A reagent kit-based method and apparatus for evaluating tomato flavor

Abstract

This invention relates to the field of food quality analysis technology, specifically to a method and apparatus for evaluating tomato flavor based on a reagent kit. The method includes: performing multi-dimensional chemical composition analysis on tomato samples collected from different sources; extracting the main flavor components from the flavor component data; generating flavor score data for each tomato sample based on the contribution of each main flavor component; classifying the tomato samples based on similarity to obtain sample groups with similar flavor characteristics; generating a central point feature vector for each sample group; generating a flavor scoring model based on the central point feature vector; generating an evaluation report based on the flavor scoring model; and, based on the flavor evaluation report, adjusting the proportions of various standardized taste solutions in the reagent kit to match the flavor characteristics of the tomato samples. This application dynamically adjusts the method based on the characteristic data of the tomato samples, making the flavor evaluation more comprehensive and providing a quantitative standard for the flavor characteristics of tomato samples.

Description

Technical Field

[0001] This invention relates to the field of food quality analysis technology, and more specifically, to a method and apparatus for evaluating tomato flavor based on a reagent kit. Background Technology

[0002] With the development of agricultural technology and the increasing demands of consumers for food quality, tomatoes have become one of the most widely grown and consumed vegetables globally. The flavor components of tomatoes directly affect their market acceptance, and factors such as different origins, growing environments, and processing methods all contribute to variations in tomato flavor. To provide high-quality tomato products, traditional flavor evaluation methods rely on manual tasting or simplified chemical analysis. While these methods can provide some flavor characteristic data, they have significant limitations in practical application. For example, manual tasting is highly subjective, heavily influenced by the experience of the evaluators, and difficult to standardize; while chemical analysis often relies on single component indicators, failing to comprehensively reflect the multi-dimensional flavor characteristics of tomatoes.

[0003] Currently, there are some technical means for evaluating tomato flavor on the market, using various instruments to detect and analyze flavor components. Although they can provide certain analytical results in theory, they have significant limitations in practical applications. They only focus on the analysis of single components and ignore the complex interactions between flavor components and the multi-dimensional characteristics of overall flavor perception. They cannot comprehensively and accurately reflect the overall feeling of tomato flavor, especially when the source and environmental conditions of tomato samples of the same variety vary greatly, making it difficult to obtain a unified and accurate flavor evaluation result.

[0004] Therefore, a new method for evaluating tomato flavor is urgently needed to solve the above problems. Summary of the Invention

[0005] The main objective of this invention is to provide a method and apparatus for evaluating tomato flavor based on a reagent kit, aiming to overcome the technical problem that existing technologies cannot comprehensively and accurately evaluate tomato flavor.

[0006] To address the aforementioned problems, this invention proposes a reagent kit-based method for evaluating tomato flavor, the method comprising:

[0007] Multidimensional chemical composition analysis was performed on tomato samples collected from different sources. The chemical composition included soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample.

[0008] Extract the main flavor components from the flavor component data, and generate flavor score data for each tomato sample based on the contribution of each main flavor component;

[0009] Identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics;

[0010] Based on the flavor component data corresponding to the sample groups, a center point feature vector is generated for each sample group, and a flavor scoring model is generated based on the center point feature vector.

[0011] Based on the flavor scoring model, an evaluation result is generated, and the sugar-acid ratio in the evaluation result is mapped to the flavor score to obtain a flavor evaluation report.

[0012] Based on the flavor evaluation report, the various standardized taste solutions in the kit were proportioned to match the flavor characteristics of the tomato sample.

[0013] Furthermore, the step of performing multi-dimensional chemical composition analysis on tomato samples collected from different sources, wherein the chemical composition includes soluble solids content, acidity, sugar-acid ratio, and aroma components, to obtain flavor component data for each tomato sample, includes:

[0014] Obtain sample images of the tomato samples and identify the surface features of the tomatoes in the sample images to obtain the geometric feature data of each tomato sample;

[0015] Chemical composition information of the tomato sample was collected by multi-dimensional sensors, and the chemical composition information was standardized to obtain a composition dataset. The chemical composition includes at least soluble solids content, acidity, sugar-acid ratio and aroma components.

[0016] The chemical components in the component dataset are fused to obtain a comprehensive component set for each tomato sample.

[0017] The comprehensive component set is associated with the geometric feature data to generate a flavor component dataset for each tomato sample.

[0018] Further, the step of extracting the main flavor components from the flavor component data and generating flavor score data for each tomato sample based on the contribution of each main flavor component includes:

[0019] Obtain the variance contribution of each flavor component in the flavor component data, generate a feature matrix containing the principal component contribution based on the variance contribution, and obtain the projection value of each tomato sample on each principal component.

[0020] The projection values ​​of each tomato sample on each principal component dimension are weighted to generate flavor component contribution data for each tomato sample.

[0021] A sliding window with a predetermined number of positions is set, and the sliding window is controlled to slide over the flavor component contribution data to obtain the potential factors of each principal component;

[0022] Calculate the score of each tomato sample on each of the latent factors, arrange the score values, and generate a latent factor score matrix;

[0023] Key factors in the latent factor score matrix are identified, the total score of the key factors is calculated based on the projection value, and the flavor score data of the tomato sample is generated based on the total score.

[0024] Further, the step of identifying the similarity between multiple flavor rating data, classifying the tomato samples based on the similarity, and obtaining sample groups with similar flavor characteristics includes:

[0025] Feature extraction was performed on the flavor score data of each tomato sample to obtain the statistical feature vector of each tomato sample;

[0026] The similarity value between each pair of tomato samples is calculated based on the cosine similarity formula, and a similarity matrix is ​​constructed based on the similarity value.

[0027] Hierarchical clustering algorithm is used to cluster the similarity matrix to obtain preliminary classified samples;

[0028] Based on a preset threshold, cohesion analysis is performed on the preliminary classified samples to obtain sample groups;

[0029] The flavor score data of each sample group are aggregated into a group representative feature vector to obtain sample groups with similar flavor characteristics.

[0030] Further, the step of generating a center point feature vector for each sample group based on the flavor component data corresponding to the sample group, and generating a flavor scoring model based on the center point feature vector, includes:

[0031] The flavor component data of each sample group are analyzed item by item to obtain the distribution pattern of each flavor component within the group;

[0032] The distribution patterns of flavor components within each sample group are analyzed for differences to obtain the local difference index of each flavor component.

[0033] Based on the local difference information, a recursive focusing analysis is performed on the flavor components of each sample group to determine the key flavor components whose influence on the central characteristics of the group is greater than a preset influence value, and to obtain the recursive weight value of each flavor component.

[0034] The weighted center point feature vector of flavor components in the sample group is calculated based on the recursive weight value, and the weighted center point feature vector is optimized based on the singular value decomposition algorithm to obtain the optimized center point feature vector.

[0035] The optimized center point feature vector is subjected to multiple mapping processing to construct a high-dimensional flavor representation model.

[0036] Further, the step of generating evaluation results based on the flavor scoring model, mapping the sugar-acid ratio in the evaluation results to the flavor score, and obtaining a flavor evaluation report includes:

[0037] The flavor score characteristics of each tomato sample are evaluated based on the flavor scoring model to obtain the corresponding evaluation dataset;

[0038] The sugar-acid ratio of each tomato sample in the evaluation dataset is sorted to obtain an ascending sequence of sugar-acid ratios;

[0039] Based on the permutation sequence, the relative position of the sugar-acid ratio of each tomato sample in the evaluation dataset is calculated, and the relative position is calibrated to obtain the calibration index of each tomato sample.

[0040] The difference between the flavor score and the calibration index for each tomato sample is calculated to obtain the difference data between the flavor score and the sugar-acid ratio position;

[0041] The difference data is divided into multiple distribution spaces. Based on the flavor characteristics of each distribution space, the flavor of each tomato sample is subdivided and evaluated to obtain the performance value of each tomato sample in the corresponding distribution space. A flavor evaluation report is generated based on the performance value.

[0042] Further, the step of proportioning the various standardized taste solutions in the kit according to the flavor evaluation report to match the flavor characteristics of the tomato sample includes:

[0043] Based on the flavor evaluation report, the correspondence between each flavor component and various standardized taste solutions in the kit is identified, and parameter values ​​that match the flavor characteristics of each standardized solution with the tomato sample are obtained.

[0044] Calculate the parameter values ​​for each type of standardized taste solution to generate a proportioning model;

[0045] The optimal ratio of each standardized taste solution is calculated based on the ratio model to minimize the difference between the flavor characteristics of the tomato sample and the flavor characteristics of the standardized taste solution in the kit.

[0046] An optimal formulation scheme is generated based on the optimal formulation, and various standardized taste solutions in the kit are adjusted based on the optimal formulation scheme.

[0047] The present invention also proposes a tomato flavor evaluation device based on a reagent kit, comprising:

[0048] The analysis module is used to perform multi-dimensional chemical composition analysis on tomato samples collected from different sources. The chemical composition includes soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample.

[0049] An extraction module is used to extract the main flavor components from the flavor component data and generate flavor score data for each tomato sample based on the contribution of each main flavor component.

[0050] The identification module is used to identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics;

[0051] The generation module is used to generate a center point feature vector for each sample group based on the flavor component data corresponding to the sample group, and to generate a flavor scoring model based on the center point feature vector.

[0052] The evaluation module is used to generate evaluation results based on the flavor scoring model, and to perform data mapping between the sugar-acid ratio and the flavor score in the evaluation results to obtain a flavor evaluation report.

[0053] The matching module is used to process the proportions of various standardized taste solutions in the kit according to the flavor evaluation report, so as to match the flavor characteristics of the tomato sample.

[0054] The present invention also proposes a computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0055] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0056] Beneficial effects:

[0057] This application proposes a reagent kit-based method and apparatus for evaluating tomato flavor. Through multi-dimensional chemical composition analysis, it can comprehensively analyze multiple flavor components in tomato samples. By employing similarity analysis of flavor score data and sample group classification technology, it enables more unified and accurate evaluation of tomato samples from different sources and cultivation environments. By extracting the main flavor components and generating flavor scores based on their contribution, the objectivity and accuracy of flavor evaluation are improved. Furthermore, this application can dynamically adjust based on the characteristic data of tomato samples, making the flavor evaluation not only more comprehensive but also highly adaptable. It can automatically optimize the evaluation results based on tomato samples from different batches or under different environmental conditions, overcoming the limitations of existing technologies that only focus on single-component indicators. This greatly improves the accuracy and adaptability of flavor evaluation, providing a quantitative standard for the flavor characteristics of tomato samples. Combined with standardized taste solutions in the reagent kit, it ensures consistency and efficiency in the flavor optimization process.

[0058] In conclusion, this application not only improves the scientific rigor and objectivity of tomato flavor assessment, but also provides effective technical support for tomato cultivation, processing, and flavor optimization, promoting the standardization and improvement of tomato flavor quality. Attached Figure Description

[0059] Figure 1 This is a schematic diagram of the steps of a tomato flavor evaluation method based on a reagent kit in one embodiment of the present invention;

[0060] Figure 2 This is a schematic block diagram of a tomato flavor evaluation device based on a reagent kit according to an embodiment of the present invention;

[0061] Figure 3 This is a schematic block diagram of a computer device according to an embodiment of the present invention.

[0062] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0064] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of features, integers, steps, operations, elements, modules, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, modules, components, and / or groups thereof. It should be understood that when an element is referred to as “connected” or “coupled” to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein may include wireless connection or wireless coupling. The term “and / or” as used herein includes all or any modules and all combinations of one or more associated listed items.

[0065] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0066] Reference Figure 1 This invention provides a method for evaluating tomato flavor based on a reagent kit, the method comprising:

[0067] S1: Perform multi-dimensional chemical composition analysis on tomato samples collected from different sources, including soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample;

[0068] In step S1, tomato samples collected from different sources undergo multi-dimensional chemical composition analysis. The analyzed chemical components include soluble solids content, acidity, sugar-acid ratio, and aroma components. These different sources can be tomatoes from different regions, varieties, or growing conditions, including systematic testing of over 3000 tomato plant materials. Soluble solids (SSC) content mainly consists of dissolved substances such as sugars, acids, organic acids, and amino acids. By measuring the soluble solids content of tomato juice, the sugar-to-acid ratio of the tomato can be roughly understood. Test results show that over 95% of tomato samples have a soluble solids content not exceeding 10°Brix; therefore, the upper limit of the sugar content in the kit is also set at 10°Brix. The acidic substances in tomatoes are mainly organic acids, such as citric acid and malic acid, which determine the intensity of the tomato's sour taste. Acidity is determined by titration. During the test, standard acid solutions such as citric acid solution are used to quantify the acid concentration in the tomatoes for accurate acidity analysis. The sugar-acid ratio, or the ratio of sugar content to acidity, is crucial for a balanced sweet and sour flavor in tomatoes. A ratio that is too high or too low results in an unbalanced flavor profile. The sugar-acid ratio can be calculated by accurately measuring the sugar and acid content. Tomatoes contain a variety of aroma components, such as those reedy, grassy, ​​and fruity notes. These volatile compounds can be analyzed using instruments like gas chromatography (GC). A comprehensive analysis of these chemical components allows for the creation of a complete database of tomato flavor components.

[0069] S2: Extract the main flavor components from the flavor component data, and generate flavor score data for each tomato sample based on the contribution of each main flavor component;

[0070] In step S2, the main flavor components are extracted from the flavor component data, and corresponding flavor score data is generated for each tomato sample based on the contribution of each main flavor component. For each flavor component, its actual contribution to flavor perception is analyzed. For example, sugar and acid are the most important components affecting tomato flavor; the concentration of sugar directly determines the intensity of sweetness, while acidity affects the degree of sourness. In practice, statistical methods such as regression analysis can be used to determine the relationship between each chemical component (such as sugar, acid, and aroma components) and flavor, quantifying the contribution of each component. For example, experimental tests have shown that when the ratio of sugar to acidity reaches a certain balance, the tomato flavor exhibits the optimal sweet-sour balance. Correspondingly, the sugar-acid ratio contributes more to the flavor score, while aroma components may affect the subtlety and complexity of the flavor. After extracting the main flavor components, flavor score data is generated for each tomato sample based on the interactions between various flavor components. For example, the sugar and acid ratios in tomato samples may vary. Some tomatoes may have a higher sugar content and lower acidity, resulting in a more pronounced sweet taste; while others may have a higher acidity, leading to a stronger sour experience. To obtain objective and reliable flavor scores, sensory test data from tasters is combined, and statistical methods are used to correlate chemical components with sensory evaluation data, thereby generating a comprehensive flavor score for each tomato sample. For instance, suppose two tomato samples are analyzed: one sample has 8% sugar and 0.3% acidity, while the other has 5% sugar and 0.5% acidity. Based on the aforementioned flavor component data, the flavor score for each tomato sample can be calculated according to the sugar-acid ratio. The sugar-acid ratio is calculated by dividing the sugar concentration by the acid concentration. The sugar-acid ratio of the first sample is 8% / 0.3% = 26.67, while that of the second sample is 5% / 0.5% = 10. Clearly, the first tomato has a higher sugar-acid ratio, indicating it may have a stronger sweetness and a more balanced flavor. According to a pre-defined scoring system, samples with higher sugar-acid ratios typically receive higher flavor scores, and vice versa. This process transforms the measurement results of chemical components into a scientific flavor score, precisely quantifying the flavor of the tomato samples.

[0071] S3: Identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics;

[0072] In step S3, based on the flavor score data obtained in step S2, these flavor score data are processed using mathematical and statistical methods to identify their similarity. Similarity identification can employ various statistical and data analysis methods, such as Euclidean distance, Pearson correlation coefficient, or cluster analysis, to measure the degree of similarity in flavor scores between different tomato samples and classify the samples based on this similarity. For example, suppose there are two tomato samples with flavor scores of (8, 6, 7) and (7, 6, 8), respectively. The difference between these two sets of scores is very small, indicating that their flavor characteristics are quite similar. In this case, based on similarity calculation, these two tomato samples will be classified into the same group. On the other hand, if a tomato has a flavor score of (3, 9, 2), then its flavor score differs significantly from the previous two tomato samples and will therefore be classified into a different group. In another embodiment, flavor score data can be mapped to different categories through automated learning based on algorithms such as K-means clustering, hierarchical clustering, or self-organizing map networks (SOM) in cluster analysis. For example, K-means clustering automatically assigns samples to different groups based on their similarity, while hierarchical clustering gradually merges similar samples, ultimately forming multiple categories with similar flavor characteristics. During classification, multiple dimensions are considered, such as sugar-acid ratio, acidity, and aroma, ensuring that each sample group represents a tomato category with similar flavor characteristics. For instance, some tomato samples might have a relatively high sugar-acid ratio, resulting in a sweeter and less acidic taste, and thus they would be classified into the sweet flavor group; while other samples might be classified into the tangy or balanced flavor group due to their higher acidity or different sugar-acid ratio.

[0073] S4: Based on the flavor component data corresponding to the sample groups, generate a center point feature vector for each sample group, and generate a flavor scoring model based on the center point feature vector;

[0074] In step S4, the sample groups with similar flavor characteristics formed in step S3, each consisting of tomato samples with similar flavor scores, are used to calculate the centroid feature vector for each group. The centroid feature vector represents the average or representative flavor characteristics of all samples within that group, obtained through statistical calculation of the flavor scores of all samples within the group. Specifically, calculating the centroid feature vector typically involves averaging the flavor scores of all samples within the group. For example, assuming a sample group contains five tomato samples with flavor scores of 8, 7, 9, 6, and 8, the centroid feature vector for this group can be obtained by calculating the average of these five scores: (8+7+9+6+8) / 5 = 7.6. This centroid feature vector reflects the overall flavor characteristics of the group, and a flavor scoring model is constructed based on these centroid feature vectors. The construction of the flavor scoring model can employ various statistical and machine learning methods, such as linear regression, support vector machines, or neural networks, to predict the flavor score of new samples by learning the relationship between the feature vectors of sample groups and their corresponding flavor scores. For example, assuming a linear regression model is used, the input is the centroid feature vector of the sample, and the output is the predicted flavor score. By training on existing sample groups, the model can identify the main factors affecting the flavor score, thus predicting the corresponding flavor score based on the feature vector of a new sample. For instance, suppose three sample groups are identified after step S3: sweet, tangy, and balanced. For the sweet group, the centroid feature vector is (9, 0.2), indicating that this group has high sugar content and low acidity; while the tangy group might have a centroid feature vector of (4, 0.6), indicating high acidity and relatively low sugar content. Based on these feature vectors, a flavor scoring model is constructed. Inputting the chemical composition data of different tomato samples, the model, after training, can predict the flavor score of a new sample.

[0075] S5: Generate evaluation results based on the flavor scoring model, perform data mapping between the sugar-acid ratio and flavor score in the evaluation results, and obtain a flavor evaluation report;

[0076] In step S5, the chemical composition data of the new sample is input into the flavor scoring model. Based on the features and patterns learned during previous training, a corresponding flavor score is output. For example, suppose a newly collected tomato sample has a soluble solids content of 6.5%, an acidity of 0.4%, and a sugar-acid ratio of 16.25. This data is input into the established flavor scoring model, and a flavor score is output based on the input data. For instance, a score of 7 indicates that the flavor characteristics of this tomato sample are at a high level within the preset scoring criteria. The sugar-acid ratio in the evaluation results is mapped to the flavor score, and the relationship between the flavor score and the sugar-acid ratio is visualized and quantified to more intuitively understand the differences between tomato samples with different flavor characteristics. This can be achieved by plotting a scatter plot or using other visualization tools to label the sugar-acid ratio of each sample with its corresponding flavor score, thus forming a flavor feature map. For example, suppose that during the evaluation process, flavor scores and sugar-acid ratios were measured for multiple tomato samples. The final data may show that samples with a sugar-acid ratio between 3 and 40 generally had higher flavor scores, while samples with a sugar-acid ratio above 40 had lower flavor scores. This phenomenon clearly reflects the importance of sugar-acid balance to tomato flavor.

[0077] S6: Based on the flavor evaluation report, the various standardized taste solutions in the kit are proportioned to match the flavor characteristics of the tomato sample.

[0078] In step S6, the flavor characteristics of the tomato sample are matched with the proportions of standardized taste solutions based on the flavor assessment report. The standardized taste solutions in the kit are formulated according to different sugar-acid ratios and flavor scores, with each solution representing a specific flavor characteristic. For example, the kit may contain nine standardized solutions, covering a flavor range from low sugar and low acid to high sugar and high acid. These solutions are derived from experimental data analysis and can simulate the diversity of tomato flavors. Different solutions in the kit are selected and proportioned according to the flavor assessment results of the tomato sample. For example, for the tomato sample with a score of 7 and a sugar-acid ratio of 28.94, the closest standardized solution is selected based on the flavor characteristics of the tomato sample to ensure that its sugar-acid ratio, sweet-acid balance, and flavor hierarchy are consistent. If the characteristics of the tomato sample are close to the standardized solution with a flavor score of 7, the sugar-acid ratio of that solution is used as a reference to further adjust the proportions of other solutions to better match the flavor of the actual tomato sample. Furthermore, in steps S1 and S2, the aroma components of the tomato sample were analyzed, identifying different aroma components such as vine-like, grassy, ​​floral, and fruity aromas. In step S6, in addition to the sugar-acid ratio, the kit includes standardized solutions formulated based on these aroma components. By adjusting the proportions of these aroma solutions, flavor and aroma can complement each other, enhancing the overall flavor profile of the tomato sample. For example, assuming a tomato sample has a flavor score of 6, a sugar-acid ratio of 26.09, and aroma characteristics primarily consisting of floral and fruity aromas, an appropriate standardized taste solution is selected based on the score and sugar-acid ratio. By adjusting the proportions of the floral and fruity aroma solutions, the aroma characteristics of the sample are fully reproduced. In this process, the proportions of each aroma component are precisely controlled using the aroma formulation solutions provided in the kit, thereby achieving accurate reproduction of the tomato sample's flavor.

[0079] In another embodiment, the sugar concentration (%) in the kit ranges from 1% to 10%, and the acid concentration (%) ranges from 0.20% to 0.50%. Nine levels are set based on these concentrations, each representing a different sugar-acid ratio and flavor characteristics. These concentration ranges are derived from previous testing and analysis of a large number of tomato samples, covering the flavor characteristics of most tomato samples and encompassing the complete flavor range from very low sugar to high acid (strong acidity and astringency) to high sugar to low acid (very sweet). The rating range is 1-9, with higher ratings indicating better flavor. The method of using the taste standard solution and aroma standard sample included in the kit is as follows: each taste sample is provided in powder or liquid form, labeled from 1 to 9; the user needs to dilute each packet of powder or liquid to 100ml of water, mix thoroughly, and experience each level sequentially. The comprehensive rating system is expressed through the sweet and acid standard solution; after dilution, the user can experience the comprehensive flavor intensity of different levels. The aroma sample is provided in a sealed glass bottle; the user needs to keep the bottle opening 5-10 cm away from the nose and gently smell to experience the aroma characteristics. Through these steps, users can scientifically and accurately complete the sensory evaluation of tomato flavor.

[0080] In another embodiment, the rating report table is shown in Table 1. X1~X6 represents the increasing sugar content range, indicating a gradual increase in sugar in tomatoes; Y1~Y6 represents the decreasing acid content range, indicating a gradual decrease in acidity in tomatoes. These numerical ranges are derived from statistical analysis of a large amount of experimental data and can accurately reflect the changing trends of tomato flavor. It is worth noting that X1~X6 is a continuous increase in sugar content across the nine levels, while Y1~Y6 is not strictly decreasing. Instead, it is appropriately divided according to the needs of flavor evaluation to ensure that the flavor characteristics between each level are significantly distinguishable. For example, in level 9, the acid content is Y4, not Y6, because the optimal tomato flavor requires a delicate balance between sweetness and acidity; too low an acid content will make the flavor monotonous.

[0081] Table 1

[0082]

[0083] In one embodiment, the step of performing multi-dimensional chemical composition analysis on tomato samples collected from different sources, wherein the chemical composition includes soluble solids content, acidity, sugar-acid ratio, and aroma components, to obtain flavor component data for each tomato sample, includes:

[0084] Obtain sample images of the tomato samples and identify the surface features of the tomatoes in the sample images to obtain the geometric feature data of each tomato sample;

[0085] Chemical composition information of the tomato sample was collected by multi-dimensional sensors, and the chemical composition information was standardized to obtain a composition dataset. The chemical composition includes at least soluble solids content, acidity, sugar-acid ratio and aroma components.

[0086] The chemical components in the component dataset are fused to obtain a comprehensive component set for each tomato sample.

[0087] The comprehensive component set is associated with the geometric feature data to generate a flavor component dataset for each tomato sample.

[0088] In the above embodiments, images of tomato samples are acquired, and surface features of the tomatoes in the images are identified to obtain geometric feature data for each tomato sample. Visual image recognition technology is used to collect the appearance information of the tomatoes, especially their surface features, such as skin color and shape. Image processing technology, especially computer vision algorithms, can extract geometric features such as the diameter, aspect ratio, and skin texture of the tomatoes. Multi-dimensional sensors are used to collect chemical composition information of the tomato samples. These sensors can include electronic noses, gas sensors, and spectrometers, enabling precise measurement of soluble solids content, acidity, sugar-acid ratio, and aroma components. During this process, the data collected by the sensors is standardized, and the chemical composition data is fused to integrate information from different chemical components, thus obtaining a comprehensive component set for each tomato sample. Data fusion algorithms, such as principal component analysis (PCA), weighted average, or other multivariate analysis methods, can be used to reasonably allocate the weights and contributions of each chemical component and integrate them into a comprehensive component set. This data is correlated with the geometric feature data of tomato samples to generate a complete flavor component dataset for each tomato sample. The external features of the tomato are combined with its internal chemical composition information to form a complete flavor component dataset.

[0089] In one embodiment, the step of extracting the major flavor components from the flavor component data and generating flavor score data for each tomato sample based on the contribution of each major flavor component includes:

[0090] Obtain the variance contribution of each flavor component in the flavor component data, generate a feature matrix containing the principal component contribution based on the variance contribution, and obtain the projection value of each tomato sample on each principal component.

[0091] The projection values ​​of each tomato sample on each principal component dimension are weighted to generate flavor component contribution data for each tomato sample.

[0092] A sliding window with a predetermined number of positions is set, and the sliding window is controlled to slide over the flavor component contribution data to obtain the potential factors of each principal component;

[0093] Calculate the score of each tomato sample on each of the latent factors, arrange the score values, and generate a latent factor score matrix;

[0094] Key factors in the latent factor score matrix are identified, the total score of the key factors is calculated based on the projection value, and the flavor score data of the tomato sample is generated based on the total score.

[0095] In the above embodiments, the variance contribution of each flavor component in the flavor component data is obtained. Variance contribution measures the magnitude of each flavor component's contribution to the overall data variability, reflecting the importance of each flavor component in the entire dataset. Statistical methods such as Principal Component Analysis (PCA) can be applied to identify the variance contribution of each flavor component by performing variance analysis on the original flavor component data. A higher variance contribution of each flavor component indicates a greater impact on the overall flavor. By calculating the variance contribution of each flavor component, it is determined which components dominate the flavor evaluation. Based on these variance contributions, a feature matrix containing the principal component contributions is generated. This feature matrix, obtained from principal component analysis, shows the contribution of each principal component in the flavor component data. Through this matrix, the projection values ​​of each tomato sample on each principal component are obtained, thus transforming the flavor characteristics of different tomato samples into numerical representations on these principal components. The calculation of projection values ​​maps the flavor characteristics of each sample to the principal component space, transforming high-dimensional flavor component data into low-dimensional principal component data through linear matrix transformation. The projected values ​​of each tomato sample across the principal component dimensions are weighted, and the contribution of each principal component is further weighted, as different principal components may have different importance in flavor evaluation. Based on the contribution of each principal component, different weights are assigned, thus generating flavor component contribution data for each tomato sample. A sliding window with a predetermined number of positions is used to control the sliding of the window across the flavor component contribution data. This sliding window extracts latent factors for each principal component; latent factors are the hidden structures or patterns in different flavor components. Through sliding window processing, potential factors influencing flavor scores can be captured at different window sizes and positions. The scores for each tomato sample on each latent factor are calculated, and these scores are arranged to generate a latent factor score matrix. The latent factor score matrix provides the performance of each tomato sample on different latent factors, reflecting the comprehensive score of each tomato sample across multiple flavor dimensions. Key factors in the latent factor score matrix are identified. Key factors are those that have the greatest impact on tomato flavor and are related to important flavor characteristics such as the sugar-acid ratio, aroma components, and acidity of tomatoes. Based on these key factors, their total scores will be calculated, and finally, flavor score data for each tomato sample will be generated using these total scores. The flavor score data can include scores for multiple dimensions such as flavor intensity, flavor balance, and flavor uniqueness, thereby comprehensively evaluating the flavor quality of the tomato.

[0096] In another embodiment, the calculation expression of the above embodiment is: Where F is the flavor score data vector of the final generated tomato sample, which includes scores for different flavor dimensions (such as flavor intensity, flavor balance, flavor uniqueness, etc.); n is the total number of tomato samples; and m is the number of principal components (obtained based on principal component analysis). The weight of each tomato sample i is determined by the number of samples or their representativeness; Let be the projection value of tomato sample i onto principal component j, representing the sample's contribution to that principal component; The eigenvalues ​​(variance contribution) of principal component j; For sliding window functions, it represents the latent factors extracted after sliding window processing on the flavor component contribution data of each sample, which in turn affect the calculation of flavor score; The contribution data of flavor components for the i-th tomato sample represents the specific contribution of the tomato sample in each flavor dimension.

[0097] In one embodiment, the step of identifying the similarity between multiple flavor rating data, classifying the tomato samples based on the similarity, and obtaining a sample group with similar flavor characteristics includes:

[0098] Feature extraction was performed on the flavor score data of each tomato sample to obtain the statistical feature vector of each tomato sample;

[0099] The similarity value between each pair of tomato samples is calculated based on the cosine similarity formula, and a similarity matrix is ​​constructed based on the similarity value.

[0100] Hierarchical clustering algorithm is used to cluster the similarity matrix to obtain preliminary classified samples;

[0101] Based on a preset threshold, cohesion analysis is performed on the preliminary classified samples to obtain sample groups;

[0102] The flavor score data of each sample group are aggregated into a group representative feature vector to obtain sample groups with similar flavor characteristics.

[0103] In the above embodiment, representative statistical features are extracted from the flavor score data of each tomato sample. Specifically, the flavor score data consists of multi-dimensional data, which reflects the scores of different flavor components of the tomato. This multi-dimensional data is transformed into a single, comparable statistical feature vector. This feature vector may include statistical measures such as mean, variance, standard deviation, skewness, and kurtosis, or other feature selection methods may be used to extract information that can effectively represent flavor characteristics. Through this process, the flavor score data of each tomato sample is transformed into a fixed-dimensional vector, and the similarity of different tomato samples in flavor characteristics is measured based on the cosine similarity formula. The similarity is measured by calculating the cosine value of the angle between two vectors. Specifically, the dot product of the two vectors is calculated, and then divided by the product of the magnitudes of the two vectors. The result is the similarity value of the two samples in the flavor feature space. The similarity value is usually in the range of 0 to 1, with a larger value indicating that the flavor characteristics of the two samples are more similar. Based on these similarity values, a similarity matrix is ​​constructed. Each element of this matrix represents the similarity between two tomato samples. The dimension of the matrix is ​​equal to the number of tomato samples, and the similarity value of each pair of samples is recorded in its corresponding position. A hierarchical clustering algorithm is employed, a clustering method that constructs a hierarchical structure of samples by continuously merging or splitting them. In this embodiment, all tomato samples are considered as independent clusters, each containing one sample. By calculating the similarity values ​​in the similarity matrix, the hierarchical clustering algorithm merges similar clusters based on their similarity, gradually building larger clusters until a complete clustering result is finally formed. Through this process, multiple preliminary classification sample groups are obtained, with samples within each group exhibiting high similarity in flavor characteristics. After clustering, cohesion analysis is performed to evaluate the tightness between samples within each sample group and the similarity of flavor characteristics among samples within the group. A preset threshold is set to analyze the cohesion in the preliminary classification samples. When the similarity between samples within a group exceeds the set threshold, this group is considered a valid group with high cohesion; otherwise, further adjustment or splitting of the group is required. Cohesion analysis can remove groups that are not consistent enough in flavor characteristics, thus obtaining more accurate and reliable sample groups. The flavor score data of each sample group is aggregated into a representative feature vector of the group. For each sample group, the flavor score data of all samples in the group are averaged or weighted averaged to obtain a feature vector representing the group.

[0104] In one embodiment, the step of generating a centroid feature vector for each sample group based on the flavor component data corresponding to the sample group, and generating a flavor scoring model based on the centroid feature vector, includes:

[0105] The flavor component data of each sample group are analyzed item by item to obtain the distribution pattern of each flavor component within the group;

[0106] The distribution patterns of flavor components within each sample group are analyzed for differences to obtain the local difference index of each flavor component.

[0107] Based on the local difference information, a recursive focusing analysis is performed on the flavor components of each sample group to determine the key flavor components whose influence on the central characteristics of the group is greater than a preset influence value, and to obtain the recursive weight value of each flavor component.

[0108] The weighted center point feature vector of flavor components in the sample group is calculated based on the recursive weight value, and the weighted center point feature vector is optimized based on the singular value decomposition algorithm to obtain the optimized center point feature vector.

[0109] The optimized center point feature vector is subjected to multiple mapping processing to construct a high-dimensional flavor representation model.

[0110] In the above embodiments, flavor component data for each sample group is analyzed item by item. The flavor component data of each tomato sample within the group is analyzed to extract the distribution pattern of each flavor component within the group. Specifically, flavor component data can include various chemical components, such as sugar, acidity, and aromatic substances, and their concentration values ​​may have different distributions in each sample. During the item-by-item analysis, the distribution of each flavor component value within the sample group is statistically analyzed, using statistical indicators such as mean, standard deviation, and skewness to describe its distribution characteristics. Through this step, the group distribution pattern of each flavor component can be obtained. Difference identification processing is performed on the flavor component distribution patterns of each sample group to analyze the degree of difference of each flavor component within the group, and based on this, a local difference index for each flavor component is obtained. The difference identification process is achieved by comparing the variability of each flavor component among different samples within the group. Flavor components with greater variability have a greater impact on the flavor score. The calculation of the local difference index involves performing a differential analysis on the distribution of each flavor component in the sample, measuring its heterogeneity within the group, and quantifying the difference value of each component accordingly. Recursive focusing analysis was employed to determine which flavor components have a significant impact on the central characteristics of the sample groups. By continuously adjusting the weights of flavor components, those with a significant influence on the group's central characteristics were identified. During this process, based on local difference information, the contribution value of each flavor component to the group's central characteristics was evaluated, and flavor components with contributions greater than a preset influence value were marked as key components. This process requires multiple iterations, recursively focusing on flavor components with a significant impact on the group's central characteristics to obtain a recursive weight value for each flavor component. After obtaining the recursive weight values ​​for each flavor component, a weighted central feature vector for each sample group was calculated. Each flavor component within the group was weighted according to its recursive weight value to obtain a weighted central feature vector. The weighted feature vector more accurately reflects the flavor characteristics of the samples within the group. Singular Value Decomposition (SVD) was introduced to decompose the matrix, identifying the most important components in the data and removing redundant information. In this embodiment, SVD was used to optimize the weighted central feature vector, thereby reducing noise and obtaining a more accurate and representative central feature vector. The optimized feature vectors are mapped to a higher-dimensional space to better capture and represent complex flavor characteristics. High-dimensional flavor representation models, by representing the flavor characteristics of each sample group in a higher-dimensional space, can more effectively reflect the differences between different sample groups and support the accurate calculation of flavor scores, thereby improving the model's ability to perceive flavor differences.

[0111] In one embodiment, the step of generating an evaluation result based on the flavor scoring model, mapping the sugar-acid ratio in the evaluation result to the flavor score, and obtaining a flavor evaluation report includes:

[0112] The flavor score characteristics of each tomato sample are evaluated based on the flavor scoring model to obtain the corresponding evaluation dataset;

[0113] The sugar-acid ratio of each tomato sample in the evaluation dataset is sorted to obtain an ascending sequence of sugar-acid ratios;

[0114] Based on the permutation sequence, the relative position of the sugar-acid ratio of each tomato sample in the evaluation dataset is calculated, and the relative position is calibrated to obtain the calibration index of each tomato sample.

[0115] The difference between the flavor score and the calibration index for each tomato sample is calculated to obtain the difference data between the flavor score and the sugar-acid ratio position;

[0116] The difference data is divided into multiple distribution spaces. Based on the flavor characteristics of each distribution space, the flavor of each tomato sample is subdivided and evaluated to obtain the performance value of each tomato sample in the corresponding distribution space. A flavor evaluation report is generated based on the performance value.

[0117] In the above embodiments, a flavor score for each tomato sample is generated based on the flavor scoring model through feature analysis of different flavor components, resulting in an evaluation dataset containing all samples. This dataset records the flavor score characteristics and related flavor index data for each tomato sample. The sugar-acid ratio of each tomato sample is sorted, with all samples arranged from smallest to largest, resulting in an ordered sugar-acid ratio sequence. The relative position of each tomato sample's sugar-acid ratio within the entire evaluation dataset is calculated, determining each sample's position within the sugar-acid ratio sequence—that is, the position interval of each sample's sugar-acid ratio compared to other samples. This is achieved through proportional analysis, such as calculating the percentage position of each sample's sugar-acid ratio within the total samples. In this way, a calibration index for each tomato sample in the sugar-acid ratio ranking can be obtained, providing a quantitative standard for each sample's relative position in the sugar-acid ratio sequence. The difference between each tomato sample's flavor score and its calibration index is calculated, reflecting the deviation between each sample's actual flavor score and its sugar-acid ratio position. Specifically, for each tomato sample, its flavor score is compared with its calibrated index, and the difference between the two is calculated. Samples with smaller differences indicate that their flavor score aligns with their sugar-acid ratio, while samples with larger differences show deviations in flavor components or different flavor characteristics. The difference data is divided into multiple distribution spaces, each corresponding to a flavor characteristic region. These regions are defined based on the differences between the sugar-acid ratio and the flavor score. For example, some distribution spaces correspond to samples with higher flavor scores, while others may be associated with samples with lower sugar-acid ratios. Within each distribution space, each tomato sample is further evaluated based on its respective flavor characteristics, combined with other flavor indicators such as sweetness, acidity, and aroma, to more comprehensively analyze the flavor characteristics of the tomato samples. In this process, each sample is calibrated according to its performance value within its corresponding distribution space, ultimately yielding the sample's overall flavor performance. The performance value is the comprehensive evaluation result of each tomato sample within that distribution space, which can be obtained through weighted calculation or comprehensive scoring. Based on the performance values ​​of each sample, a final flavor evaluation report is generated, which includes the performance of all tomato samples in different distribution spaces. Combined with the comprehensive analysis of sugar-acid ratio, flavor score, and other flavor components, a comprehensive flavor evaluation result is provided.

[0118] In one embodiment, the step of proportioning various standardized taste solutions in the kit according to the flavor evaluation report to match the flavor characteristics of the tomato sample includes:

[0119] Based on the flavor evaluation report, the correspondence between each flavor component and various standardized taste solutions in the kit is identified, and parameter values ​​that match the flavor characteristics of each standardized solution with the tomato sample are obtained.

[0120] Calculate the parameter values ​​for each type of standardized taste solution to generate a proportioning model;

[0121] The optimal ratio of each standardized taste solution is calculated based on the ratio model to minimize the difference between the flavor characteristics of the tomato sample and the flavor characteristics of the standardized taste solution in the kit.

[0122] An optimal formulation scheme is generated based on the optimal formulation, and various standardized taste solutions in the kit are adjusted based on the optimal formulation scheme.

[0123] In the above embodiments, based on the flavor assessment report, the flavor components of each tomato sample are identified, and these flavor components are mapped to various standardized taste solutions in the kit. The flavor assessment report extracts detailed data on the flavor components of the tomato samples through comprehensive analysis, including various taste characteristics such as sugar and acid. Each flavor component has a corresponding chemical composition and concentration range in the standardized taste solutions in the kit. By identifying the relationship between the flavor components in the tomato samples and the various solutions in the kit, parameter values ​​for each type of standardized solution that match the flavor characteristics of the tomato sample can be obtained. These parameter values ​​reflect how the components of the solutions in the kit match the characteristic flavor of the tomato sample, determining the required concentration and ratio of the solution during adjustment. The parameter values ​​for each type of standardized taste solution are calculated to generate a proportioning model. This model quantifies the role and influence of each standardized solution in the flavor matching process. For example, adjusting the sugar-acid ratio requires a combination of sweet and sour solutions of different concentrations. By establishing a mathematical model or using machine learning algorithms, it is possible to describe how the concentrations of these solutions produce the optimal matching effect with the flavor characteristics of the tomato sample. This model not only adjusts the proportions of different solutions but also comprehensively considers the interactions between various solutions to ensure that the final flavor ratio best matches the flavor characteristics of the tomato sample. The model calculates the optimal proportions of various standardized taste solutions, aiming to minimize the difference between the flavor characteristics of the tomato sample and the standardized taste solutions in the kit. Optimization algorithms, such as least squares and genetic algorithms, can be used to continuously adjust the solution proportions until the best flavor match is achieved. Based on the optimal proportion scheme, the concentrations and proportions of various standardized taste solutions in the kit are adjusted. This approach ensures that the flavor characteristics of each standardized taste solution in the kit more closely resemble the true flavor characteristics of the tomato sample. This process not only improves the accuracy of flavor testing but also helps researchers better understand the composition of tomato flavor, providing strong support for flavor improvement, breeding, and consumer preference research. Through these steps, the flavor characteristics of tomatoes are accurately evaluated and reproduced, providing a scientific basis for the application of the kit and promoting the development of flavor evaluation technology.

[0124] Reference Figure 2 A reagent kit-based tomato flavor evaluation device, comprising:

[0125] The analysis module 100 is used to perform multi-dimensional chemical composition analysis on tomato samples collected from different sources. The chemical composition includes soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample.

[0126] Extraction module 200 is used to extract the main flavor components from the flavor component data and generate flavor score data for each tomato sample based on the contribution of each main flavor component.

[0127] The identification module 300 is used to identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics;

[0128] The generation module 400 is used to generate a center point feature vector for each sample group based on the flavor component data corresponding to the sample group, and to generate a flavor scoring model based on the center point feature vector.

[0129] The evaluation module 500 is used to generate evaluation results based on the flavor scoring model, perform data mapping between the sugar-acid ratio and the flavor score in the evaluation results, and obtain a flavor evaluation report.

[0130] The matching module 600 is used to process the proportions of various standardized taste solutions in the kit according to the flavor evaluation report in order to match the flavor characteristics of the tomato sample.

[0131] In this embodiment, multi-dimensional chemical composition analysis comprehensively analyzes multiple flavor components in tomato samples. Similarity analysis of flavor score data and sample group classification techniques are employed to achieve more unified and accurate evaluations of tomato samples from different sources and cultivation environments. By extracting key flavor components and generating flavor scores based on their contribution, the objectivity and accuracy of flavor assessment are improved. Furthermore, this application can dynamically adjust based on the characteristic data of tomato samples, making the flavor assessment not only more comprehensive but also highly adaptable. It can automatically optimize evaluation results based on tomato samples from different batches or under different environmental conditions, overcoming the limitations of existing technologies that focus only on single-component indicators. This significantly improves the accuracy and adaptability of flavor assessment, providing a quantitative standard for the flavor characteristics of tomato samples. Combined with standardized taste solutions in the kit, consistency and efficiency are ensured during the flavor optimization process.

[0132] Reference Figure 3This application also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 3 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The database stores data and other information related to this application. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a reagent kit-based method for evaluating tomato flavor.

[0133] One embodiment of this application also provides a computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a tomato flavor evaluation method based on a reagent kit, comprising the following steps: performing multi-dimensional chemical composition analysis on tomato samples collected from different sources, the chemical composition including soluble solids content, acidity, sugar-acid ratio, and aroma components to obtain flavor component data for each tomato sample; extracting the main flavor components from the flavor component data, and generating flavor score data for each tomato sample based on the contribution of each main flavor component; identifying the similarity between multiple flavor score data, classifying the tomato samples based on the similarity to obtain sample groups with similar flavor characteristics; generating a central point feature vector for each sample group based on the flavor component data corresponding to the sample groups, and generating a flavor scoring model based on the central point feature vector; generating an evaluation result based on the flavor scoring model, mapping the sugar-acid ratio in the evaluation result to the flavor score to obtain a flavor evaluation report; and, based on the flavor evaluation report, processing the proportions of various standardized taste solutions in the reagent kit to match the flavor characteristics of the tomato samples.

[0134] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0135] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A reagent kit-based method for evaluating tomato flavor, characterized in that, The method includes: Multidimensional chemical composition analysis was performed on tomato samples collected from different sources. The chemical composition included soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample. Extract the main flavor components from the flavor component data, and generate flavor score data for each tomato sample based on the contribution of each main flavor component; Identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics; Based on the flavor component data corresponding to the sample groups, a center point feature vector is generated for each sample group, and a flavor scoring model is generated based on the center point feature vector. Based on the flavor scoring model, an evaluation result is generated, and the sugar-acid ratio in the evaluation result is mapped to the flavor score to obtain a flavor evaluation report. According to the flavor evaluation report, the various standardized taste solutions in the kit were proportioned to match the flavor characteristics of the tomato sample. The step of extracting the main flavor components from the flavor component data and generating flavor score data for each tomato sample based on the contribution of each main flavor component includes: Obtain the variance contribution of each flavor component in the flavor component data, generate a feature matrix containing the principal component contribution based on the variance contribution, and obtain the projection value of each tomato sample on each principal component. The projection values ​​of each tomato sample on each principal component dimension are weighted to generate flavor component contribution data for each tomato sample. A sliding window with a predetermined number of positions is set, and the sliding window is controlled to slide over the flavor component contribution data to obtain the potential factors of each principal component; Calculate the score of each tomato sample on each of the latent factors, arrange the score values, and generate a latent factor score matrix; Key factors in the latent factor score matrix are identified, the total score of the key factors is calculated based on the projection value, and the flavor score data of the tomato sample is generated based on the total score.

2. The tomato flavor evaluation method based on a reagent kit according to claim 1, characterized in that, The step of performing multi-dimensional chemical composition analysis on tomato samples collected from different sources, including soluble solids content, acidity, sugar-acid ratio, and aroma components, to obtain flavor component data for each tomato sample, includes: Obtain sample images of the tomato samples and identify the surface features of the tomatoes in the sample images to obtain the geometric feature data of each tomato sample; Chemical composition information of the tomato sample was collected by multi-dimensional sensors, and the chemical composition information was standardized to obtain a composition dataset. The chemical composition includes at least soluble solids content, acidity, sugar-acid ratio and aroma components. The chemical components in the component dataset are fused to obtain a comprehensive component set for each tomato sample. The comprehensive component set is associated with the geometric feature data to generate a flavor component dataset for each tomato sample.

3. The tomato flavor evaluation method based on a reagent kit according to claim 1, characterized in that, The step of identifying the similarity between multiple flavor rating data, classifying the tomato samples based on the similarity, and obtaining sample groups with similar flavor characteristics includes: Feature extraction was performed on the flavor score data of each tomato sample to obtain the statistical feature vector of each tomato sample; The similarity value between each pair of tomato samples is calculated based on the cosine similarity formula, and a similarity matrix is ​​constructed based on the similarity value. Hierarchical clustering algorithm is used to cluster the similarity matrix to obtain preliminary classified samples; Based on a preset threshold, cohesion analysis is performed on the preliminary classified samples to obtain sample groups; The flavor score data of each sample group are aggregated into a group representative feature vector to obtain sample groups with similar flavor characteristics.

4. The tomato flavor evaluation method based on a reagent kit according to claim 1, characterized in that, The step of generating a center point feature vector for each sample group based on the flavor component data corresponding to the sample group, and generating a flavor scoring model based on the center point feature vector, includes: The flavor component data of each sample group are analyzed item by item to obtain the distribution pattern of each flavor component within the group; The distribution patterns of flavor components within each sample group are analyzed for differences to obtain the local difference index of each flavor component. Based on the local difference index, a recursive focusing analysis is performed on the flavor components of each sample group to determine the key flavor components whose influence on the central characteristics of the group is greater than a preset influence value, and to obtain the recursive weight value of each flavor component. The weighted center point feature vector of flavor components in the sample group is calculated based on the recursive weight value, and the weighted center point feature vector is optimized based on the singular value decomposition algorithm to obtain the optimized center point feature vector. The optimized center point feature vector is subjected to multiple mapping processing to construct a high-dimensional flavor representation model.

5. The tomato flavor evaluation method based on a reagent kit according to claim 1, characterized in that, The steps of generating evaluation results based on the flavor scoring model, mapping the sugar-acid ratio in the evaluation results to the flavor score, and obtaining a flavor evaluation report include: The flavor score characteristics of each tomato sample are evaluated based on the flavor scoring model to obtain the corresponding evaluation dataset; The sugar-acid ratio of each tomato sample in the evaluation dataset is sorted to obtain an ascending sequence of sugar-acid ratios; Based on the permutation sequence, the relative position of the sugar-acid ratio of each tomato sample in the evaluation dataset is calculated, and the relative position is calibrated to obtain the calibration index of each tomato sample. The difference between the flavor score and the calibration index for each tomato sample is calculated to obtain the difference data between the flavor score and the sugar-acid ratio position; The difference data is divided into multiple distribution spaces. Based on the flavor characteristics of each distribution space, the flavor of each tomato sample is subdivided and evaluated to obtain the performance value of each tomato sample in the corresponding distribution space. A flavor evaluation report is generated based on the performance value.

6. The tomato flavor evaluation method based on a reagent kit according to claim 1, characterized in that, The step of proportioning the various standardized taste solutions in the kit according to the flavor evaluation report to match the flavor characteristics of the tomato sample includes: Based on the flavor evaluation report, the correspondence between each flavor component and various standardized taste solutions in the kit is identified, and parameter values ​​that match the flavor characteristics of each standardized solution with the tomato sample are obtained. Calculate the parameter values ​​for each type of standardized taste solution to generate a proportioning model; The optimal ratio of each standardized taste solution is calculated based on the ratio model to minimize the difference between the flavor characteristics of the tomato sample and the flavor characteristics of the standardized taste solution in the kit. An optimal formulation scheme is generated based on the optimal formulation, and various standardized taste solutions in the kit are adjusted based on the optimal formulation scheme.

7. A reagent kit-based tomato flavor evaluation device, applied to the method described in any one of claims 1 to 6, characterized in that, include: The analysis module is used to perform multi-dimensional chemical composition analysis on tomato samples collected from different sources. The chemical composition includes soluble solids content, acidity, sugar-acid ratio and aroma components, to obtain flavor component data for each tomato sample. The extraction module is used to extract the main flavor components from the flavor component data, and generate flavor score data for each tomato sample based on the contribution of each main flavor component; specifically, it includes: Obtain the variance contribution of each flavor component in the flavor component data, generate a feature matrix containing the principal component contribution based on the variance contribution, and obtain the projection value of each tomato sample on each principal component. The projection values ​​of each tomato sample on each principal component dimension are weighted to generate flavor component contribution data for each tomato sample. A sliding window with a predetermined number of positions is set, and the sliding window is controlled to slide over the flavor component contribution data to obtain the potential factors of each principal component; Calculate the score of each tomato sample on each of the latent factors, arrange the score values, and generate a latent factor score matrix; Identify key factors in the latent factor score matrix, calculate the total score of the key factors based on the projection values, and generate flavor score data for the tomato sample based on the total score; The identification module is used to identify the similarity between various flavor rating data, classify the tomato samples based on the similarity, and obtain sample groups with similar flavor characteristics; The generation module is used to generate a center point feature vector for each sample group based on the flavor component data corresponding to the sample group, and to generate a flavor scoring model based on the center point feature vector. The evaluation module is used to generate evaluation results based on the flavor scoring model, and to perform data mapping between the sugar-acid ratio and the flavor score in the evaluation results to obtain a flavor evaluation report. The matching module is used to process the proportions of various standardized taste solutions in the kit according to the flavor evaluation report, so as to match the flavor characteristics of the tomato sample.

8. A computer device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 6.

Images