A food supply chain-oriented multi-source data acquisition and processing method and system

Abstract

This invention discloses a method and system for multi-source data acquisition and processing in the bird's nest supply chain, relating to the field of bird's nest supply chain management technology. The system includes a multi-source adaptive acquisition module, a dynamic spatiotemporal grid module, a data conflict arbitration module, a chain fingerprint generation module, an edge federated computing module, a degradation index calculation module, and a minimum responsibility location module. The method achieves reliable data alignment, tamper-proofing, privacy protection, risk prediction, and accurate traceability across the entire bird's nest supply chain through edge acquisition normalization, dynamic spatiotemporal grid alignment, multi-source consistency verification, chain irreversible fingerprint generation, edge federated anomaly detection, degradation index calculation, and minimum responsibility grid location. This invention improves the reliability, traceability efficiency, and security management capabilities of bird's nest supply chain data and is applicable to quality supervision throughout the entire bird's nest industry chain.

Description

Technical Field

[0001] This invention relates to the field of bird's nest supply chain management technology, specifically to a method and system for multi-source data acquisition and processing in the food supply chain. Background Technology

[0002] Bird's nest is a high-end edible agricultural product, characterized by scarce raw materials, susceptibility to environmental degradation, stringent storage and cold chain logistics conditions, a long supply chain, dispersed participants, and high requirements for traceability and quality control. The bird's nest supply chain covers multiple stages, including raw material collection, cleaning and processing, sterilization, storage, cold chain transportation, import customs clearance, and end-user sales, involving various heterogeneous data such as temperature and humidity, geographical location, batch codes, processing parameters, sterilization records, sensor monitoring, video evidence, and traceability codes.

[0003] Existing technologies have many shortcomings in data processing for the bird's nest supply chain: the standards for collecting multi-source data are not uniform, the time sequence and geographical location cannot be aligned, and cross-linked data are difficult to effectively correlate; data is easily tampered with and forged, and the authenticity and integrity of traceability information cannot be guaranteed; centralized data processing poses risks of commercial privacy and leakage of core data, while distributed processing lacks an efficient collaborative mechanism; traditional solutions only achieve post-event recording and monitoring, and cannot predict deterioration of bird's nest quality in advance and implement pre-emptive measures; after an anomaly occurs, the traceability path is long and the location is vague, making it impossible to quickly identify the responsible link, equipment, and corresponding time and space range.

[0004] The aforementioned problems have resulted in weak security control in the bird's nest supply chain, low traceability efficiency, insufficient data credibility, and difficulty in controlling quality loss, failing to meet the high-quality supervision and quality assurance needs of the bird's nest industry. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for multi-source data acquisition and processing in the food supply chain, so as to solve the problems of alignment deviation, easy data tampering, insufficient privacy, post-processing, and inefficient traceability in the existing data acquisition and processing of the bird's nest supply chain.

[0006] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A method for multi-source data collection and processing in the bird's nest supply chain includes: S1: Perform edge acquisition and format normalization processing on multi-source heterogeneous data across the entire bird's nest supply chain to obtain normalized multi-source data; S2: Construct a dynamic spatiotemporal grid based on batch identifiers, time slices, and geofences; map normalized multi-source data to the corresponding dynamic spatiotemporal grid to complete spatiotemporal alignment. S3: Perform consistency checks on normalized multi-source data within the same dynamic spatiotemporal raster, identify conflicting data and perform data processing; S4: Generate a chain-like irreversible feature fingerprint according to the sequence of the entire bird's nest supply chain, and perform integrity and uniqueness verification on the feature fingerprint; S5: Performs anomaly detection on data within a dynamic spatiotemporal grid based on edge federated computing, and outputs anomaly feature vectors; S6: Calculate the food deterioration index based on the environmental parameters within the dynamic spatiotemporal grid, and compare the food deterioration index with the preset deterioration threshold; when the food deterioration index is greater than the preset deterioration threshold, locate the minimum responsible grid corresponding to the abnormality and output the handling instruction. This enables unified data collection, spatiotemporal alignment, reliable verification, privacy-preserving computation, quality prediction, and closed-loop handling of anomalies across the entire bird's nest supply chain, thereby improving data integrity, traceability reliability, and quality control efficiency.

[0007] A further proposed solution involves constructing a dynamic spatiotemporal grid, which includes dividing the logistics and warehousing areas of the bird's nest supply chain into geofence units and dividing them into time slices at fixed time intervals. Using time slices, longitude intervals, and latitude intervals as dimensions, a dynamic spatiotemporal grid that is independent and non-overlapping is constructed to constrain dispersed and heterogeneous data within a unified spatiotemporal unit, eliminate temporal and location misalignments, achieve strong correlation between multi-source data, and improve data alignment accuracy and the reliability of subsequent processing.

[0008] A further proposed solution for conflict data processing includes: employing a three-source voting mechanism to perform consistency judgment on three sets of homogeneous and heterogeneous data collected within the same dynamic spatiotemporal grid; identifying data that meets the consistency conditions as valid data and data that does not meet the consistency conditions as conflict data; and performing marking, isolation, or repair processing on conflict data to automatically remove abnormal data such as sensor drift, misrecording, and packet loss, ensuring that the data entering subsequent processes is authentic and valid, and reducing false detection rates and decision-making biases.

[0009] A further proposed solution for generating chain-like irreversible feature fingerprints includes: using the feature fingerprints of the previous link in the supply chain and the normalized multi-source data of the current link as input, calculating the feature fingerprint of the current link using a one-way hashing method; iteratively generating full-chain feature fingerprints link by link, with the feature fingerprint of the current link depending on the feature fingerprint of the previous link, forming an immutable, unreversible, and fully verifiable traceability identifier, effectively preventing data forgery and cross-selling, and improving the credibility and regulatory efficiency of the bird's nest supply chain.

[0010] A further proposed solution for calculating the food deterioration index includes: determining the food deterioration rate based on ambient temperature and collection time; performing an integral calculation based on the food deterioration rate and collection time to obtain the food deterioration index, accurately quantifying the degree of deterioration in bird's nest quality, enabling early risk prediction, avoiding post-event handling, and reducing bird's nest loss and food safety risks.

[0011] A further proposed solution, edge federated computing, includes: performing anomaly detection model training and inference locally on each collection node; transmitting only anomaly feature vectors to the cloud, without transmitting the original collection data; using regularization to constrain model parameters, improving model generalization ability, and achieving distributed collaborative detection without leaking original commercial data, thus balancing privacy and security, detection accuracy and system generalization ability.

[0012] A further proposed solution for locating the minimum responsibility grid includes: calculating the abnormal time deviation and spatial distance corresponding to each dynamic spatiotemporal grid; determining the dynamic spatiotemporal grid with the smallest sum of time deviation and spatial distance as the minimum responsibility grid, quickly locking the smallest spatiotemporal unit where the abnormality occurred, directly locating the responsible link, equipment and location, and significantly improving the efficiency of source tracing and handling.

[0013] A multi-source data acquisition and processing system for the bird's nest supply chain includes: a multi-source adaptive acquisition module for performing edge acquisition and format normalization processing on multi-source heterogeneous data across the entire bird's nest supply chain to obtain normalized multi-source data; a dynamic spatiotemporal raster module for constructing a dynamic spatiotemporal raster based on batch identifiers, time slices, and geofences, mapping the normalized multi-source data to the corresponding dynamic spatiotemporal raster to achieve spatiotemporal alignment; a data conflict arbitration module for performing consistency verification on normalized multi-source data within the same dynamic spatiotemporal raster, identifying conflicting data, and performing data processing; and a chain fingerprint generation module. The system is used to generate chain-like irreversible feature fingerprints according to the sequence of the entire bird's nest supply chain, and to perform integrity and uniqueness verification on the feature fingerprints; the edge federated computing module is used to perform anomaly detection on the data in the dynamic spatiotemporal grid based on edge federated computing, and output anomaly feature vectors; the deterioration index calculation module is used to calculate the food deterioration index based on the environmental parameters in the dynamic spatiotemporal grid, and to perform numerical comparison between the food deterioration index and the preset deterioration threshold; the minimum responsibility location module is used to locate the minimum responsibility grid corresponding to the anomaly when the food deterioration index is greater than the preset deterioration threshold, and output disposal instructions. With a modular structure, it achieves closed-loop execution of the entire process of data collection, alignment, verification, anti-counterfeiting, detection, prediction, and positioning. The structure is clear, highly scalable, and adaptable to deployment in multiple scenarios of the bird's nest supply chain.

[0014] A further solution involves deploying the edge federated computing module in a trusted execution environment to isolate the model computation process from access permissions to the raw data, thereby further enhancing data security and preventing the model and raw data from being stolen or tampered with, thus meeting the stringent privacy and security requirements of the high-end food supply chain.

[0015] A further proposed solution involves a chain-based fingerprint generation module that generates feature fingerprints that do not support reverse engineering of the original data, ensuring that traceability information is unforgeable and unbreakable, thereby strengthening the anti-counterfeiting and regulatory capabilities of the entire bird's nest supply chain and enhancing the product's brand value.

[0016] The present invention has the following beneficial effects: This invention achieves strong alignment of data across the entire bird's nest supply chain through a dynamic spatiotemporal grid, improving the accuracy of data association; ensures data authenticity and validity through multi-source consistency arbitration; achieves tamper-proof and verifiable traceability information through chain-like irreversible fingerprints; completes anomaly detection while protecting data privacy through edge federated computing; enables proactive handling of bird's nest quality risks through deterioration index prediction; and quickly locates the source of anomalies through a minimum responsibility grid, thereby comprehensively improving the data credibility, traceability efficiency, security control capabilities, and quality assurance level of the bird's nest supply chain. Attached Figure Description

[0017] Figure 1 The diagram illustrates the specific steps of the method of this invention. Figure 2 This is a logic block diagram of the system of the present invention; Figure 3 This is a schematic diagram of the generation of chain-like irreversible feature fingerprints. Detailed Implementation

[0018] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0019] refer to Figures 1-3 As shown, a multi-source data acquisition and processing method for the food supply chain achieves reliable, secure, traceable, and predictable data across the entire supply chain through standardized acquisition of multi-source data, dynamic spatiotemporal grid alignment, multi-source data consistency arbitration, chain-like irreversible feature fingerprint generation, edge federation anomaly detection, bird's nest deterioration index calculation, and least-responsibility grid positioning.

[0020] Multi-source data acquisition and format normalization: Data acquisition terminals are deployed at every stage of the bird's nest supply chain, including raw material origin, cleaning and processing, sterilization, warehousing, cold chain logistics, import customs clearance, and terminal sales. These terminals collect multi-source heterogeneous data in real time, including temperature and humidity, geographical location, batch codes, processing parameters, sterilization records, equipment status, video evidence, and traceability codes. The acquisition process is completed at the edge. The system performs unified field mapping, unit conversion, timestamp calibration, and encoding standardization on data from different sources and in different formats, resulting in normalized multi-source data with consistent structure and time sequence.

[0021] The normalization process uses a linear mapping rule to standardize the numerical values. The mapping relationship is as follows: In the formula: This represents the normalized data; This represents the original collected data; This indicates the preset minimum value for this type of data; This indicates the preset maximum value for this type of data.

[0022] The data acquisition terminal supports offline caching and automatically resumes interrupted transmission after the network is restored, ensuring that data is not lost or interrupted in weak network or mobile scenarios.

[0023] Dynamic spatiotemporal grid construction and data alignment: Using the bird's nest batch identifier as a unique index, the entire process timeline is divided into continuous time slices at fixed time intervals; at the same time, the processing workshop, storage area, and cold chain transportation route are divided into independent geofence units. Using time slices, longitude intervals, and latitude intervals as dimensions, a dynamic spatiotemporal grid that is independent and has no overlapping coverage is constructed.

[0024] The construction of a dynamic spatiotemporal raster satisfies the following set relationship: In the formula: Represents the complete set of dynamic spatiotemporal raster; Represents a single spatiotemporal grid cell; Indicates the data collection time; Indicates longitude; Indicates latitude; They represent the first The start and end times of the time slice; They represent the first The start and end boundaries of a longitude interval; They represent the first The start and end boundaries of a latitudinal interval.

[0025] The system maps all normalized data within the same batch, time slice, and geofence to the corresponding dynamic spatiotemporal grid, thus constraining the originally scattered, asynchronous, and multi-source data into a unified spatiotemporal unit, achieving forced spatiotemporal alignment, and eliminating problems such as data time sequence disorder, geographical location mismatch, and inability to correlate across stages.

[0026] Multi-source data consistency verification and conflict handling: For multi-source data falling within the same dynamic spatiotemporal raster, three sets of homogeneous but heterogeneous collected data are selected for consistency judgment. Data that meets the preset consistency conditions is determined as valid data, and data that does not meet the consistency conditions is determined as conflicting data.

[0027] The consensus determination adopts a three-source voting mechanism, and the voting rules are as follows: In the formula: The value indicates the voting result: 1 for consistent data and 0 for conflicting data. , , This represents three sets of homogeneous but heterogeneous data collected within the same grid cell; This indicates the preset data error threshold.

[0028] The system automatically marks and isolates conflicting data to prevent abnormal data from entering subsequent processing. For repairable conflicting data, it automatically repairs the data based on the trend and numerical characteristics of valid data, restoring the data to a reasonable and reliable range.

[0029] Chain-based irreversible fingerprint generation and verification: Fingerprints are generated sequentially from the beginning to the end of the bird's nest supply chain. Using the fingerprint generated in the previous stage and the normalized data of the current stage as input, a one-way calculation method is employed to generate the fingerprint of the current stage.

[0030] The iterative generation formula for chain fingerprints is as follows: In the formula: Indicates the first Characteristic fingerprints of the process; Indicates the first Characteristic fingerprints of the process; Indicates the first Normalized data for each stage; This indicates a one-way hash operation.

[0031] The entire fingerprint is generated iteratively step by step, with the fingerprint of the current step depending on the fingerprint result of the previous step, forming a chain structure that is irreversible, cannot be modified independently, and cannot be rolled back. After the fingerprint is generated at each step, the system immediately performs fingerprint integrity and uniqueness verification. Any data tampering, missing, replacement, or supplementation will result in verification failure.

[0032] Edge federated computing and anomaly detection: The anomaly detection model inference computation is completed locally at each acquisition node, without transmitting the raw acquisition data externally, only outputting anomaly feature vectors to upstream nodes or the cloud. The system uses a loss function with regularization constraints to optimize the model, as follows: In the formula: This represents the model loss value; Indicates the number of local samples; Indicates the true label of the sample; Indicates the model's predicted label; Represents the model parameter vector; Represents the regularization coefficient; This represents the L2 norm of the model parameters.

[0033] By improving the model's generalization ability and detection stability through the above constraints, detection of abnormal temperature, trajectory deviation, missing data, broken fingerprints, and abnormal processing parameters can be completed without aggregating the original data.

[0034] Bird's Nest Deterioration Index Calculation and Threshold Comparison: The system determines the bird's nest deterioration rate based on the ambient temperature and data acquisition duration within a dynamic spatiotemporal grid. Based on the deterioration rate and acquisition duration, an integral calculation is performed to obtain the deterioration index, which characterizes the degree of quality decline in the bird's nest. In the formula: This indicates the deterioration index of bird's nest; , These represent the start and end times of the data collection, respectively. express Ambient temperature at all times; This represents the temperature-dependent degradation rate coefficient.

[0035] The system compares the calculated deterioration index with a preset deterioration threshold to determine whether the current batch of bird's nest is at risk of quality deterioration. When the deterioration index does not exceed the preset deterioration threshold, the system maintains normal monitoring; when the deterioration index exceeds the preset deterioration threshold, the system immediately triggers a risk warning and initiates the anomaly location process.

[0036] Output of minimum responsibility grid location and handling instructions: When the degradation index is greater than the preset degradation threshold, the system traverses all dynamic spatiotemporal grids, calculates the abnormal time deviation and spatial distance for each grid, and determines the dynamic spatiotemporal grid with the smallest sum of time deviation and spatial distance as the minimum responsibility grid. The location formula is as follows: In the formula: G represents the minimum responsibility grid; G represents the complete set of dynamic spatiotemporal grids. This indicates the time offset corresponding to grid g; Represents grid The corresponding spatial distance.

[0037] Based on the scenario and anomaly type corresponding to the minimum responsibility grid, the system outputs handling instructions such as temperature control adjustment, route optimization, batch isolation, early warning reporting, and source tracing verification, enabling rapid risk response and closed-loop handling.

[0038] The overall working principle of this invention is as follows: Using batch identification as the main thread, edge acquisition and normalization processing are performed on multi-source heterogeneous data across the entire bird's nest supply chain. A dynamic spatiotemporal grid constrains dispersed data to a unified spatiotemporal unit, achieving strong data alignment. A three-source voting mechanism is used to complete data consistency verification, ensuring data authenticity and reliability. A chain-like irreversible feature fingerprint is iteratively generated along the supply chain links to achieve tamper-proof and verifiable traceability information. Edge federated computing is used to complete anomaly detection locally, achieving distributed collaborative monitoring while protecting data privacy. A bird's nest deterioration index is calculated based on temperature and duration integration, enabling early prediction of quality risks. When a risk is triggered, the source of the anomaly is quickly located using a minimum responsibility grid, and closed-loop handling instructions are output. The entire process forms a complete closed loop of acquisition, alignment, verification, anti-counterfeiting, detection, prediction, location, and handling, collaboratively ensuring data reliability, quality control, traceability, and risk prevention in the bird's nest supply chain from the data layer, algorithm layer, and application layer.

[0039] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for multi-source data acquisition and processing in the food supply chain, characterized in that, Includes the following steps: S1: Perform edge acquisition and format normalization processing on multi-source heterogeneous data across the entire bird's nest supply chain to obtain normalized multi-source data; S2: Construct a dynamic spatiotemporal grid based on batch identifiers, time slices, and geofences; map normalized multi-source data to the corresponding dynamic spatiotemporal grid to complete spatiotemporal alignment. S3: Perform consistency checks on normalized multi-source data within the same dynamic spatiotemporal raster, identify conflicting data and perform data processing; S4: Generate a chain-like irreversible feature fingerprint according to the sequence of the entire bird's nest supply chain, and perform integrity and uniqueness verification on the feature fingerprint; S5: Performs anomaly detection on data within a dynamic spatiotemporal grid based on edge federated computing, and outputs anomaly feature vectors; S6: Calculate the food deterioration index based on the environmental parameters within the dynamic spatiotemporal grid, and compare the food deterioration index with the preset deterioration threshold. When the food deterioration index is greater than the preset deterioration threshold, locate the minimum responsible grid corresponding to the abnormality and output the handling instruction.

2. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, The construction of a dynamic spatiotemporal grid includes: dividing the logistics and warehousing areas of the bird's nest supply chain into geofence units and dividing them into time slices according to fixed time intervals; and constructing independent and non-overlapping dynamic spatiotemporal grids using time slices, longitude intervals, and latitude intervals as dimensions.

3. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, Conflict data processing includes: using a three-source voting mechanism to perform consistency judgment on three sets of homogeneous and heterogeneous collected data within the same dynamic spatiotemporal grid; determining data that meets the consistency conditions as valid data and data that does not meet the consistency conditions as conflicting data; and performing marking, isolation, or repair processing on conflicting data.

4. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, Generating a chain-like irreversible feature fingerprint involves: taking the feature fingerprint of the previous link in the supply chain and the normalized multi-source data of the current link as input, and using a one-way hashing method to calculate the feature fingerprint of the current link; iteratively generating the full-chain feature fingerprint link by link, with the feature fingerprint of the current link depending on the feature fingerprint of the previous link.

5. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, Calculating the food deterioration index involves: determining the food deterioration rate based on ambient temperature and collection time; and performing an integral calculation based on the food deterioration rate and collection time to obtain the food deterioration index.

6. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, Edge federated computing includes: performing anomaly detection model training and inference locally at each collection node; transmitting only anomaly feature vectors to the cloud, without transmitting the original collection data; and using regularization to constrain model parameters to improve model generalization ability.

7. The method for multi-source data acquisition and processing in the food supply chain according to claim 1, characterized in that, The process of locating the minimum responsibility grid includes: calculating the abnormal time deviation and spatial distance corresponding to each dynamic spatiotemporal grid; and determining the dynamic spatiotemporal grid with the smallest sum of time deviation and spatial distance as the minimum responsibility grid.

8. A multi-source data acquisition and processing system for the bird's nest supply chain, characterized in that, include: The multi-source adaptive acquisition module is used to perform edge acquisition and format normalization processing on multi-source heterogeneous data across the entire bird's nest supply chain to obtain normalized multi-source data. The dynamic spatiotemporal raster module is used to construct dynamic spatiotemporal rasters based on batch identifiers, time slices, and geofences, and to map normalized multi-source data to the corresponding dynamic spatiotemporal rasters to complete spatiotemporal alignment. The data conflict arbitration module is used to perform consistency checks on normalized multi-source data within the same dynamic spatiotemporal raster, identify conflicting data, and perform data processing. The chain fingerprint generation module is used to generate chain-like irreversible feature fingerprints according to the sequence of the entire bird's nest supply chain, and to perform integrity and uniqueness verification on the feature fingerprints. The edge federated computing module is used to perform anomaly detection on data within a dynamic spatiotemporal grid based on edge federated computing and output anomaly feature vectors. The deterioration index calculation module is used to calculate the food deterioration index based on environmental parameters within a dynamic spatiotemporal grid, and to perform a numerical comparison between the food deterioration index and a preset deterioration threshold. The minimum responsibility location module is used to locate the minimum responsibility grid corresponding to the abnormality when the food deterioration index is greater than the preset deterioration threshold, and output the handling instructions.

9. The system according to claim 8, characterized in that, The edge federated computing module is deployed in a trusted execution environment to isolate the model computation process from access permissions to the raw data.

10. The system according to claim 8, characterized in that, The feature fingerprints generated by the chain fingerprint generation module do not support reverse derivation of the original data.

Images