A food quality and safety tracking management system based on big data

Abstract

The application discloses a food quality and safety tracking management system based on big data, and relates to the technical field of food, which comprises the following steps: collecting label identification data and storage stacking parameters of food to be tracked; extracting a label risk coefficient between a label inconsistency degree and a quality compliance risk based on historical supervision data, processing the label risk coefficient and the label identification data to obtain a label risk value; judging a stacking abnormality degree value of the food to be tracked in a storage process according to the storage stacking parameters of the food to be tracked; processing the stacking abnormality degree value and the label identification data to obtain a storage specification interference value; judging a storage specification interference degree of the food to be tracked according to the storage specification interference value to obtain first-grade storage specification interference and second-grade storage specification interference; and the effect is to improve the efficiency and effect of food quality and safety management.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of food technology, more particularly, it relates to a food quality and safety tracking management system based on big data. BACKGROUND

[0002] With the rapid development of the food industry, food quality and safety issues have attracted increasing attention. Traditional food quality and safety tracking management methods cannot comprehensively evaluate the correlation between label inconsistency and quality compliance risks, and cannot timely and accurately discover the potential impact of stacking problems on food quality. At the same time, traditional methods lack effective use of big data, cannot mine valuable information from historical regulatory data to provide a basis for food quality and safety tracking, and cannot analyze the actual quality data of food according to different warehouse specification interference levels and other conditions to generate effective early warning tracking results, which makes it difficult to guarantee the accuracy and timeliness of food quality and safety management, and therefore cannot meet the management needs of food quality and safety. SUMMARY

[0003] In view of the deficiencies in the prior art, the purpose of the present application is to provide a food quality and safety tracking management system based on big data.

[0004] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0005] A food quality and safety tracking management system based on big data, comprising:

[0006] A collection module for collecting label identification data and warehouse stacking parameters of food to be tracked;

[0007] A processing module for extracting a label risk coefficient between label inconsistency and quality compliance risk based on historical regulatory data, and processing the label risk coefficient and label identification data to obtain a label risk value;

[0008] A judgment module for determining a stacking abnormality degree value of the food to be tracked during the warehouse process according to the warehouse stacking parameters of the food to be tracked;

[0009] A processing and judgment module for processing the stacking abnormality degree value and the label identification data to obtain a warehouse specification interference value, and determining the warehouse specification interference degree of the food to be tracked to obtain a first level warehouse specification interference and a second level warehouse specification interference;

[0010] A first analysis module for processing and analyzing the warehouse stacking parameters and the label risk value to obtain first actual quality data if it is the first level warehouse specification interference;

[0011] The second analysis module: if it is a second level warehouse specification interference, the stacking deviation value and the label interference degree corresponding to the warehouse specification abnormal interference condition are processed to obtain an additional risk value, and the additional risk value, the warehouse risk coefficient and the label risk value are processed and analyzed to obtain the second actual quality data;

[0012] The generation module: generating the food quality and safety early warning tracking result according to the first actual quality data or the second actual quality data.

[0013] Preferably, the label risk coefficient between the label inconsistency degree and the quality compliance risk is extracted based on historical supervision data, specifically including the following steps:

[0014] Extracting historical label inconsistency data from historical supervision data;

[0015] Calculating the difference between the historical label inconsistency data and the compliance standard data to obtain a label deviation amplitude;

[0016] Extracting the historical compliance risk value of the historical tracking food corresponding to the label deviation amplitude from the historical supervision data;

[0017] Calculating the ratio between the historical compliance risk value and the label deviation amplitude to obtain the label risk coefficient.

[0018] Preferably, the label risk coefficient and the label identification data are processed to obtain the label risk value, specifically including the following steps:

[0019] Calculating the difference between the label identification data and the standard label identification value to obtain the label deviation value;

[0020] The label risk coefficient and the label deviation value are multiplied to obtain the label risk value.

[0021] Preferably, the stacking abnormality degree value of the to-be-tracked food in the warehouse process is determined according to the warehouse stacking parameters of the to-be-tracked food, specifically including the following steps:

[0022] Based on the warehouse stacking parameters and the category characteristic parameters of the to-be-tracked food, a stacking reference set is constructed; wherein the stacking reference set contains standard stacking pressure threshold values in different time periods, standard misplacement angle threshold values of adjacent to-be-tracked foods, and stacking stability attenuation coefficients;

[0023] The actual pressure value of the to-be-tracked food, the actual misplacement angle of adjacent to-be-tracked foods, and the real-time inclination angle of the to-be-tracked food in the stacking process are collected;

[0024] The actual pressure value is processed by ratio with the standard stacking pressure threshold value of the corresponding period to obtain a pressure deviation index; the actual misalignment angle is processed by difference with the standard misalignment angle threshold value to obtain an angle deviation index; the real-time tilt angle is processed by difference with the initial stacking tilt angle and multiplied by a stacking stability attenuation coefficient to obtain a tilt deviation index;

[0025] The damage correlation coefficient is determined according to the correlation between the historical packaging damage data of the food to be tracked and the warehouse stacking parameters.

[0026] The ground flatness of the warehouse area is detected to obtain a foundation stability coefficient.

[0027] The stacking abnormality degree value is determined according to the pressure deviation index, the angle deviation index, the tilt deviation index, the damage correlation coefficient, and the foundation stability coefficient.

[0028] Preferably, the stacking abnormality degree value and the label identification data are processed to obtain a warehouse specification interference value, specifically including the following steps:

[0029] The label identification data includes warehouse limitation annotations and traceability correlation information.

[0030] The deviation features of the actual stacking from the annotation requirements are extracted from the warehouse limitation annotations, and the warehouse area mismatch features caused by identification errors are extracted from the traceability correlation information.

[0031] The deviation features and the warehouse area mismatch features are combined into an interference feature set.

[0032] The influence weight of each feature in the interference feature set is determined according to the warehouse specification of the food category, and the label specification misleading degree is calculated by weighting the interference feature set and the influence weight.

[0033] The stacking abnormality degree value and the label specification misleading degree are input into a synergistic model to obtain a warehouse specification interference value.

[0034] Preferably, the warehouse specification interference value is used to judge the warehouse specification interference degree of the food to be tracked to obtain a first-level warehouse specification interference and a second-level warehouse specification interference, specifically including the following steps:

[0035] The warehouse specification interference value is compared with a preset warehouse specification interference threshold value.

[0036] If the warehouse specification interference value is less than the preset warehouse specification interference threshold value, it is determined that the warehouse specification interference degree of the food to be tracked is a first-level warehouse specification interference.

[0037] If the warehouse specification interference value is greater than or equal to the preset warehouse specification interference threshold value, it is determined that the warehouse specification interference degree of the food to be tracked is a second-level warehouse specification interference.

[0038] Preferably, if it is a first level warehouse specification interference, the warehouse stacking parameters and label risk values are processed and analyzed to obtain first actual quality data, specifically including the following steps:

[0039] Judging the warehouse stacking parameters affecting the warehouse risk value of the food to be tracked;

[0040] According to the warehouse risk value and the warehouse stacking parameters, the quality risk of the food to be tracked in the warehouse link is judged to obtain a first warehouse risk judgment value;

[0041] According to the label risk value and the first warehouse risk judgment value, the initial quality supervision data is corrected to obtain the first actual quality data.

[0042] Preferably, it further includes the following steps:

[0043] Detecting the stacking deviation value corresponding to the warehouse stacking parameters; detecting the label interference degree corresponding to the label identification data;

[0044] Statistically obtaining the association regulation risk value of the association regulation risk condition formed by the label interference degree and the warehouse stacking deviation;

[0045] According to the association regulation risk value, the stacking deviation value is adapted to the deviation splitting of the warehouse link to obtain the warehouse stacking deviation value to be tracked.

[0046] Preferably, the stacking deviation value and the label interference degree corresponding to the warehouse specification abnormal interference condition are processed to obtain an additional risk value, specifically including the following steps:

[0047] Extracting the additional risk value of the historical stacking deviation from the historical supervision data to obtain the historical additional risk value;

[0048] The historical additional risk value and the historical normal warehouse risk value are processed by difference to obtain the historical additional risk amplitude;

[0049] The historical additional risk amplitude and the historical stacking deviation value are processed by ratio to obtain an additional risk factor;

[0050] According to the additional risk factor and the warehouse stacking deviation value to be tracked, an additional risk value of the food to be tracked in the warehouse process is obtained.

[0051] The additional risk value, the warehouse risk coefficient and the label risk value are processed and analyzed to obtain second actual quality data, specifically including the following steps:

[0052] According to the additional risk value and the warehouse risk coefficient, a second warehouse risk judgment value of the food to be tracked in the warehouse link is judged;

[0053] The initial quality supervision data is corrected according to the second warehouse risk judgment value and the label risk value to obtain second actual quality data.

[0054] Compared with the prior art, the present application has the following beneficial effects:

[0055] The present application can comprehensively and accurately obtain food-related information by collecting label identification data and warehouse stacking parameters of the food to be tracked, laying a data foundation for subsequent quality and safety tracking. The processing module extracts a label risk coefficient with the aid of historical supervision data, thereby obtaining a label risk value, which can evaluate and predict quality compliance risks existing in the food label in advance, facilitating the adoption of measures to avoid quality and safety hazards caused by label problems. In the aspect of warehouse process monitoring, the judgment module judges a stacking abnormality degree value according to the warehouse stacking parameters, which can timely discover abnormal situations that may occur in the warehouse stacking link, guarantee the safety of food stacking in the warehouse stage, and reduce adverse effects on food quality caused by improper stacking. The processing judgment module combines and processes the stacking abnormality degree value and the label identification data to obtain a warehouse specification interference value, and distinguishes warehouse specification interference of different levels, which can more carefully analyze the interference of warehouse specification on food quality, making the quality and safety control of the warehouse link more targeted. For warehouse specification interference of different levels, the first analysis module and the second analysis module respectively adopt corresponding processing and analysis methods to obtain first actual quality data and second actual quality data, which can calculate the actual quality data of the food according to different interference situations, improving the accuracy of quality evaluation. Finally, the generation module generates food quality and safety early warning tracking results according to the actual quality data, facilitating the rapid adoption of quality and safety guarantee measures, and effectively improving the efficiency and effectiveness of food quality and safety management. BRIEF DESCRIPTION OF DRAWINGS

[0056] Fig. 1 A module schematic diagram of a food quality and safety tracking management system based on big data is provided for the present application.

[0057] Fig. 2 A step schematic diagram for obtaining a stacking abnormality degree value in a food quality and safety tracking management system based on big data is provided for the present application. DETAILED DESCRIPTION

[0058] To make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0059] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application. However, the present application can be practiced in other ways different from those described herein without departing from the scope of the present application, and those skilled in the art can make similar generalizations without departing from the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0060] Second, the "one embodiment" or "an embodiment" referred to herein means a specific feature, structure, characteristic, or combination of features and / or characteristics described herein that can be included in at least one implementation of the present application. The various appearances of "in one embodiment" or "an embodiment" in the specification are not necessarily all referring to the same embodiment.

[0061] Referring to Figs. 1-2 as shown.

[0062] An embodiment of the present application further illustrates a food quality and safety tracking management system based on big data.

[0063] A food quality and safety tracking management system based on big data, comprising:

[0064] The acquisition module adopts an SC-500 intelligent data acquisition terminal, integrates multiple sensors, and is deployed in a food storage area and a label processing station; the processing module is an SRV-8000 data processing server; the judgment module is a JD-700 intelligent analysis workstation; the processing judgment module is a DM-900 decision management server, which undertakes decision analysis tasks; the first analysis module and the second analysis module belong to an Ana-600 dual-path analysis server cluster, which respectively runs different analysis algorithms; and the generation module is a Gen-1000 early warning generation server, which is responsible for early warning result generation.

[0065] The acquisition module is connected with the processing module through an industrial Ethernet for real-time data transmission; the processing module sends data to the judgment module and the processing judgment module through a message queue and a wired network; the judgment module transmits data to the processing judgment module through a high-speed bus; the processing judgment module communicates with the first analysis module and the second analysis module by means of RESTful API; the first analysis module and the second analysis module transmit data to the generation module through a UDP protocol; and the generation module pushes early warning results through an HTTP protocol and writes them into a traceability database.

[0066] The acquisition module: acquires label identification data and storage stacking parameters of food to be tracked;

[0067] The processing module: extracts a label risk coefficient between a label inconsistency degree and a quality compliance risk based on historical supervision data, processes the label risk coefficient and the label identification data to obtain a label risk value;

[0068] The judgment module: judges a stacking abnormality degree value of the food to be tracked in the storage process according to the storage stacking parameters of the food to be tracked;

[0069] The processing judgment module processes the stacking abnormality degree value and the label identification data to obtain a storage specification interference value; and judges the storage specification interference degree of the food to be tracked according to the storage specification interference value to obtain first-grade storage specification interference and second-grade storage specification interference.

[0070] The specific types of the initial quality supervision data include physicochemical index data, microbial detection data, sensory quality data and packaging integrity data of the food, the physicochemical index data is protein and fat content, the microbial detection data is total bacterial count, and the sensory quality data is color and odor. The data are obtained by laboratory detection through a liquid chromatograph and a microbial culture device, and can also be collected in real time by an online monitoring device of a production line. The correction coefficient in the correction formula is obtained based on a large amount of historical quality data, and the coefficient values under different interference factors are determined through a regression model; so that the first actual quality data and the second actual quality data can accurately reflect the real quality condition of the food.

[0071] The first analysis module processes and analyzes the storage stacking parameters and the label risk value to obtain the first actual quality data if the first-grade storage specification interference is obtained.

[0072] The second analysis module processes the stacking deviation value and the label interference degree corresponding to the storage specification abnormal interference condition to obtain an additional risk value, and processes and analyzes the additional risk value, the storage risk coefficient and the label risk value to obtain the second actual quality data if the second-grade storage specification interference is obtained.

[0073] The generation module generates a food quality and safety early warning tracking result according to the first actual quality data or the second actual quality data.

[0074] If the first actual quality data is obtained, it is prompted that the batch of food has potential problems in the label or storage link, so the subsequent flow direction needs to be tracked.

[0075] If the second actual quality data is obtained, an early warning tracking result is generated, including the risk grade of the food and the suggestion that the food is immediately and comprehensively investigated, and detailed tracking of the food from production to sales is started to ensure that the existing quality and safety problems can be found and handled in time.

[0076] The label risk coefficient between the label inconsistency degree and the quality compliance risk is extracted based on historical supervision data, including the following steps:

[0077] The historical label inconsistency data is extracted from the historical supervision data;

[0078] The difference value between the historical label inconsistency data and the compliance standard data is calculated to obtain a label deviation amplitude value;

[0079] extracting a label deviation amplitude corresponding to a historical tracking food from historical regulatory data;

[0080] calculating a ratio between the historical compliance risk value and the label deviation amplitude to obtain a label risk coefficient.

[0081] First, historical label inconsistency data is extracted from historical regulatory data. It is assumed that a batch of food has a label indicating a shelf life of 12 months, while the compliance standard requires a shelf life of 10 months, so this is label inconsistency data. The difference between the historical label inconsistency data and the compliance standard data is calculated to obtain the label deviation amplitude, such as 2 months in this case.

[0082] extracting a label deviation amplitude corresponding to a historical tracking food from historical regulatory data, the specific value D of the label deviation amplitude needs to be determined. The specific value D of the label deviation amplitude is used as a retrieval condition to perform a matching retrieval in the historical regulatory database. The historical regulatory database stores information of each batch of food from production to circulation according to the structure of food unique identification and full-process data, including label data and compliance risk assessment results. All historical tracking food records that meet the condition of label deviation amplitude equal to D are filtered out through the retrieval function. It is assumed that each historical tracking food record in the database contains food batch number, label deviation amplitude and historical compliance risk value R. When the record with label deviation amplitude equal to D is filtered out, the corresponding historical compliance risk value R is extracted from the record. If the label deviation amplitude of a batch of biscuits is calculated to be D=2 months, the batch record with label deviation amplitude D equal to 2 is retrieved in the database, and it is assumed that the batch number B20230501 has a historical compliance risk value R=0.6 marked in the record. The batch number B20230501 is included in the result set, and R=0.6 is the historical compliance risk value to be extracted.

[0083] calculating a ratio between the historical compliance risk value and the label deviation amplitude to obtain a label risk coefficient, the label risk coefficient=historical compliance risk value÷label deviation amplitude, so the label risk coefficient is 0.6÷2=0.3.

[0084] The historical compliance risk value is a quantitative result of the quality compliance risk corresponding to the past label deviation, the label deviation amplitude is a quantitative value of the deviation degree of the actual label from the standard, and the label risk coefficient obtained by dividing the two represents the quality compliance risk degree corresponding to unit label deviation, which is a quantitative refinement of the risk and deviation correlation in the historical data.

[0085] processing the label risk coefficient and the label identification data to obtain a label risk value, specifically including the following steps:

[0086] calculating the difference between the label identification data and the standard label identification value to obtain a label deviation value;

[0087] The label risk value is obtained by multiplying the label risk coefficient and the label deviation value.

[0088] For example, the label identification data of a batch of food indicates that the shelf life is 12 months, while the standard label identification value stipulates that the shelf life of this type of food should be 10 months. The label deviation value is obtained by subtracting the standard label identification value from the label identification data, which is 2 months. The label risk value is obtained by multiplying the label risk coefficient and the label deviation value. Assuming that the label risk coefficient is 0.6, the label risk value = label risk coefficient x label deviation value, the label risk value of this batch of food is 1.2. Through such calculation, the risk degree brought by the non-compliance of the label identification with the standard can be quantified.

[0089] The stacking abnormality degree value of the food to be tracked in the storage process is determined according to the storage stacking parameters of the food to be tracked, and specifically includes the following steps:

[0090] A stacking reference set is constructed based on the storage stacking parameters and the category characteristic parameters of the food to be tracked. The stacking reference set includes standard stacking pressure threshold values in different time periods, standard misalignment angle threshold values of adjacent foods to be tracked, and stacking stability attenuation coefficients.

[0091] The actual pressure value borne by the food to be tracked, the actual misalignment angle of adjacent foods to be tracked, and the real-time inclination angle of the food to be tracked in the stacking process are collected.

[0092] The actual pressure value is processed by ratio with the standard stacking pressure threshold value in the corresponding time period to obtain a pressure deviation index. The actual misalignment angle is processed by difference with the standard misalignment angle threshold value to obtain an angle deviation index. The real-time inclination angle is processed by difference with the initial stacking inclination angle and then multiplied by the stacking stability attenuation coefficient to obtain an inclination deviation index.

[0093] A damage association coefficient is determined according to the association between the historical packaging damage data of the food to be tracked and the storage stacking parameters.

[0094] Firstly, the historical packaging damage data is stratified according to the damage degree to obtain mild damage degree, moderate damage degree and severe damage degree, and the corresponding storage stacking parameters such as stacking height, stacking pressure and stacking stability are extracted. Then, multivariate linear regression analysis is performed on each layer of data to determine the influence weight of each stacking parameter under different damage degrees, so as to obtain the preliminary association coefficient. Then, using association rule mining algorithm such as Apriori algorithm, the specific stacking parameter combination and the frequent item set of packaging damage are mined from the historical data, and the association coefficient obtained by regression is further verified and corrected, and finally the damage association coefficient reflecting the association strength between the two is determined, which provides a quantitative basis for the prediction of packaging damage risk and the optimization of storage stacking specification.

[0095] The ground flatness of the storage area is detected to obtain a foundation stability coefficient. A laser flatness detector is used to detect the ground flatness of the storage area. When operating, the detector is placed at the starting position of the storage area, and after the device is turned on, it is slowly moved along the preset detection path. The detection path is a grid-shaped path covering the entire storage area. The device will record the height difference data between the ground and the standard plane in real time by scanning the ground with a laser sensor, and then calculate the flatness deviation value of the ground. According to the comparison result of the deviation value and the preset flatness standard, the foundation stability coefficient is determined. The preset flatness standard is the industry ground flatness specification. For example, when the flatness deviation value is ≤2mm, the foundation stability coefficient is 1; when the deviation value is between 2-5mm, the coefficient is 0.8. In this way, the influence of the ground flatness on the storage stacking stability is quantified.

[0096] The stacking abnormality degree value is determined according to the pressure deviation index, the angle deviation index, the inclination deviation index, the damage correlation coefficient and the foundation stability coefficient.

[0097] First, a stacking reference set is constructed based on the storage stacking parameters and the category characteristic parameters of the food. According to the stacking requirements of the food during storage and its own category characteristics, the standard stacking pressure threshold, the standard misalignment angle threshold of the adjacent food to be tracked and the stacking stability attenuation coefficient are determined. The category characteristic parameters include the packaging material, shape, density and pressure resistance of the food. For example, for liquid food, the standard stacking pressure threshold is set to 500Pa, the standard misalignment angle threshold of adjacent food is 5°, and the stacking stability attenuation coefficient is 0.9. For solid and storage-resistant food, the standard stacking pressure threshold can be set to 1500Pa, the standard misalignment angle threshold of adjacent food is 10°, and the stacking stability attenuation coefficient is 0.7. By clearly defining these characteristic parameters of different categories of food, the construction of the stacking reference set is provided with a basis, ensuring that the subsequent stacking parameter settings conform to the actual characteristics of the food category.

[0098] For the standard stacking pressure threshold, the food category is first subjected to stacking resistance experiments under different pressures, such as selecting multiple groups of the same food category and applying different sizes of pressure, observing the damage of the food under each pressure, and recording the maximum pressure value that does not cause food damage. Combined with the historical data of the food category in the actual warehouse process, the pressure range of the food stable storage without quality problems in the normal warehouse period is determined, and the upper limit of the pressure range is taken as the standard stacking pressure threshold. The formula is: standard stacking pressure threshold = maximum non-damage experimental pressure value x historical stability coefficient. The maximum non-damage experimental pressure value is obtained by conducting pressure resistance experiments on food packaging and itself in a stacked state, and is the maximum pressure value that does not cause food packaging damage or quality loss. It is a basic reference value based on experiments. The historical stability coefficient is a correction coefficient that combines the historical data of food storage stacking to judge the influence of stacking environment and food characteristics on stacking pressure bearing capacity, and is used to reflect the influence of the difference between actual warehouse scenes and experimental scenes on the pressure threshold. The product of the two is the standard stacking pressure threshold, which is to make the standard stacking pressure threshold more suitable for the actual warehouse environment and provide a reasonable benchmark for subsequent stacking abnormality degree judgment. The historical stability coefficient is obtained by statistically correlating the pressure and quality stability of the food category in historical warehouse storage. The range of the historical stability coefficient is 0.8-0.95.

[0099] The standard misalignment angle threshold of adjacent food to be tracked is determined by experiment. The same food category is stacked at different adjacent misalignment angles and simulated warehouse environment. The stability of the food under each misalignment angle is observed, such as whether it is easy to tip over and whether the packaging is easy to break. The maximum misalignment angle that can ensure the stability of the food stacking is found. Then, the historical warehouse data of the food category caused by improper adjacent misalignment angle is referred to. The angle obtained by the experiment is corrected to obtain the standard misalignment angle threshold, that is, standard misalignment angle threshold = experimental stable maximum misalignment angle x historical correction coefficient. The historical correction coefficient is determined according to the relationship between the misalignment angle and the problem incidence rate in the historical data. The historical correction coefficient is 0.9-1.

[0100] The stacking stability decay coefficient is related to time and warehouse environment factors. By long-term monitoring of the stacking stability decay of the same food category in the warehouse process with the passage of time and environmental changes, a stacking stability model of stacking stability with time and environmental changes is established, and the decay coefficient is extracted. For example, the change of food stacking stability with time is monitored in the warehouse environment. If the stability changes with time t in accordance with the exponential decay law, that is, the stacking stability wherein, For initial stability, k is the attenuation coefficient, the attenuation coefficient k of the stacking stability is determined by fitting a plurality of groups of monitoring data, and then k is adjusted in combination with the influence of different storage environments to obtain the attenuation coefficient of the stacking stability suitable for the food of the category.

[0101] For example, the standard stacking pressure threshold value of the food in the morning period is 5 Newton; the standard misalignment angle threshold value of the adjacent food is 3 degrees; and the stacking stability attenuation coefficient is 0.8. Then, the actual data of the food to be tracked in the storage process is collected, assuming that the actual pressure value is 6 Newton; the actual misalignment angle of the adjacent food to be tracked is 5 degrees; the initial stacking inclination angle of the food to be tracked in the stacking process is 1 degree, and the real-time inclination angle of the food to be tracked in the stacking process is 3 degrees.

[0102] The actual pressure value and the standard stacking pressure threshold value of the corresponding period are processed by ratio to obtain the pressure deviation index, that is, the pressure deviation index = actual pressure value ÷ standard stacking pressure threshold value of the corresponding period. Here, the pressure deviation index is 6 ÷ 5 = 1.2, the actual pressure value is the pressure data of the food in the storage stacking process collected by the pressure sensor in real time, which can directly reflect the pressure condition of the current stacking; the standard stacking pressure threshold value of the corresponding period is the upper limit standard value of the food quality and safety pressure determined in different periods in combination with the food characteristics and the storage environment factors. The pressure deviation index obtained by dividing the actual pressure value by the standard stacking pressure threshold value of the corresponding period, if the pressure deviation index is greater than 1, it means that the actual pressure exceeds the standard requirement of the corresponding period, and there is a risk of stacking pressure abnormality; if the pressure deviation index is less than or equal to 1, it means that the actual pressure is within the standard allowable range. Through such calculation, it can be directly and quantitatively judged whether the stacking pressure conforms to the specification, and key quantitative indicators are provided for subsequent evaluation of the abnormality degree of the storage stacking. The actual misalignment angle and the standard misalignment angle threshold value are processed by difference to obtain the angle deviation index, which is 5-3 = 2 degrees. The real-time inclination angle and the initial stacking inclination angle are processed by difference and then multiplied by the stacking stability attenuation coefficient to obtain the inclination deviation index, which is (3-1) x 0.8 = 1.6.

[0103] First, a large amount of historical packaging damage data of the food to be tracked and corresponding storage stacking parameters such as stacking pressure, misalignment angle and inclination angle need to be collected. Then, statistical analysis is performed on these data to find the correlation pattern between the historical packaging damage and the storage stacking parameters. For example, the probability of food packaging damage under different combinations of stacking pressure and misalignment angle parameters is counted. Then, a damage correlation model is constructed, taking the storage stacking parameters as input and the packaging damage as output. The linear regression method is used to establish the model, assuming that the logistic regression model is used, and the formula is Here, P is the probability of packaging damage, X is the combined value of storage and stacking parameters, and a and b are model coefficients determined by fitting historical data. The model calculates the predicted probability of food packaging damage under the current storage and stacking parameters, and then normalizes the predicted probability to obtain the damage correlation coefficient. For example, if the model calculates the probability of packaging damage to be 0.3 under a certain set of storage and stacking parameters, and the damage correlation coefficient is set to a range of 0 to 1, this probability value can be directly used as the damage correlation coefficient, ensuring that it accurately reflects the strong correlation between historical packaging damage data and storage and stacking parameters.

[0104] When inspecting the flatness of a warehouse area to obtain a basic stability coefficient, ground flatness testing equipment, such as a laser flatness meter, is typically used. During operation, the laser flatness meter is moved along a grid-like path within the warehouse area. The meter emits a laser beam to measure the height deviation of various points on the ground relative to a reference plane. The entire warehouse area is divided into multiple smaller zones, and ground height deviation data is collected for each zone. These height deviation data are then statistically calculated to determine an index reflecting ground flatness, such as the International Flatness Index. Assume the height deviation of various points on the ground within a small area is . , ,..., After obtaining the international flatness index of each small area, the international flatness index of all small areas within the entire storage area is used as the basis for calculation. A comprehensive assessment of the distribution is conducted. The smoother the ground, the higher the International Roughness Index. The smaller the value, the higher the basic stability coefficient. Establish a relationship between the basic stability coefficient and the international flatness index. The corresponding relationship, such as setting the basic stability coefficient. ,in It is the international flatness index when the ground is completely flat. The test results of ground flatness are converted into basic stability coefficients, thereby quantifying the impact of ground flatness on the stability of food storage stacking.

[0105] The stacking abnormality degree value is determined according to the pressure deviation index, the angle deviation index, the inclination deviation index, the damage correlation coefficient and the foundation stability coefficient. The stacking abnormality degree value is calculated by using the weighted summation method. Assuming that the weights of the indexes are respectively 0.3 for the pressure deviation index, 0.2 for the angle deviation index, 0.2 for the inclination deviation index, 0.15 for the damage correlation coefficient and 0.15 for the foundation stability coefficient, then the stacking abnormality degree value = 1.2*0.3 + 2*0.2 + 1.6*0.2 + 0.7*0.15 + 0.9*0.15 = 0.36 + 0.4 + 0.32 + 0.105 + 0.135 = 1.32. The abnormality degree value of the food warehouse stacking to be tracked is obtained by such calculation, so as to determine whether the stacking is abnormal.

[0106] The stacking abnormality degree value and the label identification data are processed to obtain the warehouse specification interference value, specifically including the following steps:

[0107] The label identification data includes warehouse limitation annotation and traceability correlation information.

[0108] The deviation features of the actual stacking from the annotation requirements are extracted from the warehouse limitation annotation, and the warehouse area mismatch features caused by identification errors are extracted from the traceability correlation information.

[0109] The deviation features and the warehouse area mismatch features are combined into the interference feature set.

[0110] The influence weights of the features in the interference feature set are determined according to the food category warehouse specification, and the label specification misleading degree is obtained by weighted calculation of the interference feature set and the influence weights.

[0111] The stacking abnormality degree value and the label specification misleading degree are input into the synergistic model to obtain the warehouse specification interference value.

[0112] The synergy model adopts a deep neural network model. The deep neural network model can deeply mine the nonlinear synergy relationship between the stacking abnormality degree value and the label specification misleading degree by virtue of the multi-layer neuron structure, and fit the complex influence of the two on the warehouse specification interference value. The training data needs to cover no less than 1000 groups of historical samples, each group of samples including the stacking abnormality degree value, the label specification misleading degree and the corresponding warehouse specification interference value, and the data needs to cover different food categories and warehouse scenarios, such as normal temperature warehouse and cold chain warehouse, so as to guarantee the generalization ability of the deep neural network model. The core algorithm adopts the back propagation algorithm, and in the training process, the error between the predicted value and the actual warehouse specification interference value is calculated, the error is reversely transmitted from the output layer to the input layer, and the weights and biases of the neurons in each layer are continuously adjusted. The deep neural network model is set as follows: 2 neurons in the input layer, 3 hidden layers, 16 neurons in each layer, and 1 neuron in the output layer; the rectified linear unit is selected as the activation function, which can effectively alleviate the gradient disappearance problem; and the mean square error is used as the loss function, so that the model can accurately output the warehouse specification interference value after training, and provide reliable data support for subsequent interference level judgment and the like.

[0113] The label identification data includes warehouse limitation annotation and traceability association information. The deviation feature between the actual stacking and the annotation requirement can be extracted from the warehouse limitation annotation, such as that the annotation requires that each box of food is not more than 5 layers, and the actual stacking is 8 layers, so this is the deviation feature; the warehouse area mismatch feature caused by identification error can be extracted from the traceability association information, such as that the food should be stored in the refrigeration area A, but is stored in the normal temperature area B due to identification error, which is the warehouse area mismatch feature. The deviation feature and the warehouse area mismatch feature jointly constitute the interference feature set.

[0114] The warehouse area mismatch feature caused by identification error is extracted from the traceability association information. First, the content contained in the traceability association information needs to be determined, which usually records the key information of the target warehouse area where the food should be stored. The actual warehouse area where the food is stored is compared with the target warehouse area in the traceability association information. For example, a certain food is clearly stored in the refrigeration area A in the traceability association information, but is actually stored in the normal temperature area B. In order to quantify this mismatch feature, a calculation method of the mismatch feature value is constructed, and the formula is: mismatch feature value = |target warehouse area code-actual warehouse area code|×area importance coefficient, wherein the target warehouse area code and the actual warehouse area code are the encodings of different warehouse areas, such as that the refrigeration area A is encoded as 1, and the normal temperature area B is encoded as 2, and the area importance coefficient is determined according to the influence degree of the warehouse area on the food quality, such as that the refrigeration area has an importance coefficient of 1.5 for the food that needs to be refrigerated, and the normal temperature area has an importance coefficient of 1. Then, the target warehouse area code is 1, the actual warehouse area code is 2, and the area importance coefficient is 1.5, so the mismatch feature value is |1-2|×1.5=1.5.

[0115] The influence weight of each feature in the interference feature set is determined according to the food category storage specification. Assuming that the weight of the storage area mismatch feature is set to 0.6 and the weight of the deviation feature is set to 0.4. The label specification misleading degree is obtained by weighted calculation of the numerical value of each feature in the interference feature set (assuming that the numerical value of the stacking deviation feature is 3 and the numerical value of the storage area mismatch feature is 4) and the corresponding influence weight. The calculation formula is: label specification misleading degree = (stacking deviation feature numerical value x stacking deviation weight) + (storage area mismatch feature numerical value x storage area mismatch weight). Here, the label specification misleading degree is (3 x 0.4) + (4 x 0.6) = 1.2 + 2.4 = 3.6. The stacking deviation feature numerical value refers to the quantitative value of the deviation between the actual storage stacking and the stacking requirement labeled by the label, such as the difference between the actual stacking height and the label labeled height. The stacking deviation weight is a weight coefficient determined according to the influence degree of the stacking deviation on the storage specification. The greater the influence, the higher the weight. The storage area mismatch feature numerical value is a quantitative value of the degree of inconsistency between the actual storage area and the label labeled storage area, for example, the area difference quantification of actually storing in A area but the label labeling B area. The storage area mismatch weight is a weight coefficient set according to the influence degree of the area mismatch on the storage specification. The label specification misleading degree obtained by multiplying the two parts by the corresponding weight and then adding them can quantitatively reflect the misleading degree of the label identification in the stacking and storage area to the storage specification.

[0116] The synergistic model is trained by a large amount of historical data. The stacking abnormality degree value and the label specification misleading degree are input into the synergistic model to obtain the storage specification interference value, so as to quantify the interference degree of the storage specification to the food quality and safety.

[0117] According to the storage specification interference value, the storage specification interference degree of the food to be tracked is determined to obtain the first grade storage specification interference and the second grade storage specification interference, which specifically includes the following steps:

[0118] The storage specification interference value is compared with the preset storage specification interference threshold value;

[0119] If the storage specification interference value is less than the preset storage specification interference threshold value, it is determined that the storage specification interference degree of the food to be tracked is obtained as the first grade storage specification interference;

[0120] If the storage specification interference value is greater than or equal to the preset storage specification interference threshold value, it is determined that the storage specification interference degree of the food to be tracked is obtained as the second grade storage specification interference.

[0121] For example, the preset storage specification interference threshold value is set to 0.6, which is obtained based on historical data statistics. When the storage specification interference value is less than 0.6, it is determined as the first level of storage specification interference; when the storage specification interference value is greater than or equal to 0.6, it is determined as the second level of storage specification interference. In actual operation, the storage specification interference level of the food to be tracked can be clearly judged, and a clear basis is provided for subsequent quality data analysis under different levels of interference.

[0122] By such numerical comparison, the storage specification interference degree of different foods can be quickly classified, thereby providing a basis for subsequent quality analysis and early warning measures corresponding to different interference levels.

[0123] If it is the first level of storage specification interference, the storage stacking parameters and label risk value are processed and analyzed to obtain the first actual quality data, which specifically includes the following steps:

[0124] The storage stacking parameters are judged to affect the storage risk value of the food to be tracked.

[0125] The quality risk of the food to be tracked is judged according to the storage risk value and the influence of the storage stacking parameters on the food to be tracked in the storage link to obtain a first storage risk judgment value.

[0126] The initial quality supervision data is corrected according to the label risk value and the first storage risk judgment value to obtain the first actual quality data.

[0127] First, the influence of the storage stacking parameters on the storage risk value of the tracked food is judged. For example, for a batch of food that needs to be refrigerated, too dense storage stacking will affect the circulation of cold air, and this storage stacking parameter will increase the storage risk value of the food. According to the determined storage risk value and the specific storage stacking parameters, the influence of the storage link on the quality risk of the meat food is judged, thereby obtaining a first storage risk judgment value.

[0128] First, the specific content of the storage stacking parameters needs to be determined, such as stacking height, stacking density, and adjacent food spacing. Then a large amount of historical data is collected, which covers the quality changes of the food in the storage process under different storage stacking parameters and the corresponding risk events. Through analysis of these historical data, a correlation model between the storage stacking parameters and the storage risk value is established. For example, a multiple linear regression model is used, assuming that the storage risk value is , the stacking height is h, the stacking density is d, and the adjacent food spacing is s, and the formula is constructed, where , , is a coefficient obtained by fitting historical data, and e is an error term. After substituting the storage stacking parameters of the food to be tracked into the formula, the corresponding storage risk value is calculated. The model is corrected in combination with the real-time monitored storage environment data, so that the calculated storage risk value is more accurate.

[0129] The initial quality supervision data is obtained by comprehensive detection of the food by the quality supervision department or the enterprise's own quality detection department before the food enters the storage link.

[0130] Assuming that the storage risk value is represented by , and the risk coefficient related to the storage stacking parameters is k, then the first storage risk judgment value is . For example, the storage risk value is 0.6, and the risk coefficient k is 1.2, then the first storage risk judgment value is 0.72. The initial quality supervision data is corrected in combination with the label risk value to obtain the first actual quality data. For example, the initial quality supervision data is 80, then the first actual quality data = 80 - (0.5 + 0.72) * 10 = 80 - 12.2 = 67.8.

[0131] It also includes the following steps:

[0132] Detecting the stacking deviation value corresponding to the storage stacking parameters; detecting the label interference degree corresponding to the label identification data;

[0133] Statistical joint compliance risk value formed by the label interference degree and the storage stacking deviation is obtained;

[0134] In the statistics of the joint compliance risk value formed by the label interference degree and the storage stacking deviation, a multiple linear regression model is used as a statistical model. First, the label interference degree and the storage stacking deviation are used as independent variables, and the historical joint compliance risk value is used as the dependent variable. Through training of a large amount of historical data, the weight coefficients of each independent variable are determined. The current label interference degree and the storage stacking deviation value are substituted into the statistical model to obtain the joint compliance risk value. The calculation logic of the statistical model is to quantify the comprehensive influence of the label interference and the storage stacking deviation on the joint compliance risk. The weight coefficient reflects the proportion of the role of different factors in the formation of the joint risk, so that the joint compliance risk value can accurately reflect the degree of food quality compliance risk under the joint action of the two, providing data support for subsequent risk assessment and control.

[0135] According to the joint compliance risk value, the stacking deviation value is adapted to the deviation splitting of the storage link to obtain the storage stacking deviation value to be tracked.

[0136] ​First, the stacking deviation value corresponding to the storage stacking parameters is detected, and at the same time, the label interference level corresponding to the label identification data is detected. For example, for a batch of food to be tracked, the stacking situation during storage is detected, and it is found that the actual stacking height deviates from the standard stacking height, thus obtaining the stacking deviation value; the labels of the food to be tracked are also detected, and if the storage conditions are incorrectly stated on the labels, the degree of label interference is determined.

[0137] The deviation between actual stacking and labeling requirements is extracted from warehouse storage specifications. Interference-related features, such as storage area mismatch caused by labeling errors, are extracted from traceability information. For example, if a food label specifies a maximum stacking of 5 layers, but 8 layers are actually stacked, this is a deviation feature. If traceability information shows the food should be stored in the refrigerated area but was stored in the ambient temperature area due to labeling errors, this is a storage area mismatch feature. Next, the influence weights of these interference features are determined according to food category storage specifications. Assuming the deviation feature weight is w1 and the mismatch feature weight is w2, and combining this with the specific numerical values ​​of each feature, such as the deviation degree value f1 and the mismatch degree value f2, the influence weights are determined using formulas. Calculate the degree of label interference By comprehensively considering the various interference features and their influence weights in the label identification data in this way, the degree of label interference can be obtained, thereby quantifying the interference of label identification on warehousing and other processes.

[0138] Next, the associated compliance risk status formed by the degree of label interference and warehouse stacking deviation is statistically analyzed to obtain the associated compliance risk value. Assume the stacking deviation value is... The degree of label interference is indicated by... It is stated that through the constructed associated compliance risk model Calculate the associated compliance risk value ,in and The weighting coefficients are determined based on historical data, and the stacking deviation values ​​are adapted to the deviation breakdown in the warehousing process based on the associated compliance risk values ​​to obtain the warehousing stacking deviation values ​​to be tracked.

[0139] Original stacking deviation value The factors in the warehousing process are broken down according to their weights. For example, the weight of stacking height in the warehousing process is assumed to be... The spacing factor has a weight of 1. The weight of the load-bearing factor is ,and .

[0140] Through formula The calculated warehouse stacking deviation value is obtained. Where n is the number of factors affecting the warehousing process. It is the weight of the i-th factor. is the correlation coefficient of the ith factor and the joint regulation risk value.

[0141] The determination criterion of normal storage is to determine the threshold range by analyzing a large number of historical normal storage sample joint regulation risk values, for example, setting the joint regulation risk value ≤0.4 as the determination criterion of normal storage, which needs to be based on historical data statistics, such as 95% of the normal storage scene joint regulation risk value in the range, collect the label interference degree and storage stacking deviation data under the historical normal storage scene, substitute into the statistical model to get the corresponding joint regulation risk value, and then determine the distribution range of these values as the calculation basis of the historical normal storage risk value, so as to determine the determination criterion of normal storage and provide clear reference basis for risk assessment of the storage link.

[0142] For example, if the joint regulation risk value is high, it means that the storage link is greatly affected by the comprehensive influence of label interference and stacking deviation, and when splitting, more attention will be paid to analyzing the specific composition of stacking deviation from the perspective of storage, so as to obtain the to-be-traced storage stacking deviation value, so as to more accurately evaluate the influence of the storage link on food quality and safety.

[0143] The stacking deviation value and the label interference degree corresponding to the abnormal interference condition of the storage specification are processed to obtain an additional risk value, which specifically includes the following steps:

[0144] The additional risk value of the historical stacking deviation is extracted from the historical supervision data to obtain the historical additional risk value;

[0145] The difference between the historical additional risk value and the historical normal storage risk value is processed to obtain the historical additional risk amplitude;

[0146] The ratio of the historical additional risk amplitude and the historical stacking deviation value is processed to obtain the additional risk factor;

[0147] According to the additional risk factor and the to-be-traced storage stacking deviation value, the additional risk value of the to-be-traced food in the storage process is obtained.

[0148] ​The historical additional risk value is obtained by extracting the additional risk caused by the historical stacking deviation from the historical regulatory data. For example, the additional quality risk caused by the deterioration of some cookies due to stacking deviation is the historical additional risk value. The historical additional risk amplitude is obtained by subtracting the historical normal storage risk value from the historical additional risk value. Assuming that the historical normal storage risk value is the quality risk value of the cookies under standard stacking, and the historical normal storage risk value is 0.3, and the historical additional risk value is 0.5, then the historical additional risk amplitude is 0.5-0.3=0.2. The additional risk factor is obtained by dividing the historical additional risk amplitude by the historical stacking deviation value. If the historical stacking deviation value is 2, then the additional risk factor is 0.2÷2=0.1. The additional risk value of the food to be tracked in the storage process is obtained according to the additional risk factor and the storage stacking deviation value of the food to be tracked. For example, if the storage stacking deviation value of the food to be tracked is 3, then the additional risk value is 0.1×3=0.3, which is used to evaluate the additional quality risk caused by stacking deviation of the batch of cookies.

[0149] The second actual quality data is obtained by processing and analyzing the additional risk value, the storage risk coefficient, and the label risk value, specifically including the following steps:

[0150] The second storage risk judgment value of the food to be tracked in the storage process is determined according to the additional risk value and the storage risk coefficient.

[0151] The initial quality regulatory data is corrected according to the second storage risk judgment value and the label risk value to obtain the second actual quality data.

[0152] First, the second storage risk judgment value of the food to be tracked in the storage process is determined according to the additional risk value and the storage risk coefficient. For example, if the additional risk value of a batch of food to be tracked is 0.4 and the storage risk coefficient is 1.2, then the second storage risk judgment value is obtained by the formula: second storage risk judgment value=additional risk value×storage risk coefficient, where the additional risk value represents the additional risk quantization value caused by high-level interference, and the storage risk coefficient is used to reflect the influence weight of the risk level of the storage process itself on the final risk. The second storage risk judgment value obtained by multiplying the two can comprehensively reflect the storage risk degree after the additional risk and the storage risk coefficient itself under the second level of interference. Therefore, the second storage risk judgment value is 0.4×1.2=0.48, and the initial quality regulatory data is corrected in combination with the label risk value to obtain the second actual quality data. Assuming that the initial quality regulatory data is 90, the calculation formula of the second actual quality data is: second actual quality data=initial quality regulatory data-(second storage risk judgment value+label risk value)×correction coefficient, where the correction coefficient is set to 10, i.e. 90-(0.48+0.3)×10=90-7.8=82.2.

[0153] The device embodiments described above are merely illustrative, wherein the units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e., can be located in one place, or can be distributed to multiple network units. Part or all of the modules can be selected to achieve the purposes of the embodiments according to actual needs. Those skilled in the art can understand and implement without creative labor.

[0154] Through the description of the above embodiments, those skilled in the art can clearly understand that the embodiments can be realized by means of software and the necessary general hardware platform, and of course can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, etc., and includes a number of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the methods described in each embodiment or some parts of the embodiments.

[0155] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement to some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A big data-based food quality and safety traceability management system, characterized by, The application relates to a food quality supervision method and device. The application comprises: a collection module: collecting label identification data and warehouse stacking parameters of a food to be tracked; a processing module: extracting a label risk coefficient between a label inconsistency degree and a quality compliance risk based on historical supervision data, specifically comprising the following steps: extracting historical label inconsistency data from historical supervision data; calculating the difference between the historical label inconsistency data and compliance standard data to obtain a label deviation amplitude; extracting a historical compliance risk value corresponding to the label deviation amplitude from the historical supervision data, the historical compliance risk value being a quality compliance risk quantitative result corresponding to the past label deviation; calculating the ratio between the historical compliance risk value and the label deviation amplitude to obtain the label risk coefficient; processing the label risk coefficient and the label identification data to obtain a label risk value, specifically comprising the following steps: calculating the difference between the label identification data and the standard label identification value to obtain a label deviation value; multiplying the label risk coefficient and the label deviation value to obtain the label risk value; a judgment module: judging a stacking abnormality degree value of the food to be tracked in a warehouse process according to the warehouse stacking parameters of the food to be tracked; a processing judgment module: processing the stacking abnormality degree value and the label identification data to obtain a warehouse specification interference value, specifically comprising the following steps: the label identification data comprises warehouse limitation annotations and traceability association information; extracting deviation features of actual stacking from the warehouse limitation annotations and the deviation features of the warehouse area mismatch caused by identification errors from the traceability association information; wherein the deviation features and the warehouse area mismatch features are combined into an interference feature set; determining the influence weight of each feature in the interference feature set according to the warehouse specification of the food category, and performing weighted calculation on the interference feature set and the influence weight to obtain a label specification misleading degree; inputting the stacking abnormality degree value and the label specification misleading degree into a synergistic model to obtain the warehouse specification interference value; judging the warehouse specification interference degree of the food to be tracked according to the warehouse specification interference value to obtain first-grade warehouse specification interference and second-grade warehouse specification interference; The warehouse stacking parameter is judged to affect the warehouse risk value of the food to be tracked, and the warehouse risk value is , the stacking height is h, the stacking density is d, and the distance between adjacent foods is s, and the formula is constructed, wherein , , is a coefficient obtained by fitting historical data, and e is an error term; The first storage risk judgment value is obtained by judging the quality risk of the food to be tracked in the storage link according to a storage risk value and a storage stacking parameter. The storage stacking parameter includes a stacking height, a stacking density, and a distance between adjacent foods. The storage risk value is represented by , and a risk coefficient related to the storage stacking parameter is k. Therefore, the first storage risk judgment value is . a first analysis module: if it is the first-grade warehouse specification interference, processing and analyzing the warehouse stacking parameters and the label risk value to obtain first actual quality data, specifically comprising the following steps: correcting the initial quality supervision data according to the label risk value and the first warehouse risk judgment value to obtain the first actual quality data; a second analysis module: if it is the second-grade warehouse specification interference, processing the stacking deviation value corresponding to the warehouse specification abnormal interference condition and the label interference degree to obtain an additional risk value; The deviation feature weight is w1, the mismatch feature weight is w2, the deviation degree value is f1, and the mismatch degree value is f2. The formula is Calculating the label interference degree ; detecting the stacking deviation value corresponding to the warehouse stacking parameters; detecting the label interference degree corresponding to the label identification data; detecting the stacking condition during the warehouse to find that there is a deviation between the actual stacking height and the standard stacking height, so as to obtain the stacking deviation value; processing and analyzing the additional risk value, the warehouse risk coefficient and the label risk value to obtain second actual quality data, specifically comprising the following steps: According to the additional risk value and the storage risk coefficient, a second storage risk judgment value of the food to be tracked is determined, and the storage risk coefficient is used to reflect the influence weight of the risk level of the storage link itself on the final risk, and the second storage risk judgment value = additional risk value x storage risk coefficient; According to the second storage risk judgment value and the label risk value, the initial quality supervision data is corrected to obtain second actual quality data; The generation module: according to the first actual quality data or the second actual quality data, a food quality and safety early warning tracking result is generated. 2.The big data-based food quality and safety tracking management system according to claim 1, wherein, According to the storage stacking parameters of the food to be tracked, a stacking abnormality degree value of the food to be tracked in the storage process is determined, which specifically includes the following steps: Based on the storage stacking parameters and the category characteristic parameters of the food to be tracked, a stacking reference set is constructed; wherein the stacking reference set contains standard stacking pressure threshold values in different time periods, standard misplacement angle threshold values of adjacent foods to be tracked, and stacking stability attenuation coefficients; The actual pressure value borne by the food to be tracked, the actual misplacement angle of adjacent foods to be tracked, and the real-time inclination angle of the food to be tracked in the stacking process are collected; The actual pressure value is processed by ratio with the standard stacking pressure threshold value of the corresponding time period to obtain a pressure deviation index; the actual misplacement angle is processed by difference with the standard misplacement angle threshold value to obtain an angle deviation index; and the real-time inclination angle is processed by difference with the initial stacking inclination angle and then multiplied by the stacking stability attenuation coefficient to obtain an inclination deviation index; A damage correlation coefficient is determined according to the correlation between the historical packaging damage data of the food to be tracked and the storage stacking parameters; The ground flatness of the storage area is detected to obtain a basic stability coefficient; The stacking abnormality degree value is determined according to the pressure deviation index, the angle deviation index, the inclination deviation index, the damage correlation coefficient, and the basic stability coefficient. 3.The food quality and safety tracking management system based on big data according to claim 1, wherein, According to the storage specification interference value, the storage specification interference degree of the food to be tracked is determined to obtain first and second grade storage specification interferences, which specifically includes the following steps: The storage specification interference value is compared with the preset storage specification interference threshold value; If the storage specification interference value is less than the preset storage specification interference threshold value, the storage specification interference degree of the food to be tracked is determined to obtain the first grade storage specification interference; If the storage specification interference value is greater than or equal to the preset storage specification interference threshold value, the storage specification interference degree of the food to be tracked is determined to obtain the second grade storage specification interference. 4.The big data-based food quality and safety tracking management system according to claim 3, wherein, Further including the following steps: The joint regulation risk value is obtained by counting the joint regulation risk condition formed by the label interference degree and the storage stacking deviation, and the current label interference degree and the storage stacking deviation value are substituted into the statistical model to obtain the joint regulation risk value; According to the joint regulation risk value, the stacking deviation value is adapted to the deviation splitting of the storage link to obtain the tracked storage stacking deviation value; Through formula The calculated stacking deviation value of the warehouse to be tracked is obtained. Where n is the number of influencing factors in the warehousing process. It is the weight of the i-th factor. It is the i-th factor and the associated compliance risk value The correlation coefficient.

Images