CN-122022607-A - Edible agricultural product quality safety assessment method and system based on localization data integration

CN122022607ACN 122022607 ACN122022607 ACN 122022607ACN-122022607-A

Abstract

The invention provides a food agricultural product quality safety evaluation method and system based on localized data integration, wherein the method comprises the steps of constructing a space-time-ecological-social three-dimensional dynamic evaluation model; the method comprises the steps of carrying out dynamic weight distribution and data fusion on each axis to generate a comprehensive evaluation index and a risk correction coefficient, formulating a differential supervision strategy and an optimal resource distribution scheme according to the index and the coefficient, carrying out dynamic early warning on the quality and safety risk of agricultural products based on the scheme, and outputting a visual two-dimensional comprehensive evaluation report. The invention adopts a weighted geometric mean method to generate a comprehensive evaluation index, combines the correlation coefficient to construct a risk correction coefficient, and divides a quality safety grade boundary through fuzzy clustering analysis to enable an evaluation result to be highly matched with an actual risk. Finally, a full-chain closed loop is formed by combining a supervision strategy library and a resource allocation scheme with a dynamic early warning emergency linkage system, so that the timeliness, accuracy and resource utilization efficiency of agricultural product quality safety assessment are improved, and the actual risk deviation is reduced.

Inventors

LUO YAN
LIAO MIN
CHEN ZICHENG
Nie chan
Liang Hanzhe
ZHANG LING

Assignees

深圳市农产品质量安全检验检测中心(深圳市动植物疫病预防控制中心)

Dates

Publication Date: 20260512
Application Date: 20260416

Claims (8)

1. A method for evaluating the quality and safety of edible agricultural products based on localized data integration, the method comprising: Acquiring space-time quality data, ecological toughness data and social co-treatment data to form a multi-source heterogeneous data set; constructing a space-time-ecological-social three-dimensional dynamic evaluation model based on the multi-source heterogeneous data set; based on the three-dimensional dynamic evaluation model, dynamic weight distribution and data fusion are carried out on a time quality axis, an ecological toughness axis and a social co-treatment axis, and a comprehensive evaluation index and a risk correction coefficient are generated; Based on the comprehensive evaluation index and the risk correction coefficient, a differential supervision strategy and an optimal resource allocation scheme are formulated; and carrying out dynamic early warning on the quality safety risk of the agricultural products based on the differential supervision strategy and the optimal resource allocation scheme, and outputting a visual two-dimensional comprehensive evaluation report, wherein the two dimensions comprise a quality safety grade dimension and a risk early warning dimension.
2. The method for evaluating the quality and safety of edible agricultural products based on localized data integration according to claim 1, wherein constructing a space-time-ecological-social three-dimensional dynamic evaluation model comprises: based on a multi-source heterogeneous data set, constructing a three-dimensional initial framework comprising a space-time quality axis, an ecological toughness axis and a social co-treatment axis by using a space-time entropy calculation model, an ecological toughness accounting model and a blockchain intelligent contract verification model, and generating a basic model framework comprising space-time distribution characteristics of a circulation path, an ecological system service value quantization index and a social co-treatment compliance record; Based on the basic model framework, carrying out normalization processing on the space-time entropy value, the ecological service value equivalent and the intelligent contract execution rate by using a Min-Max normalization algorithm, and generating a data fusion model by adopting a multi-source data fusion algorithm fused by principal component analysis and a analytic hierarchy process; Dynamically calculating fluctuation weights of data in each dimension by utilizing an entropy weight method based on the data fusion model, and adjusting the weight distribution proportion of seasonal pollution risk coefficients and ecological toughness thresholds by combining a variation coefficient method to generate a dynamic weight distribution scheme of a space-time quality axis, an ecological toughness axis and a social co-treatment axis; Based on the dynamic weight distribution scheme, the model prediction accuracy is verified by using a K-fold cross verification method, sensitivity of each dimension index is analyzed through Monte Carlo simulation, and the weight coefficient and the threshold parameter are subjected to iterative optimization by adopting a genetic algorithm, so that a space-time-ecology-society three-dimensional dynamic evaluation model which is verified and optimized and has self-adaptive adjustment capability is generated.
3. The method for evaluating the quality and safety of edible agricultural products based on localized data integration according to claim 2, wherein the expression for dynamically calculating the fluctuation weight of each dimension data by using the entropy weight method is as follows: Wherein, the Represent the first The dynamic weight of the dimension(s), Represent the first The information entropy of the dimension(s), Represent the first The coefficient of variation in the dimensions of the dimensions, Representing the risk factor of seasonal contamination, Representing the total number of samples, Representing the normalized post-th Sample at the first The ratio of the dimensions, Represent the first The first sample is at The normalized data value of the dimension is used, Represent the first The standard deviation of the dimensional data is used, Represent the first The mean of the dimensional data.
4. The method for evaluating the quality safety of edible agricultural products based on localized data integration of claim 1, wherein generating the comprehensive evaluation index and the risk correction coefficient comprises: Based on the dynamic weight distribution scheme of the three-dimensional dynamic evaluation model, a weight self-adaptive adjustment mechanism is triggered by utilizing real-time data flow of the Internet of things, and rolling calculation is carried out on the data of the latest N time periods by combining a sliding window algorithm, so that a real-time weight calibration result which is subjected to optimization feedback adjustment by a genetic algorithm is generated; Based on the real-time weight calibration result, constructing a three-dimensional data fusion function by using a weighted geometric mean method, calculating a normalized data fusion value of a space-time entropy value, an ecological service value equivalent and an intelligent contract execution rate, and generating a comprehensive evaluation index containing a synergistic effect; Based on the comprehensive evaluation index, a risk correction coefficient calculation formula is constructed by utilizing seasonal pollution risk coefficients and ecological toughness thresholds, confidence intervals are verified through a Bootstrap sampling method, quality safety class boundaries are divided by combining fuzzy cluster analysis, and the comprehensive evaluation index and the risk correction coefficients are generated.
5. The method for evaluating the quality safety of edible agricultural products based on localized data integration according to claim 4, wherein the risk correction factor calculation formula is expressed as: Wherein, the Representing the risk correction factor(s), Representing the risk magnification factor(s), Representing the risk factor of seasonal contamination, Threshold of ecological toughness.
6. The method for evaluating the quality and safety of edible agricultural products based on localized data integration according to claim 1, wherein the formulating of the differential supervision policy and the optimal resource allocation scheme comprises: based on the comprehensive evaluation index and the risk correction coefficient, constructing a two-dimensional evaluation matrix by using a K-means clustering algorithm, dynamically dividing high/medium/low three-level quality safety grades by combining a sliding window mechanism, and generating a grade division result matched with a real-time risk level; Based on the grading result, constructing a differential strategy library of space-time quality-ecological toughness-social co-treatment dimension by utilizing a three-dimensional supervision strategy matrix, and recommending an optimal strategy combination by combining an expert system and a random forest algorithm to generate a three-dimensional supervision strategy scheme containing specific measures; based on the three-dimensional supervision strategy scheme, a resource allocation dynamic optimization model is constructed by utilizing a multi-objective optimization function, and the optimal allocation proportion of detection equipment, human resources and fund budgets is determined by combining a genetic algorithm and GIS thermodynamic diagram analysis, so that a resource allocation scheme with supervision cost, risk coverage and resource utilization rate collaborative optimization is generated; Based on the resource allocation scheme, a dynamic early warning emergency linkage system is constructed by utilizing a three-level early warning response mechanism, and a corresponding level emergency response flow is automatically triggered by combining real-time monitoring index data to generate a linkage execution scheme containing primary/secondary/three-level early warning response; Based on the linkage execution scheme, a supervision strategy closed-loop verification system is constructed by utilizing digital twin and Monte Carlo simulation, and the differential strategy library parameters are updated periodically through a virtual supervision sandbox simulation pressure test to generate a differential supervision strategy and an optimal resource allocation scheme.
7. The localized data integration-based food product quality safety assessment method of claim 6, wherein generating a three-dimensional regulatory policy scheme comprising specific measures comprises: Based on the high/medium/low three-level quality safety grade division result, constructing a differential strategy library by utilizing a space-time quality-ecological toughness-social co-treatment three-dimensional matrix, and generating a basic strategy combination containing specific measures; based on the basic strategy combination, matching agricultural regulations and supervision policies by using an expert system rule engine, and dynamically adjusting strategy parameters by combining seasonal pollution risk coefficients and ecological toughness thresholds to generate a strategy rule set conforming to industry standards; Based on the strategy rule set, constructing a strategy effect prediction model by utilizing a random forest algorithm, and optimizing a double objective function of risk coverage and resource utilization rate by combining a genetic algorithm to generate a candidate strategy combination; Based on the candidate strategy combination, a digital twin sand box is utilized to simulate pressure test, strategy robustness is verified through Monte Carlo simulation, a particle swarm optimization algorithm is triggered by combining deviation analysis to adjust strategy parameters, and a three-dimensional supervision strategy scheme containing specific measures is generated.
8. A food produce quality safety assessment system based on localized data integration, the system comprising: A processor; A memory for storing processor-executable instructions; Wherein the processor is configured to implement the agricultural product quality safety two-dimensional comprehensive assessment method of any one of claims 1 to 7 when executing the executable instructions.

Description

Edible agricultural product quality safety assessment method and system based on localization data integration Technical Field The invention relates to the technical field of agricultural product quality safety assessment, in particular to a food agricultural product quality safety assessment method and system based on localized data integration. Background Current quality safety assessment of edible agricultural products relies primarily on single dimensional data or static models. The traditional method mainly adopts laboratory detection data (such as pesticide residue and heavy metal content), circulation link spot check records and enterprise self-check reports, and generates a safety score through fixed weight weighted summation. For example, a linear model of 'qualification rate+sampling frequency' is adopted in a part of areas, or based on a static risk thermodynamic diagram of a GIS map, a threshold value is set by combining expert experience to conduct grading. The principle focuses on a passive mode of 'post detection-standard reaching judgment', the data source is mainly a structured database, and the space-time dynamic characteristic capturing capability is lacked. The ecological toughness assessment mostly adopts static indexes of agricultural ecosystem service value accounting, such as soil conservation quantity, carbon sink capacity and the like, and the fusion application of social co-treatment data (such as enterprise credit records and blockchain traceability information) is still in a test point stage, a cross-dimension collaborative analysis mechanism is not formed, and the space-time dynamic change of an agricultural product circulation path (such as seasonal transportation pollution and regional climate fluctuation influence) is difficult to reflect, so that the assessment result has deviation from the actual risk. Disclosure of Invention The invention aims at least solving the technical problems of the prior art, and particularly creatively provides a food agricultural product quality safety assessment method and system based on localized data integration. In order to achieve the above object of the present invention, the present invention provides a method for evaluating the quality safety of edible agricultural products based on localized data integration, the method comprising: s1, acquiring space-time quality data, ecological toughness data and social co-treatment data to form a multi-source heterogeneous data set; S2, constructing a space-time-ecological-social three-dimensional dynamic evaluation model based on the multi-source heterogeneous data set; s3, based on the three-dimensional dynamic evaluation model, dynamic weight distribution and data fusion are carried out on a time quality axis, an ecological toughness axis and a social co-treatment axis, and a comprehensive evaluation index and a risk correction coefficient are generated; S4, based on the comprehensive evaluation index and the risk correction coefficient, formulating a differential supervision strategy and an optimal resource allocation scheme; s5, carrying out dynamic early warning on the quality safety risk of the agricultural products based on the differential supervision strategy and the optimal resource allocation scheme, and outputting a visualized two-dimensional comprehensive evaluation report, wherein the two dimensions comprise a quality safety grade dimension and a risk early warning dimension. In another aspect, the present invention also proposes a food agricultural product quality safety assessment system based on localized data integration, the system comprising: A processor; A memory for storing processor-executable instructions; The processor is configured to implement the agricultural product quality safety two-dimensional comprehensive evaluation method when executing the executable instructions. The method has the beneficial effects that the method forms a multi-source heterogeneous data set by integrating space-time quality data (including space-time entropy values of agricultural product circulation paths and seasonal pollution risk coefficients), ecological toughness data (quantized based on an agricultural ecological system service value accounting model) and social co-treatment data (including intelligent contract automatic execution records), constructs a space-time ecological-social three-dimensional dynamic evaluation model, and effectively solves the problems of single data source and delayed updating in the prior art. The weight of each dimension is dynamically distributed through an entropy weight method and a variation coefficient method, a weight self-adaptive adjustment mechanism is triggered by the data flow of the Internet of things in combination with rolling calculation of a sliding window algorithm, the characteristics of idle dynamic change such as seasonal transportation pollution, regional climate fluctuation and the like are reflected in real time, and evaluation deviati