Agricultural Whole Industry Chain Traceability Method and System Based on Big Data

Abstract

The present invention discloses a method and system for tracing the entire agricultural industrial chain based on big data, specifically relating to the fields of agricultural Internet of Things and intelligent decision-making technologies, and is used to solve the problems of broken causal chains, mispositioning of damage-causing factors, and insufficient traceability credibility in existing agricultural traceability systems due to the neglect of the irreversibility of biological growth; by dividing time series data segments based on agronomic growth stages, restoring the time series law of crop growth, and dynamically marking the time window of key abnormal events to lock the damage-causing action interval; combining the phase space reconstruction technology to map quality inspection and transportation parameters into system trajectories, and screening the target growth stages associated with biological dynamic laws; through contribution weight analysis and stress resistance data fusion, identifying the main damage-causing factors and their synergistic relationships; finally, dynamically correcting the traceability path based on historical verification results and optimizing the thresholds and weights of the agronomic rule base, significantly improving the accuracy and credibility of tracing the entire agricultural industrial chain.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural Internet of Things and intelligent decision-making, and more specifically, to a method and system for tracing the entire agricultural industrial chain based on big data. Background Art

[0002] Currently, the agricultural traceability system generally uses technical means such as Internet of Things sensors, blockchain, and two-dimensional codes to collect and store data on various links of agricultural product production, processing, logistics, etc. By deploying field sensors, equipment log recording systems, and consumer-side scanning and query functions, a digital traceability framework covering the entire chain is constructed. For example, environmental parameters such as soil temperature and humidity and meteorological data in the production link can be transmitted to the cloud in real time, and the quality inspection reports in the processing link and the temperature and humidity records in the logistics link are ensured to be tamper-proof through blockchain technology. Consumers can view the product's full life cycle information by scanning the two-dimensional code.

[0003] However, the prior art does not fully consider the essential characteristics of the irreversible growth of agricultural organisms, resulting in a fundamental contradiction between the data backtracking logic and biological laws. Specifically, it is manifested as follows: The growth process of crops has time unidirectionality (for example, the growth period state cannot be traced back after harvesting), which will cover up the details of key growth events (such as the critical window period affected by extreme weather). When the product quality is abnormal, it is impossible to accurately locate the damage-causing factors in the production link (such as lack of irrigation or pest outbreak during a specific growth period). The data storage logic is disconnected from the dynamic biological growth law, ultimately leading to the traceability conclusion deviating from the true causal chain and reducing the accuracy and credibility of the entire industrial chain traceability. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present invention provide a method and system for tracing the entire agricultural industrial chain based on big data to solve the problems raised in the above background art.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A method for tracing the entire agricultural industrial chain based on big data includes the following steps:

[0007] S1. Obtain the original data of the entire agricultural industrial chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters;

[0008] S2. Divide the crop growth parameters and environmental parameters according to the agronomic growth stage to generate time series data segments for each stage;

[0009] S3. Dynamically mark the key abnormal events and time windows in the time series data segments based on the key event thresholds in the agronomic rule library;

[0010] S4. Reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory through phase space reconstruction, and generate a trajectory similarity index based on the reference trajectory predefined in the agronomic rule base to screen the target growth stage;

[0011] S5. Conduct a superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and identify the main loss-causing factors and synergy factors in combination with the crop stress resistance data;

[0012] S6. Modify the traceability path according to the contribution weight distribution of the main loss-causing factors, and reversely verify and iteratively update the key event thresholds and contribution weights in the agronomic rule base.

[0013] In a preferred embodiment, the crop growth parameters are collected by Internet of Things sensors deployed in the field, including leaf surface temperature, stem diameter, and photosynthetic rate; the environmental parameters include soil pH value, rainfall, and light intensity, which are synchronously obtained through a weather station and soil probes; the quality inspection parameters include the pesticide residue detection results and microbial indicators in the processing link, and the transportation environment parameters include the temperature and humidity fluctuation data of the cold chain carriage.

[0014] In a preferred embodiment, the agronomic growth stage is divided into the germination stage, vegetative growth stage, reproductive growth stage, and maturity stage according to the phenological period of the target crop; the crop growth parameters and environmental parameters are divided into stages according to the time intervals corresponding to the phenological periods to generate time series data segments for each stage;

[0015] The time series data segments are extracted by the sliding time window algorithm, and the window length of the sliding time window is dynamically adjusted according to the biological characteristics of each growth period.

[0016] In a preferred embodiment, the key event thresholds stored in the agronomic rule base include the extreme high temperature event threshold, the pest and disease outbreak threshold, and the irrigation shortage threshold, and the key event thresholds are generated based on the historical disaster record statistical model;

[0017] The key abnormal events in the time series data segments are dynamically marked by comparing the real-time data stream with the key event thresholds. When the crop growth parameters or environmental parameters continuously exceed the key event thresholds for a preset duration, they are marked as key abnormal events;

[0018] The time window of the key abnormal event is recorded as the start time and the end time; each key abnormal event is bound to the corresponding stage identifier.

[0019] In a preferred embodiment, the quality inspection parameters and transportation environment parameters are converted into a dynamic system trajectory through the phase space reconstruction method;

[0020] The trajectory similarity index of the dynamic system trajectory is generated by comparing the morphological differences between the current dynamic system trajectory and the reference trajectory predefined in the agronomic rule base;

[0021] The calculation of the trajectory similarity index integrates the overlapping ratio of the time windows of key abnormal events. The overlapping ratio of the time windows is the ratio of the overlapping duration between the time range of the current trajectory and the time window of the key abnormal event under the same phase identifier;

[0022] The screening of the target growth stage is achieved through the comprehensive determination of the difference between the trajectory similarity index and the preset similarity threshold and the overlapping ratio of the time windows. If the trajectory similarity index is lower than the preset similarity threshold and the overlapping ratio of the time windows is higher than the preset overlapping ratio threshold, it is marked as the target growth stage.

[0023] In a preferred embodiment, the phase space reconstruction method adopts the delay coordinate method, and the embedding dimension and time delay are determined according to the temporal characteristics of the temporal data segment. The temporal characteristics include the data sampling frequency and the parameter fluctuation period.

[0024] In a preferred embodiment, a superposition effect analysis is performed on the key abnormal events in the target growth stage. The key abnormal events include the event type, the time window range, and the associated phase identifier;

[0025] The contribution weight is generated through the linear combination of the trajectory similarity index and the overlapping ratio of the time windows. The weight coefficients of the trajectory similarity index and the overlapping ratio of the time windows are preset according to the statistical analysis of historical loss cases;

[0026] The crop stress resistance data is extracted from a predefined stress resistance database, and the stress resistance database stores the drought tolerance scores and disease resistance scores of crop varieties;

[0027] The main loss-causing factor is determined by the product of the contribution weight and the reciprocal of the stress resistance score. If the product value exceeds the preset product threshold, it is marked as the main loss-causing factor;

[0028] The synergy factor is identified by calculating the overlapping density of the time windows of key abnormal events under the same phase identifier. The overlapping density of the time windows is the ratio of the number of overlapping events to the total number of events. When the ratio exceeds the preset density threshold, it is marked as the synergy factor;

[0029] The main loss-causing factor and the synergy factor are associated with the key abnormal events, and the output data includes the event type, the contribution weight, and the associated phase identifier.

[0030] In a preferred embodiment, the traceability path correction is achieved by adjusting the priority order of the phase identifier corresponding to the main loss-causing factor in the traceability path. The priority order is arranged from high to low according to the contribution weight;

[0031] The reverse verification is completed by comparing the consistency of the damaged stage between the corrected traceability path and the historical actual loss cases;

[0032] The iterative update of the key event thresholds in the agronomic rule base is based on the reverse verification results. If the consistency ratio is lower than the preset verification threshold, the thresholds are recalculated according to the key abnormal event parameters in the actual damage stage in historical cases.

[0033] The iterative update of the contribution weights is achieved by statistically analyzing the distribution frequencies of the main damage-causing factors in each growth stage in the corrected traceability path. The weights of the factors with distribution frequencies higher than the preset frequency threshold are increased proportionally.

[0034] In a preferred embodiment, the consistency determination criterion is the coincidence ratio of the first N stage identifiers in the corrected path and the actual damage stage in historical cases, and the value of N is determined according to the average number of damage stages in historical cases.

[0035] On the other hand, the present invention provides a big data-based traceability system for the entire agricultural industry chain, including the following modules:

[0036] Data acquisition module: used to acquire the original data of the entire agricultural industry chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters.

[0037] Data segmentation module: used to divide the crop growth parameters and environmental parameters according to the agronomic growth stages to generate time-series data segments for each stage.

[0038] Event marking module: used to dynamically mark the key abnormal events and time windows in the time-series data segments based on the key event thresholds in the agronomic rule base.

[0039] Trajectory analysis module: used to reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory through phase space, and generate trajectory similarity indicators based on the reference trajectories predefined in the agronomic rule base to screen the target growth stages.

[0040] Factor identification module: used to perform superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and identify the main damage-causing factors and synergistic factors in combination with the crop stress resistance data.

[0041] Path optimization module: used to correct the traceability path according to the contribution weight distribution of the main damage-causing factors, and perform reverse verification and iterative update of the key event thresholds and contribution weights in the agronomic rule base.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] 1. By dividing the time-series data segments based on the agronomic growth stages, the irreversible biological growth process is transformed into discrete and traceable time-series nodes, ensuring that the data storage logic conforms to the natural rhythm of crop growth. By dynamically marking the time windows of key abnormal events, the action intervals of loss-causing factors such as extreme weather and pests and diseases are accurately captured, avoiding the problem of broken causal chains caused by data mixing in traditional methods. Further combined with the phase space reconstruction technology, the quality inspection and transportation parameters are mapped into system trajectories, and the target growth stages are screened through trajectory similarity analysis, so that the traceability path always focuses on the key nodes of biological growth, significantly improving the accuracy of anomaly location.

[0044] 2. Dynamically quantify the influence intensity of key abnormal events through contribution weights, and identify the main loss-causing factors and synergistic factors in combination with variety resistance differences, ensuring that the traceability conclusions reflect both external environmental impacts and the resistance characteristics of crops themselves. Finally, the agronomic rule base is optimized through closed-loop iteration, and the threshold setting and weight allocation are dynamically adjusted with the accumulation of historical data, forming an intelligent traceability system that continuously adapts to the actual agricultural production. It solves the problem of traceability deviation caused by the disconnection between static rules and dynamic growth in traditional technologies, providing a reliable basis for accurate decision-making in the entire industrial chain. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flowchart of the method for tracing the entire agricultural industrial chain based on big data according to the present invention;

[0046] Figure 2 It is a schematic structural diagram of the system for tracing the entire agricultural industrial chain based on big data according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] Embodiment 1: Figure 1 The method for tracing the entire agricultural industrial chain based on big data according to the present invention is given, including the following steps:

[0049] S1. Obtain the original data of the entire agricultural industrial chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters;

[0050] S2. Divide the crop growth parameters and environmental parameters according to the agronomic growth stages to generate time-series data segments for each stage;

[0051] S3. Dynamically mark the key abnormal events and time windows in the time-series data segments based on the key event thresholds in the agronomic rule base;

[0052] S4. Phase-space reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory, and generate a trajectory similarity index based on the reference trajectory predefined in the agronomic rule base to screen the target growth stage;

[0053] S5. Conduct a superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and combine the crop stress resistance data to identify the main loss-causing factors and synergistic factors;

[0054] S6. Modify the traceability path according to the contribution weight distribution of the main loss-causing factors, and inversely verify and iteratively update the key event thresholds and contribution weights in the agronomic rule base.

[0055] In step S1, obtain the original data of the entire agricultural industry chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters. The crop growth parameters are collected by Internet of Things sensors deployed in the field. The leaf surface temperature is measured by an infrared temperature sensor, and the infrared temperature sensor is installed in the range of 20 - 50 cm above the crop canopy, with a data collection interval of 30 minutes; the stem diameter is measured by a macro camera combined with an image analysis algorithm, and the macro camera is fixed on the side of the crop stem, with a measurement accuracy of 0.1 mm; the photosynthetic rate is obtained by a chlorophyll fluorometer, and the probe of the chlorophyll fluorometer is close to the leaf surface, with a data recording interval of 1 hour.

[0056] Among the environmental parameters, the soil pH value is monitored in real time by a soil probe buried at a depth of 10 - 30 cm in the crop root layer. The soil probe is equipped with an electrochemical sensor, and the data upload interval is 2 hours; the rainfall is recorded by a rain gauge at a weather station, and the weather station is deployed at the center of the farmland. The resolution of the rain gauge is 0.1 mm, and the data synchronization interval is 15 minutes; the light intensity is measured by a photosynthetically active radiation sensor, and the photosynthetically active radiation sensor is installed on the top of the crop canopy, with a data recording interval of 10 minutes.

[0057] The quality inspection parameters include the pesticide residue detection results and microbial indicators in the processing stage. The pesticide residue detection results are generated by liquid chromatography, and the detected substances include organophosphorus and pyrethroid compounds. The microbial indicators are determined by microbial culture methods, and the microbial indicators include the total number of colonies of Escherichia coli and Salmonella. The detection results are generated by the processing plant laboratory and stamped with an electronic seal. The transportation environment parameters include the temperature and humidity data of the cold chain carriage. The temperature and humidity data are recorded by digital temperature sensors and capacitive humidity sensors. The digital temperature sensors are distributed in the front, middle, and rear of the carriage, and the data acquisition interval is 5 minutes. The measurement error of the capacitive humidity sensor is ±3%. The quality inspection parameters and transportation environment parameters are uploaded through blockchain nodes. The blockchain nodes are deployed in the processing plant data center and the in-vehicle terminals of logistics vehicles. The uploaded data is appended with a timestamp. The timestamp format is UTC time and the precision is millisecond level. The timestamp is synchronized and calibrated with the national time service center time source.

[0058] The original data of crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters are stored in JSON fields and time series database tags. The JSON field names correspond to the parameter types. The field name for leaf temperature is "leaf_temperature", and the field name for stem diameter is "stem_diameter". When storing the data, a data fingerprint is generated using a hash algorithm, and the data fingerprint and the original data are stored in a distributed database.

[0059] The field IoT sensors are calibrated once every quarter. The infrared temperature sensor is linearly corrected in the range of 25°C - 40°C through a blackbody radiation source. The macro camera is calibrated for pixel resolution through a standard scale. The chlorophyll fluorometer is calibrated for the light quantum flux reading through a standard fluorescence plate. The rain gauge of the weather station is calibrated through a standard water injection method. The soil pH probe is calibrated at three points with pH 4.0, pH 7.0, and pH 10.0 standard buffer solutions. The pesticide residue detection equipment in the processing stage is certified by a third-party testing institution every year. The certification items include detection limit, repeatability, and recovery rate. The certification report number is stored in association with the detection data.

[0060] In the cold chain transportation stage, the in-vehicle temperature and humidity sensor data are uploaded to the blockchain node through a wireless transmission module. The transmission protocol is the MQTT protocol. The data packet includes the vehicle ID, sensor location number, and acquisition timestamp. If the temperature and humidity data exceed the preset range three times in a row, the in-vehicle terminal triggers an alarm and records the abnormal event. The abnormal event is marked as "transport_alert" and stored in association with the temperature and humidity data.

[0061] Crop growth parameters are continuously collected from the sowing date to the harvesting date. Quality inspection parameters are uploaded within 24 hours after processing is completed. Transportation environment parameters are recorded in real time throughout the logistics process. The growth parameters, quality inspection parameters, and transportation parameters of the same batch of crops are associated through the batch ID, and the batch ID format is "regional code_crop type_planting year_sequential number".

[0062] In step S2, crop growth parameters and environmental parameters are divided according to the agronomic growth stages to generate time-series data segments for each stage. The agronomic growth stages are divided into the germination stage, vegetative growth stage, reproductive growth stage, and maturity stage according to the phenological periods of the target crop. The division of phenological periods is based on the biological characteristics of the target crop under natural growth conditions. For example, the germination stage is defined as the period from when the seed breaks through the soil to when the first true leaf unfolds. The vegetative growth stage is defined as the period from when the stems and leaves grow rapidly to when flower bud differentiation occurs. The reproductive growth stage is defined as the period from flowering to fruit formation. The maturity stage is defined as the period from fruit maturity to harvestability.

[0063] Crop growth parameters and environmental parameters are stage-divided according to the time intervals corresponding to the phenological periods. The start and end times of the time intervals are determined according to the growth cycle of the target crop. For example, for rice crops, the time interval of the germination stage is from the 1st day to the 7th day after sowing, the vegetative growth stage is from the 8th day to the 40th day, the reproductive growth stage is from the 41st day to the 70th day, and the maturity stage is from the 71st day to the 90th day. Crop growth parameters include leaf surface temperature, stem diameter, and photosynthetic rate. Environmental parameters include soil pH value, rainfall, and light intensity. The data of these parameters within the corresponding time intervals are arranged in chronological order to form time-series data segments for each stage.

[0064] Time-series data segments are extracted through the sliding time window algorithm. The window length of the sliding time window is dynamically adjusted according to the biological characteristics of each growth stage. The window length of the germination stage is set to a shorter time span to adapt to the rapidly changing growth state. The window length of the vegetative growth stage is set to a longer time span to cover the stable growth cycle. The adjustment rule for the window length is that the difference between the start time and the end time of the window does not exceed one-third of the total time length of this growth stage. When the sliding time window algorithm extracts time-series data segments, it slides in chronological order with a fixed step size, and the step size is set to be less than one-half of the window length.

[0065] Each time-series data segment is bound to a unique stage identifier, which includes the crop variety code, the name of the growth stage, and the start date. The crop variety code is associated with the crop growth parameters collected in step S1, the name of the growth stage is consistent with the phenological period division result, and the start date is the timestamp of the first data point in this time-series data segment. The format of the stage identifier is a combination of the crop variety code, the name of the growth stage, and the start date. If the rice variety code is RC001 and the start date of the germination stage is 20231001, then the stage identifier is RC001_Germination stage_20231001.

[0066] To verify the accuracy of the stage division, the data distribution of the crop growth parameters and environmental parameters within the time interval should conform to the growth law of the target crop. The data of the stem diameter during the vegetative growth period should show a continuous upward trend. If the data trend is abnormal, the boundary of the time interval should be adjusted again. The adjustment method of the time interval boundary is to calculate the average duration of each stage according to the phenological period records of the same variety of crops in the historical planting data, and align the inflection point of the current growth parameter change with the average duration. If the growth inflection point of the current stem diameter is earlier than the historical average, the time interval of the vegetative growth period should be shortened accordingly.

[0067] The storage format of the time-series data segment is consistent with the structured format of the original data in step S1. A new field named stage_id is added to the JSON field to store the stage identifier, and new tags named the growth stage name and the start date are added to the time-series database tags. The stage identifier is associated with the batch ID in step S1. For the time-series data segments of the crop growth parameters with the same batch ID during the vegetative growth period, the start date in the stage identifier is associated with the sowing date of this batch ID.

[0068] The implementation of the sliding time window algorithm is based on a time series data processing library. When the window slides, the segments with inconsistent data collection intervals are ignored, and the missing data is linearly interpolated and filled. If the leaf surface temperature data is missing within a certain time window, the average value is calculated based on the data of the previous and subsequent time points for filling. The interpolated time-series data segments are marked as completed, and the non-interpolated segments are marked as original data. The marking information is stored in the data_status of the JSON field.

[0069] After the stage division is completed, each time-series data segment is associated with the key abnormal events in step S3 through the stage identifier. If the time-series data segment during the vegetative growth period is marked as having an irrigation missing event, the corresponding stage identifier is referenced in the event record. The stage identifier serves as the key index for data retrieval in subsequent steps. When screening the target growth stage in step S4, the dynamic system trajectory after phase space reconstruction is matched through the growth stage name in the stage identifier.

[0070] In step S3, based on the critical event thresholds in the agronomic rule library, key abnormal events and time windows in the time-series data segments are dynamically marked. The critical event thresholds stored in the agronomic rule library include extreme high temperature event thresholds, pest and disease outbreak thresholds, and irrigation shortage thresholds. The critical event thresholds are generated by analyzing the historical disaster records of the target crop. The historical disaster records include the historical growth data and disaster event data of the same crop variety in the same planting area. The extreme high temperature event threshold is set as the daily average temperature exceeding the heat tolerance critical value of the target crop for three consecutive days. The heat tolerance critical value is statistically obtained based on the temperature-sensitive interval in the historical growth data of this crop variety. The temperature-sensitive interval is the range of daily average temperatures in the historical data that causes yield reduction. The pest and disease outbreak threshold is set as the proportion of leaf lesion area exceeding 5% and the insect population density reaching 10 per plant. This threshold is statistically generated based on the critical conditions for disease spread in historical pest and disease outbreak events. The irrigation shortage threshold is set as the soil humidity being lower than the lowest threshold of the crop water requirement for five consecutive days. The lowest threshold is determined based on the transpiration experiment data of this crop variety at different growth stages.

[0071] The dynamic marking of key abnormal events in the time-series data segments is achieved by comparing the real-time data stream with the critical event thresholds. The real-time data stream is sourced from the time-series data segments generated in step S2. The time-series data segments extract crop growth parameters and environmental parameters in chronological order, and the parameter values are compared point by point with the thresholds of the corresponding events in the agronomic rule library. When the time for the parameter value to continuously exceed the threshold reaches the preset duration, a mark is triggered. The preset duration is dynamically adjusted according to the event type. The preset duration for extreme high temperature events is set as three days, the preset duration for pest and disease outbreak events is set as two days, and the preset duration for irrigation shortage events is set as five days.

[0072] The time window of the key abnormal event is recorded as the start time and the end time. The start time is the time point when the parameter first exceeds the threshold, and the end time is the time point when the parameter returns to within the threshold range. If the parameter does not return to within the threshold range, the end time is marked as the end time point of the time-series data segment; the range of the time window is associated with the stage division result of the time-series data segment in step S2, and the start time of the time window shall not be earlier than the start date of the time-series data segment.

[0073] Each key abnormal event is bound to the corresponding stage identifier. The stage identifier is sourced from the division result of the time-series data segment in step S2. The binding process of the stage identifier includes, according to the time window range of the key abnormal event, matching the stage identifiers of all time-series data segments in step S2, screening out the events whose time windows are completely contained in a certain time-series data segment, and writing the stage identifier of this segment into the event record; for example, if the time window of the extreme high temperature event is contained in the time-series data segment with the stage identifier RC001_Vegetative Growth Stage_20231008, the stage identifier is associated in the event record.

[0074] The critical abnormal events and time windows are stored in structured fields. The structured fields include event type, time window range, and associated stage identifier. The event type field records extreme high temperature, pest outbreak, or irrigation absence. The time window range field records the timestamps of the start time and end time. The associated stage identifier field directly references the identifier in step S2. The storage format of the structured fields is the same as the original data format in steps S1 and S2. In the JSON field, a new event_type field is added to store the event type, a new time_window field is added to store the time window range, and a new stage_id field is added to store the stage identifier. Event type and stage identifier tags are added to the tags of the time series database.

[0075] To verify the accuracy of the critical event markers, the marker results are cross-checked with the field monitoring records. If an irrigation absence event is marked during a certain period but the field irrigation log shows normal irrigation, the threshold recalibration process is triggered. The calibration method is to recalculate the irrigation absence threshold based on the soil moisture distribution during the normal irrigation period in the historical data. The calibrated threshold is updated to the agronomic rule library, and the version number and effective time are recorded.

[0076] The marker results of the critical abnormal events are used for the superposition effect analysis in the subsequent steps. If there are multiple critical abnormal events under the same stage identifier, the overlapping degree and intensity of the time windows of the events will affect the calculation of the contribution weight. The marker results are also associated with the phase space reconstruction in step S4. If the time window of a certain transportation environment parameter abnormal event overlaps with the time window of a critical abnormal event, they are associated to the same target growth stage through the stage identifier.

[0077] In step S4, the quality inspection parameters and transportation environment parameters are reconstructed into a dynamic system trajectory through phase space. Based on the reference trajectory predefined in the agronomic rule library, a trajectory similarity index is generated to screen the target growth stage. The quality inspection parameters and transportation environment parameters are sourced from the pesticide residue detection results, microbial indicators, and cold chain carriage temperature and humidity data collected in step S1. The quality inspection parameters and transportation environment parameters are arranged in chronological order to form a parameter sequence. The phase space reconstruction method uses the delay coordinate method. The embedding dimension and time delay of the delay coordinate method are determined according to the time series characteristics of the time series data segment in step S2. The time series characteristics include data sampling frequency and parameter fluctuation period. The data sampling frequency is the interval time for the sensor to collect data in step S1. The parameter fluctuation period is obtained by extracting the main frequency component of the parameter sequence through Fourier transform. For example, if the sampling frequency of the quality inspection parameter is once a day and the main frequency component is a once-a-week fluctuation, the time delay is set to 7 days and the embedding dimension is set to 3.

[0078] The reference trajectory predefined in the agronomic rule library is generated based on the time-series data segments of historical normal growth stages, and the historical normal growth stages are screened by the stage identifiers without marked key abnormal events in step S3. For example, screen the time-series data segments with the stage identifier RC001_Vegetative Growth Stage_20231008 and not associated with any key abnormal events, extract their quality inspection parameters and transportation environment parameter sequences, and generate a reference trajectory through the same phase space reconstruction method. The time length of the reference trajectory is the same as the length of the time-series data segment of the corresponding growth stage in step S2, the fluctuation amplitude threshold is set as the standard deviation of the reference trajectory data points, and the phase shift tolerance is set as the maximum allowable deviation range of the peak and trough positions of the reference trajectory.

[0079] The trajectory similarity index of the dynamic system trajectory is generated by comparing the morphological differences between the current trajectory and the reference trajectory. The morphological differences include the degrees of difference in trajectory shape, fluctuation amplitude, and phase shift. The calculation of the morphological differences uses the dynamic time warping algorithm. The dynamic time warping algorithm non-linearly aligns the time axes of the two trajectories and then calculates the minimum cumulative distance. The minimum cumulative distance is normalized to a value between 0 and 1 as the trajectory similarity index. The smaller the value, the higher the similarity. The calculation of the trajectory similarity index integrates the time window overlap ratio of the key abnormal events in step S3. The time window overlap ratio is the ratio of the overlapping duration of the time range of the current trajectory and the time window of the key abnormal events under the same stage identifier in step S3. For example, if the time range of the current trajectory is 10 days and the overlapping part of the time window of the key abnormal event is 4 days, then the overlap ratio is 40%.

[0080] The screening of the target growth stage is achieved through the comprehensive determination of the difference between the trajectory similarity index and the preset similarity threshold and the time window overlap ratio. The preset similarity threshold is set according to the statistical distribution of the similarity indices of historical normal trajectories. For example, take the 10% quantile of the similarity indices of historical normal trajectories as the threshold; the preset overlap ratio threshold is set according to the statistical analysis of the loss-causing cases of key abnormal events in step S3. For example, take the minimum value of the overlap ratio in the loss-causing cases as the threshold. If the trajectory similarity index is lower than the preset similarity threshold and the time window overlap ratio is higher than the preset overlap ratio threshold, then it is marked as the target growth stage. For example, when the trajectory similarity index is 0.3, the overlap ratio is 50%, the preset similarity threshold is 0.4, and the preset overlap ratio threshold is 40%, then it is determined as the target growth stage.

[0081] The target growth stage is associated with the stage identifier in step S2, and the screening result is stored as structured data containing the stage identifier, the trajectory similarity index, and the time window overlap ratio. The storage format of the structured data is consistent with the data formats in steps S1 to S3. A new field similarity_score is added to the JSON field to store the trajectory similarity index, and a new field overlap_ratio is added to store the time window overlap ratio. A target growth stage marker is added to the tags in the time series database. After the stage identifier is associated with the screening result in step S4, it is used as the input data for the superposition effect analysis in step S5. For example, if there are multiple target growth stage markers under the same stage identifier, the data segment corresponding to the marker with the highest overlap ratio is preferentially analyzed.

[0082] To verify the accuracy of the screening result, the target growth stage marker is compared with the abnormal growth stage recorded in the field investigation. If the marked result is inconsistent with the field record, the preset similarity threshold and the preset overlap ratio threshold are adjusted. The adjustment method is to reset the threshold according to the statistical distribution of historical correct marking cases. For example, the preset similarity threshold is adjusted from 0.4 to 0.35, and the preset overlap ratio threshold is adjusted from 40% to 45%. The adjusted thresholds are updated to the agronomic rule library, and the version number and the effective time are recorded.

[0083] The screening result is also associated with the traceability path correction in step S6. For example, the stage identifier in the target growth stage marker is used to reverse-locate the time series data segment in step S2, and the damage-causing factors are analyzed by combining the key abnormal events in step S3. The trajectory similarity index and the time window overlap ratio participate in the contribution weight calculation in step S5 as weight coefficients. For example, the lower the trajectory similarity index and the higher the overlap ratio, the greater the contribution weight of the corresponding key abnormal event.

[0084] In step S4, by reconstructing the quality inspection parameters and the transportation environment parameters into a dynamic system trajectory and performing a morphological difference analysis with the reference trajectory in the agronomic rule library, the problem of the disconnection between the data form and the dynamic law of biological growth in traditional agricultural traceability is solved; by integrating the time window overlap ratio of the key abnormal events in step S3, the target growth stage is screened by combining the dual judgment conditions of the trajectory similarity index and the preset threshold; the phase space reconstruction maps the discrete quality inspection parameters and transportation parameters into a continuous system trajectory to restore the dynamic characteristics of the growth process; the time window overlap ratio introduces the correlation analysis of key abnormal events to avoid isolated judgment errors; through the fusion of multi-dimensional data forms (trajectory shape, fluctuation amplitude) and event spatio-temporal correlation (window overlap), the damage-causing stage is accurately located, making the traceability conclusion more in line with the irreversible law of crop growth, significantly improving the accuracy of the abnormal growth stage screening, solving the problem of traceability path deviation caused by ignoring the biological time series characteristics in traditional methods, and providing reliable input for the superposition effect analysis and the identification of damage-causing factors.

[0085] In step S5, a superposition effect analysis is performed on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and the main loss-causing factors and synergistic factors are identified in combination with the crop stress resistance data. The key abnormal events in the target growth stage are derived from the screening results of step S4. The key abnormal events include the event type, the time window range, and the associated stage identifier. The event types are the extreme high temperature events, the pest and disease outbreak events, or the irrigation shortage events marked in step S3. The superposition effect analysis generates the contribution weights by integrating the trajectory similarity index of step S4 and the time window overlap ratio. The trajectory similarity index is the normalized value calculated in step S4, and the time window overlap ratio is the proportion of the overlapping duration between the time window of the key abnormal event in step S4 and the time range of the current target growth stage. The contribution weight is calculated using a linear combination of the trajectory similarity index and the time window overlap ratio. The weight coefficients of the trajectory similarity index and the time window overlap ratio are preset according to the statistical analysis results of historical loss-causing cases. For example, by analyzing the loss-causing impact degrees of different parameters in historical data, the weight coefficient of the trajectory similarity index is set to 0.6, and the weight coefficient of the time window overlap ratio is set to 0.4.

[0086] The crop stress resistance data is extracted from a predefined stress resistance database, which stores the drought tolerance scores and disease resistance scores of different crop varieties. The drought tolerance score is generated based on the survival rate of the target crop variety in the drought stress experiment. The experimental method is to simulate different drought intensities in a controllable environment and record the plant survival rate, and the survival rate is converted into a score from 0 to 1. For example, a survival rate of 80% corresponds to a score of 0.8. The disease resistance score is generated based on the lesion inhibition rate of the target crop variety in the pathogen inoculation experiment. The experimental method is to measure the proportion of the lesion area after inoculating the pathogen at a specified concentration and convert it into a score from 0 to 1. For example, an inhibition rate of 60% corresponds to a score of 0.6.

[0087] The determination of the main loss-causing factor is achieved through the product of the contribution weight and the reciprocal of the stress resistance score. The calculation formula for the product value is the contribution weight multiplied by the reciprocal of the stress resistance score. For example, if the contribution weight is 0.5 and the disease resistance score is 0.4, then the product value is 0.5×(1 / 0.4)=1.25. The preset product threshold is set according to the distribution of the product values of the main loss-causing factors in historical loss-causing cases. For example, the 90th percentile of the product values in historical cases is taken as the threshold. If the current product value exceeds this threshold, it is marked as the main loss-causing factor.

[0088] The identification of co-factors is achieved by calculating the time-window overlap density of key abnormal events under the same phase identifier. The time-window overlap density is the ratio of the number of time-window overlaps between events to the total number of events. For example, if there are 3 key abnormal events under the same phase identifier, and the number of time-window overlaps between every two events is 2, then the overlap density is the number of 2 overlaps divided by the total number of pairs of 3 events (3 pairs), and the calculation result is 2 / 3. The preset density threshold is set according to the statistical results of the overlap density of historical co-damage cases. For example, the 75th percentile of the overlap density in historical cases is taken as the threshold. If the current overlap density exceeds this threshold, it is marked as a co-factor.

[0089] The main damage-causing factor and co-factors are associated with the key abnormal events in step S3, and the output data includes the event type, contribution weight, and associated phase identifier. The event type field is recorded as extreme high temperature, pest and disease outbreak, or irrigation shortage. The contribution weight field is recorded as the calculation result of the linear combination. The associated phase identifier field references the phase identifier in step S2. The storage format of the output data is the same as the structured data in steps S1 to S4. In the JSON field, new fields leading_factor and synergistic_factor are added to mark the main damage-causing factor and co-factors respectively. In the time-series database tags, a new association tag between the event type and the phase identifier is added.

[0090] To verify the accuracy of the main damage-causing factor and co-factors, the output data is compared with the records of the damage-causing reasons in the field investigation. If the marked result is inconsistent with the field record, the preset product threshold and preset density threshold are adjusted. The adjustment method is to reset the threshold according to the statistical distribution of historical correctly marked cases. For example, the preset product threshold is adjusted from 1.2 to 1.1, and the preset density threshold is adjusted from 0.6 to 0.65. The adjusted thresholds are updated to the agronomic rule base and the version number and effective time are recorded.

[0091] The output data of the main damage-causing factor and co-factors is used for the traceability path correction in step S6. For example, the phase identifier corresponding to the main damage-causing factor is used to reverse-locate the time-series data segment in step S2, and the spatio-temporal correlation of the damage-causing event is analyzed by combining the time window of the key abnormal events in step S3. The marked result of the co-factor is used to supplement and correct the secondary damage-causing factors in the traceability path. For example, the event type of the co-factor within the same phase may indicate the occurrence of a compound disaster.

[0092] Step S5 solves the problem of identification deviation of damage-causing factors caused by single-parameter analysis in traditional agricultural traceability by integrating trajectory similarity indicators, time-window overlap ratios, and crop stress resistance data. Compared with existing technologies that only rely on the frequency or intensity of event occurrence, the trajectory similarity indicator quantifies the morphological differences in data, the time-window overlap ratio reflects the spatio-temporal correlation of events, and the stress resistance score reflects the varietal resistance differences; through linear combination (such as trajectory weight 0.6 + overlap ratio weight 0.4), the influence of events and the spatio-temporal coupling effect are dynamically balanced, and the reciprocal product of the stress resistance scores (such as contribution weight × 1 / stress resistance score) is used to reverse-strengthen the damage sensitivity of low-resistance varieties; based on the historical statistical distribution, preset thresholds are set (such as the product threshold taking the 90th percentile and the overlap density threshold taking the 75th percentile), so that the judgment criteria conform to the actual damage-causing rules and avoid subjective assumption errors. Compared with existing technologies, through multi-dimensional data collaborative analysis and dynamic threshold calibration, the accuracy of identifying the main damage-causing factors is significantly improved, and the problem of deviation of traceability conclusions caused by ignoring varietal resistance and event correlation in traditional methods is solved.

[0093] In step S6, the traceability path is corrected according to the contribution weight distribution of the main damage-causing factors, and the key event thresholds and contribution weights in the agronomic rule library are reversely verified and iteratively updated. The traceability path correction is achieved by adjusting the priority order of the stage identifiers corresponding to the main damage-causing factors in the traceability path. The priority order is arranged from high to low according to the contribution weights generated in step S5. The higher the contribution weight, the higher the priority of the stage identifier in the traceability path. For example, if the contribution weight of the stage identifier RC001_Vegetative Growth Stage_20231008 is 0.9 and the contribution weight of the stage identifier RC002_Reproductive Growth Stage_20231015 is 0.7, the system will first arrange RC001_Vegetative Growth Stage_20231008 at the forefront of the traceability path.

[0094] The reverse verification is completed by comparing the consistency of the damage-causing stages between the corrected traceability path and the historical actual damage-causing cases. The historical actual damage-causing cases are the cases where the main damage-causing factors verified in steps S3 to S5 match the field investigation records. The consistency judgment criterion is the coincidence ratio of the first N stage identifiers in the corrected path and the actual damage-causing stages in the historical cases. The value of N is determined according to the average number of damage-causing stages in the historical cases. For example, if the historical cases on average involve 3 damage-causing stages, then N is set to 3. The preset verification threshold is set according to the statistical distribution of the matching accuracy of the historical cases. For example, the matching ratio at the 95th percentile of the historical cases is taken as the threshold. If the coincidence ratio of the first 3 stage identifiers in the current corrected path and the historical cases is less than 70%, it is judged as inconsistent.

[0095] The iterative update of the key event thresholds in the agronomic rule library is based on the reverse verification results. If the consistency ratio is lower than the preset verification threshold, the thresholds are recalculated according to the key abnormal event parameters in the actual damage stage of historical cases. For example, for the extreme high temperature event threshold, the temperature data in the actual damage stage of historical cases is extracted, and its 95th percentile value is calculated as the new threshold; for the irrigation shortage threshold, the lowest 5th percentile value of the soil moisture data in historical cases is taken as the new threshold. After the threshold is updated, the version number is recorded, and the version number format is a combination of rule type, date, and serial number, such as "high temperature threshold_20231001_001".

[0096] The iterative update of the contribution weight is achieved by statistically analyzing the distribution frequency of the main damage-causing factors in each growth stage in the corrected traceability path. The distribution frequency is the ratio of the number of times the main damage-causing factor is marked in the historical corrected path in the same growth stage to the total number of times. For example, if the vegetative growth stage appears 60 times in 100 corrected paths, the distribution frequency is 60%. The preset frequency threshold is set according to the distribution statistics of the historical high-frequency damage stages. For example, the 80th percentile of the historical distribution frequency is taken as the threshold. If the current distribution frequency exceeds the threshold, the contribution weight is increased according to the exceeding ratio. For example, when the frequency is 85%, the weight increase ratio is (85% - 80%) / 80% = 6.25%.

[0097] The updated key event thresholds and contribution weights are synchronized to the agronomic rule library and the version number is recorded. The version number is associated with the stage identifier in step S2 and the main damage-causing factor in step S5. For example, the version number "high temperature threshold_20231001_001" is associated with the stage identifier RC001_vegetative growth stage_20231008 and the main damage-causing factor "extreme high temperature". The updated data storage format is the same as that in steps S1 to S5. The threshold_version and weight_version fields are added to the JSON fields to record the threshold and weight version numbers, and the association label between the version number and the stage identifier is added to the tags in the time series database.

[0098] To verify the effectiveness of the updated rule library, the new thresholds and weights are applied to historical cases that were not involved in the training for cross-validation. For example, the data from the previous year is used to test the recognition accuracy of the new thresholds for the damage stage. If the accuracy improvement exceeds the preset optimization threshold (e.g., 5%), the update is determined to be effective; otherwise, roll back to the previous version and readjust the parameters.

[0099] The updated agronomic rule base is used for the next round of data traceability analysis in steps S1 to S5. For example, in the next round of data collection, new thresholds are directly called to mark key abnormal events, realizing closed-loop iterative optimization. The version number is also used to trace the update history of the rule base. For example, through the version number "High Temperature Threshold_20231001_001", the phase identifier and contribution weight distribution of the corresponding historical case can be traced back.

[0100] It should be noted that the core feature of the agronomic rule base is the dynamic integration of multi-source data and iterative optimization based on a closed-loop feedback mechanism. Its construction process and data sources specifically include:

[0101] Key event threshold generation: Generated by statistical analysis of historical disaster records (such as temperature, humidity, pest and disease outbreak events). For example, the extreme high temperature threshold is determined by calculating the quantile value based on the temperature-sensitive interval in historical yield reduction cases (such as the number of consecutive high-temperature days and temperature range leading to yield reduction), rather than directly using the theoretical value of crop growth or fixed empirical values.

[0102] Reference trajectory construction: Generated based on the time-series data segments of historical normal growth stages in step S2. The normal growth stages are screened by the phase identifiers without marked key abnormal events in step S3. The fluctuation amplitude threshold and phase shift tolerance of the reference trajectory are dynamically set according to the statistical distribution of normal stage data (such as standard deviation, deviation of peak and trough positions).

[0103] Source of stress resistance data: Quantify the resistance level of crop varieties through controlled environment experiments (such as drought stress, pathogen inoculation). The experimental data is converted into standardized scores (such as survival rate → drought tolerance score, lesion inhibition rate → disease resistance score), and the scores are stored as structured fields and associated with phase identifiers.

[0104] Closed-loop iterative mechanism: Realize the dynamic optimization of the rule base through the reverse verification (historical case consistency comparison) and parameter update (threshold recalculation, weight improvement) in step S6. For example, the consistency ratio triggers threshold update, and the distribution frequency triggers weight adjustment. Version number management ensures the traceability of the update process.

[0105] Traditional agricultural rule bases are usually statically set based on fixed agronomic theories or expert experience, while this agronomic rule base realizes dynamicization and self-adaptation:

[0106] Data-driven threshold generation: The threshold is derived from the statistical distribution of historical loss-causing cases (such as quantile calculation), rather than a fixed threshold.

[0107] Cross-step data association: The rule base is dynamically bound to multi-step data such as phase identifiers, key abnormal events, and stress resistance scores. For example, the reference trajectory depends on the time-series segments of S2, and the threshold update depends on the reverse verification of S6.

[0108] Closed-loop feedback optimization: By associating the version number with the iterative process, the parameters of the rule base are dynamically adjusted according to the traceability results. For example, the version number "high temperature threshold_20231001_001" records the threshold update history and associated loss-causing factors.

[0109] This agronomic rule base integrates historical data statistics, experimental verification, and a closed-loop feedback mechanism, solving the problem of the disconnection between traditional static rule bases and the dynamic laws of agricultural organisms.

[0110] Embodiment 2: Figure 2 The structural schematic diagram of the agricultural whole-industry chain traceability system based on big data according to the present invention is given. The agricultural whole-industry chain traceability system based on big data includes the following modules:

[0111] Data acquisition module: Used to acquire the original data of the agricultural whole-industry chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters;

[0112] Data segmentation module: Used to divide crop growth parameters and environmental parameters according to the agronomic growth stage to generate time-series data segments for each stage;

[0113] Event marking module: Used to dynamically mark key abnormal events and time windows in the time-series data segments based on the key event thresholds of the agronomic rule base;

[0114] Trajectory analysis module: Used to reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory through phase space, and generate trajectory similarity indicators based on the reference trajectories predefined in the agronomic rule base to screen the target growth stage;

[0115] Factor identification module: Used to perform superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and identify the main loss-causing factors and synergistic factors in combination with crop stress resistance data;

[0116] Path optimization module: Used to correct the traceability path according to the contribution weight distribution of the main loss-causing factors, and reversely verify and iteratively update the key event thresholds and contribution weights of the agronomic rule base.

[0117] It should be noted that the present invention can be deployed on the device itself to achieve embedded applications, or can also run on a PC or other terminals with a user interface, so as to meet various hardware environments and usage requirements.

[0118] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or a data center that includes one or more collections of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0119] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0120] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division, and there can be other division methods in actual implementation. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings, direct couplings, or communication connections shown or discussed with each other can be through some interfaces. The indirect couplings or communication connections of the devices or modules can be in electrical, mechanical, or other forms.

[0121] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0122] In addition, in each embodiment of the present application, each functional module can be integrated into a processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0123] If the above-mentioned function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0124] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present application, and all of them should be covered by the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

[0125] Finally: The above is only the preferred embodiment of the present invention and is not used to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for tracing the entire agricultural industry chain based on big data, characterized in that, It includes the following steps: S1. Obtain the original data of the entire agricultural industry chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters; S2. Divide the crop growth parameters and environmental parameters according to the agronomic growth stages to generate time series data segments for each stage; S3. Dynamically mark the key abnormal events and time windows in the time series data segments based on the key event thresholds in the agronomic rule library; S4. Reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory through phase space reconstruction, and generate a trajectory similarity index based on the reference trajectory predefined in the agronomic rule library to screen the target growth stage; The trajectory similarity index of the dynamic system trajectory is generated by comparing the morphological differences between the current dynamic system trajectory and the reference trajectory predefined in the agronomic rule library; The calculation of the trajectory similarity index integrates the overlapping ratio of the time windows of the key abnormal events. The overlapping ratio of the time windows is the ratio of the overlapping duration of the time range of the current trajectory and the time window of the key abnormal events under the same stage identifier; The screening of the target growth stage is achieved through the comprehensive determination of the difference between the trajectory similarity index and the preset similarity threshold and the overlapping ratio of the time windows. If the trajectory similarity index is lower than the preset similarity threshold and the overlapping ratio of the time windows is higher than the preset overlapping ratio threshold, it is marked as the target growth stage; S5. Conduct a superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and identify the main loss-causing factors and synergistic factors in combination with the crop stress resistance data; The key abnormal events include the event type, time window range, and associated stage identifier; The contribution weights are generated through the linear combination of the trajectory similarity index and the overlapping ratio of the time windows. The weight coefficients of the trajectory similarity index and the overlapping ratio of the time windows are preset according to the statistical analysis of historical loss-causing cases; The crop stress resistance data is extracted from the predefined stress resistance database, and the stress resistance database stores the drought tolerance scores and disease resistance scores of crop varieties; The main loss-causing factors are determined by the product of the contribution weight and the reciprocal of the stress resistance score. If the product value exceeds the preset product threshold, it is marked as the main loss-causing factor; The synergistic factors are identified by calculating the overlapping density of the time windows of the key abnormal events under the same stage identifier. The overlapping density of the time windows is the ratio of the number of overlapping events to the total number of events. When the ratio exceeds the preset density threshold, it is marked as the synergistic factor; The main loss-causing factors and synergistic factors are associated with the key abnormal events, and the output data includes the event type, contribution weight, and associated stage identifier; S6. Modify the traceability path according to the contribution weight distribution of the main loss-causing factors, and reversely verify and iteratively update the key event thresholds and contribution weights of the agronomic rule library.

2. The agricultural whole industry chain traceability method based on big data according to claim 1, characterized in that, The crop growth parameters are collected by Internet of Things sensors deployed in the field, including leaf surface temperature, stem diameter, and photosynthetic rate; the environmental parameters include soil pH value, rainfall, and light intensity, which are synchronously obtained through weather stations and soil probes; the quality inspection parameters include the pesticide residue detection results and microbial indicators in the processing link, and the transportation environment parameters include the temperature and humidity fluctuation data of the cold chain carriage.

3. The method for tracing the entire agricultural industry chain based on big data according to claim 1, characterized in that, The agronomic growth stage is divided into the germination stage, vegetative growth stage, reproductive growth stage, and maturity stage according to the phenological period of the target crop; the crop growth parameters and environmental parameters are divided into stages according to the time intervals corresponding to the phenological periods to generate time-series data segments for each stage. The time-series data segments are extracted by the sliding time window algorithm, and the window length of the sliding time window is dynamically adjusted according to the biological characteristics of each growth stage.

4. The agricultural whole-industry-chain traceability method based on big data according to claim 1, wherein, The key event thresholds stored in the agronomic rule base include the extreme high temperature event threshold, the pest and disease outbreak threshold, and the irrigation shortage threshold, and the key event thresholds are generated based on the historical disaster record statistical model. The dynamic marking of key abnormal events in the time-series data segments is achieved by comparing the real-time data stream with the key event thresholds. When the crop growth parameters or environmental parameters continuously exceed the key event thresholds for a preset duration, they are marked as key abnormal events. The time window of the key abnormal event is recorded as the start time and the end time; each key abnormal event is bound to the corresponding stage identifier.

5. The method for tracing the entire agricultural industry chain based on big data according to claim 1, wherein The phase space reconstruction method uses the delayed coordinate method, and the embedding dimension and time delay are determined according to the time-series characteristics of the time-series data segments. The time-series characteristics include the data sampling frequency and the parameter fluctuation period.

6. The method for tracing the entire agricultural industry chain based on big data according to claim 1, characterized in that, The traceability path correction is achieved by adjusting the priority order of the stage identifiers corresponding to the main loss-causing factors in the traceability path. The priority order is arranged from high to low according to the contribution weight. The reverse verification is completed by comparing the consistency of the damaged stages between the corrected traceability path and the historical real damage cases. The iterative update of the key event thresholds in the agronomic rule base is based on the reverse verification results. If the consistency ratio is lower than the preset verification threshold, the thresholds are recalculated according to the key abnormal event parameters in the actual damaged stages of the historical cases. The iterative update of the contribution weight is achieved by statistically analyzing the distribution frequency of the main loss-causing factors in each growth stage in the corrected traceability path, and the weights of the factors with a distribution frequency higher than the preset frequency threshold are increased proportionally.

7. The method for tracing the entire agricultural industry chain based on big data according to claim 6, characterized in that, The consistency judgment criterion is the coincidence ratio of the first N stage identifiers in the corrected path and the actual damaged stage in the historical case. The value of N is determined according to the average number of damaged stages in the historical case.

8. A big-data-based agricultural whole-industry-chain traceability system for implementing the big-data-based agricultural whole-industry-chain traceability method according to any one of claims 1-7, characterized in that, It includes the following modules: Data acquisition module: used to acquire the original data of the entire agricultural industry chain, including crop growth parameters, environmental parameters, quality inspection parameters, and transportation environment parameters. Data segmentation module: used to divide the crop growth parameters and environmental parameters according to the agronomic growth stage to generate time-series data segments for each stage. Event marking module: used to dynamically mark the key abnormal events and time windows in the time-series data segments based on the key event thresholds in the agronomic rule base. Trajectory analysis module: used to reconstruct the quality inspection parameters and transportation environment parameters into a dynamic system trajectory through phase space, and generate a trajectory similarity index based on the reference trajectory predefined in the agronomic rule base to screen the target growth stage. Factor identification module: used to perform superposition effect analysis on the key abnormal events in the target growth stage to generate the contribution weights of each key abnormal event, and identify the main loss-causing factors and synergistic factors in combination with the crop stress resistance data. Path optimization module: used to correct the traceability path according to the contribution weight distribution of the main loss-causing factors, and to reversely verify and iteratively update the key event thresholds and contribution weights of the agronomic rule base.

Images