A fruit and vegetable fermentation process bacterial population dynamic monitoring and intelligent control method

Abstract

The application discloses a kind of fruit and vegetable fermentation process flora dynamic monitoring and intelligent control method, it is related to process monitoring and intelligent control technical field, the method includes: in fruit and vegetable fermentation process, multiple-source information is collected, forms fermentation process multidimensional data stream, determines the stage state and quality stability state where fermentation is located, corresponding process control strategy is generated based on current fermentation stage state and preset stage difference regulation rule base, the process control strategy is output to fermentation control system execution.This application carries out synchronous collection and fusion processing to multiple-source information in fruit and vegetable fermentation process, is combined with flora variation feature extraction, fermentation stage state identification and safety risk determination, and generates process compensation strategy based on stage difference regulation rule base, realizes the phased dynamic regulation and control of fermentation process and multi-parameter collaborative control, significantly improves fermentation process state regulation precision, to realize the technical effect of fermentation process intelligent monitoring and quality stability improvement.

Description

Technical Field

[0001] This application relates to the field of process monitoring and intelligent control technology, and in particular to a method for dynamic monitoring and intelligent control of microbial communities in fruit and vegetable fermentation processes. Background Technology

[0002] Fruit and vegetable fermentation, as an important process in food processing, is widely used in the production of pickles, preserved fruits and vegetables, and various fermented foods. Its essence lies in utilizing the metabolic activities of microbial communities under specific environmental conditions to achieve flavor transformation and quality formation in raw materials. During fermentation, the microbial community structure undergoes dynamic evolution over time, simultaneously influenced by environmental factors such as temperature, salinity, pH, and dissolved oxygen. However, because traditional fermentation processes rely heavily on experience-based control and lack real-time monitoring methods for microbial community structure and fermentation status, precise control of the fermentation process is difficult to achieve, resulting in poor batch-to-batch consistency of products.

[0003] In actual fermentation processes, safety issues such as the temporary accumulation of nitrites and the formation of harmful metabolites like biogenic amines can easily arise. Furthermore, significant fluctuations in fermentation quality and unstable flavor can occur. Current technologies often rely on single physicochemical indicators for monitoring or adjust processes based solely on local parameters. They lack comprehensive utilization of multi-source data, including microbial community information, image information, and spectral information, and have not established differentiated control mechanisms for different fermentation stages. Consequently, they cannot effectively achieve overall optimization and safety control of the fermentation process. Therefore, how to achieve dynamic monitoring and intelligent control of fruit and vegetable fermentation processes based on multi-source information fusion has become a pressing technical problem to be solved in this field. Summary of the Invention

[0004] This application provides a method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation. It employs multi-source information fusion acquisition, dynamic analysis of microbial community structure, safety risk assessment, and collaborative control using a stage-differentiated control rule base. This enables real-time monitoring and accurate identification of the microbial succession process and fermentation status during fruit and vegetable fermentation, thereby achieving intelligent closed-loop control of the fermentation process. This improves the accuracy and stability of fermentation process control and enhances the batch consistency and food safety assurance capabilities of fermented products.

[0005] This application provides a method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation, comprising: collecting multi-source information during the fruit and vegetable fermentation process, including fermentation environmental parameter information, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information; performing time synchronization processing, abnormal data removal, and multi-source fusion processing on the multi-source information to form a multi-dimensional data stream of the fermentation process; extracting microbial community change features based on the multi-dimensional data stream of the fermentation process to characterize the evolutionary state of the microbial community structure; identifying the state of the fermentation process based on the microbial community change features and fermentation environmental parameter information to determine the stage state and quality stability state of the fermentation; and controlling the biogenic amine content during the fermentation process based on the quality and safety detection information. The fermentation process monitors and analyzes nitrite content, generates a safety risk assessment result based on the monitoring results, and uses this result for subsequent control triggering. When an abnormal state is identified based on the safety risk assessment result, a process control trigger signal is generated, and a corresponding process control strategy is generated based on the current fermentation stage state and a preset stage-differentiated control rule library. The process control strategies include salinity compensation strategy, acidity compensation strategy, sugar source compensation strategy, temperature regulation strategy, and anaerobic environment maintenance strategy. The process control strategies are output to the fermentation control system for execution, and feedback information after execution is collected. The fermentation process is continuously monitored, and the process control strategies are corrected based on the feedback information, thereby forming a closed-loop control process. Attached Figure Description

[0006] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0007] Figure 1 This is a flowchart illustrating a method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation, as provided in an embodiment of this application. Detailed Implementation

[0008] The above description is merely an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, specific embodiments of this application are given below.

[0009] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0010] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0011] This application provides a method for monitoring and intelligently controlling the microbial community dynamics during fruit and vegetable fermentation, such as... Figure 1 As shown, the method includes: S1: Collect multi-source information during the fruit and vegetable fermentation process. The multi-source information includes fermentation environment parameter information, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information. Perform time synchronization processing, abnormal data removal, and multi-source fusion processing on the multi-source information to form a multi-dimensional data stream of the fermentation process.

[0012] Furthermore, step S1 of this application also includes: S11. Collect fermentation environmental parameter information to obtain fermentation environmental parameter data, including temperature, pH value, salinity, dissolved oxygen, and sugar content; S12. Collect microbial community structure information to obtain microbial community structure data; S13. Collect image feature information and spectral detection information to obtain image feature data and spectral detection data; S14. Perform time synchronization processing and outlier removal on the fermentation environmental parameter data, microbial community structure data, image feature data, and spectral detection data to obtain preprocessed multi-source correlated data; S15. Perform fusion processing on the preprocessed multi-source correlated data to form a multi-dimensional data stream of the fermentation process.

[0013] Specifically, during the fruit and vegetable fermentation process, multi-source information is collected from the fermentation system. This multi-source information includes fermentation environmental parameters, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information. The fermentation environmental parameters characterize the basic physicochemical conditions of the fermentation process, including parameters such as temperature, pH, salinity, dissolved oxygen, and sugar content. The microbial community structure information characterizes the composition and distribution of the microbial community in the fermentation system. The image feature information reflects the changes in the appearance and morphology of the materials during fermentation. The spectral detection information characterizes the changes in the composition and metabolites of the substances in the fermentation system. The quality and safety detection information reflects changes in safety indicators such as biogenic amines and nitrites. The collected multi-source information is time-synchronized to eliminate discrepancies in acquisition time between different data sources, achieving alignment of multi-source data under a unified time benchmark. Further, outlier removal is performed on the multi-source information to eliminate noise and abnormal fluctuations during acquisition, improving data quality and stability. Based on this, the processed multi-source information is fused, linking and integrating different types of data to form a multi-dimensional data stream characterizing the overall state of the fermentation process, providing a unified data foundation for subsequent microbial community change analysis and fermentation state identification.

[0014] Furthermore, fermentation environmental parameter information is collected to obtain fermentation environmental parameter data. This fermentation environmental parameter information is used to characterize the basic physicochemical environmental state during fruit and vegetable fermentation, including key parameters such as temperature, pH value, salinity, dissolved oxygen, and sugar content, thereby reflecting the overall environmental changes of the fermentation system. Microbial community structure information is collected to obtain microbial community structure data. This microbial community structure data is used to characterize the composition and distribution characteristics of the microbial community in the fermentation system, reflecting the existence and changes of different microbial communities during fermentation. In summary, microbial community structure information is used to characterize the composition ratio and dynamic changes of the microbial community in the fruit and vegetable fermentation system. The acquisition of microbial community structure information can be achieved through microbial detection technologies, such as high-throughput sequencing or quantitative PCR detection. Specifically, microbial DNA can be extracted from fermentation samples, and the types of microorganisms in the samples can be identified and analyzed by 16S rRNA gene sequencing or ITS sequencing to obtain the classification results of different microorganisms and their corresponding sequence reading information. Further, based on the proportion of each microbial sequence reading in the total sequence reading, the relative abundance of different microbial communities in the fermentation system is determined, thereby forming microbial community structure data. In some implementations, quantitative PCR technology can be used to quantitatively detect the quantity of specific functional bacterial groups to supplement the acquisition of abundance information of key bacterial groups, thereby improving the accuracy and completeness of bacterial community structure data. Image feature information and spectral detection information are collected to obtain image feature data and spectral detection data, respectively. The image feature data reflects the changes in color, morphology, and surface state of the fermentation material, while the spectral detection data characterizes the changes in the composition and metabolites of the fermentation system, thus achieving non-contact, multi-dimensional perception of the fermentation process. Image feature information can be acquired by continuously or periodically acquiring images of the interior of the fermentation vessel or the surface of the material using industrial cameras or CCD imaging equipment. Preprocessing of the acquired images using existing image processing techniques, including noise reduction, white balance correction, and illumination compensation, can improve image quality. Furthermore, images can be analyzed using existing machine vision methods, such as color space conversion (RGB or HSV space), edge detection, and texture feature extraction, to obtain image feature data characterizing changes in color, morphology, and surface state of the fermentation material. Spectral detection information can be acquired using existing near-infrared (NIR), mid-infrared, or visible-near-infrared spectroscopy equipment. Non-contact or contact scanning measurements are performed on the materials in the fermentation system to obtain the corresponding spectral response signals. Subsequently, the spectral data is processed using existing spectral preprocessing methods, including baseline correction, scattering correction, and noise filtering, and characteristic band information or spectral absorption characteristics are extracted to obtain spectral detection data for characterizing changes in sugars, organic acids, and metabolites during fermentation.By combining existing industrial visual inspection technologies with spectral analysis technologies, multi-dimensional, non-destructive real-time monitoring of the fruit and vegetable fermentation process is achieved, providing a data foundation for subsequent microbial community change analysis and fermentation state identification. The fermentation environment parameter data, microbial community structure data, image feature data, and spectral detection data undergo time synchronization processing and outlier removal. Time synchronization processing eliminates time-dimensional deviations between different data acquisition sources, achieving unified time alignment of multi-source data. Outlier removal removes noise and abnormal fluctuations during the acquisition process, improving data reliability and consistency, resulting in preprocessed multi-source correlated data. This preprocessed multi-source correlated data is then fused, integrating and feature-fusing different types of data within the same data space to form a multi-dimensional data stream that comprehensively characterizes the overall state of the fermentation process, providing a unified data foundation for subsequent fermentation state identification and microbial community change analysis.

[0015] S2: Extract the microbial community change characteristics based on the multidimensional data stream of the fermentation process to characterize the evolutionary state of the microbial community structure.

[0016] Furthermore, step S2 of this application also includes: S21. Analyze the changes in microbial abundance based on the microbial structure data in the multidimensional data stream of the fermentation process to obtain microbial abundance change information; S22. Analyze the fermentation state image based on the image feature data in the multidimensional data stream of the fermentation process to obtain image change feature values; S23. Analyze the spectral response based on the spectral response data in the multidimensional data stream of the fermentation process to obtain spectral response feature values; S24. Evaluate microbial activity based on the microbial abundance change information, image change feature values, and spectral response feature values ​​to obtain microbial activity evaluation values.

[0017] Furthermore, step S21 of this application also includes: S21-1. Identify bacterial species from the bacterial community composition data to obtain bacterial community classification results; S21-2. Calculate the relative abundance information of different bacterial communities based on the bacterial community classification results; S21-3. Calculate the rate of change of bacterial community abundance based on the relative abundance information over continuous time, and analyze the trend of bacterial community change in conjunction with the abundance changes at different time points to obtain bacterial community abundance change information. The bacterial community abundance change information is determined in the following way: ;in, for t Information on changes in microbial community abundance over time. For the first k Bacterial flora in t Relative abundance at any given time For the first k The relative abundance of bacterial flora at the previous time point, The time interval between adjacent samplings. The relative abundance information represents the number of bacterial community categories; whereby the relative abundance information satisfies: ;in, for t Time of the first k The count value of the bacterial flora.

[0018] Furthermore, step S22 of this application also includes: S22-1. Identify the fermentation region from the image feature data to obtain an image of the target fermentation region; S22-2. Extract color features, texture features, and bubble distribution features from the image of the target fermentation region; S22-3. Perform change analysis based on the color features, texture features, and bubble distribution features to obtain image change feature values.

[0019] Furthermore, step S23 of this application also includes: S23-1. Perform noise filtering and band correction on the spectral response data to obtain preprocessed spectral data; S23-2. Extract feature band information based on the preprocessed spectral data, perform spectral response change analysis on the feature band information, and obtain spectral response feature values.

[0020] Specifically, based on the microbial community structure data in the multidimensional data stream of the fermentation process, microbial community abundance change analysis is performed to obtain microbial community abundance change information. This information characterizes the relative proportion changes and dynamic succession processes of different microbial communities during fermentation, reflecting the changing trends of the microbial community structure. Based on the image feature data in the multidimensional data stream of the fermentation process, fermentation state image analysis is performed to obtain image change feature values. These feature values ​​characterize the changes in the appearance of the material during fermentation, such as color changes, changes in tissue morphology, and changes in bubble distribution, thereby indirectly reflecting the fermentation activity level. Based on the spectral response data in the multidimensional data stream of the fermentation process, spectral response analysis is performed to obtain spectral response feature values. These feature values ​​characterize the changes in organic matter composition and metabolites in the fermentation system, such as sugar consumption, organic acid generation, and changes in volatile components, thereby reflecting the material transformation characteristics of the fermentation process. Based on the microbial community abundance change information, image change feature values, and spectral response feature values, microbial community activity is evaluated to obtain microbial community activity evaluation values. By fusing and analyzing multi-source feature information, a comprehensive evaluation of the overall activity status of the microbial community can be achieved, reflecting the metabolic activity and evolution trend of the microbial community during the fermentation process, thereby providing a basis for subsequent fermentation status identification and regulation.

[0021] Furthermore, the microbial community composition data is analyzed to identify microbial species and obtain microbial community classification results. Specifically, by parsing and comparing the microbial sequence information or detection results contained in the microbial community composition data, different microbial species present in the fermentation system are identified, thereby forming corresponding microbial community classification results to characterize the composition of the microbial community in the fermentation system. Based on the microbial community classification results, the relative abundance information of different microbial communities is statistically analyzed. This relative abundance information is used to characterize the proportion of each type of microbial community in the overall microbial community structure. By normalizing the quantity or detection signal intensity of different microbial communities, the proportion of each microbial community at different time points is obtained, thus reflecting the distribution characteristics of the microbial community structure. Based on the relative abundance information over continuous time, the rate of change of microbial community abundance is calculated, and combined with the abundance changes at different time points, a trend analysis of microbial community change is performed to obtain information on changes in microbial community abundance. Specifically, by continuously comparing and analyzing the relative abundance information collected at different time points, the rate of change of the abundance of each bacterial community over time is determined. Furthermore, the succession trend of the bacterial community is analyzed by combining the results of changes at multiple time points, thereby obtaining bacterial community abundance change information used to characterize the dynamic change process of the bacterial community. The bacterial community abundance change information is determined in the following way: ;in, for t Information on changes in microbial community abundance over time. For the first k Bacterial flora in t Relative abundance at any given time For the first k The relative abundance of bacterial flora at the previous time point, The time interval between adjacent samplings. The relative abundance information represents the number of bacterial community categories; whereby the relative abundance information satisfies: ;in, for t Time of the first kThe detection and count values ​​of microbial communities are determined. Relative abundance information of microbial communities is continuously collected at different fermentation time points, including the initial fermentation stage, intermediate fermentation stage, and maturity stage, or periodically sampled at preset time intervals to obtain corresponding time-series data, thus forming a relative abundance data sequence arranged in chronological order. Further, the relative abundance information of the same microbial community at adjacent time points is differentially calculated to obtain the abundance change of the microbial community per unit time. This abundance change is then normalized based on the time interval to obtain the rate of change of microbial community abundance, which characterizes the growth or decline rate of the microbial community during fermentation. After obtaining the rate of change of microbial community abundance corresponding to each time point, the rates of change of abundance at multiple time points are continuously compared and analyzed to determine the directional characteristics of microbial community abundance changes. When the rate of change of abundance is consistently positive, it indicates that the microbial community is in a continuous growth state; when the rate of change of abundance gradually decreases, it indicates that the growth of the microbial community tends to stabilize; when the rate of change of abundance changes from positive to negative, it indicates that the microbial community has entered a decline stage from the growth stage. Based on this, the succession path of the microbial community during the entire fermentation cycle is analyzed by combining the overall changes in the relative abundance of the microbial community at different time points, thereby obtaining information on the changes in microbial community abundance to characterize the dynamic changes of the microbial community.

[0022] Furthermore, fermentation region identification is performed on the image feature data to obtain an image of the target fermentation region. Specifically, by performing region segmentation processing on the acquired original image, effective regions related to the fermentation material are identified, and container background, light interference areas, and irrelevant areas are removed, thereby obtaining an image of the target fermentation region that can truly reflect the fermentation state. Color features, texture features, and bubble distribution features are extracted based on the target fermentation region image. Color features are used to characterize the color changes of the material during fermentation, such as the trend from light to dark or from green to yellow; texture features are used to characterize changes in the surface structure of the material, such as softening of the tissue and changes in particle distribution; bubble distribution features are used to characterize the generation and release of gas during fermentation, thereby reflecting the degree of microbial metabolic activity. Change analysis is performed based on the color features, texture features, and bubble distribution features to obtain image change feature values. Specifically, change analysis is performed based on the color features, texture features, and bubble distribution features to obtain image change feature values. Specifically, the color features, texture features, and bubble distribution features extracted at different time points are arranged in chronological order. Changes in each type of feature are calculated based on the feature difference or feature change ratio between adjacent time points to obtain the temporal changes of various image features. Furthermore, the changes in different types of image features are normalized to eliminate the influence of differences in feature dimensions. Then, the normalized changes in each type of feature are weighted and fused based on preset weights to obtain a comprehensive image change feature value, which is used to characterize the overall trend of material appearance changes during fermentation.

[0023] Furthermore, based on spectral detection data, the changes in material composition and metabolites during fermentation are analyzed to obtain spectral response characteristic values, which characterize the changes in chemical composition within the fruit and vegetable fermentation system. Noise filtering and band correction are performed on the spectral response data to obtain preprocessed spectral data. Specifically, since spectral detection is easily affected by ambient light interference, instrument noise, and scattering effects, preprocessing of the original spectral response data is necessary to improve data quality. Noise filtering can employ smoothing filtering to suppress high-frequency noise; band correction is used to correct spectral shifts within different wavelength ranges, thereby ensuring the consistency and comparability of spectral data under different times and sampling conditions. Feature band information is extracted from the preprocessed spectral data, and spectral response change analysis is performed on the feature band information to obtain spectral response characteristic values. Specifically, based on the characteristics of changes in material composition during fermentation, characteristic bands corresponding to sugars, organic acids, and volatile metabolites are selected from the spectral data, and the absorption intensity of the characteristic bands at different time points is compared and analyzed to reflect the changing trend of material composition during fermentation, thereby obtaining spectral response characteristic values ​​for characterizing the chemical changes during fermentation.

[0024] S3. Based on the characteristics of the microbial community change and the fermentation environment parameter information, the state of the fermentation process is identified to determine the stage state and quality stability state of the fermentation.

[0025] Specifically, the fermentation process of fruits and vegetables is identified based on the characteristics of microbial community changes and fermentation environmental parameters to determine the current stage and quality stability of the fermentation process, thereby achieving dynamic judgment of the fermentation process's operational status. Specifically, the microbial community change characteristics obtained in step S2 are correlated with the fermentation environmental parameters. The microbial community change characteristics characterize the dynamic succession of the microbial community during fermentation, while the fermentation environmental parameters characterize the current physicochemical conditions of the fermentation system, including changes in parameters such as temperature, pH, salinity, dissolved oxygen, and sugar content. By comprehensively analyzing the microbial community change characteristics and fermentation environmental parameters, the current operational status of the fermentation process is identified. Furthermore, the fermentation process is further identified based on changes in microbial abundance, microbial activity assessment values, and changes in fermentation environmental parameters to determine the stage of the fermentation process. Specifically, when the relative abundance of dominant microbial communities such as lactic acid bacteria consistently exceeds a preset growth threshold at multiple consecutive time points, and the pH value maintains a downward trend within a continuous time window, the fermentation process is determined to be in an active fermentation stage. When the rate of change in microbial abundance is lower than a preset stability threshold at multiple consecutive time points, and the fermentation environmental parameters change within a preset fluctuation range, the fermentation process is determined to have entered a stable fermentation stage. When the microbial activity assessment value is consistently lower than a preset activity threshold, and the magnitude of change in fermentation environmental parameters within a continuous time window is lower than a preset change threshold, the fermentation process is determined to have entered a mature stage. Furthermore, based on changes in fermentation environmental parameters, microbial community trends, and image change characteristics, the quality stability of the fermentation process is identified. Specifically, when fermentation environmental parameters are all within the preset allowable fluctuation range of the corresponding parameters at multiple consecutive time points, and the rate of change in microbial abundance is lower than a preset stability threshold, while the magnitude of change in image change characteristic values ​​within a continuous time window is lower than a preset image change threshold, the current fermentation process is determined to be in a quality stable state. When any fermentation environment parameter exceeds the preset allowable fluctuation range of the corresponding parameter at multiple consecutive time points, or the rate of change of microbial abundance is higher than the preset abnormal threshold, or the change amplitude of image change feature value within a continuous time window is higher than the preset image change threshold, it is determined that there is a risk of quality fluctuation in the fermentation process.

[0026] Furthermore, when a quality fluctuation risk is detected, the corresponding abnormal parameter information, microbial community change information, and image anomaly information are output to the subsequent safety risk assessment module and process control module to trigger the subsequent process control process.

[0027] S4: Based on the quality and safety testing information, monitor and analyze the content of biogenic amines and nitrites during fermentation, generate a safety risk assessment result based on the monitoring results, and use the safety risk assessment result for subsequent regulation triggering.

[0028] Specifically, based on quality and safety testing information, the content of biogenic amines and nitrites during fruit and vegetable fermentation is monitored and analyzed. Safety risk assessment results are generated based on the monitoring results to achieve dynamic identification of the safety status of the fermentation process and provide triggering basis for subsequent process control. Specifically, quality and safety testing information is periodically collected during fermentation, including biogenic amine detection data and nitrite detection data. The biogenic amine detection data may include the content information of biogenic amines such as histamine, putrescine, and tyramine; the nitrite detection data reflects the accumulation of nitrites during fermentation. Furthermore, the content of biogenic amines in fermentation samples can be detected and analyzed using liquid chromatography, spectroscopic detection, or electrochemical detection techniques; simultaneously, the nitrite content can be detected and analyzed using colorimetric detection, electrochemical sensing, or ion detection techniques to obtain corresponding safety testing data. After obtaining the biogenic amine and nitrite detection data, the detection results at different time points are continuously monitored and analyzed. Specifically, by analyzing the temporal changes in biogenic amine content at continuous time points, the risk of abnormal accumulation of biogenic amines during fermentation is determined. Specifically, a biogenic amine time-series data sequence is constructed by arranging the biogenic amine content at different time points in chronological order. Difference calculations are then performed based on the changes in biogenic amine content between adjacent time points to obtain the change in biogenic amine content within each time interval. Further, the change in biogenic amine content is normalized to the adjacent time intervals to obtain the rate of biogenic amine change, which is expressed as: ;in, Indicates the first i Biogenic amine content at each time point This indicates the corresponding time point. Based on this, a sliding window smoothing process is applied to the rate of change of biogenic amines across multiple consecutive time intervals to reduce the impact of fluctuations at a single time point on the analysis results, obtaining a stable rate of change used to characterize the accumulation trend of biogenic amines, thus achieving dynamic characterization of the intensity of biogenic amine accumulation during fermentation. Furthermore, when the rate of change of biogenic amine content at multiple consecutive time points is consistently higher than a preset growth threshold, an abnormal accumulation trend of biogenic amines is determined during fermentation; when the biogenic amine content exceeds the corresponding safety threshold, a risk of biogenic amine exceeding limits is determined. Simultaneously, by performing time-series change analysis on the nitrite content detection results at consecutive time points, a risk of nitrite exceeding limits during fermentation is determined. Specifically, nitrite content at different time points is used to form nitrite time-series data, and the rate of change of nitrite is calculated based on the changes in nitrite content within consecutive time windows. When the nitrite content at multiple consecutive time points continues to rise, and the rate of change of nitrite is higher than a preset change threshold, an abnormal accumulation trend of nitrite is determined during fermentation; when the nitrite content exceeds the corresponding safety threshold, a risk of nitrite exceeding limits is determined. The results of biogenic amine content detection and nitrite content detection are compared and analyzed with corresponding preset safety thresholds. When the biogenic amine content exceeds the corresponding biogenic amine safety threshold, a biogenic amine safety risk is determined; when the nitrite content exceeds the nitrite safety threshold, a nitrite safety risk is determined. In some embodiments, the safety risk can also be dynamically determined by combining the rate of change of content within a continuous time window. When the rate of change of biogenic amine content is continuously higher than the preset growth threshold, even if the current content has not reached the safety threshold, a potential biogenic amine accumulation risk can be determined in advance; when the nitrite content continuously rises within a continuous time window and the rate of change exceeds the preset change threshold, an abnormal nitrite accumulation trend in the fermentation process can be determined. A safety risk determination result is generated based on the above monitoring and analysis results. The safety risk determination result may include a normal state, a warning state, and an abnormal state. A normal state indicates that the current fermentation process is within a safe range; a warning state indicates that the detected index has an abnormal growth trend; an abnormal state indicates that the detected index has exceeded the corresponding safety threshold. Further, the safety risk determination result is output to the subsequent process control module to trigger the corresponding process control process. When the safety risk assessment result is a warning state or an abnormal state, a corresponding process control trigger signal is generated to execute subsequent compensation and control strategies, thereby reducing the risk of abnormal accumulation of biogenic amines and nitrites during fermentation and improving the quality stability and safety of fermentation products.

[0029] S5: When an abnormal state is identified based on the safety risk assessment result, a process control trigger signal is generated, and a corresponding process control strategy is generated based on the current fermentation stage state and the preset stage-differentiated control rule library. The process control strategy includes salinity compensation strategy, acidity compensation strategy, sugar source compensation strategy, temperature regulation strategy and anaerobic environment maintenance strategy.

[0030] Furthermore, step S5 of this application also includes: S51. Invoke the corresponding stage control rules based on the current fermentation stage status and determine the parameter constraint range corresponding to the current stage; S52. Determine the target compensation type based on the safety risk assessment result. The target compensation type includes at least one of salinity compensation, acidity compensation, sugar source compensation, temperature regulation, and anaerobic environment maintenance; S53. Based on the safety risk assessment result and the current fermentation stage status, match the corresponding compensation parameters from the stage-differentiated control rule library and generate a process compensation strategy based on the compensation parameters.

[0031] Furthermore, step S53 of this application also includes: S53-1. Determine the target compensation parameters based on the safety risk assessment results; S53-2. Adapt the target compensation parameters to the current fermentation stage state and determine the allowable adjustment range of the compensation parameters under the corresponding stage; S53-3. Generate a process compensation strategy based on the adapted compensation parameters, and compensate and correct the process compensation strategy according to the coupling influence relationship between different compensation parameters, which satisfies: ;in, For the revised first i Process compensation strategy parameters, For the first i Class target compensation parameters, For the first i Class compensation parameters and the first j Coupling influence coefficients between compensation parameters This is the coupling adjustment coefficient. To compensate for the number of parameter types.

[0032] Specifically, when an abnormal state is detected in the fermentation process, a corresponding process control strategy is generated based on the safety risk assessment result to achieve dynamic intervention and stable control of the fruit and vegetable fermentation process, thereby reducing the risk of abnormal accumulation of biogenic amines and nitrites during fermentation and improving the stability of fermentation quality. Specifically, when the safety risk assessment result generated in step S4 is a warning state or an abnormal state, a process control trigger signal is generated, and the corresponding process control process is initiated. The process control process generates the corresponding process control strategy based on the current fermentation stage state and a preset stage-differentiated control rule library. The stage-differentiated control rule library stores process control rules, parameter constraint ranges, and compensation strategy information corresponding to different fermentation stages. The bacterial community activity state, allowable range of environmental parameters, and control targets differ at different fermentation stages; therefore, differentiated control methods are adopted at different stages. For example, in the active fermentation stage, the focus is on controlling temperature and sugar source supply to maintain lactic acid bacteria activity; in the stable fermentation stage, the focus is on controlling salinity and anaerobic environment to maintain bacterial community stability; and in the maturation stage, the focus is on controlling acidity changes and the risk of nitrite accumulation to avoid quality fluctuations.

[0033] Furthermore, based on the fermentation stage status identified in step S3, the process control rules corresponding to the current stage are retrieved from the stage-differentiated control rule library. These process control rules include the target control range, allowable fluctuation range, and control priority information for each fermentation environmental parameter under the current stage. Further, the corresponding parameter constraint range is determined based on the current fermentation stage. For example, in the active fermentation stage, temperature parameters are allowed to vary within a small fluctuation range to ensure the metabolic activity of lactic acid bacteria; in the stable fermentation stage, salinity and pH are maintained within their corresponding stable ranges to reduce the risk of abnormal growth of contaminating microorganisms; in the maturation stage, the range of sugar content and dissolved oxygen changes is limited to avoid abnormal fluctuations in later fermentation stages. The target compensation type is determined based on the safety risk assessment results. Specifically, based on the safety risk assessment results generated in step S4, the types of abnormal risks existing in the current fermentation process are identified, and corresponding process compensation directions are matched. The target compensation type includes at least one of salinity compensation, acidity compensation, sugar source compensation, temperature regulation, and anaerobic environment maintenance. Furthermore, when an abnormal accumulation trend of biogenic amines is detected, the acidity compensation type and temperature regulation type can be determined, and abnormal metabolic processes can be inhibited by adjusting pH and fermentation temperature. When an abnormal accumulation trend of nitrite is detected, the salinity compensation type and anaerobic environment maintenance type can be determined, and abnormal proliferation of nitrate-reducing bacteria can be inhibited by improving salinity stability and reducing dissolved oxygen levels. When a decrease in bacterial activity is detected, the sugar source compensation type can be determined to maintain the metabolic activity of the target dominant bacterial population. Based on the target compensation type, the target compensation parameters for the corresponding stage are extracted from the stage-differentiated regulation rule library. These compensation parameters include the target salinity adjustment range, target pH adjustment range, target sugar content compensation range, target temperature regulation range, and target dissolved oxygen control range. Furthermore, the compensation parameters are stage-adapted to the current fermentation stage to avoid fermentation imbalance caused by using the same regulatory intensity at different stages. For example, a smaller salinity compensation is used in the active fermentation stage to avoid inhibiting lactic acid bacteria activity; a more stable anaerobic environment maintenance strategy is used in the stable fermentation stage to reduce the risk of contaminating bacteria growth; and the sugar source compensation is reduced in the maturation stage to avoid abnormal fermentation in the later stages. After obtaining the adapted compensation parameters, a process compensation strategy is generated based on the corresponding parameters. Among them, the salinity compensation strategy is used to adjust the salt concentration in the fermentation system; the acidity compensation strategy is used to adjust the pH value; the sugar source compensation strategy is used to adjust the fermentation substrate supply; the temperature control strategy is used to control the fermentation environment temperature; and the anaerobic environment maintenance strategy is used to adjust the dissolved oxygen state in the fermentation system.

[0034] Furthermore, based on the safety risk assessment results generated in step S4, the corresponding abnormal risk type in the current fermentation process is identified, and the target compensation parameters that need to be adjusted are determined in conjunction with the risk level corresponding to the abnormal risk. The target compensation parameters include at least one of the following: target salinity parameter, target acidity parameter, target sugar content parameter, target temperature parameter, and target dissolved oxygen parameter. For example, when a continuous increase in nitrite content is detected, salinity compensation parameters and anaerobic environment maintenance parameters can be determined; when an abnormal accumulation trend of biogenic amines is detected, acidity compensation parameters and temperature regulation parameters can be determined; when a decrease in microbial activity is detected, sugar source compensation parameters can be determined. Further, the target adjustment range of the corresponding compensation parameters is determined according to the risk level corresponding to the abnormal risk. The higher the risk level, the greater the adjustment range of the corresponding compensation parameters, in order to improve the regulatory response capability under abnormal conditions. The target compensation parameters are stage-adapted based on the current fermentation stage state, and the allowable adjustment range of the compensation parameters under the corresponding stage is determined. Since the microbial activity state and environmental tolerance range are different at different fermentation stages, stage-adaptation processing of the target compensation parameters is required to avoid fermentation imbalance caused by using a uniform adjustment intensity at different stages. Furthermore, based on the current fermentation stage, the allowable adjustment range of parameters corresponding to the current stage is extracted from the stage-differentiated control rule library, and the target compensation parameters are constrained and corrected based on the allowable adjustment range. For example, in the active fermentation stage, the metabolic activity of lactic acid bacteria is high, so the salinity compensation range and temperature adjustment range are limited to avoid inhibiting the activity of dominant bacteria; in the stable fermentation stage, the intensity of anaerobic environment maintenance is appropriately increased to reduce the risk of contaminant growth; in the maturation stage, the sugar source compensation range is reduced to avoid quality fluctuations caused by continued fermentation in the later stages. Since there is a coupling effect between different process compensation parameters, when one compensation parameter is adjusted, it may affect the fermentation state corresponding to other parameters. For example, an increase in salinity may inhibit the activity of bacteria; an increase in temperature may accelerate the metabolic process and affect acidity changes; sugar source compensation may affect the subsequent organic acid production rate. Therefore, when generating process compensation strategies, it is necessary to jointly correct the coupling effects between different compensation parameters. Furthermore, the coupling effect between different compensation parameters is analyzed, and the target compensation parameters are corrected based on the coupling effect coefficient, which satisfies ;in, For the revised first i Process compensation strategy parameters, For the first i Class target compensation parameters, For the first i Class compensation parameters and the first j Coupling influence coefficients between compensation parameters This is the coupling adjustment coefficient. The number of compensation parameter types. The coupling influence coefficient is used to characterize the degree of mutual influence between different compensation parameters; the coupling influence coefficient The coupling influence coefficient can be calculated based on the changes in the corresponding fermentation state parameters before and after the compensation parameter adjustment. The calculation can be performed based on the changes in the corresponding fermentation state parameters before and after the compensation parameter adjustment, which satisfies the following: ;in, For the first The amount of adjustment change of the compensation parameter, For the first After adjusting the compensation parameters, the first The response changes corresponding to fermentation-like state parameters. To prevent the use of tiny constants with a denominator of zero, a single compensation parameter is adjusted during fermentation, and the changes in other fermentation state parameters are monitored to statistically analyze the degree of linkage between different parameters. When a change in a certain compensation parameter causes a significant change in other state parameters, the corresponding coupling effect coefficient increases, indicating a strong coupling relationship between the two types of parameters. This coupling adjustment coefficient is used to control the overall coupling correction strength. Further, the coupling adjustment coefficient... The calculation can be performed based on the overall fluctuation of various fermentation environmental parameters at the current fermentation stage, and it satisfies the following: ;in, For the first The current values ​​of fermentation-like environmental parameters. This represents the target stable value of the corresponding parameter at the current fermentation stage. The number of fermentation environmental parameters, To prevent the use of tiny constants with a denominator of zero, the fluctuation state of the current fermentation system is determined by statistically analyzing the overall deviation of each fermentation environmental parameter from the target stable value at the current stage. When the overall fluctuation is large, the coupling adjustment coefficient is increased to enhance the multi-parameter coupling correction strength; when the overall fluctuation is small, the coupling adjustment coefficient is decreased to reduce excessive linkage correction between parameters. When the coupling influence between different compensation parameters is strong, the correction amplitude of the corresponding process compensation parameters is increased to reduce the risk of mutual interference between parameters. After obtaining the corrected process compensation strategy parameters, corresponding salinity compensation strategies, acidity compensation strategies, sugar source compensation strategies, temperature regulation strategies, and anaerobic environment maintenance strategies are generated and output to the subsequent fermentation control system for execution, thereby improving the stability and adaptability of the multi-parameter synergistic regulation process. Specifically, the corrected process compensation strategy parameters are derived from the output of step S53-3. S53-3 is generated based on the safety risk assessment results and fermentation stage status, combined with stage adaptation processing and compensation parameter constraint processing. Therefore, a hierarchical mapping relationship is formed between various process compensation parameters and the preceding fermentation process data. The data source paths for each process compensation strategy parameter are as follows: Step S1 collects fermentation environment parameter information, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information to form a multidimensional data stream of the fermentation process; based on the multidimensional data stream, in step S2, microbial community change characteristics, spectral features, and image features are extracted to form feature data characterizing the fermentation process status; in step S3, the fermentation process stage status is identified based on the feature data to obtain the current fermentation stage status and quality stability status; in step S4, analysis is performed based on quality and safety detection information to generate a safety risk assessment result; in step S5, based on the safety risk assessment result and fermentation stage status, the corrected process compensation strategy parameters are generated in step S53-3 through target compensation parameter determination, stage adaptation processing, and coupling correction processing. The modified process compensation strategy parameters establish a hierarchical mapping relationship with the original multi-source data collected by S1, realizing the continuous conversion from multi-source monitoring data to process control parameters.Furthermore, the modified process compensation strategy parameters are broken down according to parameter type and mapped to corresponding process control strategies. Specifically: the salinity compensation strategy is generated based on salinity compensation parameters and corresponds to brine addition control commands; the acidity compensation strategy is generated based on acidity compensation parameters and corresponds to acid-base adjustment control commands; the sugar source compensation strategy is generated based on sugar source compensation parameters and corresponds to sugar addition control commands; the temperature adjustment strategy is generated based on temperature compensation parameters and corresponds to heating or cooling control commands; and the anaerobic environment maintenance strategy is generated based on anaerobic control parameters and corresponds to aeration or sealing control commands. Each compensation parameter originates from the unified output of step S53-3 and maintains a one-to-one correspondence with the safety risk assessment results and fermentation stage status, thereby ensuring data consistency and logical correlation between different process strategies. The fermentation control system drives the brine addition unit, acid-base adjustment unit, sugar addition unit, temperature control unit, and anaerobic adjustment unit to perform corresponding adjustment operations according to the control commands, and collects execution feedback data in real time, including information on changes in dosage, temperature, and dissolved oxygen. Furthermore, the feedback data is transmitted back to the fermentation monitoring module and correlated with the multi-source data streams formed in steps S1 to S4 to achieve closed-loop control of the fermentation process.

[0035] S6: The process control strategy is output to the fermentation control system for execution, and feedback information after execution is collected to continuously monitor the fermentation process. The process control strategy is then corrected based on the feedback information, thereby forming a closed-loop control process.

[0036] Specifically, the generated process control strategy is output to the fermentation control system for execution, and the fermentation process is continuously monitored and the strategy is corrected based on the execution feedback information, thereby forming a closed-loop control mechanism for the fermentation process to improve the stability and adaptability of multi-parameter collaborative control. Specifically, the process control strategy generated in step S5 is converted into control command data and sent to the fermentation control system for execution via a communication interface. The fermentation control system includes multiple execution units for performing salinity adjustment, acidity adjustment, sugar source compensation, temperature control, and anaerobic environment control. During execution, the fermentation control system adjusts the corresponding execution units according to the received control commands and simultaneously collects feedback information after execution. The feedback information includes actual salinity change data, pH change data, sugar content change data, temperature change data, and dissolved oxygen change data. Further, the feedback information is time-aligned and correlated with the multi-dimensional data stream of the fermentation process formed in steps S1 to S4 to establish a correspondence between the execution results and the original monitoring status.

[0037] Based on this, the changes in fermentation state before and after execution are compared to determine the execution deviation of the current process control strategy. Execution deviation includes parameter adjustment deviation, response delay deviation, and control stability deviation. When the detected execution deviation exceeds a preset threshold, the process control strategy is corrected and updated based on feedback information, and the corresponding control commands are regenerated and output to the fermentation control system for execution. When the execution deviation is within the allowable range, the current process control strategy is maintained and continues to execute. Through this method, a closed-loop control process is achieved from "multi-source data acquisition—state identification—risk assessment—process control—execution feedback—strategy correction," thereby improving the stability, controllability, and product quality consistency of the fruit and vegetable fermentation process.

[0038] This application's embodiments construct a closed-loop control method encompassing "multi-source sensing data acquisition—microbial community change feature extraction—fermentation stage state identification—safety risk assessment—generation of stage-differentiated process compensation strategies—control execution and feedback correction." This establishes a multi-parameter linkage control system based on the dynamic evolution of the microbial community and the synergistic effect of fermentation environmental parameters. This enables simultaneous perception and collaborative analysis of microbial community structure changes, quality and safety risks, and fermentation stage states during fruit and vegetable fermentation. Consequently, it achieves precise staged control and multi-parameter adaptive optimization of the fermentation process, improving its stability, safety, and product quality consistency. Specifically, this application first synchronously acquires fermentation environmental parameter information, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information from multiple sources during fruit and vegetable fermentation. The acquired heterogeneous multi-source data undergoes time alignment, outlier removal, and fusion processing to form a multi-dimensional data stream characterizing the entire fermentation process, thereby achieving comprehensive perception and unified expression of the fermentation system's operational status. Subsequently, based on the multidimensional data stream, microbial community change characteristics, image change characteristics, and spectral response characteristics are extracted, and microbial community activity is comprehensively evaluated, thereby achieving quantitative characterization of the microbial community evolution state during fermentation. On this basis, combined with fermentation environmental parameter information, the fermentation process is stage-specific, determining the current process stage and quality stability, thus achieving phased analysis of the fermentation process. Simultaneously, based on quality and safety testing information, the content of biogenic amines and nitrites is monitored and analyzed, generating safety risk assessment results to identify potential safety risks during fermentation. When abnormal risks are detected, corresponding process compensation strategies are generated according to the fermentation stage state and a preset stage-differentiated control rule library, and parameters such as salinity, acidity, sugar source, temperature, and anaerobic environment are synergistically adjusted. Finally, these strategies are executed by the fermentation control system and closed-loop correction is performed based on feedback information, thereby achieving intelligent dynamic monitoring and precise control of the fruit and vegetable fermentation process.

[0039] In summary, this application constructs a multi-dimensional data stream for the fermentation process by synchronously collecting and fusing multi-source information. It then combines microbial community change feature extraction, fermentation stage state identification, and safety risk assessment mechanisms to generate a multi-parameter collaborative process compensation strategy based on a stage-differentiated control rule base. Furthermore, it introduces an execution feedback correction mechanism to achieve closed-loop control. This significantly improves the collaborative perception and control accuracy of microbial community evolution and environmental parameter changes during fermentation, and enhances the fermentation process's rapid response and adaptive adjustment capabilities to abnormal risks. Ultimately, this achieves precise, stage-specific intelligent control and improved quality stability in the fruit and vegetable fermentation process.

[0040] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation, characterized in that, The method includes: S1. Collect multi-source information during the fruit and vegetable fermentation process. The multi-source information includes fermentation environment parameter information, microbial community structure information, image feature information, spectral detection information, and quality and safety detection information. Perform time synchronization processing, abnormal data removal, and multi-source fusion processing on the multi-source information to form a multi-dimensional data stream of the fermentation process. S2. Extract the microbial community change characteristics based on the multidimensional data stream of the fermentation process to characterize the evolutionary state of the microbial community structure; S3. Based on the microbial community change characteristics and fermentation environment parameter information, the state of the fermentation process is identified to determine the stage state and quality stability state of the fermentation. S4. Based on the quality and safety testing information, monitor and analyze the content of biogenic amines and nitrites during fermentation, generate a safety risk assessment result based on the monitoring results, and use the safety risk assessment result for subsequent regulation triggering. S5. When the safety risk assessment result indicates that the state is abnormal, a process control trigger signal is generated, and a corresponding process control strategy is generated based on the current fermentation stage state and the preset stage-differentiated control rule library. The process control strategy includes salinity compensation strategy, acidity compensation strategy, sugar source compensation strategy, temperature regulation strategy and anaerobic environment maintenance strategy. S6. The process control strategy is output to the fermentation control system for execution, and feedback information after execution is collected to continuously monitor the fermentation process. The process control strategy is then corrected based on the feedback information, thereby forming a closed-loop control process.

2. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 1, characterized in that, Multi-source information is collected during the fruit and vegetable fermentation process. This multi-source information includes fermentation environment parameters, microbial community structure, image features, spectral detection, and quality and safety testing information. The multi-source information undergoes time synchronization processing, outlier removal, and multi-source fusion processing to form a multi-dimensional data stream of the fermentation process, including: S11. Collect fermentation environment parameter information and obtain fermentation environment parameter data, including temperature, pH value, salinity, dissolved oxygen and sugar content information; S12. Collect bacterial community structure information and obtain bacterial community structure data; S13. Collect image feature information and spectral detection information to obtain image feature data and spectral detection data; S14. Perform time synchronization processing and abnormal data removal on the fermentation environment parameter data, microbial community structure data, image feature data and spectral detection data to obtain preprocessed multi-source correlation data. S15. The preprocessed multi-source associated data is fused to form a multi-dimensional data stream of the fermentation process.

3. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 1, characterized in that, Based on the multidimensional data stream of the fermentation process, the characteristics of microbial community changes are extracted to characterize the evolutionary state of the microbial community structure, including: S21. Based on the microbial community structure data in the multidimensional data stream of the fermentation process, perform microbial community abundance change analysis to obtain microbial community abundance change information; S22. Based on the image feature data in the multidimensional data stream of the fermentation process, perform fermentation state image analysis to obtain image change feature values; S23. Perform spectral response analysis based on the spectral response data in the multidimensional data stream of the fermentation process to obtain spectral response characteristic values; S24. Based on the bacterial community abundance change information, image change feature value and spectral response feature value, bacterial community activity is evaluated to obtain the bacterial community activity evaluation value.

4. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 3, characterized in that, Microbial community composition data from the multidimensional data stream of the fermentation process are analyzed to obtain information on changes in microbial community abundance, including: S21-1. Identify bacterial species from the bacterial community composition data to obtain bacterial community classification results; S21-2. Based on the bacterial community classification results, statistical analysis is performed on the relative abundance information of different bacterial communities; S21-3. Calculate the rate of change of bacterial community abundance based on relative abundance information over continuous time, and analyze the trend of bacterial community change by combining the abundance changes at different time points to obtain bacterial community abundance change information. The bacterial community abundance change information is determined in the following way: ； in, This represents the information on changes in bacterial community abundance at time t. For the first k Bacterial flora in t Relative abundance at any given time For the first k The relative abundance of bacterial flora at the previous time point, The time interval between adjacent samplings. Number of bacterial community categories; The relative abundance information satisfies: ； in, for t Time of the first k The count value of the bacterial flora.

5. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 3, characterized in that, Based on the image feature data in the multidimensional data stream of the fermentation process, fermentation state image analysis is performed to obtain image change feature values, including: S22-1. Perform fermentation region identification on image feature data to obtain an image of the target fermentation region; S22-2. Extract color features, texture features, and bubble distribution features from the target fermentation area image; S22-3. Based on the color features, texture features, and bubble distribution features, perform change analysis to obtain image change feature values.

6. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 3, characterized in that, Spectral response analysis is performed based on the spectral response data in the multidimensional data stream of the fermentation process to obtain spectral response characteristic values, including: S23-1. Perform noise filtering and band correction on the spectral response data to obtain preprocessed spectral data; S23-2. Extract feature band information based on the preprocessed spectral data, perform spectral response change analysis on the feature band information, and obtain spectral response feature values.

7. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 1, characterized in that, When an abnormal state is identified, a process control trigger signal is generated, and a corresponding process control strategy is generated based on the current fermentation stage state and a preset stage-differentiated control rule library, including: S51. Invoke the corresponding stage control rules based on the current fermentation stage status, and determine the parameter constraint range corresponding to the current stage; S52. Determine the target compensation type based on the safety risk assessment results. The target compensation type includes at least one of salinity compensation, acidity compensation, sugar source compensation, temperature regulation, and anaerobic environment maintenance. S53. Based on the safety risk assessment results and the current fermentation stage status, match the corresponding compensation parameters from the stage-differentiated control rule library, and generate a process compensation strategy based on the compensation parameters.

8. The method for dynamic monitoring and intelligent control of microbial communities during fruit and vegetable fermentation as described in claim 7, characterized in that, Based on the safety risk assessment results and the current fermentation stage status, corresponding compensation parameters are matched from the stage-differentiated control rule base, and a process compensation strategy is generated based on the compensation parameters, including: S53-1. Determine the target compensation parameters based on the security risk assessment results; S53-2. Based on the current fermentation stage state, the target compensation parameters are adapted to the current stage, and the allowable adjustment range of the compensation parameters under the corresponding stage is determined. S53-3. Generate a process compensation strategy based on the adapted compensation parameters, and compensate and modify the process compensation strategy according to the coupling influence relationship between different compensation parameters, which satisfies: ； in, For the revised first i Process compensation strategy parameters, For the first i Class target compensation parameters, For the first i Class compensation parameters and the first j Coupling influence coefficients between compensation parameters This is the coupling adjustment coefficient. To compensate for the number of parameter types.

Images