A method for tracing the origin of crayfish and preventing counterfeiting

Abstract

The present invention discloses a method for tracing the origin of crayfish and preventing counterfeiting, and relates to the technical field of origin tracing. Feature data is input into a trained deep learning model for analysis, and the crayfish is automatically identified to have a traceability error. When the traceability error is identified, the batch of crayfish is marked and a warning message is sent to the administrator. When the traceability error is not identified, the feature data is substituted into an anomaly detection algorithm for comprehensive analysis, and the crayfish is identified to have a potential traceability risk based on the analysis result. When the potential traceability risk is identified, the batch of crayfish is marked and a warning message is sent to the administrator. By learning and extracting features from the supply chain data, the time series features and complex patterns in the data can be captured, and the traceability risk of each batch of crayfish can be more comprehensively evaluated, providing a strong guarantee for the origin tracing and anti-counterfeiting of crayfish.

Description

Technical Field

[0001] The present invention relates to the technical field of origin tracing, and in particular to an origin tracing and anti-counterfeiting method for crayfish. Background Art

[0002] The crayfish industry is developing rapidly in China, with the main production areas concentrated in Hubei, Jiangsu, Anhui and other regions. As consumers' requirements for food safety and quality increase, origin traceability has become an important link in the crayfish industry. Through the traceability system, the entire process of crayfish breeding, fishing, processing and sales can be tracked to ensure that the product source is clear and transparent, thereby enhancing consumer trust;

[0003] The existing calculation has the following defects:

[0004] Traditional traceability systems usually rely on simple rules or static pattern recognition, which makes it difficult to handle complex patterns and nonlinear relationships in supply chain data. They can only perform single traceability error detection or anomaly detection, but are unable to comprehensively analyze supply chain data, resulting in an inability to fully assess risks, and thus unable to provide strong protection for the origin traceability and anti-counterfeiting of crayfish.

[0005] Based on this, the present invention proposes a method for tracing the origin and preventing counterfeiting of crayfish. By learning and extracting features from the supply chain data, the time series characteristics and complex patterns in the data can be captured, thereby improving the system's ability to recognize normal supply chain patterns and more comprehensively evaluating the traceability risks of each batch of crayfish, providing strong protection for the traceability and anti-counterfeiting of crayfish. Summary of the invention

[0006] The purpose of the present invention is to provide a crayfish origin tracing and anti-counterfeiting method to solve the shortcomings of the background technology.

[0007] In order to achieve the above object, the present invention provides the following technical solution: a method for tracing the origin of crayfish and preventing counterfeiting, the method comprising the following steps:

[0008] The system collects and processes crayfish supply chain data from the past few years, uses recursive neural networks to learn the supply chain data, extracts feature data related to the crayfish production and transportation process, and uses the feature data to train the deep learning model so that the deep learning model can identify normal supply chain patterns.

[0009] When a new batch of crayfish enters the supply chain, the system collects supply chain data in real time. After extracting feature data through a recursive neural network, the feature data is input into a trained deep learning model for analysis to automatically identify whether there are traceability errors in the crayfish. If a traceability error is identified, the batch of crayfish is marked and a warning message is sent to the administrator.

[0010] When it is identified that there is no traceability error, the feature data is substituted into the anomaly detection algorithm for comprehensive analysis. Based on the analysis results, it is determined whether there is a potential traceability risk for the crayfish. When it is identified that there is a potential traceability risk, the batch of crayfish is marked and a warning message is sent to the administrator.

[0011] In a preferred embodiment, the deep learning model is trained using feature data so that the deep learning model can recognize normal supply chain patterns, including the following steps:

[0012] The feature data is input into the input layer of the long short-term memory network model. The data point at each time point is used as the input node. The time feature data is processed through multiple LSTM layers to obtain the fluctuation value of the capture-transport correlation coefficient, the spatial feature data is processed to obtain the deviation value of the transportation distance index, and the environmental feature data is processed to obtain the environmental anomaly index.

[0013] After the processed capture-transport correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index are reduced in dimension through the pooling layer, the fully connected layer maps the reduced-dimensional correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index to the output space, and compares the correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index with the preset corresponding thresholds in the output space;

[0014] The output layer generates traceability error identification based on the comparison results through the sigmoid activation function, and classifies the batch of crayfish as having traceability errors or not, so as to identify the normal supply chain pattern.

[0015] In a preferred embodiment, the output layer maps the comparison result to a value between 0 and 1 through a sigmoid activation function to generate the final classification result. If the correlation coefficient fluctuation value in the output space is greater than the fluctuation threshold, the transportation distance index deviation value is greater than the deviation threshold, or the environmental anomaly index is greater than the anomaly threshold, the output result of the sigmoid activation function is 1, indicating that there is a traceability error in this batch of crayfish; otherwise, the output result of the sigmoid activation function is 0, indicating that there is no traceability error in this batch of crayfish.

[0016] In a preferred embodiment, when it is identified that there is no traceability error, the characteristic data is substituted into the anomaly detection algorithm for comprehensive analysis, and whether the crayfish has a potential traceability risk is identified based on the analysis results, including the following steps:

[0017] Substitute the feature data into the anomaly detection algorithm for comprehensive analysis, and the function expression is:

[0018] , where is the risk factor, is the fluctuation value of the capture-transport correlation coefficient, is the deviation value of the transport distance index, is the environmental anomaly index, , , are the deviation value of the transport distance index, the fluctuation value of the capture-transport correlation coefficient, and the proportional coefficient of the environmental anomaly index, respectively. , , All are greater than 0;

[0019] The larger the risk coefficient, the greater the potential traceability risk of this batch of crayfish. The obtained risk coefficient is compared with the preset risk threshold. If the risk coefficient is less than or equal to the risk threshold, it is identified that there is no potential traceability risk for the crayfish. If the risk coefficient is greater than the risk threshold, it is identified that there is a potential traceability risk for the crayfish. The batch of crayfish is marked and a warning message is sent to the administrator.

[0020] In a preferred embodiment, the calculation expression of the capture-transport correlation coefficient fluctuation value is:

[0021] , where is the fluctuation value of the capture-transport correlation coefficient, is the capture-transport correlation coefficient at the i-th time point, is the mean of the capture-transport correlation coefficient, is the number of time points;

[0022] The calculation expression of the transport distance index deviation value is: , where is the deviation value of the transport distance index, is the transport distance index, is the standard transport distance index;

[0023] The calculation expression of the environmental anomaly index is: , where is the environmental anomaly index, For the The deviation value of the environmental parameters, is the number of environmental parameters, and the calculation expression of the deviation value of environmental parameters is: , where is the environmental parameter deviation value, is a real-time parameter, is the parameter stability range.

[0024] In a preferred embodiment, the system collects and processes crayfish supply chain data from the past few years, uses a recursive neural network to learn the supply chain data, and extracts feature data related to the crayfish production and transportation process, including the following steps:

[0025] Collect crayfish supply chain data from all links, including breeding environment, fishing time, processing steps, transportation routes, logistics time, remove duplicate data and missing values, and fill or interpolate missing data;

[0026] Integrate data from all links into a unified data set to form supply chain data records and perform time serialization on the data;

[0027] Extract time characteristic data, which include capture and transportation correlation coefficient; extract space characteristic data, which include transportation distance index; extract environment characteristic data, which include water temperature, pH value and dissolved oxygen content.

[0028] In a preferred embodiment, the acquisition logic of the capture and transportation correlation coefficient is: obtain the correlation values ​​of multiple time periods of the farm, and the calculation expression is: , where is the correlation value, k represents the comparison between the time series and its data at k past time points, represents a time point in a time series, represents the total number of data points in the time series of catch or transport volume, represents the catch or transport volume of the time series at time point t, represents the mean of the time series of catch or transport volume, Represents the time series of catch or transport volume at time point The maximum correlation value among multiple time periods is taken as the capture correlation coefficient, and the time period corresponding to the maximum correlation value is marked as a fixed period.

[0029] In a preferred embodiment, the calculation logic of the transport distance index is:

[0030] Get the transportation distance of the transport vehicle at each transport node. The expression is:

[0031] , where d is the transportation distance between transportation nodes, is the radius of the Earth, , is the latitude of the two transport nodes, is the dimensionality difference, and , The accuracy is poor, and , , is the longitude of two points;

[0032] The transport distance index of the transport vehicle is calculated based on the transport distance between transport nodes. The expression is: , where is the transport distance index, is the distance of the i-th transport path, is the number of transport nodes.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0034] The present invention learns the supply chain data through a recursive neural network, uses feature data to train the deep learning model, enables the deep learning model to identify the normal supply chain mode, inputs the feature data into the trained deep learning model for analysis, automatically identifies whether there is a traceability error in crayfish, and when it is identified that there is a traceability error, marks the batch of crayfish and sends a warning message to the administrator, and when it is identified that there is no traceability error, substitutes the feature data into the anomaly detection algorithm for comprehensive analysis, identifies whether there is a potential traceability risk in crayfish based on the analysis results, and when it is identified that there is a potential traceability risk, marks the batch of crayfish and sends a warning message to the administrator. By learning and extracting features from the supply chain data, the time series features and complex patterns in the data can be captured, the system's ability to recognize normal supply chain modes is improved, and the traceability risk of each batch of crayfish is more comprehensively evaluated, providing a strong guarantee for the origin traceability and anti-counterfeiting of crayfish. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0036] Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] Example 1: Please refer to Figure 1 As shown, the origin tracing and anti-counterfeiting method of crayfish described in this embodiment includes the following steps:

[0039] The system collects and processes crayfish supply chain data from the past few years, uses recursive neural networks (RNNs) to learn the supply chain data, extracts feature data related to the crayfish production and transportation process, and uses the feature data to train the deep learning model so that the deep learning model can recognize normal supply chain patterns. When a new batch of crayfish enters the supply chain, the system collects the supply chain data of the batch in real time. After extracting the feature data through the recursive neural network, the feature data is input into the trained deep learning model for analysis, and it automatically identifies whether there are traceability errors in the crayfish. When it is identified that there are traceability errors, the batch of crayfish is marked and a warning message is sent to the administrator. When it is identified that there are no traceability errors, the feature data is substituted into the anomaly detection algorithm for comprehensive analysis. Based on the analysis results, it is identified whether there are potential traceability risks for crayfish. When it is identified that there are potential traceability risks, the batch of crayfish is marked and a warning message is sent to the administrator.

[0040] This application uses a recursive neural network to learn the supply chain data, uses feature data to train the deep learning model, so that the deep learning model can recognize the normal supply chain mode, and the feature data is input into the trained deep learning model for analysis, and automatically identifies whether there is a traceability error in the crayfish. When it is identified that there is a traceability error, the batch of crayfish is marked and a warning message is sent to the administrator. When it is identified that there is no traceability error, the feature data is substituted into the anomaly detection algorithm for comprehensive analysis. Based on the analysis results, it is identified whether there is a potential traceability risk in the crayfish. When it is identified that there is a potential traceability risk, the batch of crayfish is marked and a warning message is sent to the administrator. By learning and extracting features from the supply chain data, the time series characteristics and complex patterns in the data can be captured, which improves the system's ability to recognize normal supply chain patterns, and more comprehensively evaluates the traceability risk of each batch of crayfish, providing strong protection for the origin traceability and anti-counterfeiting of crayfish.

[0041] Example 2: The system collects and processes the crayfish supply chain data of the past few years, uses a recurrent neural network (RNN) to learn the supply chain data, and extracts feature data related to the crayfish production and transportation process, including the following steps:

[0042] Collect crayfish supply chain data from all links, including breeding environment (such as temperature and humidity), fishing time, processing steps, transportation routes, logistics time, etc. Data sources can include sensor data, RFID tag data, logistics management system data, etc., remove duplicate data and missing values, fill or interpolate missing data, standardize data formats, and ensure that the timestamps and units of all data are consistent.

[0043] Integrate data from each link into a unified data set to form a complete supply chain data record, and perform time serialization on the data to ensure that each record has a clear time stamp.

[0044] The gated recurrent unit (GRU) is used to process time series data, extract time-dependent features related to crayfish production and transportation, identify key time points, such as fishing time, processing time, arrival time, etc., extract trends and periodic features in the time series, and use recursive neural networks to extract spatial features related to the transportation path, such as transportation distance, changes in transportation routes, etc., and extract features related to the breeding environment, such as water temperature, pH value, dissolved oxygen content, etc. These features have an important impact on the growth and quality of crayfish.

[0045] Extract time characteristic data, which include capture and transportation correlation coefficient; extract space characteristic data, which include transportation distance index; extract environment characteristic data, which include water temperature, pH value and dissolved oxygen content.

[0046] The logic for obtaining the capture and transportation correlation coefficient is: obtain the correlation values ​​of multiple time periods of the farm, and the calculation expression is: , where is the correlation value, k represents the comparison between the time series and its data at k past time points. For example, lag 1 represents the correlation between the catch or transport volume series and its data from the previous day or time period. Represents a time point in a time series, such as a time point corresponding to daily or weekly catch or transport data. Indicates the total number of data points in the time series of catch or transport volume, that is, how many days or weeks of observation data are there in this period. represents the catch or transport volume of the time series at time point t, represents the mean of the time series of catch or transport volume, Represents the time series of catch or transport volume at time point The maximum correlation value among multiple time periods is taken as the fishing and transportation correlation coefficient. If significant autocorrelation is observed at the lag value of 7 days or 30 days (i.e., the ACF value is high), it means that there is a fixed cycle of fishing or transportation on a weekly or monthly basis, and the time period corresponding to the maximum correlation value is marked as a fixed cycle.

[0047] The calculation logic of the transport distance index is: obtain the transport distance of the transport vehicle at each transport node, and the expression is:

[0048] , where d is the transportation distance between transportation nodes, is the radius of the Earth (about 6371 km), , are the latitudes of the two transport nodes (in radians), is the dimensionality difference, and , The accuracy is poor, and , , is the longitude of the two points (in radians);

[0049] The transport distance index of the transport vehicle is calculated based on the transport distance between transport nodes. The expression is: , where is the transport distance index, is the distance of the i-th transport path (i.e. the transport distance between two transport nodes), is the number of transport nodes.

[0050] Using feature data to train a deep learning model so that it can recognize normal supply chain patterns includes the following steps:

[0051] Extracting time characteristic data, including the capture and transportation correlation coefficient, extracting space characteristic data, including the transportation distance index, extracting environment characteristic data, including water temperature, pH value, and dissolved oxygen content;

[0052] Input the feature data into the input layer of the long short-term memory network model. The data point at each time point is used as the input node. The time feature data is processed through multiple LSTM layers to obtain the fluctuation value of the capture-transport correlation coefficient, the spatial feature data is processed to obtain the deviation value of the transport distance index, and the environmental feature data is processed to obtain the environmental anomaly index. The LSTM layer captures the long-term and short-term dependencies through memory units and forget gates, input gates, and output gates.

[0053] The calculation expression of the fluctuation value of the capture and transport correlation coefficient is:

[0054] , where is the fluctuation value of the capture-transport correlation coefficient, is the capture-transport correlation coefficient at the i-th time point, is the mean of the capture-transport correlation coefficient, is the number of time points. The greater the fluctuation value of the capture and transportation correlation coefficient, the greater the fluctuation of the crayfish catch or transportation volume of the farm in the past period of time, and the more likely there is a traceability error, specifically:

[0055] Inconsistent fluctuations: Under normal circumstances, the catch and transportation volume of crayfish should maintain a relatively stable relationship with production and market demand. If the catch-transport correlation coefficient fluctuates greatly, it means that the relationship between the catch and transportation volume has experienced abnormal fluctuations, which may be inconsistent with normal supply chain operations. Such abnormal fluctuations may indicate that an abnormal situation has occurred in a certain link, such as data falsification, record errors, or illegal operations.

[0056] Cyclic disruptions: Cyclic fluctuations in the fishing and transportation processes are normal, such as seasonal changes or cyclical adjustments in market demand. However, large fluctuations may indicate that these cyclical patterns are broken, suggesting that abnormal behavior or operations have occurred in a certain link of the supply chain.

[0057] Concealment of abnormal operations: When the catch or transportation volume in the supply chain fluctuates greatly, it may be because some links are trying to cover up abnormal operations or illegal behaviors. Abnormal catch and transportation data may come from false reports, unrecorded fishing activities, unauthorized transportation, etc., which may reduce the credibility of traceability data.

[0058] Record errors or tampering: Large fluctuations may mean that some data in the supply chain has been recorded incorrectly or deliberately tampered with to conceal the actual operation. These anomalies may cause the traceability system to be unable to accurately track the true source of crayfish, thereby causing traceability errors.

[0059] Market anomalies: Large fluctuations may also be related to market anomalies, such as a sudden surge or decrease in market demand. However, if such fluctuations are not related to normal market changes but are caused by human intervention or operational errors, it may lead to traceability errors.

[0060] Environmental changes: Drastic changes in the environment (such as sudden changes in water quality and abnormal climate) may lead to abnormal fluctuations in catch and transport volumes. If such fluctuations are not caused by environmental factors but are related to human factors or inconsistent data, the traceability system will identify potential errors.

[0061] The calculation expression of the transport distance index deviation value is: , where is the deviation value of the transport distance index, is the transport distance index, is the standard transport distance index. The larger the deviation value of the transport distance index is, the more the actual total transport distance of the transport vehicle deviates from the standard total transport distance, and the more likely there is a traceability error, specifically:

[0062] Unusual changes in transportation routes

[0063] Deviation from the scheduled route: Under normal circumstances, the transport vehicle should be transported according to the scheduled optimal route, and the actual transport distance should be consistent with the expected distance. If the deviation value of the transport distance index is large, it may indicate that the vehicle has deviated from the scheduled route. The deviation from the route may be due to the driver's detour, unexpected circumstances during the transportation process (such as road closures or traffic accidents), etc., which may cause abnormal behavior or data errors during the transportation process.

[0064] Illegal operation or tampering: If the transport vehicle does not follow the normal transport route, but deliberately chooses an unnecessary detour or non-standard route, it may be to cover up illegal operations or tamper with the transport data. For example, the vehicle may stop at an unauthorized location or transfer goods midway, causing a large deviation between the actual transport distance and the standard distance, which may mean that the origin of the crayfish or other key information has been tampered with or hidden.

[0065] GPS data error: A large deviation in the transport distance index may also indicate a GPS data recording error or equipment failure in the logistics system. This error may cause the transport path recorded by the system to be inconsistent with the actual path, thus affecting the accuracy of the traceability data.

[0066] Manual recording errors: In some cases, transport data may be recorded manually. If the operator makes an error when entering the transport route or distance data, or deliberately tampers with the data, it may also cause a large deviation from the expected value of the transport distance, which may also cause traceability errors.

[0067] Unexpected events: Unexpected events during transportation (such as vehicle failure, bad weather, etc.) may force the vehicle to change the original transportation route, causing the actual transportation distance to exceed the standard distance. Although this situation may be uncontrollable, if it is not updated or reported in time, it may still cause the traceability system to detect anomalies and believe that there may be traceability errors.

[0068] Illegal stops or mid-transfers: If a transport vehicle makes an illegal stop or mid-transfer without permission, the actual transport distance may also be greatly increased. This abnormal behavior may attempt to cover up illegal operations during transportation, ultimately resulting in inaccurate traceability data.

[0069] False origin labeling: If the actual origin of crayfish does not match the labeled origin, in order to cover up this fact during transportation, the transporter may deliberately take a detour or increase the transportation distance to match the false origin labeling. Such operations will increase the deviation value of the transportation distance index, thus causing the system to question the traceability data.

[0070] Substitution of goods in transit: If goods are replaced or tampered with in transit, this may also lead to abnormal changes in the transportation route and distance. Larger deviation values ​​for transportation distance may be an important indicator of these illegal behaviors, helping the system identify potential traceability errors.

[0071] The calculation expression of the environmental anomaly index is: , where is the environmental anomaly index, For the The deviation value of the environmental parameters, is the number of environmental parameters, and the calculation expression of the deviation value of environmental parameters is: , where is a real-time parameter, The larger the environmental abnormality index is, the more the comprehensive environment of the farm deviates from the stable environmental condition, and the more likely there is a traceability error. Specifically:

[0072] Abnormal changes in water quality: Crayfish have high requirements for water quality. Drastic changes in water quality (such as pH value, dissolved oxygen content, pollutant concentration, etc.) may affect the growth, survival rate and health of crayfish. If the environmental abnormality index is too high, it indicates that the water quality may have serious fluctuations or pollution. This may lead to inconsistencies between actual production data and recorded data, resulting in traceability errors.

[0073] Sudden changes in temperature and climate: The temperature and climate conditions of the farm directly affect the growth cycle and yield of crayfish. If the environmental anomaly index is too large, it means that the climate conditions (such as temperature, humidity, rainfall, etc.) have changed abnormally, which may cause the crayfish production to deviate significantly from historical records. Such abnormal conditions may affect the traceability system's accurate prediction of production and lead to errors in traceability data.

[0074] Environmental data tampering: In some cases, in order to cover up unfavorable breeding conditions, farms may deliberately tamper with environmental data to meet ideal environmental standards. If the environmental anomaly index is too high, it may be caused by these tampering behaviors. When the system detects these anomalies, it will be considered that there is a possibility of traceability error.

[0075] Illegal use of chemicals: In order to cope with environmental degradation or increase production, farms may illegally use certain chemicals or additives. These operations may increase production in the short term, but will cause abnormal increases in environmental indicators. When the traceability system detects these anomalies, it may doubt the authenticity of the data and then believe that there is a traceability error.

[0076] Abnormally high production: If a farm still reports a higher production when the environmental abnormality index is high, this may be unreasonable because adverse conditions usually suppress production. This mismatch may indicate that the actual production data has been tampered with or is inaccurate, leading to traceability errors.

[0077] Abnormally low production: Similarly, if the environmental anomaly index is high, but the farm reports a much lower production than expected, this may indicate potential problems, such as disease outbreaks, theft or other illegal activities, which may not be reported truthfully, resulting in the data recorded by the traceability system not matching the actual situation.

[0078] Conflict of multi-source data: The environmental anomaly index reflects the comprehensive situation of multiple environmental factors. If there is a conflict or serious inconsistency in the data between these factors, the traceability system may be unable to correctly identify and match data from different sources, resulting in traceability errors.

[0079] Data sensor failure: Failure of environmental monitoring equipment or sensor errors may cause large fluctuations in the environmental anomaly index. These errors may be mistaken by the traceability system as environmental anomalies, resulting in incorrect traceability judgments.

[0080] After the processed capture-transport correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index are reduced in dimension through the pooling layer, the fully connected layer maps the reduced-dimensional correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index to the output space, and compares the correlation coefficient fluctuation value, transportation distance index deviation value and environmental anomaly index with the preset corresponding thresholds in the output space;

[0081] The output layer generates traceability error identification based on the comparison results through the sigmoid activation function, and classifies the batch of crayfish as having traceability errors or not having traceability errors;

[0082] The output layer maps the comparison results to a value between 0 and 1 through the sigmoid activation function to generate the final classification results. If the correlation coefficient fluctuation value in the output space is greater than the fluctuation threshold, the transportation distance index deviation value is greater than the deviation threshold, or the environmental anomaly index is greater than the anomaly threshold, the output result of the sigmoid activation function is 1, indicating that there is a traceability error in this batch of crayfish. Otherwise, the output result of the sigmoid activation function is 0, indicating that there is no traceability error in this batch of crayfish. When a traceability error is identified, the batch of crayfish is marked and a warning message is sent to the administrator.

[0083] When it is identified that there is no traceability error, the feature data is substituted into the anomaly detection algorithm for comprehensive analysis, and whether there is a potential traceability risk for crayfish is identified based on the analysis results. When it is identified that there is a potential traceability risk, the batch of crayfish is marked and a warning message is sent to the administrator, including the following steps:

[0084] Substitute the feature data into the anomaly detection algorithm for comprehensive analysis, and the function expression is:

[0085] , where is the risk factor, is the fluctuation value of the capture-transport correlation coefficient, is the deviation value of the transport distance index, is the environmental anomaly index, , , are the deviation value of the transport distance index, the fluctuation value of the capture-transport correlation coefficient, and the proportional coefficient of the environmental anomaly index, respectively. , , All are greater than 0;

[0086] The larger the risk coefficient, the greater the potential traceability risk of this batch of crayfish. The obtained risk coefficient is compared with the preset risk threshold. If the risk coefficient is less than or equal to the risk threshold, it is identified that there is no potential traceability risk for the crayfish. If the risk coefficient is greater than the risk threshold, it is identified that there is a potential traceability risk for the crayfish. The batch of crayfish is marked and a warning message is sent to the administrator.

[0087] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0088] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0089] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for tracing the origin of crayfish and preventing counterfeiting, characterized by: The method comprises the following steps: The system collects and processes crayfish supply chain data from the past few years, uses recursive neural networks to learn from the supply chain data, extracts feature data related to the crayfish production and transportation process, and uses this feature data to train a deep learning model to identify normal supply chain patterns. When a new batch of crayfish enters the supply chain, the system collects supply chain data in real time. After extracting feature data through a recursive neural network, the feature data is input into a trained deep learning model for analysis. The system automatically identifies whether there are traceability errors in the crayfish. If a traceability error is identified, the new batch of crayfish is marked and a warning message is sent to the administrator. When no traceability errors are identified, the feature data is fed into an anomaly detection algorithm for comprehensive analysis. Based on the analysis results, it is determined whether there is a potential traceability risk for the crayfish. If a potential traceability risk is identified, a new batch of crayfish is marked and a warning message is sent to the administrator. Using feature data to train a deep learning model so that it can identify normal supply chain patterns involves the following steps: The feature data is input into the input layer of the long short-term memory network model. The data point at each time point is used as the input node. The time feature data is processed through multiple LSTM layers to obtain the fluctuation value of the capture-transport correlation coefficient, the spatial feature data is processed to obtain the deviation value of the transportation distance index, and the environmental feature data is processed to obtain the environmental anomaly index. After the processed capture-transport correlation coefficient fluctuation value, transport distance index deviation value, and environmental anomaly index are reduced in dimension through the pooling layer, the fully connected layer maps the reduced-dimensional correlation coefficient fluctuation value, transport distance index deviation value, and environmental anomaly index to the output space. In the output space, the correlation coefficient fluctuation value, transport distance index deviation value, and environmental anomaly index are compared with the preset corresponding thresholds; The output layer generates traceability error identification based on the comparison results through the sigmoid activation function, and classifies a new batch of crayfish as having traceability errors or not, so as to identify the normal supply chain pattern.

2. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 1, characterized in that: The output layer maps the comparison results to a value between 0 and 1 through the sigmoid activation function to generate the final classification results. If the correlation coefficient fluctuation value in the output space is greater than the fluctuation threshold, the transportation distance index deviation value is greater than the deviation threshold, or the environmental anomaly index is greater than the anomaly threshold, the output result of the sigmoid activation function is 1, indicating that there is a traceability error in the new batch of crayfish. Otherwise, the output result of the sigmoid activation function is 0, indicating that there is no traceability error in the new batch of crayfish.

3. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 2, characterized in that: When no traceability error is found, the feature data is substituted into the anomaly detection algorithm for comprehensive analysis. Based on the analysis results, it is determined whether there is a potential traceability risk for crayfish, including the following steps: Substitute the feature data into the anomaly detection algorithm for comprehensive analysis, and the function expression is: , where is the risk factor, is the fluctuation value of the capture-transport correlation coefficient, is the transport distance index deviation value, is the environmental anomaly index, 、 、 are the deviation value of the transport distance index, the fluctuation value of the capture-transport correlation coefficient, and the proportional coefficient of the environmental anomaly index, respectively. 、 、 All greater than 0; The larger the risk coefficient, the greater the potential traceability risk of a new batch of crayfish. The obtained risk coefficient is compared with the preset risk threshold. If the risk coefficient is less than or equal to the risk threshold, it is identified that the crayfish does not have a potential traceability risk. If the risk coefficient is greater than the risk threshold, it is identified that the crayfish has a potential traceability risk. The new batch of crayfish is marked and a warning message is sent to the administrator.

4. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 3, characterized in that: The calculation expression of the capture-transport correlation coefficient fluctuation value is: , where is the fluctuation value of the capture-transport correlation coefficient, is the capture-transport correlation coefficient at the i-th time point, is the mean of the capture-transport correlation coefficient, is the number of time points; The calculation expression of the transport distance index deviation value is: , where is the transport distance index deviation value, is the transport distance index, is the standard transport distance index; The calculation expression of the environmental anomaly index is: , where is the environmental anomaly index, For the The deviation value of the environmental parameters, is the number of environmental parameters, and the calculation expression of the deviation value of the environmental parameters is: , where is the environmental parameter deviation value, is a real-time parameter, is the parameter stability range.

5. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 4, characterized in that: The system collects and processes crayfish supply chain data from the past few years, uses a recurrent neural network to learn from the supply chain data, and extracts feature data related to the crayfish production and transportation process. The system includes the following steps: Collect data from all aspects of the crayfish supply chain, including the breeding environment, fishing time, processing steps, transportation routes, and logistics time, remove duplicate data and missing values, and fill or interpolate missing data; Integrate data from all links into a unified data set to form supply chain data records and perform time serialization on the data; Extract time characteristic data, which include capture and transportation correlation coefficient; extract space characteristic data, which include transportation distance index; extract environment characteristic data, which include water temperature, pH value and dissolved oxygen content.

6. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 5, characterized in that: The logic for obtaining the capture and transportation correlation coefficient is: obtain the correlation values ​​of multiple time periods of the farm, and the calculation expression is: , where is the correlation value, k represents the comparison between the time series and its data at k past time points, represents a time point in a time series, represents the total number of data points in the time series of catch or transport volume, represents the catch or transport volume of the time series at time point t, represents the mean of the time series of catch or transport volume, Represents the time series of catch or transport volume at time point The maximum correlation value among multiple time periods is used as the capture correlation coefficient, and the time period corresponding to the maximum correlation value is marked as a fixed period.

7. The method for tracing the origin of crayfish and preventing counterfeiting according to claim 6, characterized in that: The calculation logic of the transport distance index is: Get the transport distance of the transport vehicle at each transport node. The expression is: , where d is the transportation distance between transportation nodes, is the radius of the Earth, 、 is the latitude of the two transport nodes, is the dimensionality difference, and , The accuracy is poor, and , 、 is the longitude of two points; The transport distance index of the transport vehicle is calculated based on the transport distance between transport nodes. The expression is: , where is the transport distance index, is the distance of the i-th transport path, is the number of transport nodes.

Images