A method and system for tracing the origin of agricultural products in e-commerce

Abstract

This invention relates to the field of Internet of Things (IoT) information processing technology, specifically to a method and system for tracing the origin of agricultural products in e-commerce. The method includes: acquiring agricultural product traceability information and order numbers; writing traceability labels and binding the labels to the order numbers; and performing hierarchical sampling of logistics environment data to obtain historical environmental data; responding to verification requests to verify authenticity and determining the logistics stage sequence based on logistics node data; calculating the total damage factor and remaining shelf life based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold; completing responsibility attribution and credit assessment updates; and outputting the authenticity verification results, remaining shelf life, and seller credit score. This invention can simultaneously support authenticity judgment, quality assessment, and responsibility identification.

Description

Technical Field

[0001] This invention relates to the field of Internet of Things (IoT) information processing technology, specifically to a method and system for tracing the origin of agricultural products through e-commerce. Background Technology

[0002] As the cross-regional circulation and online transaction volume of fresh fruits and vegetables continue to expand, the industry's requirements for credible sources, transparent circulation, verifiable quality, and traceable responsibility are constantly increasing. Existing technologies typically handle label verification, environmental records, and operational evaluations separately, lacking a unified data loop. This makes it difficult to link product identity, logistics environment, quality changes, and operational responsibility, easily leading to problems such as unclear causes of quality changes, difficulty in distinguishing responsibility for anomalies, and insufficient basis for platform handling. Summary of the Invention

[0003] This invention provides a method and system for tracing agricultural products in e-commerce, which addresses at least how to simultaneously achieve authenticity verification, quality degradation quantification, stage responsibility differentiation, and credit result linkage updates during the circulation of agricultural products.

[0004] In a first aspect, the present invention provides a method for tracing the origin of agricultural products through e-commerce, the method comprising: Obtain agricultural product traceability information and order numbers, write agricultural product traceability information into traceability labels, establish a binding relationship between traceability label identifiers and order numbers, collect logistics environment data and perform hierarchical sampling to obtain historical environmental data; In response to the verification request, the system performs authenticity verification based on the binding relationship and sends historical environmental data, logistics node data, and verification behavior data to the server. The server determines the logistics stage sequence based on the historical environmental data and logistics node data. The total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence. The responsibility for the total damage factor is then attributed based on the logistics stage sequence. The seller credit assessment model is updated using Bayesian methods based on the attribution results and verification behavior data to obtain the seller credit score. The results of authenticity verification, remaining shelf life, and seller credit score are then output.

[0005] In one possible implementation, the agricultural product traceability information includes product identity information, place of origin information, harvest date, and batch number; the traceability label identifier is bound to the order number, including: generating a binding record corresponding to the order number; writing the traceability label identifier into the binding record; and storing the binding record on a server.

[0006] In one possible implementation, the logistics environment data includes temperature data and humidity data; the logistics environment data is collected and sampled in a tiered manner, including: collecting temperature data and humidity data; determining the intensity of environmental change based on the rate of temperature change and the rate of humidity change at adjacent sampling times; determining the sampling frequency based on the intensity of environmental change, and recording the logistics environment data according to the sampling frequency to obtain historical environmental data.

[0007] In one possible implementation, the sampling frequency is determined based on the intensity of environmental change, including: using a first sampling frequency when the intensity of environmental change is less than a first preset threshold; using a second sampling frequency when the intensity of environmental change is greater than or equal to the first preset threshold and less than a second preset threshold; and using a third sampling frequency when the intensity of environmental change is greater than or equal to the second preset threshold; wherein the first sampling frequency is less than the second sampling frequency, and the second sampling frequency is less than the third sampling frequency.

[0008] In one possible implementation, in response to a verification request, an authenticity verification is performed based on the binding relationship, and historical environmental data, logistics node data, and verification behavior data are sent to the server. This includes: reading the traceability label identifier; verifying whether the traceability label identifier matches the order number according to the binding relationship; if the verification passes, sending the historical environmental data, logistics node data, and verification behavior data to the server; and if the verification fails, outputting an authenticity verification failure message.

[0009] In one possible implementation, the server determines the logistics stage sequence based on historical environmental data and logistics node data, including: aligning the historical environmental data and logistics node data by time; dividing the logistics stages according to the time-aligned historical environmental data and logistics node data; and arranging the logistics stages in chronological order to obtain a logistics stage sequence; wherein the logistics stages include at least two of the following: warehousing stage, transportation stage, transit stage, and last-mile delivery stage.

[0010] In one possible implementation, the total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise thresholds corresponding to the logistics stage sequence. This includes: determining the basic damage factor based on temperature data from historical environmental data; determining the thermal shock damage factor based on the rate of temperature change and the allowable temperature rise thresholds corresponding to each logistics stage; determining the total damage factor based on the basic damage factor and the thermal shock damage factor; and determining the remaining shelf life based on the total damage factor and a preset reference shelf life.

[0011] In one possible implementation, attributing responsibility for the total damage factor based on the logistics stage sequence includes: decomposing the total damage factor into multiple stage damage values ​​based on the logistics stage sequence; determining the seller's responsibility weight based on the multiple stage damage values; and using the seller's responsibility weight as the responsibility attribution result.

[0012] In one possible implementation, the seller credit assessment model is updated using Bayesian methods based on the attribution results and verification behavior data. This includes: determining the observed values ​​based on the verification behavior data; updating the posterior parameters of the seller credit assessment model based on the attribution results, the observed values, and the time decay factor; determining the seller credit score based on the posterior parameters; and outputting an anomaly warning message when the observed values ​​are continuously below a preset threshold. The seller credit assessment model is a beta distribution model.

[0013] Secondly, this invention provides an agricultural e-commerce traceability system for implementing agricultural e-commerce traceability methods, the system comprising: The binding sampling module is used to obtain agricultural product traceability information and order number, write agricultural product traceability information into traceability label, establish a binding relationship between traceability label identifier and order number, collect logistics environment data and perform hierarchical sampling to obtain historical environmental data; The verification and identification module is used to respond to verification requests, perform authenticity verification based on the binding relationship, and send environmental historical data, logistics node data and verification behavior data to the server. The server determines the logistics stage sequence based on the environmental historical data and logistics node data. The damage attribution module is used to determine the total damage factor and remaining shelf life based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence, and to attribute the responsibility for the total damage factor based on the logistics stage sequence. The credit output module is used to perform a Bayesian update on the seller credit assessment model based on the responsibility attribution results and verification behavior data, obtain the seller credit score, and output the authenticity verification results, remaining shelf life, and seller credit score.

[0014] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows: By establishing binding relationships, using environmental grading sampling, staged damage calculation, and liability attribution updating techniques, the system achieves an integrated correlation between source information, logistics status, quality changes, and credit results. By combining the temperature change rate with the staged allowable temperature rise threshold, it enables dynamic assessment of the remaining shelf life during the distribution process. By incorporating liability attribution results into Bayesian updates, it achieves an evaluation function that correlates credit scores with the actual performance process. Attached Figure Description

[0015] Figure 1 This is a schematic flowchart of the method of the present invention; Figure 2 This is a block diagram of the module combination of the system of the present invention. Detailed Implementation

[0016] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0017] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0018] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0019] E-commerce traceability refers to establishing a full-process digital tracking mechanism for online transactions, encompassing product information registration, order association, circulation records, status collection, and terminal verification feedback. Its core lies in establishing a stable correspondence between physical goods, transaction orders, and process data, enabling the continuous recording, querying, and verification of the product's identity, process, and outcome information during circulation. For agricultural products, e-commerce traceability involves not only source identification and traceable circulation but also the objective characterization of quality changes during transportation and delivery. Therefore, it is necessary to integrate label binding, environmental data collection, stage identification, quality assessment, and credit feedback within the e-commerce traceability framework to create a traceability solution tailored to agricultural e-commerce scenarios.

[0020] like Figure 1 As shown, an e-commerce traceability method for agricultural products includes: Obtain agricultural product traceability information and order numbers, write agricultural product traceability information into traceability labels, establish a binding relationship between traceability label identifiers and order numbers, collect logistics environment data and perform hierarchical sampling to obtain historical environmental data; In one embodiment, after an order is generated, the e-commerce platform retrieves the traceability information and order number of the corresponding agricultural product. The tag writing terminal writes the agricultural product traceability information into the traceability tag and reads the tag identifier, sending the tag identifier and the order number to the server to establish a binding relationship. After writing is completed, the traceability tag enters the logistics process with the agricultural product, and the environmental collection unit begins to collect logistics environmental data. Initially, temperature and humidity are recorded according to a basic sampling frequency. Subsequently, the sampling frequency is automatically adjusted according to environmental changes, forming historical environmental data arranged in chronological order. The historical environmental data can be stored in the traceability tag or cached in the collection terminal before being synchronized to the server, thereby ensuring a continuous and traceable data foundation for subsequent verification, stage identification, and shelf-life calculation.

[0021] Agricultural product traceability information includes product identity information, place of origin information, harvest date and batch number; establishing a binding relationship between traceability label identifiers and order numbers includes: generating a binding record corresponding to the order number; writing the traceability label identifiers into the binding record; and storing the binding record on the server.

[0022] In one embodiment, agricultural product traceability information is used to characterize the source and circulation attributes of agricultural products for sale, including at least product identity information, place of origin information, harvest date, and batch number. Product identity information may include product name, category, specifications, and packaging form, used to distinguish different product items from the same merchant; place of origin information may include production area, production entity name, or production entity code, used to characterize the source location; harvest date reflects the start time before the agricultural product enters circulation; batch number distinguishes the differences of the same product in different harvesting batches, sorting batches, or shipping batches. After receiving the order number, the label writing terminal first generates a corresponding binding record based on the order number, then writes the label identifier into the binding record, and stores the binding record on the server.

[0023] The binding record includes at least the order number, tag identifier, creation time, and binding status. If necessary, it can also include the merchant identifier, product item number, and the most recent update time. To prevent a tag from being bound to multiple orders, the server searches the historical records corresponding to the tag identifier before writing the binding record. When the tag identifier is unbound, creating a new binding record is allowed. When the tag identifier is already validly bound to another incomplete order, the server refuses to write it again and returns an exception message to the tag writing terminal. To ensure rapid matching in subsequent verification stages, the server can establish a bidirectional search table with the order number as the primary index and the tag identifier as the secondary index. With this setup, the order number can directly locate the unique binding record, and the tag identifier can also be used to locate the corresponding order. After the tag is written, the tag writing terminal can write a successful binding marker back to the status area of ​​the traceability tag, indicating that the current traceability tag has entered a valid circulation state. In this way, a one-to-one correspondence is formed between the basic identity information of agricultural products, order flow information, and the physical carrier of the tag. Subsequent authenticity verification only requires reading the tag identifier and calling the binding record in the server to determine whether the current tag matches the current order, providing a basis for the association between historical environmental data and specific orders.

[0024] Logistics environment data includes temperature and humidity data; collecting logistics environment data and performing hierarchical sampling includes: collecting temperature and humidity data; determining the intensity of environmental change based on the rate of temperature and humidity change between adjacent sampling times; determining the sampling frequency based on the intensity of environmental change, and recording the logistics environment data according to the sampling frequency to obtain historical environmental data.

[0025] In one embodiment, the logistics environment data includes temperature data and humidity data. Temperature data reflects changes in the thermal state of the environment in which the agricultural products are located during cold chain transportation, while humidity data reflects changes in the moisture content of the environment surrounding the packaging. Temperature and humidity data acquisition can be performed by a miniature temperature and humidity sensing unit embedded in the traceability label, or by a standalone data collector used in conjunction with the traceability label.

[0026] Before data collection begins, the data acquisition terminal performs time calibration and adds a sampling time marker to each sampling record to ensure that subsequent historical environmental data can be arranged in chronological order. During the data collection process, temperature and humidity values ​​are acquired simultaneously in each sampling cycle, and the current sampling result is compared with the result of the previous sampling cycle to determine the intensity of environmental changes.

[0027] To facilitate direct implementation by those skilled in the art, the intensity of environmental change can be determined using the following formula: in, For the first The intensity of environmental change corresponding to each sampling period For the first Temperature values ​​for each sampling period This is the temperature value from the previous sampling period. For the first Humidity values ​​for each sampling period, The humidity value is from the previous sampling period. The time interval between two adjacent sampling periods. Weighted by temperature change, The weighting is based on humidity changes.

[0028] The weights for temperature and humidity changes can be pre-set based on the sensitivity of the agricultural product category. For fresh fruits, fresh-cut vegetables, or chilled meats that are more sensitive to temperature, the weight for temperature changes can be higher than that for humidity changes; for leafy vegetables that are more sensitive to water loss, the weight for humidity changes can be appropriately increased. The data collection terminal adjusts the sampling frequency according to the intensity of environmental changes and records temperature data, humidity data, and sampling time according to the adjusted sampling frequency, forming historical environmental data. Historical environmental data can be stored in a time-series format, with each record including at least the sampling time, temperature value, humidity value, and sampling frequency marker. After this processing, the rate of environmental change at different logistics stages can be continuously recorded, and subsequent stage identification and remaining shelf life calculation can directly call upon historical environmental data without further data entry.

[0029] The sampling frequency is determined based on the intensity of environmental change, including: using a first sampling frequency when the intensity of environmental change is less than a first preset threshold; using a second sampling frequency when the intensity of environmental change is greater than or equal to the first preset threshold and less than a second preset threshold; and using a third sampling frequency when the intensity of environmental change is greater than or equal to the second preset threshold; wherein the first sampling frequency is less than the second sampling frequency, and the second sampling frequency is less than the third sampling frequency.

[0030] In one embodiment, the sampling frequency is determined based on the intensity of environmental change using a tiered switching method. When the intensity of environmental change is less than a first preset threshold, it indicates that the current environment is in a relatively stable state, and the first sampling frequency can be used for recording. When the intensity of environmental change is greater than or equal to the first preset threshold and less than a second preset threshold, it indicates that the current environment is fluctuating to some extent but has not yet entered a significant abnormal state, and the second sampling frequency can be used for recording. When the intensity of environmental change is greater than or equal to the second preset threshold, it indicates that the current environment has undergone significant changes, and a third sampling frequency is needed to increase the recording density. The first sampling frequency is less than the second sampling frequency, and the second sampling frequency is less than the third sampling frequency. The first and second preset thresholds can be comprehensively set based on historical transportation data, cold chain requirements for agricultural products, and tag power capacity.

[0031] A feasible setting principle is to first statistically analyze the daily fluctuation range of environmental change intensity under normal transportation conditions, and then use the upper limit of this fluctuation range as the base value of the first preset threshold; then, combined with the upper limit of permissible short-term fluctuations for agricultural products, determine the second preset threshold. With this setting, the first preset threshold is mainly used to distinguish between stable transportation conditions and general fluctuation conditions, while the second preset threshold is mainly used to distinguish between general fluctuation conditions and high-risk change conditions. To avoid frequent switching of sampling frequency around the threshold, it can be required that the sampling frequency only switches to a higher level when the environmental change intensity meets the upgrade condition for two consecutive sampling cycles, and then reverts to a lower sampling frequency when it falls back to a lower range for several consecutive sampling cycles.

[0032] The first, second, and third sampling frequencies can be set to once every 30 minutes, once every 10 minutes, and once every 1 minute, respectively, or adjusted to other progressive relationships based on different agricultural product categories. For agricultural products with short shelf lives and sensitivity to temperature fluctuations, the three sampling frequencies can be increased overall; for agricultural products with higher stability and shorter transportation times, the three sampling frequencies can be decreased overall. Through this tiered switching method, data collection during stable phases will not consume excessive storage and power, while data with higher temporal resolution can be obtained during periods of environmental abrupt changes. This ensures that implementation costs are controllable while improving the ability of historical environmental data to characterize real logistics processes.

[0033] In response to the verification request, the system performs authenticity verification based on the binding relationship and sends historical environmental data, logistics node data, and verification behavior data to the server. The server determines the logistics stage sequence based on the historical environmental data and logistics node data. In one embodiment, after the user initiates a verification request, it reads the traceability tag identifier and verifies the binding relationship by combining it with the order number. If the verification passes, the user sends historical environmental data, logistics node data, and verification behavior data to the server. Upon receiving the data, the server first completes a unified timeline organization, then divides the logistics into stages based on the correspondence between historical environmental data and logistics node data, and generates a logistics stage sequence in chronological order. This logistics stage sequence characterizes the stage transitions of agricultural products from warehousing to receipt, providing a stage basis for subsequent calculations of the total damage factor, liability attribution, and credit score updates.

[0034] In response to the verification request, the system performs authenticity verification based on the binding relationship and sends historical environmental data, logistics node data, and verification behavior data to the server. This includes: reading the traceability label identifier; verifying whether the traceability label identifier matches the order number based on the binding relationship; if the verification passes, sending the historical environmental data, logistics node data, and verification behavior data to the server; and if the verification fails, outputting an authenticity verification failure message.

[0035] In one embodiment, the verification request is initiated by a user terminal with near-field communication (NFC) capability. The user terminal can be a mobile phone, a dedicated verification terminal, or a receiving device with integrated NFC read / write functionality. After receiving the verification command, the user terminal first reads the tag identifier from the traceability tag, then obtains the order number from the order page, order cache record, or order interface, and compares the tag identifier with the binding record corresponding to the order number.

[0036] Binding records can be stored on the server or cached locally and then synchronized with the server. During comparison, the currently valid binding record on the server is used as the primary reference. When the tag identifier, order number, and binding status match, the authenticity verification is considered successful. When the tag identifier is missing, the order number does not match, the binding status is invalid, or the tag has already been used by another order, the authenticity verification is considered unsuccessful. When the authenticity verification fails, the user terminal outputs a failure message. This message may include at least one of the following: tag and order mismatch, tag not registered, abnormal binding status, or tag reused, indicating that the current agricultural product does not meet the normal verification conditions.

[0037] To prevent abnormal data from entering subsequent analysis processes, complete data uploads are not performed after verification failures; only the failure time and type are retained for anomaly recording. After successful verification, the user terminal begins assembling and uploading data. Environmental historical data comes from traceability tags or their associated data collectors, and includes at least temperature, humidity, and sampling times recorded in chronological order. Logistics node data comes from the order fulfillment system or logistics service interface, and includes at least the node type and time of outbound, transit, delivery, and receipt nodes. Verification behavior data characterizes the verification action itself and may include verification time, verification terminal identifier, verification location marker, verification result, and number of verifications. Verification location markers do not require high-precision positioning; city-level, station-level, or delivery area markers can be used to meet subsequent business analysis needs. The user terminal encapsulates the environmental historical data, logistics node data, and verification behavior data into a single verification message and sends it to the server. The verification message may also include the order number and tag identifier, facilitating the server's rapid establishment of the association between the order, tag, and verification behavior upon receipt. With this processing method, the authenticity verification, data collection results and user verification actions are completed in a closed loop in the same request, and the server does not need to initiate supplementary queries again to enter the logistics stage for identification and processing.

[0038] The server determines the logistics stage sequence based on historical environmental data and logistics node data, including: aligning the historical environmental data and logistics node data by time; dividing the logistics stages according to the time-aligned historical environmental data and logistics node data; and arranging the logistics stages in chronological order to obtain the logistics stage sequence. The logistics stages include at least two of the following: warehousing stage, transportation stage, transit stage, and last-mile delivery stage.

[0039] In one embodiment, after receiving historical environmental data and logistics node data, the server first performs time alignment on the two types of data, then divides the logistics into stages and generates a logistics stage sequence. The purpose of time alignment is to eliminate the time granularity differences between different data sources, so that the environmental change process can correspond to the logistics fulfillment process segment by segment. In specific processing, the server first unifies the time format, converting the sampling time in the historical environmental data and the node time in the logistics node data to the same time base, and then sorts all records according to time sequence. For historical environmental data, the server retains the original sampling order; for logistics node data, the server constructs a node timetable based on node type and node time.

[0040] If the sampling frequency of historical environmental data is higher than the recording frequency of logistics node data, the server uses the logistics node time as the stage segmentation anchor point, merging the environmental records before and after the node into the corresponding time interval. If there are missing logistics node data, the server performs a completion judgment based on the interval between preceding and following nodes, the continuity of environmental changes, and the order fulfillment status. However, the completion is only used for stage boundary estimation and does not change the original environmental record values. After completing the time alignment, the server divides the logistics stages based on node type and environmental change characteristics. The warehousing stage typically corresponds to the time interval from packaging completion to outbound shipment, characterized by relatively few changes in nodes and relatively smooth environmental fluctuations. The transportation stage typically corresponds to the interval from outbound shipment to arrival at the transit point or delivery point, characterized by continuous movement, a long time span, and fluctuations in environmental data due to vehicle movement. The transit stage typically corresponds to the interval from goods entering the transit station, completing sorting, and being re-shipped, characterized by concentrated node updates, a short duration, and potential short-term fluctuations in environmental data. The last-mile delivery stage typically corresponds to the interval from the start of delivery to receipt completion, characterized by short delivery routes, rapid changes in nodes, and environmental fluctuations easily affected by unpacking, loading and unloading, and short-term stops.

[0041] For orders without intermediate stops, the logistics stages can include at least two of the following: warehousing, transportation, and last-mile delivery. For same-city instant delivery orders, only warehousing and last-mile delivery stages may be formed. After completing the stage division, the server generates a stage name, stage start time, stage end time, and associated node information for each stage, and arranges them in chronological order to form a logistics stage sequence. Once the logistics stage sequence is formed, subsequent temperature and humidity records from historical environmental data can be assigned to the corresponding stages, providing a unified basis for subsequent calls to allowable temperature rise thresholds, calculation of total damage factors, and attribution of responsibility.

[0042] The total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence. The responsibility for the total damage factor is then attributed based on the logistics stage sequence. In one embodiment, after obtaining historical environmental data and the logistics stage sequence, the server first extracts temperature change information for each time period according to the sampling time. Then, combining this information with the allowable temperature rise thresholds corresponding to different logistics stages, it calculates the basic damage and thermal shock damage of agricultural products during the current logistics process and summarizes them to obtain the total damage factor. Next, the server determines the remaining shelf life based on the correspondence between the total damage factor and a preset reference shelf life. After completing the shelf life calculation, the server further breaks down the total damage factor into each logistics stage according to the logistics stage sequence, forming stage damage values. Based on the proportion of stage damage values ​​corresponding to the stages controllable by the seller, the server determines the responsibility attribution result. Through this process, environmental changes, stage boundaries, and responsibility allocation can be unified on the same data chain, facilitating direct access for subsequent credit score updates.

[0043] The total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise thresholds corresponding to the logistics stage sequence. This includes: determining the basic damage factor based on temperature data from historical environmental data; determining the thermal shock damage factor based on the rate of temperature change and the allowable temperature rise thresholds corresponding to each logistics stage; determining the total damage factor based on the basic damage factor and the thermal shock damage factor; and determining the remaining shelf life based on the total damage factor and the preset reference shelf life.

[0044] In one embodiment, the additional limiting point for determining the total damage factor and remaining shelf life involves quantifying the natural quality decline of agricultural products and the short-term abnormal temperature rise during logistics separately before performing a combined calculation. This approach aims to distinguish between persistent temperature deviations and sudden temperature shocks, avoiding distorted results due to relying solely on average temperature. Specifically, the server first sorts the temperature data according to the sampling time in the historical environmental data and calculates the interval temperature and temperature change rate for each sampling interval.

[0045] The basic damage factor is calculated using a simplified form of the Arrhenius equation: in, The temperature is The rate of deterioration at that time, The baseline metamorphic rate at the reference temperature. This is the rate of change multiplier for every ten degrees Celsius increase in temperature. For reference temperature, Basic damage factor, For the first Temperature values ​​corresponding to each sampling interval For the first The duration of each sampling interval For the first Temperature change rate in each sampling interval For the first The allowable temperature rise threshold corresponding to the logistics stage of each sampling interval. This is the thermal shock correction factor. It is a thermal shock damage factor. The total damage factor, For the remaining shelf life, For reference shelf life.

[0046] Reference temperature, reference shelf life, and multiplication factor can be determined based on the preservation standards of agricultural products, historical test data, or preset parameters by the merchant. The allowable temperature rise threshold is set separately for each stage of logistics, with a smaller value for the warehousing stage and a more relaxed value for the transportation and last-mile delivery stages. The principle is to reflect the actual environmental tolerance at different stages while avoiding misjudging normal short-term fluctuations as abnormal temperature rises.

[0047] The thermal shock correction factor is used to adjust the impact of the temperature change rate exceeding the threshold on the total damage factor. Its value can be determined based on the correspondence between quality anomalies and temperature mutations in historical orders. After the total damage factor is determined, the server then calculates the remaining shelf life based on the reference shelf life. When the total damage factor is close to one, the remaining shelf life is close to zero; when the total damage factor is small, the remaining shelf life is close to the reference shelf life. This allows historical environmental data to be directly converted into readable remaining quality time results.

[0048] The responsibility attribution for the total damage factor is based on the logistics stage sequence, including: decomposing the total damage factor into multiple stage damage values ​​according to the logistics stage sequence; determining the seller's responsibility weight based on the multiple stage damage values; and using the seller's responsibility weight as the responsibility attribution result.

[0049] In one embodiment, the new limitation in liability attribution is that instead of directly determining the seller's liability based on the overall order result, the total damage factor is first broken down by logistics stage, and then the liability attribution result is formed using the damage percentage corresponding to the seller's controllable stage. This limitation is introduced because agricultural product quality degradation can occur at any stage of warehousing, trunk transportation, transshipment, or last-mile delivery. Without stage breakdown, subsequent credit updates could easily attribute all damage simply to the seller. Specifically, the server first determines the stage to which each sampling interval belongs based on the logistics stage sequence, and then accumulates the basic damage and thermal shock damage to the corresponding stage according to the sampling interval, obtaining multiple stage damage values. These multiple stage damage values ​​include at least two of the following: warehousing stage damage value, transportation stage damage value, transshipment stage damage value, and last-mile delivery stage damage value. Then, the server identifies the seller-controllable stages from these multiple stages. Seller-controllable stages typically include the warehousing stage after packaging is completed, the pre-shipment stage, or the initial stage of the merchant's self-operated delivery; the transportation stage, transshipment stage, and third-party last-mile delivery stage are usually classified as stages not directly controlled by the seller.

[0050] For ease of direct implementation, the seller's responsibility weight can be determined using the following formula: in, Weighting of seller responsibility For the first The stage damage value corresponding to each logistics stage. A set of stages that are controllable by the seller. This represents the total number of stages.

[0051] A higher seller responsibility weight indicates a higher proportion of total damage originating from stages within the seller's control; a lower seller responsibility weight indicates more damage occurring in logistics stages beyond the seller's direct control. To avoid bias caused by differences in stage length, stage damage values ​​are calculated based on the cumulative results of the actual sampling interval, rather than directly replacing stage duration. For orders without intermediate steps, only the actual stages are counted; for orders with missing logistics nodes, stage boundaries can be supplemented based on the temporal continuity and node relationships in historical environmental data before further segmentation, but this supplementation only changes the stage segmentation position and does not alter the recorded temperature data. After the seller responsibility weight is determined, the server writes it as the responsibility attribution result into the order analysis record for direct use in subsequent credit score updates. This processing ensures that responsibility attribution is no longer based on general judgments but on the actual damage value of each logistics stage, making it easier for the platform to distinguish between merchant quality control issues and logistics fulfillment issues.

[0052] The seller credit assessment model is updated using Bayesian methods based on the attribution results and verification behavior data to obtain the seller credit score. The results of authenticity verification, remaining shelf life, and seller credit score are then output.

[0053] In one embodiment, after obtaining the responsibility attribution result, the server performs a Bayesian update on the seller credit assessment model based on the verification behavior data corresponding to the current order, resulting in a seller credit score. After the update, the server writes the authenticity verification result, remaining shelf life, and seller credit score into the order analysis result and returns it to the order page or verification page. In this way, the authenticity judgment, remaining shelf life, and seller credit status formed in a single verification can be presented synchronously in the same output, making it convenient for users to view and for the platform to continue processing subsequent orders from the same seller using the updated credit status.

[0054] The seller credit assessment model is updated using Bayesian methods based on the attribution results and validation behavior data. This includes: determining the observations based on the validation behavior data; updating the posterior parameters of the seller credit assessment model based on the attribution results, the observations, and the time decay factor; determining the seller credit score based on the posterior parameters; and outputting anomaly warning information when the observations are continuously below a preset threshold. The seller credit assessment model is a beta distribution model.

[0055] In one embodiment, the seller credit assessment model employs a beta distribution model. A key limitation is that credit updates do not directly use the entire order's abnormal results. Instead, observations are first extracted from verification behavior data, and then the impact ratio of this observation on the seller's credit is adjusted based on the responsibility attribution results. This approach aims to link user verification behavior with prior responsibility analysis, preventing damage not caused by the seller's control from being included in the seller's credit score. Specifically, the server first extracts verification results, verification counts, verification times, and user feedback information from the verification behavior data, and determines the observations for the current order accordingly. The observations characterize the positive or negative contribution of this order to the seller's credit, and their range can be set between zero and one. When the authenticity verification passes, the number of verifications is within the normal range, and user feedback is normal, the observation is close to one; when duplicate verification, abnormal feedback, or verification failure records occur, the observation decreases towards zero.

[0056] To prevent a single isolated incident from causing excessive fluctuations in seller credit, the server introduces a time decay factor and liability attribution results to jointly update posterior parameters. The calculation process can be expressed as follows: in, For the updated positive posterior parameters, These are the updated negative posterior parameters. These are the initial positive prior parameters. These are the initial negative prior parameters. These are the positive parameters before the update. The negative parameter before the update. The time decay factor, The time interval between the current order and the last credit update. For the decay period, The seller's liability weight corresponding to the liability attribution results. For the observed values, Rate the seller's credit score. The variance of the ratings.

[0057] The principle for setting the time decay factor is that the longer the time interval, the weaker the impact of historical orders on the current credit status. The decay period is set based on the platform's statistical period, order frequency, and category fluctuation characteristics. A higher responsibility weight indicates a higher proportion of the quality damage occurred during a phase controllable by the seller, and a more significant impact of this observation on credit parameters; a lower responsibility weight results in a correspondingly weaker impact on credit parameters. The server obtains the seller's credit score based on the updated posterior parameters and combines this with the score variance to determine the score's stability. A smaller score variance indicates that the seller's current credit status is relatively stable; a larger score variance indicates significant recent order fluctuations, and the platform can continue to observe subsequent orders.

[0058] To implement anomaly alerts, the server simultaneously maintains a count of consecutive low observation values. When an observation value continuously falls below a preset threshold, an anomaly alert is output. The preset threshold can be set according to the lower limit of the historical normal order observation value distribution, and the number of consecutive occurrences can be set to two or three according to the platform's risk control strategy. The anomaly alert information can include the seller identifier, anomaly time, anomaly order number, and anomaly type, for subsequent risk control review. After completing the credit update, the server writes the authenticity verification result, remaining shelf life, and seller credit score into the output, for display on the order page and for subsequent platform processing.

[0059] like Figure 2 As shown, an agricultural e-commerce traceability system is used to implement agricultural e-commerce traceability methods. The system includes: The binding sampling module is used to acquire agricultural product traceability information and order numbers, write the traceability information into traceability tags, and establish a binding relationship between the traceability tag identifier and the order number. It also collects logistics environment data and performs tiered sampling to obtain historical environmental data. The binding sampling module mainly consists of a tag writing terminal, traceability tags, and environmental acquisition hardware. The tag writing terminal can be a handheld device, industrial tablet, or industrial control terminal with near-field communication (NFC) read / write capabilities, used to acquire agricultural product traceability information and order numbers and write the relevant information into the traceability tag. The traceability tag itself may include a tag chip, a tag identifier storage unit, a NFC antenna, and a basic storage unit for storing the tag identifier and traceability information. The environmental acquisition hardware can integrate a temperature sensor, a humidity sensor, a low-power control chip, a clock unit, and a storage unit. The low-power control chip is responsible for performing tiered sampling according to preset rules, and the storage unit is used to generate historical environmental data. This module emphasizes the integrated coordination of "writing capability + sensing capability + local recording capability" in its hardware.

[0060] The verification and identification module responds to verification requests, performs authenticity verification based on the binding relationship, and sends historical environmental data, logistics node data, and verification behavior data to the server. The server determines the logistics stage sequence based on the historical environmental data and logistics node data. The verification and identification module mainly consists of a user verification terminal, communication interface hardware, and server access hardware. The user verification terminal can be a smartphone with near-field communication (NFC) capabilities, a self-service verification terminal, or a dedicated handheld device, used to read traceability tags, initiate verification requests, and complete authenticity verification. The communication interface hardware can include a cellular communication unit, a wireless LAN unit, or a wired network interface, used to upload historical environmental data, logistics node data, and verification behavior data to the server. The server access side typically includes a network interface card, a gateway device, and an application server host. The application server host is responsible for receiving uploaded data and performing logistics stage sequence identification based on the processor and memory. This module emphasizes a closed-loop link of "near-field reading + remote transmission + server-side identification" in its hardware.

[0061] The damage attribution module determines the total damage factor and remaining shelf life based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise thresholds corresponding to the logistics stage sequence. It then assigns responsibility for the total damage factor based on the logistics stage sequence. This module is primarily deployed on server-side hardware, typically consisting of a server processor, RAM, database storage, and necessary computing acceleration resources. The server processor performs calculations related to the Arrhenius equation, the rate of temperature change, the allowable temperature rise threshold, the total damage factor, and the remaining shelf life conversion. RAM caches historical environmental data, logistics stage sequences, and stage threshold parameters. The database storage stores historical environmental records, logistics node records, category parameters, and stage attribution results. For large orders, multi-core processors or distributed computing nodes can be configured to improve concurrent processing capabilities. This module is hardware-oriented towards "centralized computing and data support," essentially using server computing power and storage capacity to complete quality assessment and responsibility allocation.

[0062] The credit output module is used to perform Bayesian updates on the seller credit assessment model based on the responsibility attribution results and verification behavior data, obtaining the seller credit score and outputting the authenticity verification results, remaining shelf life, and seller credit score. The credit output module mainly consists of server computing hardware, result storage hardware, and front-end output hardware. The server computing hardware continues to perform Bayesian update calculations on the seller credit assessment model to generate the seller credit score. The result storage hardware stores the authenticity verification results, remaining shelf life, credit score, and anomaly warning information. The front-end output hardware can include an e-commerce platform server, user terminal display unit, merchant back-end terminal, and operations management terminal. The display unit presents the authenticity verification results, remaining shelf life, and seller credit score, while the operations management terminal can also receive anomaly warning information. This module is hardware-wise a combination of "server-side computation + result persistence + multi-terminal display output."

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0064] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. A method for tracing the origin of agricultural products through e-commerce, characterized in that, The method includes: Obtain agricultural product traceability information and order number, write agricultural product traceability information into traceability label, establish a binding relationship between traceability label identifier and order number, collect logistics environment data and perform hierarchical sampling to obtain historical environmental data; In response to the verification request, the authenticity of the binding relationship is verified, and the environmental history data, logistics node data and verification behavior data are sent to the server. The server determines the logistics stage sequence based on the environmental history data and logistics node data. The total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence, and the responsibility for the total damage factor is attributed based on the logistics stage sequence. The seller credit assessment model is updated using Bayesian methods based on the attribution results and the verification behavior data to obtain the seller credit score. The results of the authenticity verification, the remaining shelf life, and the seller credit score are then output.

2. The method according to claim 1, characterized in that, The agricultural product traceability information includes product identity information, place of origin information, harvest date, and batch number; the traceability label is linked to the order number, including: Generate a binding record corresponding to the order number; Write the traceability tag identifier into the binding record; The binding record is stored on the server.

3. The method according to claim 1, characterized in that, The logistics environment data includes temperature data and humidity data; collecting the logistics environment data and performing tiered sampling includes: Collect temperature and humidity data; The intensity of environmental change is determined based on the rate of temperature change and the rate of humidity change at adjacent sampling times; The sampling frequency is determined based on the intensity of the environmental change, and the logistics environment data is recorded according to the sampling frequency to obtain the historical environmental data.

4. The method according to claim 3, characterized in that, Determining the sampling frequency based on the intensity of the environmental changes includes: When the intensity of the environmental change is less than a first preset threshold, a first sampling frequency is used; When the intensity of the environmental change is greater than or equal to the first preset threshold and less than the second preset threshold, a second sampling frequency is used; When the intensity of the environmental change is greater than or equal to the second preset threshold, a third sampling frequency is used; Wherein, the first sampling frequency is less than the second sampling frequency, and the second sampling frequency is less than the third sampling frequency.

5. The method according to claim 1, characterized in that, In response to the verification request, a verification of authenticity is performed based on the binding relationship, and the historical environmental data, the logistics node data, and the verification behavior data are sent to the server, including: Read the traceability label identifier; Verify whether the traceability tag identifier is consistent with the order number based on the binding relationship; If the verification passes, the historical environmental data, the logistics node data, and the verification behavior data will be sent to the server. If the verification fails, output a message indicating that the verification failed.

6. The method according to claim 1, characterized in that, The server determines the logistics stage sequence based on the historical environmental data and the logistics node data, including: Time alignment is performed on the historical environmental data and the logistics node data; The logistics stages are divided based on the time-aligned historical environmental data and the logistics node data. The logistics stages are arranged in chronological order to obtain the logistics stage sequence. The logistics stages include at least two of the following: warehousing, transportation, transit, and last-mile delivery.

7. The method according to claim 6, characterized in that, The total damage factor and remaining shelf life are determined based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence, including: The basic damage factor is determined based on the temperature data in the aforementioned historical environmental data. The thermal shock damage factor is determined based on the temperature change rate and the allowable temperature rise threshold corresponding to each logistics stage. The total damage factor is determined based on the basic damage factor and the thermal shock damage factor. The remaining shelf life is determined based on the total damage factor and the preset reference shelf life.

8. The method according to claim 1, characterized in that, Attributing responsibility for the total damage factor based on the logistics stage sequence includes: The total damage factor is decomposed into multiple stage damage values ​​according to the logistics stage sequence; The seller's liability weight is determined based on the damage values ​​at the multiple stages. The seller's liability weight is used as the liability attribution result.

9. The method according to claim 8, characterized in that, The seller credit assessment model is updated using Bayesian methods based on the attribution results and the verification behavior data, including: The observations are determined based on the verification behavior data; The posterior parameters of the seller credit assessment model are updated based on the attribution results, the observed values, and the time decay factor. The seller credit score is determined based on the posterior parameters; If the observed value is continuously lower than a preset threshold, an abnormal warning message will be output; The seller credit assessment model is a beta distribution model.

10. An agricultural e-commerce traceability system, used to implement the agricultural e-commerce traceability method according to any one of claims 1-9, characterized in that, The system includes: The binding sampling module is used to obtain agricultural product traceability information and order number, write agricultural product traceability information into traceability label, establish a binding relationship between traceability label identifier and order number, collect logistics environment data and perform hierarchical sampling to obtain historical environmental data; The verification and identification module is used to respond to the verification request, perform authenticity verification based on the binding relationship, and send the environmental history data, logistics node data and verification behavior data to the server. The server determines the logistics stage sequence based on the environmental history data and logistics node data. The damage attribution module is used to determine the total damage factor and remaining shelf life based on the Arrhenius equation, the rate of temperature change, and the allowable temperature rise threshold corresponding to the logistics stage sequence, and to attribute the responsibility for the total damage factor based on the logistics stage sequence. The credit output module is used to perform a Bayesian update on the seller credit assessment model based on the responsibility attribution results and the verification behavior data, obtain the seller credit score, and output the authenticity verification results, the remaining shelf life, and the seller credit score.

Images