CN-122022693-A - Dynamic monitoring and intelligent grading pricing system and method for post-harvest quality of agricultural products

Abstract

The application provides a system and a method for dynamically monitoring the quality of an agricultural product after harvest and intelligently grading and pricing, comprising the steps of extracting a high-intensity conduction area from an influence coefficient matrix, judging the quality degradation risk level of the agricultural product, obtaining a risk distribution map, using a classification algorithm to evaluate the current quality state of the agricultural product according to the risk distribution map, determining grading categories for distinguishing agricultural product batches with different maturity and freshness degree, updating a storage layout optimization model by combining ethylene conduction path data after the grading categories are obtained, obtaining optimized coexistence configuration, adjusting a shelf placement strategy to reduce ethylene cross contamination, iteratively updating a local environment distribution diagram through a prevention adjustment scheme, quantifying the efficiency improvement effect of an overall supply chain, and determining final grading fixed values for guiding storage and logistics decision of the agricultural product.

Inventors

• XU QIONG
• XIONG WEI

Assignees

• 武汉舟文科技有限公司

Dates

Publication Date

20260512

Application Date

20260226

Claims (9)

1. 1. A system and a method for dynamically monitoring and intelligently grading pricing post-harvest quality of agricultural products are characterized in that the method comprises the following steps: The ethylene gas concentration, the ambient temperature and the air humidity data of each shelf position are collected in real time through sensor array equipment arranged in the warehouse, and a local environment distribution map in the warehouse is generated and used for reflecting microclimate conditions of an agricultural product storage area; according to the local environment distribution diagram, analyzing a gas concentration gradient change mode by adopting a neural network algorithm, and determining the starting point position and the propagation direction of an ethylene conduction path, wherein the starting point position and the propagation direction are used for tracking a potential agricultural product maturation influence source; If the starting point of the ethylene conduction path is positioned on a specific agricultural product shelf which is easy to be influenced by ethylene, quantifying the ethylene influence intensity of the path on the adjacent shelf by an integrated learning algorithm to obtain an influence coefficient matrix which is used for representing the interaction degree among the areas; Extracting a high-strength conduction region from the influence coefficient matrix, judging the quality degradation risk level of the agricultural products, and obtaining a risk distribution map for visualizing a potential quality degradation hot spot region in a warehouse; Aiming at the risk distribution mapping, a classification algorithm is adopted to evaluate the current quality state of the agricultural products, and classification categories are determined and used for distinguishing agricultural product batches with different maturity and freshness; After the classification category is obtained, the storage layout optimization model is updated by combining the ethylene conduction path data to obtain an optimized coexistence configuration, and the optimized coexistence configuration is used for adjusting a shelf placement strategy to reduce ethylene cross contamination; Simulating a propagation process of a future ethylene conduction effect in a warehouse according to the optimized coexistence configuration, judging potential linkage risks, and obtaining a preventive adjustment scheme for providing targeted storage environment intervention measures; and iteratively updating a local environment distribution map through the prevention adjustment scheme, quantifying the efficiency improvement effect of the whole supply chain, and determining the final grading fixed value for guiding the storage and logistics decision of agricultural products.
2. 2. The system and method for dynamically monitoring and intelligently grading pricing post-harvest quality of agricultural products according to claim 1, wherein the real-time collection of ethylene gas concentration, ambient temperature and air humidity data for each shelf location by sensor array devices deployed in the warehouse generates a local environmental profile within the warehouse for reflecting microclimate conditions of the storage area of the agricultural products, comprising: Acquiring initial data of ethylene concentration, temperature and humidity of each shelf position acquired by a sensor array; obtaining a concentration sequence, a temperature sequence and a humidity sequence of each position synchronization by a time sequence alignment method; calculating a three-dimensional environment state vector of each shelf position at the current moment according to the concentration sequence, the temperature sequence and the humidity sequence; Generating a continuous ethylene concentration distribution field, a continuous temperature distribution field and a continuous humidity distribution field in a warehouse based on three-dimensional environment state vectors of all shelf positions by adopting a Kriging interpolation method; extracting the position coordinates of the region with the current concentration higher than a preset threshold value from the ethylene concentration distribution field to obtain a high-concentration region set; Carrying out space superposition analysis on the high-concentration region set, the temperature distribution field and the humidity distribution field, and judging whether a subarea with higher temperature and lower humidity exists in the high-concentration region; if a subarea with higher temperature and lower humidity exists in the high-concentration area, marking the subarea as a key attention area and recording the space range of the subarea; extracting a concentration change sequence of a corresponding position from an ethylene concentration distribution field according to the spatial range of a key attention area, and calculating the rising rate of the sequence in a last period of time; if the rising rate exceeds a preset rate threshold, determining that the important attention area is in an ethylene acceleration accumulation state; extracting a temperature change sequence and a humidity change sequence of corresponding positions from a temperature distribution field and a humidity distribution field aiming at a key attention area in an ethylene acceleration accumulation state; Judging whether the current microclimate condition further promotes ethylene release according to the temperature change sequence and the humidity change sequence; if the judgment result is that release is promoted, local environment abnormality prompting information comprising the position of the important attention area, the ethylene concentration value, the temperature value, the humidity value and the acceleration accumulation state is generated.
3. 3. The system and method for dynamic monitoring and intelligent grading pricing of post-harvest quality of agricultural products according to claim 1, wherein the analyzing the gas concentration gradient change pattern based on the local environmental profile using a neural network algorithm to determine a starting point location and propagation direction of an ethylene conduction path for tracking potential sources of agricultural product maturation effects comprises: acquiring the position of each monitoring point in the local environmental distribution diagram and the corresponding ethylene gas concentration data; calculating concentration difference values between adjacent points by adopting a neural network algorithm according to the positions of all monitoring points and ethylene gas concentration data to obtain a gas concentration gradient field; Determining the opposite direction of the local maximum gradient of each point through the direction and the size of the gradient vector in the gas concentration gradient field, and obtaining the initial ethylene conduction path direction; Selecting a region with highest gradient vector continuity from the primary ethylene conduction path direction to obtain a candidate ethylene conduction path set; for each path in the candidate ethylene conduction path set, backtracking from the tail end along the opposite gradient direction point by point, judging whether the concentration is continuously increased, if the concentration is continuously increased, reserving the path, and if the concentration is reduced, discarding the path to obtain a carefully selected ethylene conduction path; Searching a starting end point of concentration gradient vector convergence in the carefully selected ethylene conduction path, and determining the position of the end point as the starting position of the ethylene conduction path; and taking the initial position of the ethylene conduction path as a source point, and extending along the direction of the carefully selected ethylene conduction path to obtain the ethylene propagation direction and the potential agricultural product maturation influence source position.
4. 4. The system and method for dynamically monitoring and intelligently grading pricing post-harvest quality of agricultural products according to claim 1, wherein if the starting point of the ethylene conduction path is located on a shelf for storing a specific agricultural product susceptible to ethylene, quantifying the ethylene impact strength of the path on the adjacent shelf by an ensemble learning algorithm to obtain an impact coefficient matrix for representing the interaction degree between areas, comprising: acquiring real-time ethylene concentration data of each shelf in the shelf layout; judging whether an ethylene conduction path exists or not through a preset concentration difference value, if the concentration difference of the adjacent goods shelves exceeds a preset threshold value, determining that the path exists and recording the position of the initial goods shelf; extracting a downstream adjacent shelf sequence of the initial shelf of the determined ethylene conduction path to form a path-associated shelf group; training the ethylene concentration sequence of the path-associated shelf group by adopting an integrated learning algorithm to obtain the influence coefficient of each path on the adjacent shelf; constructing a coefficient matrix through influence coefficients, wherein matrix elements represent the ethylene action strength of the initial shelf on the target shelf; determining the interaction relation strength distribution among the goods shelf areas according to the positions of non-zero elements in the coefficient matrix; And (3) performing threshold screening on the coefficient matrix, and reserving the interaction relation with the action intensity higher than a preset threshold value to form a final region interaction intensity characterization result.
5. 5. The system and method for dynamic monitoring and intelligent grading pricing of post-harvest quality of agricultural products according to claim 1, wherein the extracting high-intensity conduction regions from the influence coefficient matrix, judging the risk level of deterioration of the quality of agricultural products, and obtaining a risk distribution map for visualizing potential quality degradation hot spot regions in a warehouse comprises: Acquiring a high-strength conduction region from the influence coefficient matrix; judging the region with the conduction intensity larger than a preset threshold value according to the conduction intensity values of all the positions in the high-intensity conduction region to obtain a high-risk conduction sub-region; aiming at the high risk conduction subregion, spatial clustering is carried out on the subregion position by adopting a Kmeans clustering method, so as to obtain a quality degradation aggregation cluster; Ordering the risk conduction intensity of each position in the quality degradation cluster, and determining the risk level of the cluster to obtain a high risk cluster and a medium risk cluster; generating corresponding risk score grids according to the space coordinates of the high risk cluster and the medium risk cluster; Dividing a risk level interval according to the score of each grid point in the risk score grid to obtain a risk distribution matrix; and mapping the risk distribution matrix to a warehouse space layout coordinate system to obtain the visual distribution of the quality-reduced hot spot areas in the warehouse.
6. 6. The system and method for dynamically monitoring and intelligently grading pricing for post-harvest quality of agricultural products according to claim 1, wherein the step of evaluating current quality status of agricultural products for the risk distribution map using a classification algorithm to determine grading categories for differentiating batches of agricultural products with different maturity and freshness levels comprises: Acquiring related acquisition data of the agricultural product batch from a storage system, and starting preliminary analysis on the quality state to obtain initial distribution data; according to the initial distribution data, adopting a classification algorithm to extract characteristics of the maturity and freshness of the agricultural product batches, and determining the state evaluation result of each batch; aiming at the state evaluation result, classifying a certain agricultural product batch into a category to be processed if the maturity of the agricultural product batch is lower than a preset threshold value, and acquiring detailed distribution data of the category to be processed; The difference of each batch to be processed in the fresh-keeping degree is judged by carrying out secondary analysis on the detailed distribution data, so that a specific class division basis is obtained; according to the classification basis, if the freshness of a certain batch to be processed accords with a preset standard, adjusting the batch to be processed to a normal class, and determining the final batch class; after the final batch type is obtained, corresponding quality state records are generated for the agricultural product batches of the normal type and the to-be-processed type, and classification processing is completed; And storing classification results of each batch by adopting a uniform format through sorting the quality state records, judging whether the storage is complete, and obtaining a final analysis file.
7. 7. The system and method for post-harvest quality dynamic monitoring and intelligent hierarchical pricing of agricultural products of claim 1, wherein after the hierarchical categories are obtained, a storage layout optimization model is updated in conjunction with the ethylene conduction path data to obtain an optimized coexistence configuration for adjusting shelf placement policies to reduce ethylene cross-contamination, comprising: constructing an initial coexistence constraint matrix through the classification category and the ethylene conduction path data; Calculating a conduction intensity sequence between each cargo pair according to the initial coexistence constraint matrix and the path data; Carrying out iterative optimization on the transmission intensity sequence by adopting a genetic algorithm to obtain an optimized cargo coexistence configuration table; extracting a recommended goods placement set of each goods shelf unit from the optimized goods coexistence configuration table; Generating a goods shelf adjustment instruction sequence aiming at the recommended goods set and the current goods shelf occupation state; After the shelf adjustment instruction sequence is executed, updating the actual occupation record of the warehouse layout; and re-acquiring the classification and ethylene conduction path data according to the updated actual occupation record to form closed loop update.
8. 8. The system and method for dynamic monitoring and intelligent grading pricing of post-harvest quality of agricultural products according to claim 1, wherein the simulating the future ethylene conduction effect propagation process in the warehouse according to the optimized coexistence configuration, judging potential linkage risk, and obtaining a preventive adjustment scheme for providing targeted storage environment intervention measures comprises: Acquiring an initial distribution state of ethylene conduction by constructing a data model of a warehouse environment, and determining concentration change trends of the ethylene conduction in different areas; Simulating a propagation path of ethylene in a warehouse environment by adopting a space grid dividing method according to the concentration change trend to obtain a dynamic diffusion range of each region; aiming at the dynamic diffusion range, analyzing trigger points of potential threats and linkage risks by combining layout data of coexistence configuration, and judging the position distribution of a high-risk area; if the position distribution of the high-risk areas exceeds a preset threshold value, calculating the linkage risk probability of each area through a risk assessment module, and determining a key area with preferential intervention; Acquiring storage environment parameters of a key area, and generating a targeted adjustment scheme by combining the results of simulation analysis to obtain an optimized layout configuration; Generating specific intervention measures through the optimized layout configuration, adjusting environmental parameters aiming at weak links of ethylene conduction, and judging a final prevention and control effect; And updating real-time monitoring data of the warehouse environment according to data feedback of the prevention and control effect, and continuously tracking the change of potential threats to obtain a long-term prevention strategy.
9. 9. The system and method for dynamically monitoring and intelligently grading pricing post-harvest quality of agricultural products according to claim 1, wherein the iterative updating of local environmental profiles via the preventive adjustment scheme quantifies overall supply chain efficiency improvement effects, determines final grading values, and is used for guiding agricultural product storage and logistics decisions, comprising: acquiring an initial local environment distribution map; carrying out one-time iterative update on the local environment distribution map through a preventive adjustment scheme to obtain an updated local environment distribution map; calculating the storage suitability score of the agricultural products in each region according to the updated local environment distribution map to obtain storage suitability distribution; Matching calculation is carried out according to the warehouse fitness distribution and the logistics path distance data, so that a full-link efficiency score sequence of the supply chain is obtained; carrying out regression analysis on the full-link efficiency score sequence of the supply chain by adopting a random forest algorithm to obtain an efficiency improvement quantized value; if the efficiency improvement quantized value reaches a preset threshold, the current grading fixed value is reserved, otherwise, the prevention adjustment scheme is continuously executed to update the local environment distribution map in the next iteration; And determining the final grading fixed value and outputting the final grading fixed value to the agricultural product storage position distribution module and the logistics path planning module.

Description

Dynamic monitoring and intelligent grading pricing system and method for post-harvest quality of agricultural products Technical Field The invention relates to the technical field of information, in particular to a system and a method for dynamic monitoring and intelligent grading pricing of post-harvest quality of agricultural products. Background The research field of dynamic monitoring and intelligent grading pricing system and method for the quality of the picked agricultural products relates to the whole fresh-keeping and value realization process from the picking to the selling of the agricultural products such as fresh fruits and vegetables, and the core importance is to directly determine the income and circulation loss of farmers and whether consumers can buy high-quality products. In modern agriculture storage and supply chains, how to enable different agricultural products to coexist in a limited space and keep the best state as far as possible has become a key link for influencing the efficiency of the whole industry. Most current monitoring and pricing schemes are mainly designed for independent storage of single varieties, and practical situations when multiple agricultural products are mixed and placed are rarely considered. This approach ignores the true interactions between different agricultural products, resulting in inconsistent rates of quality change often occurring in actual warehouses. For example, some fruits release ethylene gas, which has a significant effect of accelerating ripening or advancing aging of other surrounding fruits and vegetables, but the existing systems often cannot capture the gradual diffusion process of the effect from one shelf to an adjacent shelf in time, so that systematic deviation between quality evaluation and price formulation occurs. The gas interaction such as ethylene release is only represented by a surface layer, and the technical difficulty of the deeper layer is that the interaction between agricultural products has strong spatial dependence and conduction characteristics. The quality change of the product is gradually transferred to the surroundings through the tiny gradient difference of local temperature, humidity and gas concentration, and the transfer path and strength are obviously changed according to different shelf placement positions, stacking densities and ventilation conditions. The existing monitoring means are difficult to accurately describe the correlation strength which is gradually increased or reduced in space, and also difficult to quantify how much and how far a certain agricultural product has negative conduction effect on adjacent products under a specific combination, so that storage management staff cannot pre-judge linkage risks caused by mixed storage. Therefore, how to identify and quantify the mutual influence relationship formed by the spatial distribution of gas, temperature and humidity among different kinds of agricultural products in real time in the actual scene of mixed storage, especially capture the conduction path and specific strength of adjacent products generated by the deterioration of the quality of a product, become the key problem for realizing accurate quality dynamic monitoring and reasonable intelligent grading pricing. Disclosure of Invention The invention provides a system and a method for dynamic monitoring and intelligent grading pricing of post-harvest quality of agricultural products, which mainly comprise the following steps: The ethylene gas concentration, the ambient temperature and the air humidity data of each shelf position are collected in real time through sensor array equipment arranged in the warehouse, and a local environment distribution map in the warehouse is generated and used for reflecting microclimate conditions of an agricultural product storage area; according to the local environment distribution diagram, analyzing a gas concentration gradient change mode by adopting a neural network algorithm, and determining the starting point position and the propagation direction of an ethylene conduction path, wherein the starting point position and the propagation direction are used for tracking a potential agricultural product maturation influence source; If the starting point of the ethylene conduction path is positioned on a specific agricultural product shelf which is easy to be influenced by ethylene, quantifying the ethylene influence intensity of the path on the adjacent shelf by an integrated learning algorithm to obtain an influence coefficient matrix which is used for representing the interaction degree among the areas; Extracting a high-strength conduction region from the influence coefficient matrix, judging the quality degradation risk level of the agricultural products, and obtaining a risk distribution map for visualizing a potential quality degradation hot spot region in a warehouse; Aiming at the risk distribution mapping, a classification algorithm is adopted to ev