorithm combined with the training set and the validation set.

[0011] In some embodiments, the key spectral fingerprint characteristics are input into the flavor component prediction model to output the flavor score of the flavor components in the target plant beverage sample, specifically including: standardizing the key spectral fingerprint characteristics, and inputting the standardized key spectral fingerprint characteristics into the flavor component prediction model; The standardized key spectral fingerprint features inputted are calculated by the flavor ingredient prediction model to output a flavor score of the flavor ingredients in the target plant beverage sample.

[0012] In some embodiments, the content of each flavor substance in the target plant beverage sample is detected by using gas chromatography-mass spectrometry and high-performance liquid chromatography instrument analysis methods.

[0013] In some embodiments, the target plant beverage sample is detected by near-infrared spectroscopy.

[0014] In a second aspect, the present application provides a plant beverage flavor ingredient detection system based on artificial intelligence, which comprises: The detection module is configured to detect the content of each flavor substance in the target plant beverage sample to obtain a multi-dimensional original data matrix of the content of the flavor substances in the target plant beverage sample. The processing module is configured to analyze the sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample based on the sensory evaluation data of the target plant beverage sample and the multi-dimensional original data matrix, and to perform dimension reduction processing on the multi-dimensional original data matrix by principal component analysis to obtain a principal component vector of the flavor substances in the target plant beverage sample, the principal component vector being used to reflect the synergistic change characteristics between the flavor substances. The processing module is further configured to analyze the contribution degree of the core ingredients affecting the flavor of the plant beverage in the target plant beverage sample by the sensory contribution spectrum and the synergistic change characteristics between the flavor substances reflected by the principal component vector to obtain key flavor markers of the flavor ingredients in the target plant beverage sample. The processing module is further configured to perform spectral detection on the target plant beverage sample, and extract key spectral fingerprint features of the flavor ingredients in the target plant beverage sample from the results of the spectral detection based on the key flavor markers. The execution module is configured to construct a flavor ingredient prediction model of the target plant beverage sample according to historical spectral data of the plant beverage, and input the key spectral fingerprint features into the flavor ingredient prediction model to output a flavor score of the flavor ingredients in the target plant beverage sample.

[0015] The technical scheme provided by the embodiments disclosed in the present application has the following beneficial effects: The plant beverage flavor ingredient detection system and method based on artificial intelligence provided in the application first detects the content of each flavoring substance in the target plant beverage sample to obtain a multi-dimensional original data matrix of the flavoring substance content in the target plant beverage sample; based on the sensory evaluation data of the target plant beverage sample and the multi-dimensional original data matrix, the sensory contribution spectrum of the volatile and non-volatile flavoring substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample is analyzed, the multi-dimensional original data matrix is processed by dimension reduction using principal component analysis to obtain the principal component vector of the flavoring substances in the target plant beverage sample; the contribution degree of the core ingredients affecting the flavor of the plant beverage in the target plant beverage sample is analyzed through the sensory contribution spectrum combined with the synergistic change characteristics of the flavoring substances reflected by the principal component vector, and the key flavor markers of the flavor ingredients in the target plant beverage sample are obtained; the target plant beverage sample is subjected to spectral detection, and the key spectral fingerprint characteristics of the flavor ingredients in the target plant beverage sample are extracted from the results of spectral detection based on the key flavor markers; a flavor ingredient prediction model of the target plant beverage sample is constructed according to the historical spectral data of the plant beverage, and the key spectral fingerprint characteristics are input into the flavor ingredient prediction model to output the flavor score of the flavor ingredients in the target plant beverage sample.

[0016] It can be seen that, in the process of detecting the flavor ingredients of the plant beverage, the application first comprehensively detects the content of the flavoring substances to obtain a multi-dimensional original data matrix, thereby providing a solid material content data basis for subsequent analysis and ensuring that the analysis is based on real and comprehensive material composition information; secondly, the sensory evaluation data and the original data matrix are combined to obtain a sensory contribution spectrum, thereby directly establishing the correlation between the content of the flavoring substances and the overall sensory attributes, quantifying the specific sensory contribution of the volatile and non-volatile flavoring substances, and clearly defining the influence weight of different substances on the sensory experience; the multi-dimensional data is reduced by dimension reduction using principal component analysis to obtain a principal component vector, thereby simplifying the complex data structure, eliminating redundant information, highlighting the core influencing factors, and reducing analysis interference; the core ingredients affecting the flavor are accurately locked through the joint analysis of the sensory contribution spectrum and the principal component vector, the key flavor markers are successfully obtained, and the accurate identification of the key flavoring substances is realized; the spectral fingerprint characteristics are extracted based on the key flavor markers, thereby establishing the correspondence between the key substances and the spectral signals and providing specific characteristic basis for rapid detection; finally, the prediction model is constructed by combining the artificial intelligence algorithm and the historical spectral data, and the flavor score is output, thereby realizing the goal of rapidly and accurately predicting the flavor using spectral technology and improving the efficiency and objectivity of flavor evaluation. The above scheme can be used to analyze the specific contribution of each substance to the overall sensory attributes to identify the key flavoring substances in the traditional plant beverage flavor evaluation. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1is an exemplary flow chart of an artificial intelligence-based plant beverage flavor ingredient detection method according to some embodiments of the present application; Figure 2 is an exemplary flow chart of determining a sensory contribution profile according to some embodiments of the present application; Figure 3 is an exemplary flow chart of determining a key spectral fingerprint feature according to some embodiments of the present application; Figure 4 is a structural schematic diagram of an artificial intelligence-based plant beverage flavor ingredient detection system according to some embodiments of the present application; Figure 5 is a structural schematic diagram of a computer device implementing an artificial intelligence-based plant beverage flavor ingredient detection method according to some embodiments of the present application. DETAILED DESCRIPTION

[0018] In order to better understand the technical solutions of the present application, the technical solutions of the present application will be described in detail below in combination with the drawings of the specification and specific embodiments.

[0019] Reference Figure 1 The figure is an exemplary flow chart of an artificial intelligence-based plant beverage flavor ingredient detection method according to some embodiments of the present application, which mainly includes the following steps: In step 101, the content of each flavor substance in the target plant beverage sample is detected to obtain a multi-dimensional original data matrix of the content of the flavor substance in the target plant beverage sample.

[0020] In specific implementation, first, representative target plant beverage samples are selected, and gas chromatography-mass spectrometry, high-performance liquid chromatography instrument analysis methods are used to qualitatively and quantitatively detect volatile flavor substances (such as aldehydes, esters, terpenes) and non-volatile flavor substances (such as sugars, organic acids, amino acids) in the target plant beverage samples, record the specific content values of each detected flavor substance, and then organize these data into a multi-dimensional original data matrix containing multiple samples, multiple flavor substances and their corresponding contents according to the three-dimensional relationship of “sample-flavor substance-content”, wherein each row in the multi-dimensional original data matrix represents a sample, each column represents a flavor substance, and the matrix element is the content of the corresponding flavor substance in the sample. For example, a detection result containing 5 samples and 8 flavor substances will form a 5-row 8-column multi-dimensional original data matrix, and the value in the 3rd row and the 4th column of the matrix represents the content of the 4th flavor substance in the 3rd sample. The target plant beverage sample contains multiple samples.

[0021] In step 102, based on the sensory evaluation data of the target plant beverage sample, the sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample is analyzed in combination with the multi-dimensional original data matrix, and principal component analysis is used to reduce the dimension of the multi-dimensional original data matrix to obtain the principal component vector of the flavor substances in the target plant beverage sample, which is used to reflect the synergistic change characteristics between the flavor substances.

[0022] In some embodiments, the reference Figure 2 As shown in the figure, which is an exemplary flowchart for determining the sensory contribution spectrum in some embodiments of the present application, the analysis of the sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample based on the sensory evaluation data of the target plant beverage sample in combination with the multi-dimensional original data matrix can be achieved by the following steps: First, in step 1021, the sensory evaluation data of the target plant beverage sample is obtained; Second, in step 1022, the correlation coefficients between the volatile and non-volatile flavor substances and the sensory attributes are calculated according to the sensory evaluation data and the multi-dimensional original data matrix; Finally, in step 1023, the sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample is constructed according to all the correlation coefficients.

[0023] Among them, the sensory evaluation data of the target plant beverage sample is extracted from the sensory evaluation database corresponding to the target plant beverage sample, wherein the sensory evaluation data represents the quantitative data of the overall sensory attributes of the target plant beverage sample, and the subjective perception of the human sensory system to the flavor of the beverage is converted into objective data that can be analyzed, including two aspects: one is the evaluation index system, covering the core sensory attributes of aroma intensity, sweetness, acidity, and fullness, which directly reflect the flavor characteristics of the beverage in terms of olfaction, taste, and mouthfeel; the second is the specific data results based on the evaluation index, which are generated by professional evaluators through blind scoring (such as 1-9 point system) or fuzzy mathematical evaluation method to form standardized scoring data for subsequent correlation analysis of the relationship between flavor substances and sensory experience; in other embodiments, other ways can also be used, which are not limited here.

[0024] In a specific implementation, the correlation coefficient between volatile flavor substances, non-volatile flavor substances and each sensory attribute can be calculated according to the sensory evaluation data and the multi-dimensional original data matrix in the following manner: taking the content data of volatile flavor substances and non-volatile flavor substances in the multi-dimensional original data matrix as independent variables, taking the scores of each sensory attribute in the sensory evaluation data as dependent variables, and using partial least squares regression or Pearson correlation analysis to calculate the correlation coefficient between each flavor substance and each sensory attribute. The correlation coefficient represents the degree of correlation between the flavor substance and the sensory attribute. The greater the absolute value of the coefficient, the stronger the correlation. A positive or negative sign is also marked to reflect the direction of the correlation (positive or negative correlation). In other embodiments, other methods can also be used for calculation, which are not limited here.

[0025] In addition, in a specific implementation, the sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample can be constructed according to all the correlation coefficients in the following manner: arranging the calculated correlation coefficients of each type of flavor substance and each sensory attribute, and constructing a matrix diagram with flavor substance categories (volatile / non-volatile) as the horizontal axis and sensory attributes as the vertical axis. The contribution intensity of different substances to each attribute is directly reflected by color gradient (such as heat map) or numerical annotation. The substance and attribute position corresponding to the high absolute value coefficient is displayed with dark color or highlighted marker, forming a sensory contribution spectrum that can clearly reflect the correlation strength of “flavor substance-sensory attribute”. In other embodiments, other methods can also be used for construction, which are not limited here.

[0026] It should be noted that the sensory contribution spectrum in this application represents the specific influence degree of volatile and non-volatile flavor substances on the overall sensory attributes of beverage aroma intensity, sweetness, acidity, etc. The contribution intensity of different substances is directly reflected by color gradient or numerical annotation (high absolute value coefficient corresponds to dark color or highlighted marker), which can be used to locate the key flavor driving factor and determine which substances play a major role in specific sensory characteristics (such as sweetness, aroma), providing a targeted target for beverage flavor optimization.

[0027] In some embodiments, the principal component analysis is used to reduce the dimension of the multi-dimensional original data matrix to obtain the principal component vector of the flavor substances in the target plant beverage sample in the following steps: The multi-dimensional original data matrix is standardized to obtain a standardized multi-dimensional original data matrix. Eigenvalue calculation is performed on the standardized multi-dimensional original data matrix using a principal component analysis algorithm to obtain eigenvalue data of the flavor substances in the target plant beverage sample. A characteristic threshold of the flavor substances in the target plant beverage sample is determined. The characteristic value data is judged according to the characteristic threshold value, and a principal component vector of the flavor substance in the target plant beverage sample is obtained.

[0028] In a specific implementation, first, in view of the dimension difference of the content data of different flavor substances in the multidimensional original data matrix, the Z-score standardization method is used to calculate the mean and standard deviation of the content of each flavor substance, the original data is converted into standardized numerical values with a mean of 0 and a standard deviation of 1, and finally a standardized multidimensional original data matrix eliminating the dimension influence is obtained; second, the standardized multidimensional original data matrix is imported into the statistical product and service solutions or the data analysis tool of the R language, a principal component analysis function is called to perform eigenvalue decomposition or singular value decomposition on the standardized multidimensional original data matrix, and the characteristic value corresponding to each principal component, the characteristic vector (that is, the principal component loading coefficient, reflecting the correlation strength between each flavor substance and the principal component) and the score of each sample on each principal component are calculated. The set of all characteristic values is taken as the characteristic value data of the flavor substance in the target plant beverage sample, wherein the characteristic value in the characteristic value data reflects the explanation ability of the principal component to the total variation of the data, and is a core index for measuring the importance of the principal component. The greater the characteristic value is, the stronger the explanation ability of the principal component to the overall variation of the data is, and the richer the information amount of the original flavor substance content data contained is. In other embodiments, other ways can also be used for implementation, which are not limited here.

[0029] In addition, the specific implementation of determining the characteristic threshold value of the flavor substance in the target plant beverage sample is as follows: based on the characteristic value data obtained by the principal component analysis, the classical criterion of “characteristic value greater than 1” (that is, the principal component with a characteristic value greater than or equal to 1 is retained, because its explanation ability exceeds the average level of a single original variable) or the cumulative variance contribution rate threshold (such as setting the cumulative variance contribution rate to be greater than or equal to 85%, to ensure that the main information of the data is retained) is used, a scree plot (the horizontal axis is the principal component number, and the vertical axis is the characteristic value) is drawn to intuitively judge the downward trend of the characteristic value, so as to determine the characteristic threshold value for screening the core principal component, wherein the characteristic threshold value represents the quantitative limit of whether the principal component has sufficient information retention value, and can be used as a critical judgment standard for screening the core principal component. In other embodiments, other ways can also be used for determination, which are not limited here.

[0030] In addition, the specific implementation of judging the feature value data according to the feature threshold value to obtain the principal component vector of the flavor substance in the target plant beverage sample is as follows: comparing the feature value data with the set feature threshold value, screening out the principal components with feature values greater than the feature threshold value as core principal components, extracting the flavor substances corresponding to the core principal components (i.e., the substances that significantly contribute to the core principal components), and then combining the score distribution of the sample on the core principal components to integrate to form a principal component vector with the core principal components as dimensions and the key flavor substances as cores. The principal component vector specifically includes three aspects: first, the core principal components, i.e., the principal components (such as the first two to three principal components) screened out by the feature threshold value and having feature values greater than the threshold value. These principal components collectively carry the main variation information (usually cumulative variance contribution rate ≥ 85%) of the original flavor substance data, and each principal component represents a core flavor dimension (such as "aroma dominant dimension" and "sweet and sour taste dimension"). Second, the key flavor substances, i.e., the flavor substances with high absolute values of loading coefficients corresponding to the core principal components. They are the core contributors to the core flavor dimensions and directly determine the flavor characteristics of the dimensions (such as aldehydes and ester substances with high loadings in a core principal component representing aroma-related characteristics). Third, the sample score distribution, i.e., the score data of each sample on the core principal components. The score distribution can directly reflect the differences or commonalities of different samples on the core flavor dimensions, helping to distinguish the flavor characteristic types of the samples. In other embodiments, other ways can be used to determine the principal component vector, which is not limited here.

[0031] It should be noted that the principal component vector in the present application reflects the synergistic change characteristics between flavor substances. This vector not only retains the main distribution characteristics of flavor substances in the original data, but also directly reflects the flavor differences and commonalities between samples.

[0032] It should be noted that the principal component vector is a comprehensive feature vector obtained by reducing the dimension of a multi-dimensional original data matrix through principal component analysis. Its core function is to capture the most representative change pattern in the flavor substance content data, thereby directly reflecting the synergistic change characteristics between different flavor substances. For example, when some volatile aroma components (such as esters and alcohols) and non-volatile taste components (such as sugars and organic acids) show a significant synchronous increasing or decreasing trend in content change, the principal component vector will condense this correlation into a few comprehensive dimensions through feature values and loadings. This vector not only retains the inherent correlation information between flavor substances in the original data, but also eliminates the interference of redundant noise, so that subsequent analysis can more clearly focus on the core component combinations that truly drive the flavor synergistic change.

[0033] In step 103, the key flavor markers of the flavor components in the target plant beverage sample are obtained by analyzing the contribution of the core components affecting the flavor of the plant beverage in the target plant beverage sample according to the synergistic change characteristics between the flavor substances reflected by the sensory contribution spectrum combined with the principal component vector.

[0034] In some embodiments, the key flavor markers of the flavor components in the target plant beverage sample can be obtained by analyzing the contribution of the core components affecting the flavor of the plant beverage in the target plant beverage sample according to the synergistic change characteristics between the flavor substances reflected by the sensory contribution spectrum combined with the principal component vector by the following steps: extracting a plurality of high-contribution flavor substances affecting the flavor of the plant beverage from the target plant beverage sample according to the sensory contribution spectrum; screening a plurality of core flavor substances from the target plant beverage sample according to the synergistic change characteristics between the flavor substances reflected by the principal component vector; performing intersection analysis on each high-contribution flavor substance and each core flavor substance to obtain a plurality of candidate core components of the target plant beverage sample; calculating the contribution score of each candidate core component affecting the flavor of the plant beverage, and further obtaining the key flavor markers of the flavor components in the target plant beverage sample.

[0035] It should be noted that the sensory contribution spectrum clearly quantifies the specific contribution of different flavor substances (including volatile and non-volatile) to the overall sensory properties (such as sweetness, acidity, aroma intensity, etc.) of the plant beverage, and determines which substances are directly related to the sensory experience. The principal component vector captures the synergistic change characteristics between the flavor substances, revealing the inherent correlation patterns between the substances (such as which substances always change synchronously, which substances are in an inverse correlation). When the two are combined for core component contribution analysis, first, the candidate substances that have a substantial impact on the sensory experience are locked by using the sensory contribution spectrum, and then the positions of these candidate substances in the synergistic change are analyzed by using the principal component vector. Those substances that are in a strong synergistic combination and have a significant contribution to the sensory experience will be identified as the core components driving the overall flavor characteristics. By quantitatively sorting the contribution of these core components, the substances that are most representative and can stably reflect the flavor characteristics of the beverage are finally screened as the key flavor markers.

[0036] In the specific implementation, the plurality of high-contribution flavoring substances affecting the flavor of the target plant beverage sample are extracted from the sensory contribution spectrum according to the following procedure: based on the correlation intensity matrix of the “flavoring substance-sensory attribute” in the sensory contribution spectrum, flavoring substances with a high correlation coefficient absolute value for each sensory attribute (such as aroma, sweetness, acidity, etc.) are screened out (for example, 30% of the flavoring substances with the highest correlation coefficient absolute value are selected), and these flavoring substances have a significant contribution to the flavor of the beverage in the sensory perception level. Each of the screened flavoring substances is a high-contribution flavoring substance, and the high-contribution flavoring substance represents a flavoring substance that has a significant impact on each sensory attribute of the plant beverage in the sensory contribution spectrum. In other embodiments, other methods can also be used for extraction, which are not limited here.

[0037] In the specific implementation, the plurality of core flavoring substances screened from the target plant beverage sample according to the synergistic change characteristics of the flavoring substances reflected by the principal component vectors are as follows: first, flavoring substances with a significant contribution to the principal component are screened out according to the load value and sign of each flavoring substance in the principal component vector (usually, flavoring substances with the highest load absolute value are selected), and these substances are potential carriers of the synergistic change characteristics; then, the sign consistency of these substances in the principal component vector is analyzed. If the load signs of multiple substances are the same, it indicates that they have a synergistic increase-decrease trend in content change, that is, the flavoring substances have a positive synergistic change characteristic through synchronous fluctuation, and the opposite sign indicates a negative synergistic change characteristic, which represents a dynamic balance relationship of this consumes that. Finally, in combination with the proportion of total variation explained by different principal components, a substance combination that has a strong synergistic characteristic in a principal component with a high explanation degree is preferentially selected. The synergistic change characteristic of this combination can more accurately reflect the overall fluctuation rule of the flavor of the plant beverage, so as to determine the core flavoring substance that drives the overall flavor characteristic of the plant beverage. The core flavoring substance represents a flavoring substance that has a significant contribution to the core principal component in the target plant beverage sample. In other embodiments, other methods can also be used for screening, which are not limited here.

[0038] In the specific implementation, the plurality of candidate core components of the target plant beverage sample are obtained by performing intersection analysis on the high-contribution flavoring substances and the core flavoring substances. The flavoring substances that appear in both sets are screened out by comparing the substance names or chemical identifiers of the high-contribution flavoring substances and the core flavoring substances. These flavoring substances have a significant impact on the sensory attributes in the sensory contribution spectrum and are key components of the core principal component in the principal component vector, so as to be determined as candidate core components, and then a plurality of candidate core components are obtained. The candidate core component refers to a flavoring substance that has both high sensory contribution and core chemical characteristics and is screened out by integrating the sensory contribution spectrum and the principal component vector. In other embodiments, other methods can also be used for analysis, which are not limited here.

[0039] Among them, the specific implementation method of calculating the contribution score of each candidate core component to the flavor of the plant beverage, and then obtaining the key flavor markers of the flavor components in the target plant beverage sample is: using weighted summation and other algorithms (such as weighting the sensory contribution weight and the principal component load coefficient according to a preset ratio) to calculate the comprehensive contribution score of each candidate core component, and after sorting them from high to low according to the score, select flavor substances with a score exceeding the set threshold (such as the top 20% or a score ≥0.6). These flavor substances perform outstandingly in both the "sensory impact" and "chemical characteristics" dimensions, and are ultimately determined as key flavor markers; other calculation methods can also be used in other embodiments, which are not limited here.

[0040] It should be noted that the key flavor markers in this application refer to the characteristic flavor substances that have a decisive influence on the flavor and are screened from the flavor components of the target plant beverage samples. They can characterize the core flavor characteristics of the beverage and accurately associate the sensory experience with the chemical composition. They are the core indicators for analyzing the flavor formation mechanism, optimizing product flavor and constructing flavor detection models.

[0041] In step 104 , a spectral detection is performed on the target plant beverage sample, and key spectral fingerprint features of flavor components in the target plant beverage sample are extracted from the spectral detection results based on the key flavor markers.

[0042] In some embodiments, reference Figure 3 As shown in FIG. 1 , this figure is an exemplary flow chart for determining key spectral fingerprint features in some embodiments of the present application. In this embodiment, a target plant beverage sample is subjected to spectral detection, and the key spectral fingerprint features of the flavor components in the target plant beverage sample are extracted from the spectral detection results based on the key flavor markers. This can be achieved by using the following steps: First, in step 1041, a spectral detection is performed on the target plant beverage sample to obtain a spectral detection result; Next, in step 1042, the spectrum detection result is subjected to denoising processing to obtain a denoised spectrum detection result; Then, in step 1043, based on the key flavor marker, characteristic spectral intervals, absorption peak intensities, and peak shape changes that are highly correlated with changes in the key flavor marker content are extracted from the denoised spectral detection results; Finally, in step 1044, the characteristic spectral interval, absorption peak intensity and peak shape change are used as key spectral fingerprint features of the flavor components in the target plant beverage sample.

[0043] It should be noted that as a specific chemical substance, the functional groups in the molecular structure of the key flavor marker will produce characteristic absorption, scattering or vibration signals (such as absorption peaks of specific wavelengths) in spectral detection, and these signals have quantitative correlation with the content of the marker; because the molecular structure of different chemical substances is different, it will present a unique "fingerprint" in the spectrum, by focusing on the characteristic spectral range of the known key flavor marker, the signals directly related to these markers can be screened out from the complex overall spectrum, excluding the interference of other ingredients, so as to extract the key spectral fingerprint characteristics that can specifically reflect the target flavor ingredients, and establish the direct correlation between the spectral signal and the flavor ingredients.

[0044] In the method, the spectral detection of the target plant beverage sample is performed to obtain a result, and the specific implementation manner of the spectral detection is as follows: an appropriate amount of the target plant beverage sample is taken and placed in a quartz cuvette, an appropriate spectral detection technology (such as near-infrared spectroscopy, mid-infrared spectroscopy or Raman spectroscopy) is selected according to the chemical properties of the key flavor marker, reasonable detection parameters (including wavelength range, scanning times, resolution, etc.) are set, a spectrometer is used to scan the sample, the absorbance, transmittance or scattering intensity data of the sample at different wavelengths are recorded, and a result of the spectral detection containing two-dimensional information of wavelength-signal intensity is formed; in other embodiments, other detection methods can also be used, which are not limited here. In the method, the spectral detection of the target plant beverage sample is performed to obtain a result, and the specific implementation manner of the spectral detection is as follows: an appropriate amount of the target plant beverage sample is taken and placed in a quartz cuvette, an appropriate spectral detection technology (such as near-infrared spectroscopy, mid-infrared spectroscopy or Raman spectroscopy) is selected according to the chemical properties of the key flavor marker, reasonable detection parameters (including wavelength range, scanning times, resolution, etc.) are set, a spectrometer is used to scan the sample, the absorbance, transmittance or scattering intensity data of the sample at different wavelengths are recorded, and a result of the spectral detection containing two-dimensional information of wavelength-signal intensity is formed; in other embodiments, other detection methods can also be used, which are not limited here. Among them, the specific implementation method of extracting characteristic spectral intervals, absorption peak intensities and peak shape changes that are highly correlated with changes in the content of key flavor markers from the denoised spectral detection results based on key flavor markers is as follows: based on the known molecular structure of the key flavor marker (such as the functional groups of hydroxyl, carbonyl, and benzene ring) and its characteristic absorption wavelength (such as hydroxyl has characteristic absorption in the 3200-3600 cm⁻¹ range in the mid-infrared region), combined with partial least squares regression or correlation analysis, the correlation coefficient between the signal intensity of each wavelength point in the denoised spectrum and the actual content of the marker is calculated, and then a correlation coefficient absolute value threshold (such as ≥0.8) is set, and the wavelength range that continuously meets the threshold is defined as a candidate characteristic interval; then, the moving window method (such as a window size of 10-20 wavelength points) is used to calculate the average correlation coefficient within the candidate interval, and the continuous interval with the highest average correlation coefficient is retained as the characteristic spectral interval. For the identification of absorption peaks within the interval, the spectral curve can be processed in combination with the first-order derivative method or the second-order derivative method. The absorption peak in the original spectrum can be located by the peak position of the derivative spectrum, and then a linear regression analysis is performed on the intensity value of each absorption peak and the marker content. The absorption peaks showing a significant linear relationship are screened out using the coefficient of determination (R²) as an indicator (e.g., R² ≥ 0.7). At the same time, parameters such as the peak position and half-peak width are recorded to fully characterize the peak shape characteristics, and the intensity value and peak shape change characteristics such as peak position shift and half-peak width are recorded. Among them, the characteristic spectral interval refers to a specific wavelength range that is highly correlated with the chemical structure or content change of the key flavor marker in the target botanical beverage in spectral detection. The absorption peak intensity refers to the signal strength of the absorption peak in the spectral curve within the characteristic spectral interval (usually expressed as a numerical value of absorbance, transmittance or scattering intensity). The peak shape change refers to the morphological characteristics of the absorption peak within the characteristic spectral interval and its change trend.

[0045] It should be noted that the key spectral fingerprint features in this application represent a set of spectral signals extracted from the spectral detection results of the target plant beverage that can specifically characterize the core characteristics of its flavor components. It is a "bridge" between spectral data and the chemical composition and sensory characteristics of flavor substances.

[0046] In step 105, a flavor component prediction model of the target botanical beverage sample is constructed based on the historical spectral data of the botanical beverage, and the key spectral fingerprint features are input into the flavor component prediction model to output a flavor score of the flavor components in the target botanical beverage sample.

[0047] In some embodiments, constructing a flavor component prediction model for a target botanical beverage sample based on historical spectral data of botanical beverages can be achieved by using the following steps: Obtain historical spectral data of plant-based beverages; Dividing the historical spectral data into a training set and a validation set; constructing a flavor component prediction model of the target plant beverage sample based on an artificial intelligence algorithm combined with the training set and the validation set.

[0048] In a specific implementation, first, the historical spectrum of the plant beverage corresponding to the target plant beverage sample is obtained from the spectrum database; second, the specific implementation of dividing the historical spectrum data into a training set and a validation set is as follows: the historical spectrum dataset is divided by using stratified sampling or random sampling, wherein the training set accounts for 70%-80%, which is used for parameter learning and model construction, and the validation set accounts for 20%-30%, which is used for evaluating the generalization ability of the model; when dividing, the content distribution and sample batch distribution of the key flavor markers in the training set and the validation set need to be consistent to avoid affecting the model performance due to data distribution deviation; the rationality of the division can be verified by calculating the mean, standard deviation, and content range of the two groups of data, and if necessary, k-fold cross-validation is used to further optimize the data division.

[0049] In a specific implementation, the specific implementation of constructing a flavor component prediction model of the target plant beverage sample based on an artificial intelligence algorithm combined with the training set and the validation set is as follows: an appropriate artificial intelligence algorithm (such as partial least squares regression, support vector regression, random forest, or a deep learning model such as a convolutional neural network) is selected, the feature spectrum data (such as the extracted key spectral fingerprint features) in the training set is taken as the input, and the content and flavor score of the corresponding key flavor marker are taken as the output; the model parameters (such as adjusting the number of layers of the neural network and the kernel function parameters of the support vector machine) are optimized through iterative training; the performance of the trained model is evaluated by using the validation set, and the root mean square error and the determination coefficient (R²) are taken as the core indicators; if the model accuracy does not reach the preset threshold (such as R²≥0.85), the algorithm parameters are adjusted or the feature engineering steps (such as spectrum pretreatment optimization) are added for retraining until a stable and generalizable flavor component prediction model is obtained.

[0050] It should be noted that the flavor component prediction model in the present application is a mathematical model constructed based on an artificial intelligence algorithm, which aims to quickly predict the flavor component-related features of a plant beverage through the spectrum data of the plant beverage, and can directly predict the content of the key flavor marker and further output a comprehensive or subdivided flavor score, thereby realizing the rapid detection, quality evaluation, and optimization guidance of the flavor components of the plant beverage, and avoiding the tediousness and time consumption of traditional chemical analysis methods.

[0051] In some embodiments, the key spectral fingerprint features are input into the flavor component prediction model to output the flavor score of the flavor components in the target plant beverage sample, which can be achieved by the following steps: The key spectral fingerprint features are standardized, and the standardized key spectral fingerprint features are input into the flavor component prediction model; The standardized key spectral fingerprint features are calculated by the flavor ingredient prediction model to output a flavor score of a flavor ingredient in the target plant beverage sample.

[0052] The specific implementation of standardizing the key spectral fingerprint features and inputting the standardized key spectral fingerprint features into the flavor ingredient prediction model is as follows: the extracted key spectral fingerprint features (including signal values of feature spectral intervals, absorption peak intensities, and peak shape parameters, etc.) are processed by using the same standardization method (such as Z-score standardization or minimum-maximum value standardization) as that of the historical spectral data in model construction, the feature values are converted into standardized data of a uniform dimension by calculating the feature mean and standard deviation or the maximum and minimum value range, and the scale difference between different features is eliminated; then, the standardized key spectral fingerprint features are imported into the model in the form of a vector or a matrix according to the input format requirements of the flavor ingredient prediction model, and it is ensured that the input data structure is consistent with the feature dimension in model training.

[0053] The specific implementation of calculating the standardized key spectral fingerprint features by the flavor ingredient prediction model to output a flavor score of a flavor ingredient in the target plant beverage sample is as follows: the flavor ingredient prediction model calls the internally trained algorithm parameters (such as the weight matrix of a neural network, the kernel function parameters of a support vector regression), performs a series of operations such as feature mapping, weight calculation, and nonlinear conversion on the input standardized key spectral fingerprint features; the model outputs a quantitative score of the corresponding flavor ingredient (such as a comprehensive score obtained by comprehensively considering the predicted contents and contribution weights of the key flavor markers, or a subdivided score of a single flavor attribute such as aroma and sweet-sour taste) according to the correlation between the key spectral fingerprint features and the flavor ingredients, and the mapping relationship between the spectral features and the flavor marker contents in the historical data, and takes the comprehensive score as the flavor score of the flavor ingredient in the target plant beverage sample; in other embodiments, other calculation methods can also be used, which are not limited here.

[0054] It should be noted that the flavor score in the present application represents the overall characteristics and degree of excellence of the flavor ingredients of the sample, and the numerical size is directly related to the actual performance of the sample flavor: a high score usually corresponds to a sample with appropriate contents of key flavor substances, outstanding and coordinated flavor characteristics, and a low score may reflect the absence of flavor substances, unbalanced proportions, or the presence of undesirable flavors, which provides a quantifiable reference standard for the flavor quality evaluation, product optimization, and quality control of plant beverages.

[0055] In addition, another aspect of the present application, in some embodiments, the present application provides an artificial intelligence-based plant beverage flavor ingredient detection system, which is described with reference to Figure 4, which is a schematic diagram of the structure of a plant beverage flavor component detection system based on artificial intelligence according to some embodiments of the present application. The plant beverage flavor component detection system based on artificial intelligence 400 includes: a detection module 401, a processing module 402 and an execution module 403, which are described as follows: Detection module 401, in this application, detection module 401 is mainly used to detect the content of each flavor substance in the target plant beverage sample, and obtain a multi-dimensional raw data matrix of the flavor substance content in the target plant beverage sample; Processing module 402, in the present application, is used to analyze the sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample based on the sensory evaluation data of the target plant beverage sample in combination with the multidimensional raw data matrix, and use principal component analysis to perform dimensionality reduction processing on the multidimensional raw data matrix to obtain the principal component vectors of the flavor substances in the target plant beverage sample; It should be noted that the processing module 402 in the present application is also used to analyze the contribution of the core components that affect the flavor of the target botanical beverage sample by combining the sensory contribution spectrum with the synergistic variation characteristics between the flavor substances reflected by the principal component vector, and obtain the key flavor markers of the flavor components in the target botanical beverage sample; In addition, it should be noted that the processing module 402 in the present application is also used to perform spectral detection on the target plant beverage sample, and extract key spectral fingerprint features of the flavor components in the target plant beverage sample from the spectral detection results based on the key flavor markers; Execution module 403. In this application, execution module 403 is mainly used to construct a flavor component prediction model of the target plant beverage sample based on the historical spectral data of the plant beverage, input the key spectral fingerprint features into the flavor component prediction model, and output the flavor score of the flavor component in the target plant beverage sample.

[0056] In addition, the present application also provides a computer device, which includes a memory and a processor, the memory storing a code, and the processor being configured to obtain the code and execute the above-mentioned artificial intelligence-based plant beverage flavor component detection method.

[0057] In some embodiments, reference Figure 5 , which is a schematic diagram of the structure of a computer device for implementing an artificial intelligence-based plant beverage flavor component detection method according to some embodiments of the present application. The artificial intelligence-based plant beverage flavor component detection method in the above embodiment can be Figure 5 The computer device 500 shown in FIG. 5 is implemented as shown in FIG. 5 . The computer device 500 includes at least one processor 501 , a communication bus 502 , a memory 503 , and at least one communication interface 504 .

[0058] The processor 501 may be a general-purpose central processing unit (CPU) or an application-specific integrated circuit (ASIC).

[0059] The communication bus 502 may be used to transmit information between the aforementioned components.

[0060] The memory 503 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, a random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, an optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto. The memory 503 may be independent and connected to the processor 501 via the communication bus 502. The memory 503 may also be integrated with the processor 501.

[0061] The memory 503 is used to store program code for executing the solution of the present application, and is controlled by the processor 501. The processor 501 is used to execute the program code stored in the memory 503. The program code may include one or more software modules. The method used in the above embodiment can be implemented by the processor 501 and one or more software modules in the program code in the memory 503.

[0062] The communication interface 504 uses any device such as a transceiver to communicate with other devices or communication networks, such as Ethernet, radio access network (RAN), wireless local area network (WLAN), etc.

[0063] In a particular implementation, as one embodiment, the computer device can include multiple processors, each of which can be a single-CPU processor or a multi-CPU processor. A processor herein can refer to one or more devices, circuits, and / or processing cores for processing data (e.g., computer program instructions).

[0064] The computer device described above can be a general-purpose computer device or a special-purpose computer device. In a particular implementation, the computer device can be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of the computer device.

[0065] In addition, the present application also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the plant beverage flavor ingredient detection method based on artificial intelligence described above.

[0066] Although the preferred embodiments of the present application have been described, those skilled in the art who understand the basic inventive concept can make additional changes and modifications to the embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0067] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A method for detecting flavor components of plant beverages based on artificial intelligence, characterized in that: The steps include: Detecting the content of each flavor substance in the target plant beverage sample to obtain a multidimensional raw data matrix of the flavor substance content in the target plant beverage sample; Based on the sensory evaluation data of the target plant beverage sample and the multidimensional raw data matrix, the sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample is analyzed, and the multidimensional raw data matrix is ​​subjected to dimensionality reduction processing using principal component analysis to obtain the principal component vectors of the flavor substances in the target plant beverage sample, and the principal component vectors are used to reflect the synergistic change characteristics between the flavor substances; The contribution of the core components that affect the flavor of the target botanical beverage sample is analyzed by combining the sensory contribution spectrum with the synergistic variation characteristics between the flavor substances reflected by the principal component vector, thereby obtaining the key flavor markers of the flavor components in the target botanical beverage sample; Performing spectral detection on the target plant beverage sample, and extracting key spectral fingerprint features of the flavor components in the target plant beverage sample from the spectral detection results based on the key flavor markers; A flavor component prediction model of a target botanical beverage sample is constructed based on historical spectral data of the botanical beverage, and the key spectral fingerprint features are input into the flavor component prediction model to output a flavor score of the flavor components in the target botanical beverage sample.

2. The method according to claim 1, wherein The sensory contribution spectrum of the volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample is analyzed based on the sensory evaluation data of the target plant beverage sample in combination with the multidimensional raw data matrix, specifically including: Obtain sensory evaluation data of target plant beverage samples; Calculating correlation coefficients between volatile flavor substances and non-volatile flavor substances and various sensory attributes based on the sensory evaluation data and the multidimensional raw data matrix; Based on all the correlation coefficients, a sensory contribution spectrum of volatile and non-volatile flavor substances in the target plant beverage sample to the overall sensory attributes of the target plant beverage sample was constructed.

3. The method according to claim 1, wherein The principal component analysis is used to perform dimensionality reduction processing on the multidimensional original data matrix to obtain the principal component vectors of the flavor substances in the target plant beverage sample, specifically including: performing standardization processing on the multidimensional original data matrix to obtain a standardized multidimensional original data matrix; Using a principal component analysis algorithm to calculate eigenvalues ​​of the standardized multidimensional raw data matrix, to obtain eigenvalue data of flavor substances in the target plant beverage sample; Determine the characteristic thresholds of flavor substances in target plant beverage samples; The characteristic value data is judged according to the characteristic threshold value, thereby obtaining the principal component vector of the flavor substance in the target plant beverage sample.

4. The method according to claim 1, wherein The contribution of the core components that affect the flavor of the target botanical beverage sample is analyzed by combining the sensory contribution spectrum with the synergistic variation characteristics between the flavor substances reflected by the principal component vector, and the key flavor markers of the flavor components in the target botanical beverage sample are obtained, specifically including: extracting a plurality of high-contribution flavor substances that affect the flavor of the plant beverage from the target plant beverage sample according to the sensory contribution spectrum; a plurality of core flavor substances screened from the principal component loadings of the principal component vector; Conduct intersection analysis on each high-contribution flavor substance and each core flavor substance to obtain multiple candidate core components of the target plant beverage sample; The contribution score of each candidate core ingredient to the flavor of the plant beverage is calculated, and the key flavor markers of the flavor components in the target plant beverage sample are obtained.

5. The method according to claim 1, wherein Spectral detection is performed on the target plant beverage sample. Based on the key flavor markers, the key spectral fingerprint features of the flavor components in the target plant beverage sample are extracted from the spectral detection results. Specifically, the following features are included: Performing spectral detection on the target plant beverage sample to obtain spectral detection results; Performing denoising on the result of the spectrum detection to obtain a denoised spectrum detection result; Extracting characteristic spectral intervals, absorption peak intensities, and peak shape changes that are highly correlated with changes in key flavor marker content from the denoised spectral detection results based on the key flavor markers; The characteristic spectral range, absorption peak intensity and peak shape changes are used as key spectral fingerprint features of flavor components in the target plant beverage sample.

6. The method according to claim 1, wherein The flavor component prediction model of the target plant beverage sample is constructed based on the historical spectral data of the plant beverage, specifically including: Obtain historical spectral data of plant-based beverages; Dividing the historical spectral data into a training set and a validation set; A flavor component prediction model for the target plant beverage sample is constructed based on an artificial intelligence algorithm combined with the training set and the validation set.

7. The method according to claim 1, wherein Inputting the key spectral fingerprint features into the flavor component prediction model to output a flavor score of the flavor components in the target botanical beverage sample specifically includes: Standardizing the key spectral fingerprint features, and inputting the standardized key spectral fingerprint features into the flavor component prediction model; The flavor component prediction model is used to calculate the input standardized key spectral fingerprint features to output the flavor score of the flavor components in the target plant beverage sample.

8. The method according to claim 1, wherein The content of various flavor substances in the target plant beverage samples was detected by gas chromatography-mass spectrometry and high performance liquid chromatography instrument analysis methods.

9. The method according to claim 1, wherein Spectral detection of target plant beverage samples was performed by near-infrared spectroscopy.

10. A plant beverage flavor component detection system based on artificial intelligence, characterized in that: The system includes: A detection module is used to detect the content of each flavor substance in the target plant beverage sample to obtain a multidimensional raw data matrix of the flavor substance content in the target plant beverage sample; a processing module for analyzing, based on the sensory evaluation data of the target botanical beverage sample and the multidimensional raw data matrix, a sensory contribution spectrum of volatile and non-volatile flavor substances in the target botanical beverage sample to the overall sensory attributes of the target botanical beverage sample, and performing dimensionality reduction processing on the multidimensional raw data matrix using principal component analysis to obtain principal component vectors of the flavor substances in the target botanical beverage sample, wherein the principal component vectors are used to reflect the synergistic variation characteristics among the flavor substances; The processing module is further configured to analyze the contribution of the core components that affect the flavor of the target botanical beverage sample by combining the sensory contribution spectrum with the synergistic variation characteristics between the flavor substances reflected by the principal component vector, and obtain key flavor markers of the flavor components in the target botanical beverage sample; The processing module is further configured to perform spectral detection on the target plant beverage sample, and extract key spectral fingerprint features of flavor components in the target plant beverage sample from the spectral detection results based on the key flavor markers; The execution module is used to construct a flavor component prediction model of the target plant beverage sample based on the historical spectral data of the plant beverage, input the key spectral fingerprint features into the flavor component prediction model, and output the flavor score of the flavor component in the target plant beverage sample.

Images