Electronic quarantine mark traceability management system

Abstract

The present application relates to the technical field of identification coding, and in particular to an electronic quarantine identification traceability management system. The present application ensures the authenticity and reliability of the data source by obtaining the birth date and immunization date of animals; calculates the age difference of adjacent individuals and immunization, combines historical data to dynamically construct batch consistency mean and dispersion, and realizes the quantitative characterization of the batch biological attribute consistency; uses attribute outlier coefficient to quantify the standardized deviation of individuals relative to the normal level of the batch, and introduces the scanning time for correction, which unifies the physical operation rate and the biological attribute deviation into a dimensionless operation abnormality evaluation index; the simulator extremely fast single factor is amplified due to the scanning time approaching zero, and the heterogeneous individual mixing factor is directly reflected due to the increase of the attribute outlier coefficient, which solves the two types of violation problems without increasing the hardware cost; finally, the reliability and accuracy of the electronic quarantine identification traceability are significantly improved through the index.

Description

Technical Field

[0001] This invention relates to the field of identification and coding technology, and specifically to an electronic quarantine identification traceability management system. Background Technology

[0002] In the modern animal health supervision system, origin quarantine declaration is the first line of defense to ensure the safety and traceability of meat products. The current digital quarantine process mainly relies on a one-animal-one-tag identity verification model, where farmers scan the electronic ear tag (such as an RFID tag) worn by the animal using a handheld terminal. The system backend verifies whether the ear tag code is in the database and whether its status is normal (alive, not yet sold). Once the verification is successful, an electronic quarantine certificate is generated.

[0003] However, in actual operations, especially in large-scale farms operating under an all-in, all-out (AIO) model, quarantine declarations face two data authenticity challenges that are difficult to resolve through simple identity verification: First, mixed declarations of heterogeneous individuals. A proper AIO system requires animals from the same batch to come from the same building and have highly consistent age and immunization history. However, some farmers, in order to reach the required number of animals to be sold, may mix culled livestock and poultry of varying ages and origins into the same declared batch. Since the ear tags of these individuals are all legal when individually searched in the database, the existing system struggles to identify this seemingly compliant but actually biosafety-violating mixed-group behavior. Second, mass false declarations using simulators. Software simulators or scripts can send large amounts of legal ear tag data to the backend in a very short time, bypassing the actual on-site scanning process and achieving instant approval for false declarations. This behavior not only leads to discrepancies between certificates and goods but also severely undermines the spatiotemporal authenticity of quarantine data.

[0004] Currently, the above problems are usually solved by adding GPS positioning or video surveillance, but this significantly increases hardware costs and is limited by network access in remote areas. Existing technologies cannot effectively identify the above-mentioned logical violations using only existing barcode scanning data streams, resulting in insufficient reliability of electronic quarantine label traceability. Summary of the Invention

[0005] To address the technical problem that existing technologies struggle to effectively identify logical violations using only existing barcode scanning data streams, leading to insufficient reliability in electronic quarantine label traceability, this invention aims to provide an electronic quarantine label traceability management system. The specific technical solution adopted is as follows: The data acquisition module is used to obtain the birth date and most recent immunization date of the inspected animal at each quarantine end time, as well as the scanning duration at each quarantine end time; The feature extraction module is used to determine the attribute difference value based on the relative historical offset of the birth date and the most recent immunization date at each quarantine end time; to determine the corresponding batch consistency mean based on the relative historical distribution of the birth date at each quarantine end time; and to determine the corresponding batch dispersion based on the relative historical temporal fluctuation of the most recent immunization date at each quarantine end time. The feature fusion module is used to determine the corresponding absolute attribute deviation based on the magnitude of the difference between the attribute difference value and the batch consistency mean at each quarantine end time; determine the corresponding attribute outlier coefficient based on the relative numerical offset between the absolute attribute deviation and the batch dispersion at each quarantine end time; and correct the attribute outlier coefficient at each quarantine end time based on the scan duration to determine the corresponding operational anomaly evaluation index. The decision-making module performs electronic quarantine identification traceability based on the aforementioned operational anomaly evaluation indicators.

[0006] Furthermore, the method for obtaining the attribute difference value includes: The birth outlier value is determined based on the birth date corresponding to each quarantine end time and the local outlier factor among all birth dates; The immunization age at each quarantine end time is determined based on the absolute value of the difference between the most recent immunization date and the date of birth corresponding to each quarantine end time; the immunization outlier value is determined based on the local outlier factor between the immunization age at each quarantine end time and all immunization ages. The birth outlier value and the immune outlier value are weighted and summed to determine the attribute difference value at each quarantine end time.

[0007] Furthermore, the method for obtaining the batch consistency mean includes: All quarantine end times within a preset historical window for each quarantine end time are used as the target time corresponding to each quarantine end time; the average value of the attribute difference values ​​corresponding to all target times is used as the quarantine reference mean for each quarantine end time; the quarantine reference mean is weighted and summed with the preset quarantine mean to determine the batch consistency mean at each quarantine end time.

[0008] Furthermore, the method for obtaining the batch dispersion includes: The standard deviation of the attribute difference values ​​corresponding to all target times is used as the quarantine reference dispersion for each quarantine end time; the quarantine reference dispersion is weighted and summed with the preset quarantine standard deviation to determine the batch dispersion at each quarantine end time.

[0009] Furthermore, the method for obtaining the absolute attribute deviation includes: The absolute attribute deviation at each quarantine end time is determined based on the absolute value of the difference between the attribute difference value and the batch consistency mean.

[0010] Furthermore, the method for obtaining the outlier coefficient of the attribute includes: The outlier coefficient of each quarantine end time is determined based on the ratio between the absolute attribute deviation and the batch dispersion.

[0011] Furthermore, the method for obtaining the scan duration includes: The time interval between each quarantine end time and the previous quarantine end time is used as the scanning duration for each quarantine end time.

[0012] Furthermore, the method for obtaining the operational anomaly evaluation index includes: The scanning duration at each quarantine end time is negatively correlated to determine the corresponding gain coefficient; the gain coefficient and the attribute outlier coefficient are weighted and summed to determine the operational anomaly evaluation index at each quarantine end time.

[0013] Furthermore, the process of tracing the electronic quarantine label includes: If the operational anomaly evaluation index is less than or equal to the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be normal. If the operational anomaly evaluation index is greater than the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be abnormal. The quarantine end time corresponding to all those identified as abnormal will be designated as abnormal time. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is less than the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be an occasional mixed cluster abnormality. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is greater than or equal to the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be a high-risk order-brushing abnormality.

[0014] Furthermore, the size of the preset history window is set to accommodate 10 quarantine end times; each quarantine end time is included in the preset history window of each quarantine end time.

[0015] The present invention has the following beneficial effects: This invention obtains the birth and immunization dates of different animals at the end of each quarantine period, providing an immutable raw data foundation for subsequent analysis and ensuring the authenticity and reliability of the data source. By calculating the differences in age and immunization background between adjacent individuals and dynamically constructing batch consistency mean and dispersion based on historical data, it achieves a quantitative characterization of the consistency of biological attributes of individuals within a batch, providing a reliable statistical reference system for accurate identification of heterogeneous individuals. The invention quantifies the standardized deviation of the current individual from the normal level of the batch using attribute outlier coefficients, and simultaneously utilizes scan duration to analyze attribute outlier coefficients. The algorithm is modified to creatively map the speed of physical operations and the partial uniformity of biological attributes into a dimensionless operational anomaly evaluation index. This causes the simulator's rapid order-brushing behavior to be dramatically amplified due to the near-zero scanning time, while the heterogeneous individual infiltration behavior is directly reflected due to the increased attribute outlier coefficient. Thus, a single algorithm solves two completely different technical problems—false declarations and mixed declarations—without adding any additional hardware costs. Finally, by using the operational anomaly evaluation index for quarantine label traceability, the credibility of quarantine certification—namely, the consistency between certificates and goods and the homogeneity of batches—can be guaranteed from the source, significantly improving the reliability and accuracy of electronic quarantine label traceability. Attached Figure Description

[0016] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a structural block diagram of an electronic quarantine mark traceability management system provided in one embodiment of the present invention; Figure 2 This is a flowchart of a method for obtaining attribute difference values ​​according to an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of an electronic quarantine identification traceability management system proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0020] The specific solution of the electronic quarantine mark traceability management system provided by the present invention will be described in detail below with reference to the accompanying drawings.

[0021] Please see Figure 1 The diagram illustrates a structural block diagram of an electronic quarantine mark traceability management system provided by an embodiment of the present invention. The system includes: a data acquisition module S101, a feature extraction module S102, a feature fusion module S103, and a decision module S104.

[0022] The data acquisition module S101 is used to obtain the birth date and most recent immunization date of the inspected animal corresponding to each quarantine end time, as well as the scanning duration corresponding to each quarantine end time.

[0023] In real-world animal husbandry, for ease of management, the ages (actual survival days) of animals in the same batch should be the same or approximately the same. That is, the birth dates of all animals in the same pen should be similar. Furthermore, since animals need to be vaccinated (immunized) at specific ages after birth (e.g., 20 days or 60 days), the most recent immunization dates for animals in the same batch should also be the same or close. Therefore, this embodiment of the invention first obtains the birth dates and most recent immunization dates of different animals at the end of each quarantine period, providing reliable data support for subsequent data analysis and feature extraction. It should be noted that the data upload interval for simulator-based order manipulation is extremely short, while the data scanning interval for normal human operation is longer and unpredictable. Therefore, obtaining the scanning duration at the end of each quarantine period is beneficial for subsequently distinguishing simulator-based order manipulation.

[0024] As an example, in a specific implementation of this invention, the process of obtaining the birth date and most recent immunization date of different animals at each quarantine end time is as follows: First, the system directly reads the hardware clock value at the moment the RF module triggers an interrupt by registering the interrupt callback function of the underlying driver. This hardware clock value represents the device response time when the operator scans the animal's ear tag each time, and is recorded as the quarantine end time, with the value accurate to the millisecond level (0.001 seconds). The purpose of this operation is to avoid the delay interference caused by the operating system scheduling, where the quarantine end time has a monotonically increasing characteristic. Then, the system uses the decoded ear tag ID to retrieve the animal's electronic file in the local database, and extracts its birth date (e.g., March 3, 2015) and most recent immunization date (i.e., the most recent immunization date). Finally, for ease of calculation and storage, the birth date and most recent immunization date of each animal are encapsulated into a tuple and stored in a temporary cache.

[0025] In other specific implementations of the present invention, the most recent immunization date can also be replaced with the first immunization date of the animal's first vaccination or other dates, as long as the immunization dates of all the animals obtained correspond in the time dimension (e.g., all are the first or all are the second). It can be adjusted according to the specific implementation environment and is not limited here.

[0026] It should be noted that the quarantine process involves the operator using a handheld scanning device to scan the ear tag of each animal leaving the enclosure in sequence at the exit of the building. Therefore, each quarantine end time corresponds to one and only one different animal, and each animal also corresponds to only one quarantine end time.

[0027] The birth date and the date of most recent immunization output by the data acquisition module S101 are input into the feature extraction module S102.

[0028] The feature extraction module S102 is used to determine the attribute difference value based on the relative historical offset of the birth date and the most recent immunization date at each quarantine end time; to determine the corresponding batch consistency mean based on the relative historical distribution of the birth date at each quarantine end time; and to determine the corresponding batch dispersion based on the relative historical temporal fluctuation of the most recent immunization date at each quarantine end time.

[0029] If animals from different batches are mixed into a particular enclosure, there will be significant differences in the birth dates of the mixed-in animals and the animals in that enclosure, as well as significant differences in the most recent immunization dates of the mixed-in animals and the animals in that enclosure. Therefore, by analyzing the relative historical offsets of the birth dates and most recent immunization dates at each quarantine end point (each quarantine end point corresponds to a different animal), it is helpful to screen out the animals from different batches that may have mixed into the enclosure, laying the foundation for subsequent traceability operations.

[0030] Different farms have different management levels (for example, the age difference in intensive farms is very small, while the age difference in extensive farms is slightly larger). If a fixed threshold is used, it is easy to cause false alarms or false alarms. Therefore, it is necessary to establish dynamic benchmark features that can follow the characteristics of the current batch, namely the batch consistency mean and batch dispersion, which can provide a unified and reliable benchmark for subsequent feature fusion, and help improve the engineering practicality and reliability of traceability. On the one hand, since animals in the same batch have the same or similar birth dates, if an animal at the end of the quarantine period belongs to that batch, its birth date should match the birth date distribution of animals in that batch. Therefore, this embodiment of the invention analyzes the relative historical distribution of the animal's birth date at each quarantine end point to determine a batch consistency mean, which can provide a data basis for further feature fusion. On the other hand, since animals in the same batch should also have the same or similar most recent immunization date, if an animal at the end of the quarantine period belongs to that batch, its most recent immunization date should have little fluctuation compared to the most recent immunization date of animals in that batch. Therefore, this embodiment of the invention analyzes the relative historical temporal fluctuation of the most recent immunization date at each quarantine end point to determine a batch dispersion, which can provide a scale benchmark for subsequent feature fusion. In summary, by calculating the batch consistency mean and batch dispersion as benchmarks for subsequent analysis, it is possible to ensure that uniform and accurate anomaly determination can be achieved for farms with different management levels.

[0031] The attribute difference values, batch consistency mean, and batch dispersion output by the feature extraction module S102 are input into the feature fusion module S103.

[0032] The feature fusion module S103 is used to determine the corresponding absolute attribute deviation based on the magnitude of the difference between the attribute difference value and the batch consistency mean at each quarantine end time; determine the corresponding attribute outlier coefficient based on the relative numerical offset between the absolute attribute deviation and the batch dispersion at each quarantine end time; determine the corresponding scan duration based on the relative historical time offset at each quarantine end time; and correct the attribute outlier coefficient at each quarantine end time based on the scan duration to determine the corresponding operational anomaly evaluation index.

[0033] In an all-in, all-out farming model, the age and immunization background of animals in the same batch should be highly consistent. However, this consistency is not absolutely zero in actual operation, but rather there is a certain range of normal fluctuations (such as slight age differences within a building). In contrast to these normal fluctuations, there are also abnormal fluctuations. If we rely solely on absolute values ​​(the age difference itself, i.e., the attribute difference value) for judgment, we cannot simultaneously distinguish between intensive farming (where the normal fluctuation range is small, and even small absolute differences may represent abnormalities) and extensive farming (where the normal fluctuation range is large, and even large absolute differences are within the normal range). Therefore, we must compare the attribute difference of the current individual (the attribute difference value) with the average level of the batch itself (the batch consistency mean) in order to make an accurate judgment that conforms to the characteristics of the batch.

[0034] After unifying the differences between different farms under the same benchmark (absolute attribute deviation), further differences exist between different batches within the same farm: that is, the same absolute deviation has completely different degrees of anomaly in batches with different fluctuation levels. Therefore, in order to eliminate the dimensional differences in fluctuation levels between different batches and achieve unified judgment across batches and farms, it is necessary to analyze the relative numerical offset between absolute attribute deviation and batch dispersion to obtain dimensionless attribute outlier coefficients, providing a unified and comparable data basis for subsequent fusion with time-series features (i.e., scan time).

[0035] When there are instances of using simulators to make batch false declarations in a short period of time, the time interval between each two adjacent ear tag data uploads is extremely short, much shorter than the time interval between two adjacent scans performed by a normal operator using a handheld scanning device. This embodiment of the invention takes this situation into account. By analyzing the relative historical time offset of each quarantine end time, a scanning duration can be determined. This allows for the accurate quantification of the differences between manual scanning by a normal operator and false declarations made by simulators, significantly improving the reliability of quarantine tag traceability.

[0036] Using only a single attribute outlier coefficient is insufficient to identify perfectly homogeneous data generated by simulators. Therefore, in order to accurately distinguish between potential heterogeneous individual mixed declarations (i.e., animals from other batches mixed into the current batch) and false declarations by simulators at the source, it is necessary to modify the attribute outlier coefficient in conjunction with the scanning time. This allows for a unified mapping of physical operation rate (i.e., scanning time) and biological attribute characteristics (i.e., attribute outlier coefficient) into a dimensionless operational anomaly evaluation index. This results in the simulator's rapid order-brushing behavior being drastically amplified due to the near-zero scanning interval, and the heterogeneous individual mixing behavior being directly reflected due to the change in the attribute outlier coefficient. Thus, the same algorithm can simultaneously identify two distinct types of violations. Without increasing any additional hardware costs, this helps to block false declarations and mixed declarations from the data source, effectively improving the reliability and anti-counterfeiting capabilities of electronic quarantine labels.

[0037] The operational anomaly evaluation index output by the feature fusion module S103 is input into the decision module S104.

[0038] The decision module S104 performs electronic quarantine identification traceability based on the operational anomaly evaluation index.

[0039] The operational anomaly evaluation index includes both the attribute deviation component (attribute outlier coefficient) caused by the mixing of heterogeneous individuals and the time compression component (scanning time) caused by simulator order brushing. Therefore, by using the operational anomaly evaluation index for the traceability of electronic quarantine labels, the authenticity and compliance of quarantine behavior can be verified simultaneously from two dimensions: biological attributes and physical operations. In addition to revealing the underlying logic corresponding to the normal and compliant quarantine end time, it can also accurately distinguish the underlying operational logic of the random mixing of heterogeneous individuals and the batch order brushing by simulators, which significantly improves the reliability and authenticity of electronic quarantine label traceability.

[0040] In summary, this invention provides an immutable raw data foundation for subsequent analysis by acquiring the birth and immunization dates of different animals at the end of each quarantine period, ensuring the authenticity and reliability of the data source. By calculating the differences in age and immunization background between adjacent individuals and dynamically constructing batch consistency mean and dispersion based on historical data, it achieves a quantitative characterization of the consistency of biological attributes among individuals within a batch, providing a reliable statistical reference system for accurate identification of heterogeneous individuals. Furthermore, it quantifies the standardized deviation of the current individual from the normal level of the batch through the attribute outlier coefficient, and utilizes the scan duration to analyze attribute outliers. The coefficients are modified to creatively map the speed of physical operations and the partial uniformity of biological attributes into a dimensionless operational anomaly evaluation index. This causes the simulator's rapid order-brushing behavior to be dramatically amplified due to the near-zero scanning time, while the heterogeneous individual mixing behavior is directly reflected due to the increased attribute outlier coefficient. Thus, a single algorithm solves two completely different technical problems—false declarations and mixed declarations—without adding any additional hardware costs. Finally, by using the operational anomaly evaluation index for quarantine label traceability, the credibility of quarantine can be guaranteed from the source for the consistency between certificates and goods and the homogeneity of batches, significantly improving the reliability and accuracy of electronic quarantine label traceability.

[0041] Preferably, in some possible implementations of the embodiments of the present invention, the specific implementation process of obtaining the attribute difference value at each quarantine end time in the feature extraction module S102 includes: Please refer to Figure 2 The diagram illustrates a flowchart of an attribute difference value acquisition method according to an embodiment of the present invention, the method comprising: Step S201: Determine the birth outlier value based on the birth date corresponding to each quarantine end time and the local outlier factor among all birth dates.

[0042] If the animal scanned at the end of the current quarantine period belongs to the current batch, its birth date should not differ significantly from the birth dates of all animals in the same batch. Therefore, this embodiment of the invention considers calculating the Local Outlier Factor (LOF) of each animal relative to all animals, which can initially distinguish heterogeneous individuals (animals from other sheds) that may have been mixed into the current shed. In a specific implementation of this embodiment, the Local Outlier Factor (LOF) is calculated after arranging the immunization ages of all animals in chronological order of quarantine time. The K-neighborhood range of the Local Outlier Factor is set to 5, but can be adjusted according to the specific implementation scenario. Where there are fewer than 5 data points in the K-neighborhood range, the average value of the existing data in the K-neighborhood range is used to make up the difference.

[0043] It should be noted that the method for calculating the local outlier factor is a well-known technique and will not be elaborated here.

[0044] Step S202: Determine the immunization age at each quarantine end time based on the absolute value of the difference between the most recent immunization date and the date of birth corresponding to each quarantine end time; determine the immunization outlier value based on the immunization age at each quarantine end time and the local outlier factor among all immunization ages.

[0045] Immunization age reflects the immunization compliance of each animal. The higher the value, the more likely the animal is not from the same batch. This is because the immunization age of animals born in the same batch should be relatively consistent. Therefore, by performing outlier analysis on immunization age, it is possible to determine whether each immunization age is an outlier relative to the immunization age of all animals, which helps to distinguish heterogeneous individuals mixed into the current building.

[0046] In one specific implementation of this invention, the local outlier factor (LOF) is calculated after arranging the immunization age of all animals in chronological order of quarantine time. The K-neighborhood range of the LOF is set to 5, but can be adjusted according to the specific implementation scenario. If there are fewer than 5 data points in the K-neighborhood, the average value of the existing data in the K-neighborhood is used to make up the difference.

[0047] It should be noted that the method for calculating the local outlier factor is a well-known technique and will not be elaborated here.

[0048] Step S203: The birth outlier value and the immune outlier value are weighted and summed to determine the attribute difference value at each quarantine end time.

[0049] By weighted summing the birth outlier value and the immune outlier value, the differentiated characteristics of heterogeneous individuals can be further highlighted. In a specific implementation of this invention, the method for calculating the attribute difference value is as follows: ,in, This represents the attribute difference value at the i-th quarantine end time; This represents the date of birth at the end of the i-th quarantine period; Indicates the most recent immunization date at the end of the i-th quarantine period. A function that converts a date to an absolute number of days, such as the number of days between the date of birth and the current date; This function represents the method for finding local outlier coefficients. The weight representing the birth outlier value can be 0.5; The weight representing the immune outlier value can be 0.5. and It can be adjusted according to the specific implementation scenario. It should be noted that the formula superimposes the difference in immunization age and the difference in birth time. The larger the value of the result (attribute difference value), the greater the difference between the current scanned animal and the previous animal in terms of growth stage or immunization background. In the ideal all-in all-out mode, the attribute difference value should be close to 0 (animals in the same batch have the same age). Conversely, if the attribute difference value increases significantly (e.g., 30 days), it may indicate that mixed grouping or cross-building operations may have occurred.

[0050] It should be noted that, at the initial stage of system startup, i.e., the end time of the first quarantine, there is no previous end time of quarantine. At this time, the system does not trigger the calculation of attribute difference values ​​or any subsequent calculation processes, but only locks the end time of quarantine as the first anchor point (session start point).

[0051] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the batch consistency mean includes: All quarantine end times within a preset historical window for each quarantine end time are used as the target time corresponding to each quarantine end time; the average value of the attribute difference values ​​corresponding to all target times is used as the quarantine reference mean for each quarantine end time; the quarantine reference mean is weighted and summed with the preset quarantine mean to determine the batch consistency mean at each quarantine end time.

[0052] Because management levels vary among different farms (e.g., intensive farms have minimal age differences, while extensive farms have slightly larger age differences), using a fixed threshold can easily lead to false alarms or missed alarms. Therefore, it is necessary to establish a dynamic reference benchmark that reflects the normal state of the current batch. This embodiment of the invention considers that the average value represents the combined level of age and immune differences between adjacent individuals during recent operations. Therefore, a preset historical window is set to provide a benchmark for judging whether subsequent operations deviate from batch characteristics. Thus, in a specific implementation of this embodiment, the average batch consistency value can be expressed by the following formula: ,in, This represents the mean batch consistency at the i-th quarantine end time. represents the quarantine reference mean at the i-th quarantine end time, meaning the average of all attribute differences within the preset historical window at the i-th quarantine end time; m represents the weight of the quarantine reference mean at the i-th quarantine end time, specifically: if the number of all quarantine end times before the i-th quarantine end time is insufficient to fill the preset historical window, i.e., there are missing values ​​within the preset historical window, then only the actual quarantine end times within the preset historical window are used for calculation, and in this case, m is taken as the number of actual quarantine end times within the preset historical window (e.g., only 4 actually exist within the preset historical window). If there are no missing values ​​in the preset historical window, then the value of m is 4. (In this invention, the size of the preset historical window is set to accommodate a maximum of 10 quarantine end times, and the preset historical window for the i-th quarantine end time contains the i-th quarantine end time, i.e., m is 10). In summary, the value of m is between [1, 10]. The advantage of this setting is that it can make full use of the limited samples and effectively avoid the baseline vacuum period. The size of the preset historical window can be adjusted according to the specific implementation environment and is not limited here. This represents the preset quarantine average value. It can be 0, which means that the animals in the same batch should ideally be of the same age, or 1, which means that there can be a maximum difference of 1 day in age among all animals in the same batch. The implementer can set it according to the specific implementation environment, and there is no limitation here. The value represents the weight of the preset quarantine mean, which can be 5. It should be noted that the preset quarantine mean is introduced in this formula because: when the actual sample size is very small (e.g., ... ), in the denominator The dominant factor is the calculation result, which is forcibly pulled towards the preset ideal value (i.e., the preset quarantine mean) to prevent the initial individual abnormal data from biasing the benchmark; as the operation progresses, the real sample size gradually increases (m is greater than w). The impact is diluted, and the benchmark will gradually and smoothly transition to the statistical characteristics of the real data. It should be noted that the size of the above-mentioned preset historical window can also be adjusted by the implementer according to the specific real-time scenario, and is not limited here.

[0053] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the batch dispersion includes: The standard deviation of the attribute difference values ​​corresponding to all target times is used as the quarantine reference dispersion for each quarantine end time; the quarantine reference dispersion is weighted and summed with the preset quarantine standard deviation to determine the batch dispersion at each quarantine end time.

[0054] Standard deviation reflects the degree of consistency between the current tested animal and the current batch of animals. The larger the standard deviation, the more likely the current tested animal and the current batch of animals belong to the same batch. Conversely, the smaller the standard deviation, the less likely the current tested animal and the current batch of animals belong to the same batch. Therefore, calculating the batch dispersion helps to accurately trace heterogeneous individuals (animals that do not belong to the current building) that have mixed into the current building.

[0055] In one specific implementation of this invention, the formula for calculating batch dispersion is as follows: ,in, This represents the batch dispersion at the i-th quarantine end time. Represents the quarantine reference dispersion at the i-th quarantine end time, which means the standard deviation of all attribute difference values ​​within the preset historical window at the i-th quarantine end time; The preset quarantine standard deviation is defined as follows, with a value in the range [0, 3]. For example, a positive integer of 1 represents a reasonable fluctuation range of 1 day, used to tolerate minor input errors that may exist in the records. Alternatively, a value of 0 indicates an error tolerance of 0. m represents the weight of the quarantine reference dispersion at the end of the i-th quarantine period. Its specific value is the same as the weight of the quarantine reference mean at the end of the i-th quarantine period, which will not be elaborated further here. Similarly, the weight w of the preset quarantine standard deviation is set to the same as the weight of the preset quarantine mean, both being 5. This can be adjusted according to the specific implementation environment, as long as the value of w does not exceed half of m. The aim is to ensure that the preset values ​​(preset quarantine standard deviation and preset quarantine mean) only play a smooth transition role in the early stage of quarantine, avoiding the preset values ​​dominating the entire quarantine process. It should be noted that the role of the preset quarantine standard deviation is to enable the system to use the preset quarantine standard deviation to create a certain tolerance space in the early stage, avoiding the standard deviation being 0 due to the first few data being exactly the same, which would lead to division by zero overflow in subsequent calculations or the system being too sensitive. In other specific implementations of the present invention, the standard deviation can also be used to represent the batch dispersion. In this case, it is only necessary to take the arithmetic square root of the batch dispersion obtained above. The implementer can adjust it according to the specific implementation scenario.

[0056] It should be noted that, to prevent malicious attackers from using poisoning attacks (i.e., deliberately inputting data with extremely high dispersion or even completely non-compliant data before the operation begins, attempting to tame the system into accepting abnormal benchmarks), the system must perform security verification logic before outputting the final benchmarks (batch dispersion and batch consistency mean). Specifically, the system compares the batch consistency mean with a preset tolerance mean (which can be taken as 30 days) and the batch dispersion with a preset tolerance dispersion (which can be taken as 15 days). Only when the batch consistency mean is less than the preset tolerance mean and the batch dispersion is less than the preset tolerance dispersion is the current final benchmark deemed reliable. The system further determines whether the batch dispersion is less than a preset minimum noise floor. If it is less, the final batch dispersion is set to the preset minimum noise floor (which can be taken as 0.5 days). If it is greater than or equal to, the current batch dispersion is kept unchanged as the final batch dispersion. This is to prevent the calculated batch dispersion from being extremely small due to the possible existence of perfectly homogeneous data (i.e., all attribute differences are 0), hence the need for noise floor correction. When the batch consistency mean is greater than or equal to the preset tolerance mean, or the batch dispersion is greater than or equal to the preset tolerance dispersion, the system determines that the current historical sample data has serious logical contamination or anomalies. In this case, the system refuses to update the baseline, but instead directly rolls back to use the preset values ​​(preset quarantine standard deviation and preset quarantine mean) or the previous valid baseline as the current calculation basis, and records a baseline anomaly event in the background log. This ensures that the system will not be swayed by malicious dirty data streams and always maintains the minimum safety baseline. It should be further noted that the preset tolerance mean, preset tolerance dispersion, and preset minimum noise floor can all be adjusted by the implementer according to the specific implementation scenario; unless otherwise specified, all batch consistency mean and batch dispersion mentioned later in this invention are the final batch consistency mean and batch dispersion after the above-mentioned security verification logic.

[0057] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the absolute attribute deviation includes: The absolute attribute deviation at each quarantine end time is determined based on the absolute value of the difference between the attribute difference value and the batch consistency mean.

[0058] As explained in the feature fusion module S103, simply using the fluctuations in the attribute difference values ​​themselves is insufficient to adapt to the different farming models of different farms. Therefore, it is necessary to convert them into deviation values ​​from the average level of the batch itself (batch consistency mean), i.e., the absolute value of the difference. The larger the absolute value of the difference, the greater the fluctuation of the tested animal relative to the average level of the batch, and vice versa. Therefore, in a specific implementation of this invention, the absolute attribute deviation can be expressed by the following formula: ,in, This represents the absolute attribute deviation at the i-th quarantine end time; This represents the absolute value function; where the larger the absolute attribute deviation, the more likely the tested animal is to be a heterogeneous individual, and vice versa.

[0059] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the attribute outlier coefficient includes: The outlier coefficient of each quarantine end time is determined based on the ratio between the absolute attribute deviation and the batch dispersion.

[0060] Because the group characteristics of all animals in different enclosures are different—for example, the average age difference of all animals in one enclosure is 1 day, while the average age difference of all animals in another enclosure is 3 days—and the number of heterogeneous individuals that may be mixed in between different enclosures is also different, it is necessary to further transform the absolute attribute deviation into a value that can indicate how much the current individual's deviation deviates from the overall batch deviation level (batch dispersion). Therefore, by calculating the ratio between the absolute attribute deviation and the batch dispersion, an attribute outlier coefficient can be determined, which can indicate how many standard deviations the current individual has deviated from the average level.

[0061] In one specific implementation of this invention, the attribute outlier coefficient can be expressed as: ,in, Represents the outlier coefficient of the attribute at the i-th quarantine end time; This represents the batch dispersion at the end of the i-th quarantine period. Since it has already undergone security check logic, its value will not be 0, preventing the denominator from being 0. The formula is: ; The absolute attribute deviation at the end of the i-th quarantine period is expressed by the formula: If the outlier coefficient of this attribute is small, it indicates that the individual is highly consistent with the characteristics of the group and belongs to a homogeneous individual. If the outlier coefficient of this attribute is large, it indicates that the individual is a significant heterogeneous outlier and may be a mixed-in culled livestock or poultry (heterogeneous individual).

[0062] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the scanning duration includes: The time interval between each quarantine end time and the previous quarantine end time is used as the scanning duration for each quarantine end time.

[0063] Since the actual scanning speed of operators and the speed of rapid order placement by simulators differ significantly, the time interval between two adjacent quarantine completion times can be used to characterize the difference between manual scanning and simulator order placement. That is, the longer the scanning time, the greater the probability of manual scanning, and the shorter the scanning time, the greater the probability of simulator order placement. This is helpful for accurately tracing the source of any potential simulator order placement behavior.

[0064] It should be noted that in practice, operators may take breaks or change work areas, resulting in a longer scan time. While this closely matches the requirements of manual scanning, the physical continuity is broken. Directly incorporating this large time interval into subsequent calculations would distort the results. Therefore, after obtaining the scan time, the system compares it with a preset maximum idle threshold (e.g., 5 minutes, which can be adjusted). If the scan time exceeds the preset maximum idle threshold, the end time of the quarantine is used as the starting point of a new session and is not included in any subsequent calculations. Conversely, if the scan time is less than or equal to the preset maximum idle threshold, the normal operation process can proceed.

[0065] Preferably, in some possible implementations of the embodiments of the present invention, the method for obtaining the operational anomaly evaluation index includes: The scanning duration at each quarantine end time is negatively correlated to determine the corresponding gain coefficient; the gain coefficient and the attribute outlier coefficient are weighted and summed to determine the operational anomaly evaluation index at each quarantine end time.

[0066] Scanning time is negatively correlated with the probability of simulator-based fraudulent activity. A longer scan time increases the likelihood of human intervention, while a shorter scan time increases the likelihood of simulator-based fraud. Therefore, a negative correlation mapping is needed to ensure the resulting scan mapping value is positively correlated with the probability of simulator-based fraud. It should be noted that in reality, different scanning devices have different response speeds (i.e., different minimum response times). Using only the scan time is insufficient for accurate judgment. Therefore, a preset minimum device response time must be incorporated into the negative correlation mapping process. This converts the absolute value into a relative rate multiple of the device's physical limits, eliminating judgment bias caused by differences in device response speeds and achieving unified anomaly identification across devices and scenarios.

[0067] The larger the attribute outlier coefficient, the more likely it is that heterogeneous individuals have been mixed in (indicating an anomaly). That is, the attribute outlier coefficient is positively correlated with the probability of heterogeneous individuals being mixed in. Therefore, in this embodiment of the invention, a final operational anomaly evaluation index can be determined based on the weighted summation between the gain coefficient and the attribute outlier coefficient. The larger the value, the greater the possibility of anomalies (simulator order manipulation and heterogeneous individual mixing in), and vice versa. This can effectively provide accurate and reliable discrimination criteria for quarantine identification and traceability, which is conducive to improving the reliability of traceability results.

[0068] In one specific implementation of this invention, the formula for calculating the operational anomaly evaluation index is as follows: ,in, This represents the operational anomaly evaluation index at the i-th quarantine end time. Represents the outlier coefficient of the attribute at the i-th quarantine end time; This represents a basic logic noise floor constant, which can be set to 0.1. It is designed to prevent the product result from always being 0 when the attribute outlier coefficient is 0, thus causing the result to fail. The basic logic noise floor constant ensures that even if the data content is perfect, the total score will still increase as long as the data entry speed is abnormal. This represents the scan time at the end of the i-th quarantine period; This represents a preset scan duration threshold, which can be set to 0.01 seconds. The reason for this setting is to prevent division by zero errors when the scan duration is extremely small or equal to 0. This represents a preset minimum device response time. It is a physical constant, representing the theoretical lower limit of the time required for RF module charging, data demodulation and transmission, and manual reset. It can be set to 0.5 seconds; a positive integer. Intended as a gain coefficient The baseline value ensures that when the scan duration equals the device's minimum response time, the gain coefficient is 2; when the scan duration exceeds the device's minimum response time, the gain term approaches 1 but remains greater than 1, thus guaranteeing... It should always be greater than a positive integer 1 and minimize its impact on the evaluation value during normal operation. It should be noted that... This reflects an anomaly in biological properties. This indicates its weight (which can be 0.5), representing whether heterogeneous individuals have been mixed in. This reflects an anomaly in the physical operation. This indicates its weight (which can be 0.5, representing that the determination of biological abnormalities and physical operation abnormalities is of equal importance). and The algorithm can be adjusted according to the specific implementation scenario, as long as the sum of the two terms is a positive integer of 1, indicating whether simulator order fraud has occurred. When the operation is normal, the scanning time is greater than the minimum response time of the device, and the gain coefficient is close to a positive integer of 1. At this time, the operation anomaly evaluation index is mainly determined by the attribute outlier coefficient, and the system checks the biological attributes normally. When simulator order fraud occurs, the scanning time is much less than the minimum response time of the device. At this time, the gain coefficient is extremely large. Even if the attribute outlier coefficient is very small, the operation anomaly evaluation index obtained by the final weighted sum will be greatly amplified, thereby triggering anomaly judgment. When both mixed groups and simulator order fraud occur at the same time, both the attribute outlier coefficient and the gain coefficient are large. The sum of the two terms causes the operation anomaly evaluation index to rise extremely, which is very easy to trigger anomaly judgment.

[0069] Preferably, in some possible implementations of the embodiments of the present invention, the process of tracing the source of electronic quarantine identification includes: If the operational anomaly evaluation index is less than or equal to the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be normal; at this time, the system will provide feedback on the terminal interface (such as a green indicator light). If the operational anomaly evaluation index is greater than the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be abnormal; at this time, it is further determined whether the anomaly is due to mixed groups or simulator order brushing, and a logic anomaly prompt is displayed on the terminal interface. The quarantine end time corresponding to all those identified as abnormal will be designated as abnormal time. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is less than the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be an occasional mixed cluster abnormality. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is greater than or equal to the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be a high-risk order-brushing abnormality.

[0070] The higher the operational anomaly evaluation index, the greater the likelihood of anomalies (heterogeneous individual mixing and simulator order manipulation). Therefore, by setting a preset evaluation threshold, normal and abnormal situations can be distinguished. In a specific implementation of this invention, the preset evaluation threshold is set to 3, based on the statistical principle of 3- The principle is that, under the assumption of normal distribution, the probability of a compliant operation's score deviating from the mean by more than 3 standard deviations is extremely low (less than 0.3%). This can also be set according to the specific implementation scenario.

[0071] To prevent large-scale systemic fraud (e.g., entire batches fabricated using simulators), the system maintains an anomaly frequency counter to accumulate the number of anomaly moments within a preset historical quarantine window at the current quarantine end time. Each time an anomaly is detected, the number of anomaly moments is incremented. Since heterogeneous individual mixing occurs occasionally or on a small scale, the corresponding anomaly moments are few and do not appear consecutively. Simulator-generated fraud, however, is generated in batches, resulting in numerous and consecutive anomaly moments. Therefore, a preset anomaly threshold can be set to accurately identify occasional mixing anomalies and consecutively occurring high-risk fraud anomalies. In one specific implementation of this invention, the preset anomaly threshold is set to 4. Implementers can set this threshold according to their specific implementation scenarios, as long as it does not exceed the size of the preset historical quarantine window (set to 5 in this embodiment, but can also be adjusted). Further details and limitations are not provided here.

[0072] In summary, the two-level judgment process described above significantly improves the reliability and accuracy of electronic tagging traceability. It should be noted that, to ensure subsequent scanning operations can continue after the anomaly judgment is completed, the system performs a reset operation after the high-risk fraud anomaly judgment is finished. This involves setting the anomaly frequency counter to 0, allowing the system to resume normal operation and preventing permanent system blockage due to a single high-risk event.

[0073] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0074] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. An electronic quarantine label traceability management system, characterized in that, The system includes: The data acquisition module is used to obtain the birth date and most recent immunization date of the inspected animal at each quarantine end time, as well as the scanning duration at each quarantine end time; The feature extraction module is used to determine the attribute difference value based on the relative historical offset of the birth date and the most recent immunization date at each quarantine end time; to determine the corresponding batch consistency mean based on the relative historical distribution of the birth date at each quarantine end time; and to determine the corresponding batch dispersion based on the relative historical temporal fluctuation of the most recent immunization date at each quarantine end time. The feature fusion module is used to determine the corresponding absolute attribute deviation based on the magnitude of the difference between the attribute difference value and the batch consistency mean at each quarantine end time; determine the corresponding attribute outlier coefficient based on the relative numerical offset between the absolute attribute deviation and the batch dispersion at each quarantine end time; and correct the attribute outlier coefficient at each quarantine end time based on the scan duration to determine the corresponding operational anomaly evaluation index. The decision-making module performs electronic quarantine identification traceability based on the aforementioned operational anomaly evaluation indicators.

2. The electronic quarantine label traceability management system according to claim 1, characterized in that, The method for obtaining the attribute difference value includes: The birth outlier value is determined based on the birth date corresponding to each quarantine end time and the local outlier factor among all birth dates; The immunization age at each quarantine end time is determined based on the absolute value of the difference between the most recent immunization date and the date of birth corresponding to each quarantine end time; the immunization outlier value is determined based on the local outlier factor between the immunization age at each quarantine end time and all immunization ages. The birth outlier value and the immune outlier value are weighted and summed to determine the attribute difference value at each quarantine end time.

3. The electronic quarantine label traceability management system according to claim 1, characterized in that, The method for obtaining the batch consistency mean includes: All quarantine end times within a preset historical window for each quarantine end time are used as the target time corresponding to each quarantine end time; the average value of the attribute difference values ​​corresponding to all target times is used as the quarantine reference mean for each quarantine end time; the quarantine reference mean is weighted and summed with the preset quarantine mean to determine the batch consistency mean at each quarantine end time.

4. The electronic quarantine label traceability management system according to claim 3, characterized in that, The method for obtaining the batch dispersion includes: The standard deviation of the attribute difference values ​​corresponding to all target times is used as the quarantine reference dispersion for each quarantine end time; the quarantine reference dispersion is weighted and summed with the preset quarantine standard deviation to determine the batch dispersion at each quarantine end time.

5. The electronic quarantine label traceability management system according to claim 1, characterized in that, The method for obtaining the absolute attribute deviation includes: The absolute attribute deviation at each quarantine end time is determined based on the absolute value of the difference between the attribute difference value and the batch consistency mean.

6. The electronic quarantine label traceability management system according to claim 1, characterized in that, The method for obtaining the outlier coefficient of the attribute includes: The outlier coefficient of each quarantine end time is determined based on the ratio between the absolute attribute deviation and the batch dispersion.

7. The electronic quarantine label traceability management system according to claim 1, characterized in that, The method for obtaining the scan duration includes: The time interval between each quarantine end time and the previous quarantine end time is used as the scanning duration for each quarantine end time.

8. The electronic quarantine label traceability management system according to claim 7, characterized in that, The method for obtaining the operational anomaly evaluation index includes: The scanning duration at each quarantine end time is negatively correlated to determine the corresponding gain coefficient; the gain coefficient and the attribute outlier coefficient are weighted and summed to determine the operational anomaly evaluation index at each quarantine end time.

9. The electronic quarantine label traceability management system according to claim 5, characterized in that, The process of tracing the electronic quarantine label includes: If the operational anomaly evaluation index is less than or equal to the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be normal. If the operational anomaly evaluation index is greater than the preset evaluation threshold, the root cause operation logic corresponding to the current quarantine end time is determined to be abnormal. The quarantine end time corresponding to all those identified as abnormal will be designated as abnormal time. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is less than the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be an occasional mixed cluster abnormality. If the number of abnormal moments within the preset historical quarantine window at the current quarantine end time is greater than or equal to the preset abnormal threshold, then the abnormal root cause operation logic corresponding to the current quarantine end time is determined to be a high-risk order-brushing abnormality.

10. The electronic quarantine label traceability management system according to claim 4, characterized in that, The preset history window is set to accommodate 10 quarantine end times; each preset history window for a quarantine end time contains each quarantine end time.

Images