A food full-link quality deterioration tracing method, system, medium and product

Abstract

The application discloses a food full-link quality deterioration tracing method and system, a medium and a product, and relates to the technical field of food quality tracing.The method is implemented, and a sensory description label in consumer feedback information is extracted by a tracing system; with the aid of a preset food quality knowledge graph, accurate mapping of subjective sensory information and objective physical and chemical indexes is realized, and the problem that subjective feedback and objective data are disconnected in a traditional tracing system is solved.The tracing system carries out dynamic deduction on food full-link monitoring data through a preset quality deterioration dynamics model, and upgrades food quality deterioration from static threshold judgment to dynamic numerical evolution track analysis, so that the cumulative deterioration effect of each circulation link can be quantified.The tracing system screens food quality deterioration responsibility links based on average deterioration rate and link deterioration increment, traces deterioration inducement of the responsibility links in combination with the preset food quality knowledge graph, accurately defines the deterioration responsibility and traces the reasons, and improves the tracing efficiency and accuracy.

Description

Technical Field

[0001] This application relates to the field of food quality traceability technology, and in particular to a method, system, medium and product for tracing the quality deterioration of food throughout the entire supply chain. Background Technology

[0002] With the improvement of people's living standards and the enhancement of food safety awareness, consumers have placed higher demands on food quality. Throughout the entire supply chain (from production to sales), food is susceptible to environmental factors such as temperature and humidity, which can lead to quality deterioration and result in consumers purchasing substandard food. Therefore, in the event of food quality problems, quickly tracing the responsible link and the cause of the deterioration is of great significance for improving food supply chain management and ensuring food quality.

[0003] Currently, food quality deterioration traceability methods mainly rely on Internet of Things (IoT) technology and information management systems. This involves deploying RFID tags, QR codes, and temperature and humidity sensors on food packaging or transport vehicles to record the time, location, and environmental monitoring data of the food at each stage of its distribution. When a consumer complaint about food quality is received, the traceability system retrieves historical monitoring records from the entire supply chain based on the food's batch information. Subsequently, technicians or the traceability system compare the recorded environmental monitoring parameters (such as temperature and humidity) with preset safety ranges. If the environmental monitoring data at a particular stage of distribution exceeds the preset safety range, that stage is identified as the responsible party for the food problem.

[0004] However, the above methods have significant limitations in practical applications. First, food quality deterioration is often a continuous, cumulative, and dynamic biochemical process, not merely a matter of whether environmental monitoring parameters exceed standards at a single moment. Methods based on static threshold comparisons struggle to quantify the long-term cumulative effects on food quality within compliant limits, making it impossible to pinpoint responsibility when food spoils even when environmental monitoring parameters are not significantly exceeded. Second, consumer feedback is typically based on subjective sensory descriptions (such as off-odors, discoloration, or deteriorated taste), while traceability systems record objective physical environmental data. Related technologies lack analytical mechanisms to translate subjective sensory descriptions into changes in specific physicochemical indicators, making it difficult to determine which physicochemical property has changed and to accurately analyze which specific operational step or environmental fluctuation in the distribution chain caused the decline in that particular sensory quality, thus leading to ambiguity in liability determination. Summary of the Invention

[0005] This application provides a method, system, medium, and product for tracing food quality deterioration across the entire supply chain, which can improve the accuracy of tracing food quality deterioration.

[0006] Firstly, this application provides a method for tracing food quality deterioration across the entire supply chain, applied to a traceability system. The method includes: receiving consumer feedback information and extracting sensory description labels from the consumer feedback information; determining abnormal physicochemical indicators based on the sensory description labels and a pre-defined food quality knowledge graph. The pre-defined food quality knowledge graph includes sensory description nodes, physicochemical indicator nodes, and deterioration inducing node nodes. Sensory description nodes are connected to physicochemical indicator nodes through a pre-defined sensory-physicochemical-biochemical mapping relationship, and physicochemical indicator nodes are connected to deterioration inducing node nodes through a pre-defined causal relationship of food quality deterioration. Sensory description nodes store consumer descriptions of food sensory attributes, physicochemical indicator nodes store physicochemical testing indicator information corresponding to food quality, and deterioration inducing node nodes store information on the causes of food physicochemical indicator deterioration; acquiring food supply chain monitoring data and inputting the food supply chain monitoring data into a pre-defined food quality knowledge graph. A food quality deterioration dynamics model was used to obtain food quality deterioration data. The food supply chain monitoring data included spatiotemporal trajectory information, environmental parameter information, and operational record information corresponding to multiple circulation stages. The food quality deterioration data included different numerical evolution trajectories corresponding to different food physicochemical indicators. Abnormal numerical evolution trajectories corresponding to abnormal physicochemical indicators were extracted from the food quality deterioration data, and these abnormal numerical evolution trajectories were divided into multiple segmented numerical evolution trajectories corresponding to different circulation stages. The average deterioration rate and the increment of deterioration at each stage were calculated. The first circulation stage with the highest average deterioration rate and the second circulation stage with the largest increment of deterioration were identified as the responsible stages for food quality deterioration. Based on the spatiotemporal trajectory information, environmental parameter information, and operational record information corresponding to the responsible stages, combined with abnormal physicochemical indicators and a pre-set food quality knowledge graph, the deterioration causes at the responsible stages were determined.

[0007] By adopting the above technical solutions, firstly, the traceability system extracts sensory description labels from consumer feedback information. Utilizing a pre-set food quality knowledge graph, it achieves a precise mapping between subjective sensory information and objective physicochemical indicators, alleviating the problem of the disconnect between subjective feedback and objective data in traditional traceability systems and providing a clear target for food quality deterioration analysis. Simultaneously, the traceability system dynamically extrapolates the monitoring data across the entire food supply chain through a pre-set quality deterioration dynamics model, upgrading food quality deterioration judgment from static threshold judgment to dynamic numerical evolution trajectory analysis, thereby quantifying the cumulative deterioration effect at each stage of circulation. Furthermore, the traceability system screens responsible links for food quality deterioration based on both average deterioration rate and the incremental deterioration at each stage. Combined with the pre-set food quality knowledge graph, it traces the deterioration causes at these responsible links, achieving precise definition of deterioration responsibility and tracing the causes. This provides a scientific basis for targeted optimization of the food supply chain, improves the efficiency and accuracy of food quality deterioration traceability, and effectively safeguards food quality and safety.

[0008] In conjunction with some embodiments of the first aspect, in some embodiments, physicochemical index nodes correspond to physicochemical indicators; food full-chain monitoring data is input into a preset quality deterioration kinetic model to obtain food quality deterioration data. The food full-chain monitoring data includes spatiotemporal trajectory information, environmental parameter information, and operation record information corresponding to multiple circulation links. The food quality deterioration data includes different numerical evolution trajectories corresponding to different food physicochemical indicators. Specifically, this includes: obtaining the kinetic reaction equation, benchmark activation energy parameter, and initial state detection value for each physicochemical indicator; calculating the time-varying reaction rate constant of each physicochemical indicator at different time points using the Arrhenius equation based on the spatiotemporal trajectory information and environmental parameter information, combined with the benchmark activation energy parameter of each physicochemical indicator; starting from the initial state detection value, performing parallel differential iterative calculations on multiple physicochemical indicators based on the kinetic reaction equation and the time-varying reaction rate constant to obtain a set of full numerical evolution trajectories including the changes of multiple physicochemical indicators over time; and determining the set of full numerical evolution trajectories as food quality deterioration data.

[0009] By adopting the above technical solution, the traceability system clarifies the complete path from obtaining basic parameters of physicochemical indicators to generating full numerical evolution trajectories. It introduces the Arrhenius equation to calculate the time-varying reaction rate constant of each physicochemical indicator at different time points, fully combining the influence of spatiotemporal trajectory information and environmental parameter information on the reaction rate of physicochemical indicators. This achieves dynamic quantitative simulation of the deterioration process, overcoming the shortcomings of traditional traceability methods in depicting the dynamic process of food quality deterioration. Through kinetic reaction equations and time-varying reaction rate constants, the traceability system performs parallel differential iterative calculations on multiple physicochemical indicators, simultaneously generating numerical evolution trajectories corresponding to each indicator. This comprehensively reflects the multi-dimensional deterioration state of food quality, avoiding the one-sidedness of single-indicator analysis. This refined calculation method provides high-precision data support for subsequent abnormal trajectory extraction and responsibility identification, ensuring the accuracy and reliability of traceability results. It provides a solid technical foundation for end-to-end deterioration traceability and also provides more valuable quantitative data for food quality prediction and control.

[0010] In conjunction with some embodiments of the first aspect, in some embodiments, obtaining the initial state detection value of each physicochemical indicator specifically includes: parsing the food production batch identifier from the operation record information; searching in a preset production quality inspection database for a quality inspection record uniquely associated with the production batch identifier; if it exists, extracting the measured value corresponding to the physicochemical indicator from the quality inspection record and determining the measured value as the initial state detection value; if it does not exist, obtaining the preset standard formula benchmark data or historical average detection data corresponding to the food to determine the initial state detection value of the physicochemical indicator.

[0011] By adopting the above technical solution, the traceability system prioritizes extracting the measured values ​​corresponding to the physicochemical indicators from the quality inspection records uniquely associated with the production batch identifier. These measured values ​​are then used as the initial state detection values, maximizing the restoration of the quality state of food in its early stages of production. When no quality inspection records are available, they are supplemented using preset standard formula benchmark data or historical average detection data. This effectively avoids the problem of model failure or distorted calculation results due to missing data, ensuring the continuity and integrity of the traceability process. The initial state detection value serves as the starting point for calculating deterioration data, and its accuracy directly affects the reliability of subsequent numerical evolution trajectories. This method, based on the principle of "measurement priority, alternative supplementation," balances data accuracy and the practicality of the solution. It fully utilizes existing production and quality inspection data resources while providing a reasonable and feasible solution for scenarios lacking data. This allows the preset quality deterioration dynamics model to be applied to food supply chains with varying degrees of production record completeness, expanding the applicability of the traceability method and improving the accuracy and credibility of the end-to-end deterioration traceability results.

[0012] In conjunction with some embodiments of the first aspect, in some embodiments, the preset food quality knowledge graph also includes physicochemical index coupling relationships, which are used to define the cascading influence paths between different physicochemical index nodes; starting from the initial state detection value, based on the kinetic reaction equation and the time-varying reaction rate constant, parallel differential iterative calculations are performed on multiple physicochemical indices to obtain a set of full numerical evolution trajectories of multiple physicochemical indices over time, specifically including: when performing iterative calculations at the current time step, based on the physicochemical index coupling relationships, identifying the driving dependency relationships between the target physicochemical index and other physicochemical indices to determine the driving index and response index, where the target physicochemical index is any physicochemical index, and the other physicochemical indices are physicochemical indices other than the target physicochemical index; obtaining the driving index at the current time step. Instantaneous simulated values ​​within time steps are obtained, and based on a pre-defined coupling influence coefficient model, the degradation rate coupling gain coefficient of the driving index on the response index is determined. The time-varying response rate constant of the response index at the current time step is corrected using the degradation rate coupling gain coefficient to obtain the corrected time-varying response rate constant. Based on the corrected time-varying response rate constant, the predicted simulated values ​​of the response index at the next time step are deduced. By iteratively executing the steps of identifying the driving dependency between the target physicochemical index and other physicochemical indices, determining the degradation rate coupling gain coefficient of the driving index on the response index, and correcting the time-varying response rate constant of the response index at the current time step using the degradation rate coupling gain coefficient within all time steps, a set of full numerical evolution trajectories including multiple physicochemical indices changing over time is obtained.

[0013] By adopting the above technical solution, the traceability system introduces the coupling relationship of physicochemical indicators and optimizes the parallel differential iterative calculation of multiple physicochemical indicators, significantly improving the simulation realism and accuracy of food quality deterioration data. In the actual food deterioration process, the various physicochemical indicators do not change in isolation, but rather have a cascading effect of mutual influence. Traditional traceability methods ignore the coupling relationship of physicochemical indicators, leading to significant deviations between simulation results and actual conditions. This method, in the iterative calculation, identifies driving and response indicators based on the coupling relationship of physicochemical indicators. It determines the coupling gain coefficient of the deterioration rate of the response indicator through the instantaneous simulated value of the driving indicator, and then corrects the time-varying reaction rate constant of the response indicator, realizing a dynamic characterization of the interaction between indicators. This refined coupling correction mechanism enables the entire numerical evolution trajectory set to more realistically reflect the complex process of food quality deterioration, avoiding the one-sidedness caused by independent calculation of a single indicator.

[0014] In conjunction with some embodiments of the first aspect, in some embodiments, the average deterioration rate and the incremental deterioration at each stage of the numerical evolution trajectory are calculated respectively. The first circulation stage with the largest average deterioration rate and the second circulation stage with the largest incremental deterioration are identified as the responsible stages for food quality deterioration. Specifically, this includes: calculating the initial remaining quality life and the final remaining quality life, the average deterioration rate, and the incremental deterioration at each circulation stage based on the segmented numerical evolution trajectory; calculating the difference ratio between the initial remaining quality life and the final remaining quality life to obtain the quality life loss rate; and using the quality life loss rate as a weighting factor to weight and correct the average deterioration rate and the incremental deterioration at each stage to determine the responsible stages for food quality deterioration.

[0015] By adopting the above technical solution, the traceability system introduces the quality lifespan loss rate as a weighting factor, overcoming the limitations of traditional single-dimensional methods for determining responsibility for food quality deterioration and achieving comprehensiveness and accuracy in responsibility identification. The traceability system not only calculates the average deterioration rate and the increment of deterioration at each stage of the process, but also derives the start and end points of the remaining quality lifespan through segmented numerical evolution trajectories, thereby obtaining the quality lifespan loss rate. This rate is then used to weight and correct the core judgment indicators. This method fully considers the impact of different stages of the process on the overall quality lifespan of the food, avoiding potential misjudgments that may occur when relying solely on the rate or increment of deterioration—for example, some stages may have a fast rate of deterioration but a short duration, with limited impact on the overall quality lifespan, while weighting can accurately measure their actual responsibility proportion. Through comprehensive weighted analysis of multi-dimensional indicators, the system can more objectively locate the responsible stage with the greatest impact on food quality deterioration, providing a precise basis for subsequent targeted rectification and effectively improving the quality management efficiency of the food supply chain.

[0016] In conjunction with some embodiments of the first aspect, in some embodiments, based on the segmented numerical evolution trajectory, the starting remaining quality lifespan and the ending remaining quality lifespan of each circulation stage are calculated. Specifically, this includes: extracting the initial quality feature value corresponding to the start time of the current circulation stage and the ending quality feature value corresponding to the end time of the current circulation stage from the segmented numerical evolution trajectory; obtaining a preset food quality kinetic model and a preset quality failure threshold; using the initial quality feature value as an input parameter, substituting it into the preset food quality kinetic model, calculating the first duration required for evolution to the preset quality failure threshold, and recording the first duration as the starting remaining quality lifespan; using the ending quality feature value as an input parameter, substituting it into the preset food quality kinetic model, calculating the second duration required for evolution to the preset quality failure threshold, and recording the second duration as the ending remaining quality lifespan.

[0017] By employing the aforementioned technical solution, the traceability system extracts quality characteristic values ​​of key time nodes from the segmented numerical evolution trajectory. Combined with a preset food quality dynamics model and a preset quality failure threshold, it quantifies the remaining quality lifespan of food before and after each stage of distribution. This quantification method transforms abstract quality states into concrete duration data, allowing for direct comparison of the degree of food quality loss at different stages of distribution. Compared to traditional, fuzzy quality assessment methods, this approach accurately captures the impact of each stage of distribution on food lifespan, ensuring the scientific validity and consistency of the quality lifespan loss rate calculation. Weighting the average deterioration rate and the incremental deterioration at each stage based on the quality lifespan loss rate makes the determination of responsible stages more convincing, while providing a quantifiable evaluation standard for the quality control effectiveness of each stage of the food supply chain.

[0018] In conjunction with some embodiments of the first aspect, in some embodiments, the method for constructing a pre-defined food quality knowledge graph specifically includes: acquiring academic literature, industry standard documents, and historical customer complaint records in the field of food science as a raw corpus; using natural language processing technology to extract entities and relationships from the raw corpus, identifying sensory description entities, physicochemical indicator entities, and deterioration cause entities; based on co-occurrence frequency analysis and semantic similarity calculation, constructing a mapping probability matrix between sensory description entities and physicochemical indicator entities, and a causal probability matrix between physicochemical indicator entities and deterioration cause entities; establishing connection edges between sensory description entities and physicochemical indicator entities, and between physicochemical indicator entities and deterioration cause entities, according to the mapping probability matrix and the causal probability matrix, and determining the weights of the connection edges to form an initial food quality knowledge graph; sending the initial food quality knowledge graph to experts for manual correction to obtain the pre-defined food quality knowledge graph.

[0019] By adopting the above technical solutions, a scientific and comprehensive pre-defined food quality knowledge graph generation system was constructed, providing core data support and logical foundation for the entire food quality deterioration traceability method. This effectively solves the problem of the disconnect between subjective feedback and objective indicators and deterioration causes in traditional traceability systems. This method uses academic literature, industry standard documents, and historical customer complaint records in the field of food science as the original corpus. Through natural language processing technology, it accurately extracts three core entities: sensory descriptions, physicochemical indicators, and deterioration causes. Then, based on co-occurrence frequency analysis and semantic similarity calculation, a causal probability matrix is ​​constructed, establishing relationships between entities and assigning weights. Finally, it is manually revised by experts to form the pre-defined food quality knowledge graph. This "data-driven + expert verification" construction model ensures both the comprehensiveness and objectivity of the pre-defined food quality knowledge graph and compensates for the limitations of machine extraction through manual revision, ensuring the accuracy of entity associations. The pre-defined food quality knowledge graph can efficiently map sensory description labels to abnormal physicochemical indicators and trace abnormal physicochemical indicators to deterioration causes, providing a clear logical link for the entire traceability process and improving traceability efficiency and intelligence.

[0020] In a second aspect, embodiments of this application provide a traceability system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, and the memory is used to store computer program code, which includes computer instructions, and the one or more processors call the computer instructions to cause the traceability system to perform the method described in the first aspect and any possible implementation thereof.

[0021] Thirdly, embodiments of this application provide a computer program product containing instructions that, when the computer program product is run on a traceability system, cause the traceability system to perform the method described in the first aspect and any possible implementation thereof.

[0022] Fourthly, embodiments of this application provide a computer-readable storage medium including instructions that, when executed on a traceability system, cause the traceability system to perform the method described in the first aspect and any possible implementation thereof.

[0023] Understandably, the traceability system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the methods provided in the embodiments of this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0025] 1. By adopting the above technical solution, firstly, the traceability system extracts sensory description labels from consumer feedback information. Utilizing a pre-set food quality knowledge graph, it achieves a precise mapping between subjective sensory information and objective physicochemical indicators, alleviating the problem of the disconnect between subjective feedback and objective data in traditional traceability systems, and providing a clear target for food quality deterioration analysis. Simultaneously, the traceability system dynamically extrapolates the monitoring data across the entire food supply chain through a pre-set quality deterioration dynamics model, upgrading food quality deterioration judgment from static threshold judgment to dynamic numerical evolution trajectory analysis, thereby quantifying the cumulative deterioration effect at each stage of circulation. Furthermore, the traceability system screens responsible links for food quality deterioration based on both average deterioration rate and the incremental deterioration at each stage. Combined with the pre-set food quality knowledge graph, it traces the deterioration causes at these responsible links, achieving precise definition of deterioration responsibility and tracing the causes. This provides a scientific basis for targeted optimization of the food supply chain, improves the efficiency and accuracy of food quality deterioration traceability, and effectively safeguards food quality and safety.

[0026] 2. By adopting the above technical solution, the traceability system clarifies the complete path from obtaining basic parameters of physicochemical indicators to generating the full numerical evolution trajectory. It introduces the Arrhenius equation to calculate the time-varying reaction rate constant of each physicochemical indicator at different time points, fully combining the influence of spatiotemporal trajectory information and environmental parameter information on the reaction rate of physicochemical indicators. This achieves dynamic quantitative simulation of the deterioration process, overcoming the shortcomings of traditional traceability methods in depicting the dynamic process of food quality deterioration. The traceability system performs parallel differential iterative calculations on multiple physicochemical indicators through kinetic reaction equations and time-varying reaction rate constants, simultaneously generating the numerical evolution trajectories corresponding to each indicator. This comprehensively reflects the multi-dimensional deterioration state of food quality, avoiding the one-sidedness of single-indicator analysis. This refined calculation method provides high-precision data support for subsequent abnormal trajectory extraction and responsibility identification, ensuring the accuracy and reliability of traceability results. It provides a solid technical foundation for end-to-end deterioration traceability and also provides more valuable quantitative data for the prediction and control of food quality.

[0027] 3. By adopting the above technical solution, the traceability system introduces the coupling relationship of physicochemical indicators, optimizing the parallel differential iterative calculation of multiple physicochemical indicators, significantly improving the simulation realism and accuracy of food quality deterioration data. In the actual food deterioration process, the various physicochemical indicators do not change in isolation, but rather have a cascading effect of mutual influence. Traditional traceability methods ignore the coupling relationship of physicochemical indicators, leading to significant deviations between simulation results and actual conditions. This method, in the iterative calculation, identifies driving and response indicators based on the coupling relationship of physicochemical indicators. It determines the coupling gain coefficient of the deterioration rate of the response indicator through the instantaneous simulated value of the driving indicator, and then corrects the time-varying reaction rate constant of the response indicator, achieving a dynamic characterization of the interaction between indicators. This refined coupling correction mechanism allows the entire numerical evolution trajectory set to more realistically reflect the complex process of food quality deterioration, avoiding the one-sidedness caused by independent calculation of a single indicator. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a method for tracing food quality deterioration across the entire supply chain, as described in this application.

[0029] Figure 2 This is a schematic diagram of the physical device structure of the traceability system in the embodiments of this application. Detailed Implementation

[0030] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to any or all possible combinations including one or more of the listed items.

[0031] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0032] The following describes the process of the method provided in this implementation. Please refer to [link / reference]. Figure 1 This is a flowchart illustrating a method for tracing food quality deterioration across the entire supply chain, as described in this application.

[0033] S101. Receive consumer feedback information and extract sensory description labels from the consumer feedback information;

[0034] Consumer feedback refers to complaints or evaluations of food quality issues submitted by consumers through various channels (such as telephone, apps, websites, etc.). Sensory description labels are descriptive words used by consumers to describe abnormal characteristics of food, such as "off-flavor," "discoloration," and "deteriorated taste."

[0035] Specifically, the traceability system first obtains food feedback information submitted by consumers (i.e., consumer feedback information) through a pre-defined information interface. Then, it uses natural language processing technology to segment and tag the consumer feedback information, identifying adjectives and noun phrases that describe the sensory characteristics of the food. The traceability system then standardizes these adjectives and noun phrases, mapping them to predefined sensory description labels. For example, "has a strange taste" will be extracted as the "off-taste" label, and "sticky texture" will be extracted as the "abnormal texture" label.

[0036] S102. Based on the sensory description labels and the preset food quality knowledge graph, determine the abnormal physicochemical indicators. The preset food quality knowledge graph includes sensory description nodes, physicochemical indicator nodes, and deterioration cause nodes. Sensory description nodes are connected to physicochemical indicator nodes through preset sensory-physicochemical-biochemical mapping relationships. Physicochemical indicator nodes are connected to deterioration cause nodes through preset food quality deterioration causal relationships. Sensory description nodes are used to store consumers' descriptions of the sensory attributes of food. Physicochemical indicator nodes are used to store the physicochemical testing indicator information corresponding to the food quality. Deterioration cause nodes are used to store the information on the reasons for the deterioration of the food's physicochemical indicators.

[0037] The pre-defined food quality knowledge graph refers to a database storing food quality-related knowledge in a graph structure, including various nodes and their relationships, such as sensory descriptions, physicochemical indicators, and deterioration causes. Sensory description nodes represent food quality characteristics perceptible to consumers. Physicochemical indicator nodes represent physicochemical indicators of food that can be detected by instruments. Deterioration cause nodes represent various reasons that lead to a decline in food quality. Abnormal physicochemical indicators refer to specific physicochemical testing indicators associated with abnormal sensory phenomena. The pre-defined sensory-physicochemical-biochemical mapping relationship refers to the correspondence rules between sensory descriptions and physicochemical indicators. The pre-defined causal relationship of food quality deterioration refers to the association rules between abnormal physicochemical indicators and causes of deterioration.

[0038] Specifically, the traceability system uses a pre-defined sensory physicochemical mapping relationship within a pre-defined food quality knowledge graph to find physicochemical indicator nodes that are connected to the sensory description labels. The system comprehensively considers the weight and confidence level of the connecting edges, selecting one or more physicochemical indicators most likely related to the current sensory abnormality as the abnormal physicochemical indicators. For example, the "off-odor" label might map to physicochemical indicators such as "peroxide value" and "acid value," while "abnormal taste" might map to indicators such as "viscosity" and "moisture content."

[0039] Optionally, in general, the construction of a pre-defined food quality knowledge graph can be achieved through the following methods, which are not limited here: Acquire academic literature, industry standard documents, and historical customer complaint records in the field of food science as the original corpus; use natural language processing technology to extract entities and relationships from the original corpus, identifying sensory description entities, physicochemical indicator entities, and deterioration cause entities; based on co-occurrence frequency analysis and semantic similarity calculation, construct a mapping probability matrix between sensory description entities and physicochemical indicator entities, and a causal probability matrix between physicochemical indicator entities and deterioration cause entities; based on the mapping probability matrix and the causal probability matrix, establish connection edges between sensory description entities and physicochemical indicator entities, and between physicochemical indicator entities and deterioration cause entities, and determine the weights of the connection edges to form an initial food quality knowledge graph; send the initial food quality knowledge graph to experts for manual correction to obtain the pre-defined food quality knowledge graph.

[0040] The original corpus refers to the basic text data set used to construct a pre-defined food quality knowledge graph. Academic literature refers to research papers, monographs, and other academic materials in the field of food science. Industry standard documents refer to normative documents such as quality standards and testing specifications in the food industry. Historical customer complaint records refer to records of processed consumer complaint cases. Entity extraction refers to the process of identifying specific types of named entities from text. Relation extraction refers to the process of determining the semantic relationships between entities. Co-occurrence frequency refers to the number of times two entities co-occur in the same text segment. Semantic similarity refers to the degree of semantic closeness between words or phrases. The mapping probability matrix is ​​used to represent the correspondence probability between sensory descriptions and physicochemical indicators. The causal probability matrix is ​​used to represent the causal relationship probability between physicochemical indicators and deterioration causes.

[0041] Specifically, the traceability system first establishes a comprehensive corpus (raw corpus) containing academic literature, industry standard documents, and historical customer complaints in the field of food science through literature collection and data crawling. Then, using natural language processing techniques, including named entity recognition and dependency parsing, it extracts three core entities from the raw corpus: sensory descriptive entities (such as "off-odor" and "discoloration"), physicochemical indicator entities (such as "peroxide value" and "pH value"), and deterioration-causing entities (such as "excessively high temperature" and "excessively strong light"). The traceability system then analyzes the relationships between entities: by statistically analyzing the frequency of two types of entities appearing together in the same text paragraph, and combining this with semantic similarity calculated from word vectors, it constructs two probability matrices.

[0042] (1) Sensory description-physicochemical index mapping probability matrix: representing the probability of association between each pair of sensory descriptions and physicochemical indexes;

[0043] (2) Physicochemical index-deterioration cause causal probability matrix: representing the strength of the causal relationship between each pair of physicochemical indexes and deterioration causes;

[0044] Based on these two probability matrices, the traceability system establishes weighted connection edges between entity nodes in the pre-defined food quality knowledge graph, with the weights determined according to the corresponding probabilities.

[0045] The traceability system automatically constructs an initial food quality knowledge graph, which is then submitted to experts in the field of food science for review and correction. This includes deleting erroneous associations, supplementing missing relationships, and adjusting weight values, ultimately forming a reliable, pre-defined food quality knowledge graph. This hybrid construction method of "data-driven + expert verification" ensures both the breadth of knowledge coverage and the accuracy of knowledge associations.

[0046] For ease of understanding, taking "fresh meat products" as an example, the specific storage structure of the food quality knowledge graph is pre-defined as follows:

[0047] Sensory description node: contains node ID="S01", tag name="rancid smell", thesaurus set={"rancid smell", "stale smell"};

[0048] Physicochemical index nodes: including node ID="P01", index name="peroxide value (POV)", detection unit="g / 100g", standard threshold="0.25"; including node ID="P02", index name="volatile basic nitrogen (TVB-N)";

[0049] Deterioration inducing nodes: including node ID="R01", inducing name="photo-oxidation"; node ID="R02", inducing name="excessive time for normal temperature transportation".

[0050] The connection relationships and weights are set as follows: A mapping connection is established between the "rancid smell" node and the "peroxide value" node, with an edge attribute weight of 0.92 (meaning that when consumers report a rancid smell, there is a 92% probability that the peroxide value exceeds the standard); a causal connection is established between the "peroxide value" node and the "light oxidation" node, with an edge attribute type of "positive correlation" and a weight of 0.85.

[0051] S103. Obtain food full-chain monitoring data, input the food full-chain monitoring data into the preset quality deterioration dynamic model, and obtain food quality deterioration data. The food full-chain monitoring data includes spatiotemporal trajectory information, environmental parameter information and operation record information corresponding to multiple circulation links. The food quality deterioration data includes different numerical evolution trajectories corresponding to different food physicochemical indicators.

[0052] Among them, food end-to-end monitoring data refers to various data collected throughout the entire process of food production and sales. Spatiotemporal trajectory information represents the location and time records of food at each stage of its circulation. Environmental parameter information refers to environmental condition data such as temperature and humidity. Operational record information includes specific operational data for each stage of production, storage, and transportation. The preset quality deterioration kinetic model is a mathematical model describing the law of food quality change over time. Food quality deterioration data refers to the change data of various physicochemical indicators of food calculated through the preset quality deterioration kinetic model. Numerical evolution trajectory represents the numerical sequence of a certain physicochemical indicator changing over time.

[0053] Specifically, firstly, the traceability system retrieves the corresponding end-to-end food monitoring data from the IoT monitoring system and business systems based on food batch information. This data includes the time and location, environmental parameters, and operational records for each stage of the process. Then, the traceability system inputs this end-to-end food monitoring data into a pre-defined quality deterioration kinetic model. This model, based on food science theory, considers the impact of factors such as temperature and humidity on food quality and describes the changing patterns of various physicochemical indicators through mathematical equations. By calculating using this model, the numerical trajectory of each physicochemical indicator from production to sales can be obtained, forming complete data on food quality deterioration.

[0054] Optionally, under normal circumstances, physicochemical index nodes correspond to physicochemical indicators. The food whole-chain monitoring data is input into a preset quality deterioration kinetic model to obtain food quality deterioration data. This data includes spatiotemporal trajectory information, environmental parameter information, and operation record information corresponding to multiple circulation stages. The food quality deterioration data, including different numerical evolution trajectories corresponding to different food physicochemical indicators, can be achieved in the following ways, without limitation: Obtain the kinetic reaction equation, baseline activation energy parameter, and initial state detection value for each physicochemical indicator; based on the spatiotemporal trajectory information and environmental parameter information, and combined with the baseline activation energy parameter of each physicochemical indicator, calculate the time-varying reaction rate constant of each physicochemical indicator at different time points using the Arrhenius equation; starting from the initial state detection value, perform parallel differential iterative calculations on multiple physicochemical indicators based on the kinetic reaction equation and time-varying reaction rate constant to obtain a set of full numerical evolution trajectories including the changes of multiple physicochemical indicators over time; and determine this set of full numerical evolution trajectories as the food quality deterioration data.

[0055] Among these, the kinetic reaction equation refers to the mathematical equation describing the change of physicochemical indicators over time. The baseline activation energy parameter represents the minimum energy threshold required to induce changes in physicochemical indicators. The initial state detection value refers to the initial value of the physicochemical indicators at the completion of production. The Arrhenius equation is the fundamental equation describing the relationship between chemical reaction rate and temperature. The time-varying reaction rate constant represents the rate of change of physicochemical indicators under different environmental conditions. Parallel differential iterative calculation refers to the process of simultaneously performing continuous numerical calculations on multiple physicochemical indicators. The complete numerical evolution trajectory set refers to the complete set of change data for all physicochemical indicators.

[0056] Specifically, the traceability system's calculation process consists of four main steps:

[0057] (1) Obtaining basic parameters:

[0058] Kinetic reaction equation: A mathematical expression describing the change law of this physicochemical index;

[0059] Reference activation energy parameter: A parameter reflecting the sensitivity of this physicochemical index to environmental factors;

[0060] Initial state test value: The measured value or standard value of this physicochemical index at the end of the production process;

[0061] (2) Calculation of time-varying reaction rate:

[0062] The tracing system uses the Arrhenius equation and combines the monitored environmental parameter information (mainly temperature) to calculate the actual reaction rate of physicochemical indicators at each time point: by substituting the benchmark activation energy parameter into the Arrhenius equation and considering environmental factors such as temperature and humidity corresponding to the spatiotemporal trajectory, a time-varying reaction rate constant reflecting the actual environmental impact is obtained.

[0063] (3) Parallel differential iterative calculation:

[0064] The traceability system simultaneously performs numerical evolution calculations on all physicochemical indicators: starting from the initial state detection value, the calculation rules are set according to the kinetic reaction equation, and the time-varying reaction rate constant is used for step-by-step iteration to obtain the complete trajectory of each physicochemical indicator changing over time.

[0065] (4) Data set integration:

[0066] The traceability system integrates the numerical evolution trajectory of all physicochemical indicators into a unified data set: ensuring that the time points of different physicochemical indicators are aligned, forming the final data on food quality deterioration.

[0067] Optionally, parallel differential iterative calculations are performed using the Euler method or the fourth-order Runge-Kutta method at discrete time steps. Taking the Euler method as an example, for any physicochemical index i, the calculation logic for deducing the value C_i(t_n+1) at the nth time step (time point t_n) at the n+1th time step (time point t_n) is as follows: C_i(t_n+1) = C_i(t_n) ± k_i(T_n) × [C_i(t_n)]^m × Δt;

[0068] in:

[0069] C_i(t_n) is the current accumulated value of the physicochemical index;

[0070] Δt is the time step (e.g., set to 5 minutes);

[0071] T_n represents the ambient temperature at that moment;

[0072] m is the reaction order (usually 0th or 1st order);

[0073] k_i(T_n) is the reaction rate constant at the current temperature calculated based on the Arrhenius equation, i.e.: k_i(T_n) = k_ref × exp[(-Ea / R) × (1 / T_n - 1 / T_ref)];

[0074] The tracing system continuously executes the above formula in units of Δt along the entire time axis until the end of the time axis, thereby generating a continuous numerical evolution trajectory composed of discrete points.

[0075] Optionally, under normal circumstances, the initial state test value of each physicochemical indicator can be obtained in the following ways, without limitation: parse the food production batch identifier from the operation record information; search the preset production quality inspection database for a quality inspection record uniquely associated with the production batch identifier; if it exists, extract the measured value corresponding to the physicochemical indicator from the quality inspection record and determine the measured value as the initial state test value; if it does not exist, obtain the preset standard formula benchmark data or historical average test data corresponding to the food to determine the initial state test value of the physicochemical indicator.

[0076] Among these, the production batch identifier refers to the coding information used to uniquely identify a batch of food. The pre-set production quality inspection database refers to a database system that stores quality inspection data from the food production process. Quality inspection records represent the original records of physicochemical index testing of food. Measured values ​​refer to the actual physicochemical index values ​​measured by instruments. Pre-set standard formula benchmark data refers to the standard values ​​of physicochemical indexes determined based on the product formula. Historical average test data refers to the statistical average of historical quality inspection data for similar products.

[0077] Specifically, the traceability system employs a layered acquisition strategy of "prioritizing actual testing and supplementing with alternatives": the system extracts food batch-related fields from operation record information, parses them to obtain standardized production batch identifiers, and ensures the uniqueness and accuracy of these identifiers. The system then accesses a pre-set production quality inspection database, using the production batch identifier as a keyword for precise matching to determine if a corresponding quality inspection record exists.

[0078] If quality inspection records exist, the traceability system locates the specific item, extracts the measured values ​​corresponding to the physicochemical indicators, and directly uses the measured values ​​as the initial state test values. If no quality inspection records exist, the traceability system obtains the standard formula information of the food or queries the statistical results of historical test data, and selects an appropriate replacement value after comprehensive evaluation.

[0079] Optionally, in general, the preset food quality knowledge graph also includes physicochemical index coupling relationships. These relationships are used to define the cascading influence paths between different physicochemical index nodes. Starting from the initial state detection value, based on the kinetic reaction equation and the time-varying reaction rate constant, parallel differential iterative calculations are performed on multiple physicochemical indices to obtain a set of full numerical evolution trajectories of multiple physicochemical indices over time. This can be achieved in the following way, without limitation: When performing iterative calculations at the current time step, based on the physicochemical index coupling relationships, the driving dependency relationships between the target physicochemical index and other physicochemical indices are identified to determine the driving and response indices. The target physicochemical index is any one of the physicochemical indices, and the other physicochemical indices are those other than the target physicochemical index. The driving index at the current time step is obtained. Instantaneous simulated values ​​within time steps are obtained, and based on a pre-defined coupling influence coefficient model, the degradation rate coupling gain coefficient of the driving index on the response index is determined. The time-varying response rate constant of the response index at the current time step is corrected using the degradation rate coupling gain coefficient to obtain the corrected time-varying response rate constant. Based on the corrected time-varying response rate constant, the predicted simulated values ​​of the response index at the next time step are deduced. By iteratively executing the steps of identifying the driving dependency between the target physicochemical index and other physicochemical indices, determining the degradation rate coupling gain coefficient of the driving index on the response index, and correcting the time-varying response rate constant of the response index at the current time step using the degradation rate coupling gain coefficient within all time steps, a set of full numerical evolution trajectories including multiple physicochemical indices changing over time is obtained.

[0080] Among these, the coupling relationship of physicochemical indicators refers to the mutual influence relationship between different physicochemical indicators. The cascaded influence path represents the process by which a change in one physicochemical indicator transmits and affects other physicochemical indicators. The driving dependency relationship refers to the dominant and passive relationship among physicochemical indicators. The driving indicator is the dominant physicochemical indicator that influences other physicochemical indicators. The response indicator is the passive physicochemical indicator that changes under the influence of the driving indicator. The instantaneous simulated value refers to the calculated value at a specific point in time. The coupling influence coefficient model is a mathematical model that quantitatively describes the degree of influence of the driving indicator on the response indicator. The degradation rate coupling gain coefficient represents the degree of influence of the change in the driving indicator on the degradation rate of the response indicator. The corrected time-varying reaction rate constant refers to the actual reaction rate after considering the coupling effect.

[0081] Specifically, the traceability system employs the following steps for coupled computation:

[0082] (1) Coupling analysis:

[0083] At each time step: a physicochemical indicator is selected as the target physicochemical indicator. Based on the coupling relationship of physicochemical indicators in the preset food quality knowledge graph, the interaction between the target physicochemical indicator and other physicochemical indicators is analyzed to determine the relationship network between driving indicators and response indicators. For example, in meat products, free radicals generated by lipid oxidation will attack proteins, leading to an accelerated protein denaturation rate. At this time, the lipid oxidation indicator is the driving indicator, and the protein denaturation indicator is the response indicator. The coupling coefficient reflects this accelerating effect.

[0084] (2) Calculation of driving effect:

[0085] For each pair of coupled indices: obtain the current value of the driving index, apply the preset coupling influence coefficient model, calculate the influence intensity of the change of the driving index on the response index, and obtain the degradation rate coupling gain coefficient.

[0086] (3) Rate constant correction:

[0087] For each response index: obtain the original time-varying response rate constant, apply the deterioration rate coupling gain coefficient to correct it, and obtain the corrected time-varying response rate constant considering the coupling effect;

[0088] (4) Iterative deduction and calculation:

[0089] Using the corrected time-varying reaction rate constant, calculate the physicochemical parameters for the next time step, and repeat the coupling analysis and correction process until all time steps have been calculated.

[0090] In this embodiment, in order to quantify the mutual influence between physicochemical indicators, the preset coupling influence coefficient model is specifically calculated using the following nonlinear correction formula:

[0091] Assuming that an accelerating coupling effect is identified between a lipid oxidation indicator (driving indicator A) and a protein denaturation indicator (response indicator B), the tracing system first obtains the instantaneous simulated value C_A(t) of driving indicator A at the current time step t, and then calculates the degradation rate coupling gain coefficient γ, using the following formula:

[0092] γ=1+α×tanh[β×(C_A(t)-C_A_thresh) / C_A_thresh];

[0093] in:

[0094] C_A_thresh is the preset safety threshold for the driving indicator A;

[0095] α is the maximum gain amplitude coefficient (e.g., 0.5).

[0096] β is a sensitivity adjustment factor (e.g., 2.0).

[0097] tanh is the hyperbolic tangent function, used to limit the degradation rate coupling gain coefficient within a reasonable range.

[0098] Next, the tracing system uses the deterioration rate coupling gain coefficient γ to correct the time-varying response rate constant k_B(t) of the response index B. The correction formula is: k_B'(t) = k_B(t) × γ;

[0099] When C_A(t) is within the limit, γ approaches 1, and the coupling effect is not obvious; when C_A(t) is significantly exceeded, γ increases, thus simulating the chain reaction phenomenon of "lipid oxidation accelerating protein deterioration".

[0100] S104. Extract the abnormal value evolution trajectory corresponding to the abnormal physicochemical indicators from the food quality deterioration data, and divide the abnormal value evolution trajectory into multiple segmented value evolution trajectories corresponding to different circulation links.

[0101] Among them, the abnormal value evolution trajectory refers to the complete numerical change curve of the abnormal physicochemical index throughout the entire monitoring period. The segmented numerical evolution trajectory represents the sub-trajectory obtained by dividing the complete abnormal value evolution trajectory according to the time period corresponding to different circulation links. Circulation links refer to different stages in the food supply chain, including production, warehousing, transportation, and sales.

[0102] Specifically, firstly, the traceability system filters out the abnormal numerical evolution trajectories corresponding to the identified abnormal physicochemical indicators from the food quality deterioration data. Then, based on the switching points in the food's distribution process, the system divides the complete abnormal numerical evolution trajectory into multiple sub-trajectory segments (i.e., segmented numerical evolution trajectories). Each sub-trajectory segment corresponds to a specific distribution stage and includes a continuous numerical sequence from the start to the end of that stage. For example, if a batch of food has gone through four distribution stages—"production-warehouse-transportation-sales"—and "peroxide value" is identified as an abnormal physicochemical indicator, the traceability system will divide the complete abnormal numerical evolution trajectory of the peroxide value into segmented numerical evolution trajectories corresponding to these four distribution stages.

[0103] S105. Calculate the average deterioration rate and the deterioration increment of each segment's numerical evolution trajectory, and identify the first circulation link with the largest average deterioration rate and the second circulation link with the largest deterioration increment as the responsible links for food quality deterioration.

[0104] The average rate of deterioration refers to the average rate of change of a physicochemical indicator within a certain processing stage, expressed as the change in value per unit time. The increment of deterioration at a particular processing stage represents the overall change in the physicochemical indicator within that stage, and is the difference between the final and initial values ​​of that stage. The first processing stage refers to the stage with the highest average rate of deterioration. The second processing stage refers to the stage with the largest increment of deterioration. The responsible stage for food quality deterioration refers to the critical processing stage that has the greatest impact on food quality deterioration.

[0105] Specifically, the traceability system first analyzes and calculates the segmented numerical evolution trajectory of each circulation stage. By calculating the slope of the segmented numerical evolution trajectory, the average deterioration rate is obtained; by calculating the difference between the start and end points of the segmented numerical evolution trajectory, the deterioration increment at each stage is obtained. The traceability system then ranks and compares these two indicators across all circulation stages, identifying the circulation stage with the highest average deterioration rate (first circulation stage) and the circulation stage with the largest deterioration increment (second circulation stage). Both of these circulation stages have a significant impact on food quality deterioration: a high deterioration rate indicates a drastic decline in quality at that stage, and a large deterioration increment indicates significant cumulative losses at that stage. Therefore, the traceability system jointly identifies these two circulation stages as the responsible stages for food quality deterioration.

[0106] Optionally, under normal circumstances, the average degradation rate and the increment of degradation at each stage of the numerical evolution trajectory can be calculated separately. The first stage with the highest average degradation rate and the second stage with the largest increment of degradation can be identified as the responsible stages for food quality degradation. This can be achieved in the following way, without limitation: Based on the segmented numerical evolution trajectory, calculate the initial remaining quality life and the final remaining quality life, the average degradation rate, and the increment of degradation at each stage; calculate the ratio of the difference between the initial remaining quality life and the final remaining quality life to obtain the quality life loss rate; use the quality life loss rate as a weighting factor to weight and correct the average degradation rate and the increment of degradation at each stage to determine the responsible stages for food quality degradation.

[0107] Here, "Starting Remaining Quality Life" represents the expected remaining shelf life of the food at the beginning of the distribution process. "Ending Remaining Quality Life" represents the expected remaining shelf life of the food at the end of the distribution process. Quality Life Loss Rate refers to the percentage of quality life lost at a particular distribution stage. The weighting factor is a coefficient used to adjust the importance of the evaluation indicators.

[0108] Specifically, firstly, the traceability system will calculate basic indicators for each circulation stage based on the segmented numerical evolution trajectory: calculate the average deterioration rate, reflecting the speed of quality change through the slope of the segmented numerical evolution trajectory; at the same time, calculate the deterioration increment of each stage, obtaining the overall change by subtracting the initial value from the final value; in addition, it also needs to predict the remaining quality life at the beginning and end of the circulation stage based on the preset quality deterioration dynamics model.

[0109] Next, the traceability system calculates the difference between the beginning and end of the remaining quality lifespan, and then divides it by the beginning of the remaining quality lifespan to obtain the quality lifespan loss rate in the circulation process (which can well reflect the actual impact of the circulation process on food quality).

[0110] Then, the traceability system uses this quality life loss rate as a weighting factor to adjust the average deterioration rate and the incremental deterioration of each process, resulting in a weighted index value that more accurately reflects the impact of each process.

[0111] Finally, when determining the responsible link, the traceability system comprehensively compares the weighted index values ​​of all circulation links. By ranking the weighted average deterioration rate and the increment of deterioration at each link, the circulation links with the greatest impact in these two dimensions are identified.

[0112] This multi-dimensional weighted evaluation method considers both the actual loss of quality lifespan and the impact of both rate and incremental dimensions, providing a more objective and comprehensive evaluation standard. It not only improves the accuracy of liability determination but also provides reliable data support for supply chain optimization and the improvement of quality control systems.

[0113] Optionally, in general, based on the segmented numerical evolution trajectory, the calculation of the initial remaining quality lifespan and the final remaining quality lifespan of each circulation stage can be achieved in the following ways, without limitation: Extract the initial quality feature value corresponding to the start time of the current circulation stage and the final quality feature value corresponding to the end time of the current circulation stage from the segmented numerical evolution trajectory; obtain a preset food quality kinetic model and a preset quality failure threshold; use the initial quality feature value as an input parameter, substitute it into the preset food quality kinetic model, calculate the first time required for evolution to the preset quality failure threshold, and record the first time as the initial remaining quality lifespan; use the final quality feature value as an input parameter, substitute it into the preset food quality kinetic model, calculate the second time required for evolution to the preset quality failure threshold, and record the second time as the final remaining quality lifespan.

[0114] The initial quality characteristic value refers to the numerical state of the physicochemical indicators at the beginning of the circulation process. The final quality characteristic value refers to the numerical state of the physicochemical indicators at the end of the circulation process. The preset quality failure threshold represents the critical value of the physicochemical indicators at which the food is judged to be invalid. The first duration represents the estimated time from the initial state of the circulation process to the point of invalidation. The second duration represents the estimated time from the final state of the circulation process to the point of invalidation.

[0115] Specifically, the traceability system operates according to the following steps:

[0116] (1) Extraction of quality feature values:

[0117] Locate the starting point of the process from the segmented numerical evolution trajectory: determine the starting timestamp of the circulation process, extract the quality value of the starting timestamp, and record it as the initial quality characteristic value;

[0118] Locate the endpoint of the process from the segmented numerical evolution trajectory: determine the end timestamp of the circulation process, extract the quality value of the end timestamp, and record it as the termination quality characteristic value.

[0119] (2) Preset parameter acquisition:

[0120] Call the preset food quality dynamics model (including mathematical equations of quality change, parameters of key influencing factors, and standard storage condition settings).

[0121] Obtain the preset quality failure threshold (determined based on food safety standards, industry norms, and consumer acceptability).

[0122] (3) Begin calculating remaining quality life:

[0123] Substitute the initial quality characteristic values ​​into the preset food quality dynamics model (set standard storage environment conditions and initial calculation parameters), perform quality evolution prediction calculation (gradually extrapolate quality changes until the preset quality failure threshold is reached), record the required time as the first duration, and determine the first duration as the start of the remaining quality life.

[0124] (4) End the remaining quality life calculation:

[0125] The termination quality characteristic value is substituted into the preset food quality kinetic model to perform quality evolution prediction calculation, and the required time is recorded as the second duration. The second duration is then determined as the end of the remaining quality life.

[0126] S106. Based on the spatiotemporal trajectory information, environmental parameter information, and operation record information corresponding to the responsible links for food quality deterioration, and combined with abnormal physicochemical indicators and preset food quality knowledge graphs, determine the causes of deterioration in the responsible links.

[0127] Among these, the deterioration causes refer to the specific reasons leading to the deterioration of food quality, such as excessively high temperatures or damaged packaging. Spatiotemporal trajectory information includes the specific time period and geographical location of the responsible link in the food quality deterioration process. Environmental parameter information includes environmental monitoring data such as temperature, humidity, and light intensity. Operational record information includes relevant records of manual operations and equipment operation.

[0128] Specifically, firstly, the traceability system extracts detailed monitoring records of the responsible links in the food quality deterioration process, including spatiotemporal information, environmental data, and operational records. Then, based on preset causal relationships for food quality deterioration within a pre-defined food quality knowledge graph, the system analyzes possible causes of abnormal physicochemical indicators. By matching the spatiotemporal trajectory information, environmental parameter information, and operational record information corresponding to the responsible links with the preset causal relationships in the food quality knowledge graph, the system can identify the most likely specific cause of the current quality problem. For example, if an abnormally high temperature is recorded in the responsible link for abnormal "peroxide value," and the preset food quality knowledge graph indicates that high temperatures accelerate the rise in peroxide value, the traceability system will identify "excessive temperature" as the cause of deterioration in the responsible link. Finally, the traceability system outputs a complete traceability conclusion, including the responsible link for food quality deterioration and the cause of deterioration in that link.

[0129] By adopting the above technical solutions, firstly, the traceability system extracts sensory description labels from consumer feedback information. Utilizing a pre-set food quality knowledge graph, it achieves a precise mapping between subjective sensory information and objective physicochemical indicators, alleviating the problem of the disconnect between subjective feedback and objective data in traditional traceability systems and providing a clear target for food quality deterioration analysis. Simultaneously, the traceability system dynamically extrapolates the monitoring data across the entire food supply chain through a pre-set quality deterioration dynamics model, upgrading food quality deterioration judgment from static threshold judgment to dynamic numerical evolution trajectory analysis, thereby quantifying the cumulative deterioration effect at each stage of circulation. Furthermore, the traceability system screens responsible links for food quality deterioration based on both average deterioration rate and the incremental deterioration at each stage. Combined with the pre-set food quality knowledge graph, it traces the deterioration causes at these responsible links, achieving precise definition of deterioration responsibility and tracing the causes. This provides a scientific basis for targeted optimization of the food supply chain, improves the efficiency and accuracy of food quality deterioration traceability, and effectively safeguards food quality and safety.

[0130] The traceability system in the embodiments of this invention is described below from the perspective of hardware processing. Please refer to [link / reference]. Figure 2 This is a schematic diagram of the physical device structure of the traceability system in this application embodiment.

[0131] It should be noted that, Figure 2 The structure of the traceability system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of the present invention.

[0132] like Figure 2 As shown, the traceability system includes a CPU 201, which can perform various appropriate actions and processes based on a program stored in the read-only memory ROM 202 or a program loaded from the storage section 208 into the random access memory RAM 203, such as performing the methods described in the above embodiments. The RAM 203 also stores various programs and data required for system operation. The CPU 201, ROM 202, and RAM 203 are interconnected via a bus 204. An I / O interface 205 is also connected to the bus 204.

[0133] The following components are connected to I / O interface 205: input section 206 including audio input devices, push-button switches, etc.; output section 207 including a liquid crystal display (LCD) and audio output devices, indicator lights, etc.; storage section 208 including a hard disk, etc.; and communication section 209 including a network interface card such as a LAN (Local Area Network) card, modem, etc. Communication section 209 performs communication processing via a network such as the Internet. Drive 210 is also connected to I / O interface 205 as needed. Removable media 211, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 210 as needed so that computer programs read from them can be installed into storage section 208 as needed.

[0134] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing computer programs for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 209, and / or installed from removable medium 211. When the computer program is executed by CPU 201, it performs the various functions defined in the present invention.

[0135] It should be noted that specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0136] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. Each block in a flowchart or block diagram may represent a module, program segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those shown in the drawings.

[0137] Specifically, the traceability system in this embodiment includes a processor and a memory. The memory stores a computer program, and when the computer program is executed by the processor, it implements the food quality deterioration traceability method provided in the above embodiment.

[0138] In another aspect, the present invention also provides a computer-readable storage medium, which may be included in the traceability system described in the above embodiments; or it may exist independently and not be assembled into the traceability system. The storage medium carries one or more computer programs, which, when executed by a processor of the traceability system, enable the traceability system to implement the food end-to-end quality deterioration traceability method provided in the above embodiments.

[0139] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

[0140] As used in the above embodiments, depending on the context, the term "when..." can be interpreted as meaning "if...", "after...", "in response to determining...", or "in response to detecting...". Similarly, depending on the context, the phrase "when determining..." or "if (the stated condition or event) is interpreted as meaning "if determining...", "in response to determining...", "when (the stated condition or event) is detected", or "in response to detecting (the stated condition or event)".

[0141] 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. This program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for tracing food quality deterioration across the entire supply chain, characterized in that, Applied to a traceability system, the method includes: Receive consumer feedback information and extract sensory description tags from the consumer feedback information; Based on the sensory description labels and the preset food quality knowledge graph, abnormal physicochemical indicators are determined. The preset food quality knowledge graph includes sensory description nodes, physicochemical indicator nodes, and deterioration cause nodes. The sensory description nodes are connected to the physicochemical indicator nodes through a preset sensory-physicochemical-biochemical mapping relationship. The physicochemical indicator nodes are connected to the deterioration cause nodes through a preset food quality deterioration causal relationship. The sensory description nodes are used to store consumers' descriptions of the sensory attributes of food. The physicochemical indicator nodes are used to store the physicochemical testing indicator information corresponding to the food quality. The deterioration cause nodes are used to store the cause information of the deterioration of the food's physicochemical indicators. Acquire food full-chain monitoring data, input the food full-chain monitoring data into a preset quality deterioration dynamic model to obtain food quality deterioration data. The food full-chain monitoring data includes spatiotemporal trajectory information, environmental parameter information and operation record information corresponding to multiple circulation links. The food quality deterioration data includes different numerical evolution trajectories corresponding to different food physicochemical indicators. Extract the abnormal value evolution trajectory corresponding to the abnormal physicochemical index from the food quality deterioration data, and divide the abnormal value evolution trajectory into multiple segmented value evolution trajectories corresponding to different circulation links. Calculate the average deterioration rate and the deterioration increment of each segment's numerical evolution trajectory, and identify the first circulation link with the largest average deterioration rate and the second circulation link with the largest deterioration increment as the responsible links for food quality deterioration. Based on the spatiotemporal trajectory information, environmental parameter information, and operation record information corresponding to the responsible links for food quality deterioration, combined with the abnormal physicochemical indicators and the preset food quality knowledge graph, the causes of deterioration in the responsible links are determined.

2. The method according to claim 1, characterized in that, The physicochemical index nodes correspond to each other; the food full-chain monitoring data is input into a preset quality deterioration dynamic model to obtain food quality deterioration data. The food full-chain monitoring data includes spatiotemporal trajectory information, environmental parameter information, and operation record information corresponding to multiple circulation links. The food quality deterioration data includes different numerical evolution trajectories corresponding to different food physicochemical indicators, specifically including: Obtain the kinetic reaction equation, baseline activation energy parameter, and initial state detection value for each of the physicochemical indicators; Based on the spatiotemporal trajectory information and the environmental parameter information, and combined with the baseline activation energy parameter of each physicochemical index, the time-varying reaction rate constant of each physicochemical index at different time points is calculated using the Arrhenius equation. Starting from the initial state detection value, based on the kinetic reaction equation and the time-varying reaction rate constant, parallel differential iterative calculations are performed on multiple physicochemical indicators to obtain a set of full numerical evolution trajectories of multiple physicochemical indicators over time. The set of all numerical evolution trajectories is determined as the data on food quality deterioration.

3. The method according to claim 2, characterized in that, Obtaining the initial state detection value for each of the physicochemical indicators specifically includes: The production batch identifier of the food is obtained by parsing the operation record information; Search the preset production quality inspection database to see if there is a quality inspection record uniquely associated with the production batch identifier; If it exists, the measured values ​​corresponding to the physicochemical indicators are extracted from the quality inspection records, and the measured values ​​are determined as the initial state detection values. If not, obtain the preset standard formula benchmark data or historical average test data corresponding to the food to determine the initial state test value of the physicochemical index.

4. The method according to claim 2, characterized in that, The preset food quality knowledge graph also includes physicochemical index coupling relationships, which are used to define the cascading influence paths between different physicochemical index nodes. Starting from the initial state detection value, based on the kinetic reaction equation and the time-varying reaction rate constant, parallel differential iterative calculations are performed on multiple physicochemical indices to obtain a set of complete numerical evolution trajectories of multiple physicochemical indices over time, specifically including: When performing iterative calculations at the current time step, based on the coupling relationship of the physicochemical indicators, the driving dependency relationship between the target physicochemical indicator and the other physicochemical indicators is identified to determine the driving indicator and the response indicator. The target physicochemical indicator is any physicochemical indicator, and the other physicochemical indicators are physicochemical indicators other than the target physicochemical indicator. Obtain the instantaneous simulated value of the driving index within the current time step, and determine the degradation rate coupling gain coefficient of the driving index on the response index based on a preset coupling influence coefficient model. The time-varying reaction rate constant of the response index at the current time step is corrected by the degradation rate coupling gain coefficient to obtain the corrected time-varying reaction rate constant. Based on the corrected time-varying reaction rate constant, the predicted simulated value of the response index at the next time step is derived. By iteratively executing the steps of identifying the driving dependency between the target physicochemical index and other physicochemical indices within all time steps, determining the degradation rate coupling gain coefficient of the driving index on the response index, and correcting the time-varying reaction rate constant of the response index at the current time step using the degradation rate coupling gain coefficient, a set of full numerical evolution trajectories of multiple physicochemical indices over time is obtained.

5. The method according to claim 1, characterized in that, The process involves calculating the average deterioration rate and the increment of deterioration at each stage of the numerical evolution trajectory for each segment, and identifying the first stage of the process with the highest average deterioration rate and the second stage of the process with the largest increment of deterioration as the responsible stages for food quality deterioration. Specifically, this includes: Based on the segmented numerical evolution trajectory, the starting remaining quality life and ending remaining quality life, average degradation rate and degradation increment of each of the circulation links are calculated. The difference ratio between the starting remaining quality lifetime and the ending remaining quality lifetime is calculated to obtain the quality lifetime loss rate. Using the quality life loss rate as a weighting factor, the average deterioration rate and the incremental deterioration at each stage are weighted and corrected to determine the responsible stage for the food quality deterioration.

6. The method according to claim 5, characterized in that, The calculation of the initial remaining quality life and the final remaining quality life of each of the circulation stages based on the segmented numerical evolution trajectory specifically includes: From the segmented numerical evolution trajectory, extract the initial quality feature value corresponding to the start time of the current circulation stage, and the final quality feature value corresponding to the end time of the current circulation stage. Obtain a preset food quality dynamics model and a preset quality failure threshold; The initial quality characteristic value is used as an input parameter and substituted into the preset food quality dynamics model to calculate the first time required for evolution to the preset quality failure threshold. The first time is recorded as the starting remaining quality life. The termination quality characteristic value is used as an input parameter and substituted into the preset food quality dynamics model to calculate the second time required for the evolution to the preset quality failure threshold. The second time is recorded as the end of the remaining quality life.

7. The method according to claim 1, characterized in that, The method for constructing the preset food quality knowledge graph specifically includes: We acquired academic literature, industry standard documents, and historical customer complaint records in the field of food science as the original corpus. Natural language processing technology is used to extract entities and relations from the original corpus, and to identify sensory description entities, physicochemical indicator entities, and deterioration cause entities. Based on co-occurrence frequency analysis and semantic similarity calculation, a mapping probability matrix between the sensory description entity and the physicochemical index entity, and a causal probability matrix between the physicochemical index entity and the deterioration cause entity are constructed. Based on the mapping probability matrix and the causal probability matrix, establish connection edges between the sensory description entity and the physicochemical indicator entity, and between the physicochemical indicator entity and the deterioration cause entity, and determine the weight of the connection edges to form an initial food quality knowledge graph. The initial food quality knowledge graph is sent to experts for manual correction to obtain the preset food quality knowledge graph.

8. A traceability system, characterized in that, The traceability system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code including computer instructions, and the one or more processors call the computer instructions to cause the traceability system to perform the method as described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instruction is executed on the traceability system, it causes the traceability system to perform the method as described in any one of claims 1-7.

10. A computer program product, characterized in that, When the computer program product is run on the traceability system, the traceability system performs the method as described in any one of claims 1-7.

Images