Agricultural digital twin construction method and system based on data fusion

Abstract

The application discloses a kind of agricultural digital twin construction method and system based on data fusion, it is related to agricultural platform technical field.The application constructs farmland three-dimensional grid data model and initializes environment benchmark data;Constructing farmer digital business entity;According to agricultural operation instruction, the farmer skill weight parameter is mapped to the corresponding farmland grid, and the farmland process state data that fuses environment benchmark and operation weight is generated;Respond to the harvesting event of agricultural product, extract the corresponding grid data set from farmland process state data, and instantiate to generate agricultural product digital object;Finally, farmland model, farmer entity and agricultural product object are encapsulated by data inheritance relationship, and agricultural digital twin is output.It realizes the fusion and two-way feedback of farmland, farmer and agricultural product, achieves the whole link accurate traceability from production to product, solves the problem of isolated modeling of each element of agricultural digital twin in the prior art, lack of data correlation and feedback mechanism.

Description

Technical Field

[0001] This invention relates to the field of agricultural platform technology, and in particular to a method and system for constructing agricultural digital twins based on data fusion. Background Technology

[0002] With the rapid development of digital agriculture, agricultural production processes generate a large amount of multi-source data involving farmland environment, farmer operations, and agricultural product quality. In the context of digital rural construction, achieving precise management and traceability across the entire supply chain from production to product has become a key focus for the industry. Existing digital twin technologies mainly focus on the digital modeling of single elements, such as building standalone farmland growth models or farmer information management systems, lacking a complete technical solution that organically integrates farmland, farmers, and agricultural products.

[0003] In related technologies, the construction of agricultural digital twins mainly adopts an independent modeling approach, that is, establishing separate farmland environmental monitoring systems, farmer management databases, and agricultural product traceability platforms. For example, in farmland monitoring, sensors are deployed to collect soil and meteorological data; in farmer management, basic information and agricultural operations are recorded; and in agricultural product traceability, technologies such as QR codes are used to record circulation information. However, these systems lack effective data correlation and feedback mechanisms, resulting in data silos. It is difficult to accurately trace who influenced the land at what level, and it is also impossible to optimize the production process based on the final product quality, thus limiting the accuracy and practicality of the digital twins. Summary of the Invention

[0004] This application provides a data fusion-based method and system for constructing agricultural digital twins, which solves the problems of isolated modeling of various elements and lack of data association and feedback mechanisms in existing agricultural digital twins. It realizes the integration and two-way feedback of farmland, farmers, and agricultural products, and achieves precise traceability of the entire chain from production to product.

[0005] This application provides a method for constructing an agricultural digital twin based on data fusion. This method is applied to an agricultural digital twin construction system based on data fusion, and includes: Construct a three-dimensional raster data model of farmland with spatial coordinate attributes, and initialize the environmental reference data for each raster. Construct a digital business entity for farmers that includes skill weight parameters, which are used to reflect the credibility of farmers' operations; Receive agricultural operation instructions, and based on the spatial scope and executing entity in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights; In response to agricultural product harvesting events, based on the spatial coordinates of the harvested area, a set of raster data for the corresponding time period is extracted from the farmland process state data, and agricultural product digital objects are instantiated and generated. The three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products are encapsulated through data inheritance relationships to output an agricultural digital twin.

[0006] Furthermore, the construction of a three-dimensional raster data model of farmland with spatial coordinate attributes, and the initialization of environmental reference data for each raster, specifically includes: Obtain geographic information data and historical soil monitoring data of the target farmland; Based on geographic information data, the target farmland is logically divided into multiple three-dimensional raster units with independent spatial indexes; Historical soil monitoring data is mapped to each three-dimensional raster cell using an interpolation algorithm, serving as the initial environmental baseline data to form a static attribute layer for farmland.

[0007] Furthermore, the construction of the farmer's digital business entity, which includes skill weight parameters, specifically includes: Establish basic farmer profiles in the database and set initial skill weight parameters; Obtain the farmer's historical planting records and corresponding historical yield and quality data; Based on the deviation rate between historical yield and quality data and the regional average, a correction coefficient is calculated. This correction coefficient is then used to weight the initial skill weight parameters to obtain the current skill weight parameters for the farmer's digital business entity.

[0008] Furthermore, the generation of farmland process state data that integrates environmental benchmarks and operational weights specifically includes: Parse agricultural operation instructions to identify the operation type, operation time, covered grid coordinate set, and the farmer ID that executed the instruction; Retrieve the corresponding skill weight parameters based on the farmer ID index; Multiply the skill weight parameter by the standard influence factor corresponding to the operation type to obtain the operation efficiency value; The operational efficiency value is accumulated and written into the three-dimensional raster cell corresponding to the raster coordinate set, and the value of the three-dimensional raster cell is updated to form farmland process state data that changes dynamically over time.

[0009] Furthermore, the instantiation and generation of agricultural product digital objects specifically includes: Listen for harvest trigger signals and lock onto the target grid set corresponding to the harvest area; Traverse the target raster set and extract the farmland process state data accumulated by each raster cell during the crop growth cycle; The extracted farmland process data are processed by weighted averaging to generate a set of feature data vectors; Create a unique digital object ID for agricultural products and write the feature data vector into the object as an initial attribute to complete the instantiation of the digital object for agricultural products.

[0010] Furthermore, after encapsulating the 3D raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships and outputting an agricultural digital twin, the process also includes: Obtain actual testing data for agricultural products; Calculate the deviation between the test data and the feature data vector of the agricultural product digital object; If the deviation value exceeds the preset threshold, the skill weight parameters of the farmer's digital business entity associated with the agricultural product digital object will be adjusted in the opposite direction according to the deviation direction, and the environmental baseline data in the farmland 3D raster data model will be updated simultaneously.

[0011] Furthermore, the step of accumulating the operational performance value and writing it into the corresponding three-dimensional raster cell of the raster coordinate set, and updating the value of that cell, specifically includes: Read the current state value of the corresponding 3D raster unit before the operation time; The residual value of the current state value over the operation time is calculated using the time decay function; The operational performance value and the residual value are linearly superimposed to obtain the updated raster state value, and the update timestamp is recorded.

[0012] Furthermore, the weighted average processing of the extracted farmland process data to generate a set of feature data vectors specifically includes: Identify the growth stages within the crop growth cycle; Different time importance coefficients are assigned to different growth stages; By using time importance coefficients to perform weighted summation on farmland process data, a feature data vector reflecting the growth quality of crops throughout their entire life cycle is generated.

[0013] Furthermore, the output agricultural digital twin specifically includes: Build a unified interface for the visual rendering engine; Map the three-dimensional raster data model of farmland to the basic geometry of the visualization layer; Map the associated information of farmers' digital business entities to the interactive attributes of the basic geometry; Map the circulation trajectory of digital agricultural product objects as a dynamic layer on a timeline; By combining the above layers through the visualization rendering engine interface, an agricultural digital twin that supports spatiotemporal backtracking queries is output.

[0014] This application provides an agricultural digital twin construction system based on data fusion, which is used to implement an agricultural digital twin construction method based on data fusion, including: a farmland data analysis module, a farmer data analysis module, a farmland-farmer association module, an agricultural product analysis module, and a comprehensive association module; The farmland data analysis module is used to construct a three-dimensional raster data model of farmland with spatial coordinate attributes and initialize the environmental reference data of each raster. The farmer data analysis module is used to construct a digital business entity for farmers that includes skill weight parameters, which reflect the credibility of the farmer's operations. The farmland-farmer association module is used to receive agricultural operation instructions, and based on the spatial range and execution subject in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights. The agricultural product analysis module is used to respond to agricultural product harvesting events, extract raster data sets for the corresponding time period from farmland process state data based on the spatial coordinates of the harvested area, and instantiate agricultural product digital objects. The integrated association module is used to encapsulate the three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships, and output an agricultural digital twin.

[0015] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: This application provides a data fusion-based method for constructing agricultural digital twins. The method involves: constructing a 3D raster data model of farmland with spatial coordinate attributes and initializing environmental baseline data; constructing a digital business entity for farmers containing skill weight parameters; mapping the farmer skill weight parameters to corresponding farmland rasters according to agricultural operation instructions to generate farmland process-state data that integrates environmental baselines and operation weights; responding to agricultural harvesting events by extracting the corresponding raster data set from the farmland process-state data and instantiating agricultural product digital objects; and finally encapsulating the farmland model, farmer entity, and agricultural product objects through data inheritance relationships to output an agricultural digital twin.

[0016] In this process, by dividing farmland into three-dimensional spatiotemporal grids and using them as containers to receive farmer operation events, precise spatial fusion of environmental data and operational behaviors was achieved. Furthermore, during crop harvesting, object instantiation was used to extract accumulated data from corresponding grids, enabling digital objects of agricultural products to automatically inherit environmental and operational information from the production process, achieving end-to-end data connectivity from production to product. Moreover, by comparing actual detection data with the feature values ​​of digital objects, the farmer skill weights and farmland baseline parameters were dynamically adjusted, effectively improving the accuracy and practicality of the digital twin. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a data fusion-based method for constructing an agricultural digital twin, as provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of an agricultural digital twin construction system based on data fusion, provided in an embodiment of this application. Detailed Implementation

[0018] This application provides a data fusion-based method and system for constructing agricultural digital twins, which solves the problems of isolated modeling of various elements and lack of data association and feedback mechanisms in existing agricultural digital twins. By constructing a three-dimensional raster data model of farmland with spatial attributes, a digital entity of farmers containing skill weights, and an event-triggered instantiation mechanism for agricultural product digital objects, the organic integration and two-way feedback of the three major elements of farmland, farmers, and agricultural products are achieved, realizing precise traceability of the entire chain from production to product.

[0019] In related technologies, the construction of agricultural digital twins adopts an independent modeling approach for each element, failing to consider the inherent relationships between these elements in the actual production process. Farmland monitoring systems only record environmental data, farmer management systems only record operational information, and agricultural product traceability systems only record circulation data; there is a lack of effective data interaction mechanisms between these systems. This isolated modeling approach makes it impossible to accurately trace who affected the land at what level, or to reverse-engineer production parameters based on the final product quality, making it difficult for the digital twin to truly reflect the entire agricultural production process.

[0020] Based on the aforementioned technical issues, this application provides a data fusion-based agricultural digital twin construction method. This method achieves the fusion and storage of environmental data and operational behaviors through a spatially containerized three-dimensional raster data model, uses skill weight parameters to calibrate farmers' operational levels, automatically instantiates agricultural product digital objects through event triggering, and establishes a parameter correction mechanism based on feedback from detection data. This effectively solves the problem of data silos among various elements and achieves precise traceability and process optimization across the entire chain.

[0021] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0022] like Figure 1 As shown, this application provides a method for constructing an agricultural digital twin based on data fusion. This method is applied to an agricultural digital twin construction system based on data fusion, and includes: Construct a three-dimensional raster data model of farmland with spatial coordinate attributes, and initialize the environmental reference data for each raster. Construct a digital business entity for farmers that includes skill weight parameters, which are used to reflect the credibility of farmers' operations; Receive agricultural operation instructions, and based on the spatial scope and executing entity in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights; In response to agricultural product harvesting events, based on the spatial coordinates of the harvested area, a set of raster data for the corresponding time period is extracted from the farmland process state data, and agricultural product digital objects are instantiated and generated. The three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products are encapsulated through data inheritance relationships to output an agricultural digital twin.

[0023] Furthermore, the construction of a three-dimensional raster data model of farmland with spatial coordinate attributes, and the initialization of environmental reference data for each raster, specifically includes: Obtain geographic information data and historical soil monitoring data of the target farmland; Based on geographic information data, the target farmland is logically divided into multiple three-dimensional raster units with independent spatial indexes; Historical soil monitoring data is mapped to each three-dimensional raster cell using an interpolation algorithm, serving as the initial environmental baseline data to form a static attribute layer for farmland.

[0024] Furthermore, the construction of the farmer's digital business entity, which includes skill weight parameters, specifically includes: Establish basic farmer profiles in the database and set initial skill weight parameters; Obtain the farmer's historical planting records and corresponding historical yield and quality data; Based on the deviation rate between historical yield and quality data and the regional average, a correction coefficient is calculated. This correction coefficient is then used to weight the initial skill weight parameters to obtain the current skill weight parameters for the farmer's digital business entity.

[0025] Furthermore, the generation of farmland process state data that integrates environmental benchmarks and operational weights specifically includes: Parse agricultural operation instructions to identify the operation type, operation time, covered grid coordinate set, and the farmer ID that executed the instruction; Retrieve the corresponding skill weight parameters based on the farmer ID index; Multiply the skill weight parameter by the standard influence factor corresponding to the operation type to obtain the operation efficiency value; The operational efficiency value is accumulated and written into the three-dimensional raster cell corresponding to the raster coordinate set, and the value of the three-dimensional raster cell is updated to form farmland process state data that changes dynamically over time.

[0026] Furthermore, the instantiation and generation of agricultural product digital objects specifically includes: Listen for harvest trigger signals and lock onto the target grid set corresponding to the harvest area; Traverse the target raster set and extract the farmland process state data accumulated by each raster cell during the crop growth cycle; The extracted farmland process data are processed by weighted averaging to generate a set of feature data vectors; Create a unique digital object ID for agricultural products and write the feature data vector into the object as an initial attribute to complete the instantiation of the digital object for agricultural products.

[0027] Furthermore, after encapsulating the 3D raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships and outputting an agricultural digital twin, the process also includes: Obtain actual testing data for agricultural products; Calculate the deviation between the test data and the feature data vector of the agricultural product digital object; If the deviation value exceeds the preset threshold, the skill weight parameters of the farmer's digital business entity associated with the agricultural product digital object will be adjusted in the opposite direction according to the deviation direction, and the environmental baseline data in the farmland 3D raster data model will be updated simultaneously.

[0028] Furthermore, the step of accumulating the operational performance value and writing it into the corresponding three-dimensional raster cell of the raster coordinate set, and updating the value of that cell, specifically includes: Read the current state value of the corresponding 3D raster unit before the operation time; The residual value of the current state value over the operation time is calculated using the time decay function; The operational performance value and the residual value are linearly superimposed to obtain the updated raster state value, and the update timestamp is recorded.

[0029] Furthermore, the weighted average processing of the extracted farmland process data to generate a set of feature data vectors specifically includes: Identify the growth stages within the crop growth cycle; Different time importance coefficients are assigned to different growth stages; By using time importance coefficients to perform weighted summation on farmland process data, a feature data vector reflecting the growth quality of crops throughout their entire life cycle is generated.

[0030] Furthermore, the output agricultural digital twin specifically includes: Build a unified interface for the visual rendering engine; Map the three-dimensional raster data model of farmland to the basic geometry of the visualization layer; Map the associated information of farmers' digital business entities to the interactive attributes of the basic geometry; Map the circulation trajectory of digital agricultural product objects as a dynamic layer on a timeline; By combining the above layers through the visualization rendering engine interface, an agricultural digital twin that supports spatiotemporal backtracking queries is output.

[0031] Example Step 1: Construct a 3D raster data model of farmland: Geographic information data and historical soil monitoring data were acquired: Boundary coordinates and elevation points of the target farmland were obtained from remote sensing satellites and GPS measuring equipment, with an accuracy of 1:500. The data was stored in GIS format, including latitude, longitude, and altitude information. Monitoring data from soil sampling points over the past three years were obtained from agricultural monitoring stations, including pH value, organic matter content (%), and total nitrogen content (%). Sampling points were evenly distributed, covering the entire farmland area.

[0032] 3D raster unit division: Based on geographic information data, farmland is logically divided into raster units of 1 meter × 1 meter × 0.5 meters (depth). Each raster unit is assigned a unique spatial index; for example, the index of raster A-01 is (x=1, y=1, z=0.5).

[0033] Interpolation algorithm maps environmental baseline data: The environmental baseline data for each raster cell is calculated using an inverse distance-weighted interpolation algorithm. The formula is: ; First, determine the coordinates of the center point of each grid cell; then calculate the distance from the center point to all monitoring points. Next, weight the monitoring point data values ​​according to the inverse square of the distance. Finally, sum and normalize to obtain the environmental baseline data of the grid cell.

[0034] in, This represents the environmental baseline data for the grid cells, including pH value, organic matter content, and total nitrogen content. This represents the soil data value at the i-th monitoring point. This represents the distance from the i-th monitoring point to the center of the grid, in meters. n represents the total number of monitoring points.

[0035] After calculation, the environmental baseline data for grid A-01 are: pH=6.8, organic matter=2.1%, and total nitrogen=0.11%. These data constitute the static property layer of farmland, used to represent the initial state of the soil.

[0036] Step 2: Construct digital business entities for farmers: Establish basic farmer profiles and acquire historical data: Create a profile for farmer Zhang San in the relational database, with fields including farmer ID (F-001), name, and contact information. Set initial skill weight parameters. Extract farmer Zhang San's planting records for the past three years from the agricultural management system, including crop type (e.g., corn), planting area (unit: mu), and agricultural operation log (e.g., sowing, fertilization, and irrigation dates). Extract the corresponding crop yield per mu (unit: kg / mu) and the rate of high-quality products (unit: %) from the harvest records. For example, Zhang San's corn high-quality product rate is 78%.

[0037] Calculate the regional average and correction factor: Obtain the average quality rate of similar crops (maize) from the regional agricultural database. The correction factor is calculated based on the deviation rate between historical yield and quality data and the regional average. The deviation rate formula is: ; in, This indicates the deviation rate, expressed as a percentage. This indicates the historical rate of high-quality products from farmers, for example, 78%. This represents the regional average, for example, 72%.

[0038] Calculated Correction factor .

[0039] Weighted calculation of current skill weight parameters: The initial skill weight parameters are weighted using the following formula: ; in, This represents the current skill weight parameter, reflecting the credibility of the farmer's operations. This represents the initial skill weight parameter, for example, 0.8.

[0040] Calculated This parameter characterizes the reliability of farmers' operations; a higher value indicates more stable quality of agricultural operations.

[0041] Step 3: Process agricultural operation instructions and generate farmland process state data: Receiving and parsing agricultural operation instructions: The system receives instructions from the agricultural management platform, including the operation type (e.g., pesticide spraying), operation time (e.g., 2023-06-01 10:00:00), spatial range (grid coordinate set, such as A-01 to A-20), and executing entity (farmer IDF-001). The system identifies the operation type, the covered grid coordinate set, and the executing entity.

[0042] Obtain skill weight parameters and standard impact factors: Based on the farmer IDF-001 index database, obtain the current skill weight parameters. Query the standard impact factor corresponding to the operation type from the predefined table. (Standard values ​​for pesticide spraying).

[0043] Calculate the operational performance value using the following formula: ; in, This represents the operational efficiency value, used to quantify the effectiveness of agricultural operations.

[0044] Calculated .

[0045] Update raster cell state: For each target raster cell (e.g., A-01), read the current state value. (For example, a growth index of 0.62), this value is obtained from a farmland process database. The time decay function is calculated using an exponential decay model, and the residual value formula is: ; in, This represents the residual value, which is the state value after decay. This represents the attenuation coefficient, with a value of 0.1. It represents a time interval, with the unit being days, such as 1 day, and e is a natural constant.

[0046] Calculated The state value is updated linearly and by way of formula: ; in, This indicates the updated state value.

[0047] Calculated Update the status value The data is written to the farmland process state database, and update timestamps are recorded. Farmland process state data is represented as a set of time-series state values, used to track dynamic changes in farmland.

[0048] Step 4: Instantiate the digital object for agricultural products: Monitoring agricultural harvesting events and locking target grids: When the combine harvester's GPS sensors enter a preset geofence (e.g., grids B-01 to B-50), the system triggers a harvesting event signal. Based on the spatial coordinates of the harvested area, grids B-01 to B-50 are locked as the target set.

[0049] Extracting raster data and identifying growth stages: Extracting data sequences of target rasters within the crop growth cycle (e.g., 120 days) from the farmland process state database, with each raster containing state values ​​at multiple time points. The growth cycle is divided into seedling stage (days 1-30), jointing stage (days 31-70), and grain-filling stage (days 71-120).

[0050] Assigning time importance coefficients and weighted summation: Assigning coefficients to each growth stage: Seedling stage Jointing period Grouting period For each raster, calculate the feature data vector. The formula is: ; in, This represents a feature data vector used to characterize the quality of agricultural products. This represents the time importance coefficient of the j-th growth stage, such as the seedling stage. . This represents the average state value of the j-th growth stage, such as the seedling stage. .

[0051] For example, the average seedling status value of grid B-01. Jointing period Grouting period ,but .

[0052] Instantiate agricultural product digital objects: Create a unique agricultural product digital object ID (e.g., P-202410001) and assign the feature data vector. Write it to the object as an initial property.

[0053] Step 5: Encapsulate and output the agricultural digital twin: Build a visualization rendering engine interface: Develop a unified interface using WebGL technology to support 3D rendering and interactive queries.

[0054] Mapping farmland models and farmer entities: The 3D raster data model of farmland is converted into basic geometry, with each raster cell rendered as a 3D cube, and the color gradient based on environmental baseline data. The associated information of the farmer's digitized business entity (such as skill weight parameters) is bound to the interactive attributes of the basic geometry; clicking on a raster displays the farmer's details.

[0055] Mapping agricultural product objects and outputting digital twins: The circulation trajectory of agricultural product digital objects is mapped as a dynamic layer on a timeline, supporting the viewing of historical states by sliding the timeline. By combining these layers through an interface, an agricultural digital twin supporting spatiotemporal retrospective queries is output. Users can query the status of any raster at a specific point in time, or track the entire lifecycle data of agricultural products.

[0056] Step 6: Self-updating of the 3D raster data model of farmland: Obtaining testing data and calculating deviation values ​​for agricultural products: Obtaining testing data from actual corn samples from quality testing laboratories. (Quality score, range 0-1). Feature data vector of agricultural product digital objects. The deviation formula is as follows: ; in, This indicates the deviation value. This represents the actual testing data for agricultural products, such as 0.69.

[0057] Calculated .

[0058] Determine the preset threshold and adjust parameters: Preset threshold .because No adjustment will be triggered. If If so, the farmer's skill weight parameter will be adjusted in the opposite direction. The formula is: ; in, This indicates the adjusted skill weight parameters. This represents the learning rate, with a value of 0.1.

[0059] At the same time, the environmental baseline data in the farmland three-dimensional raster data model is updated synchronously according to the direction of deviation, such as adjusting the pH value or organic matter content.

[0060] like Figure 2 As shown, this application provides an agricultural digital twin construction system based on data fusion, used to implement the agricultural digital twin construction method based on data fusion, including: a farmland data analysis module, a farmer data analysis module, a farmland-farmer association module, an agricultural product analysis module, and a comprehensive association module; The farmland data analysis module is used to construct a three-dimensional raster data model of farmland with spatial coordinate attributes and initialize the environmental reference data of each raster. The farmer data analysis module is used to construct a digital business entity for farmers that includes skill weight parameters, which reflect the credibility of the farmer's operations. The farmland-farmer association module is used to receive agricultural operation instructions, and based on the spatial range and execution subject in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights. The agricultural product analysis module is used to respond to agricultural product harvesting events, extract raster data sets for the corresponding time period from farmland process state data based on the spatial coordinates of the harvested area, and instantiate agricultural product digital objects. The integrated association module is used to encapsulate the three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships, and output an agricultural digital twin.

[0061] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0062] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other 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.

[0063] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0064] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0065] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0066] In conclusion, 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 constructing an agricultural digital twin based on data fusion, characterized in that, Includes the following steps: Construct a three-dimensional raster data model of farmland with spatial coordinate attributes, and initialize the environmental reference data for each raster. Construct a digital business entity for farmers that includes skill weight parameters, which are used to reflect the credibility of farmers' operations; Receive agricultural operation instructions, and based on the spatial scope and executing entity in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights; In response to agricultural product harvesting events, based on the spatial coordinates of the harvested area, a set of raster data for the corresponding time period is extracted from the farmland process state data, and agricultural product digital objects are instantiated and generated. The three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products are encapsulated through data inheritance relationships to output an agricultural digital twin.

2. The method for constructing an agricultural digital twin based on data fusion as described in claim 1, characterized in that, The construction of a three-dimensional raster data model of farmland with spatial coordinate attributes, and the initialization of environmental reference data for each raster, specifically includes: Obtain geographic information data and historical soil monitoring data of the target farmland; Based on geographic information data, the target farmland is logically divided into multiple three-dimensional raster units with independent spatial indexes; Historical soil monitoring data is mapped to each three-dimensional raster cell using an interpolation algorithm, serving as the initial environmental baseline data to form a static attribute layer for farmland.

3. The method for constructing an agricultural digital twin based on data fusion as described in claim 1, characterized in that, The construction of the farmer's digital business entity, which includes skill weight parameters, specifically includes: Establish basic farmer profiles in the database and set initial skill weight parameters; Obtain the farmer's historical planting records and corresponding historical yield and quality data; Based on the deviation rate between historical yield and quality data and the regional average, a correction coefficient is calculated. This correction coefficient is then used to weight the initial skill weight parameters to obtain the current skill weight parameters for the farmer's digital business entity.

4. The method for constructing an agricultural digital twin based on data fusion as described in claim 1, characterized in that, The generation of farmland process state data that integrates environmental benchmarks and operational weights specifically includes: Parse agricultural operation instructions to identify the operation type, operation time, covered grid coordinate set, and the farmer ID that executed the instruction; Retrieve the corresponding skill weight parameters based on the farmer ID index; Multiply the skill weight parameter by the standard influence factor corresponding to the operation type to obtain the operation efficiency value; The operational efficiency value is accumulated and written into the three-dimensional raster cell corresponding to the raster coordinate set, and the value of the three-dimensional raster cell is updated to form farmland process state data that changes dynamically over time.

5. The method for constructing an agricultural digital twin based on data fusion as described in claim 1, characterized in that, The instantiation and generation of agricultural product digital objects specifically includes: Listen for harvest trigger signals and lock onto the target grid set corresponding to the harvest area; Traverse the target raster set and extract the farmland process state data accumulated by each raster cell during the crop growth cycle; The extracted farmland process data are processed by weighted averaging to generate a set of feature data vectors; Create a unique digital object ID for agricultural products and write the feature data vector into the object as an initial attribute to complete the instantiation of the digital object for agricultural products.

6. The method for constructing an agricultural digital twin based on data fusion as described in claim 5, characterized in that, After encapsulating the three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships, and outputting an agricultural digital twin, the process also includes: Obtain actual testing data for agricultural products; Calculate the deviation between the test data and the feature data vector of the agricultural product digital object; If the deviation value exceeds the preset threshold, the skill weight parameters of the farmer's digital business entity associated with the agricultural product digital object will be adjusted in the opposite direction according to the deviation direction, and the environmental baseline data in the farmland 3D raster data model will be updated simultaneously.

7. The method for constructing an agricultural digital twin based on data fusion as described in claim 4, characterized in that, The step of accumulating and writing the operational performance value into the corresponding three-dimensional raster cell of the raster coordinate set, and updating the value of that cell, specifically includes: Read the current state value of the corresponding 3D raster unit before the operation time; The residual value of the current state value over the operation time is calculated using the time decay function; The operational performance value and the residual value are linearly superimposed to obtain the updated raster state value, and the update timestamp is recorded.

8. The method for constructing an agricultural digital twin based on data fusion as described in claim 5, characterized in that, The process of extracting farmland process data and performing weighted average processing to generate a set of feature data vectors specifically includes: Identify the growth stages within the crop growth cycle; Different time importance coefficients are assigned to different growth stages; By using time importance coefficients to perform weighted summation on farmland process data, a feature data vector reflecting the growth quality of crops throughout their entire life cycle is generated.

9. The method for constructing an agricultural digital twin based on data fusion as described in claim 1, characterized in that, The output agricultural digital twin specifically includes: Build a unified interface for the visual rendering engine; Map the three-dimensional raster data model of farmland to the basic geometry of the visualization layer; Map the associated information of farmers' digital business entities to the interactive attributes of the basic geometry; Map the circulation trajectory of digital agricultural product objects as a dynamic layer on a timeline; By combining the above layers through the visualization rendering engine interface, an agricultural digital twin that supports spatiotemporal backtracking queries is output.

10. A data fusion-based agricultural digital twin construction system, used to implement the data fusion-based agricultural digital twin construction method according to any one of claims 1-9, characterized in that, include: Farmland data analysis module, farmer data analysis module, farmland-farmer association module, agricultural product analysis module, comprehensive association module; The farmland data analysis module is used to construct a three-dimensional raster data model of farmland with spatial coordinate attributes and initialize the environmental reference data of each raster. The farmer data analysis module is used to construct a digital business entity for farmers that includes skill weight parameters, which reflect the credibility of the farmer's operations. The farmland-farmer association module is used to receive agricultural operation instructions, and based on the spatial range and execution subject in the agricultural operation instructions, map the skill weight parameters of the farmer's digital business entity to the corresponding three-dimensional raster data model of farmland, and generate farmland process state data that integrates environmental benchmarks and operation weights. The agricultural product analysis module is used to respond to agricultural product harvesting events, extract raster data sets for the corresponding time period from farmland process state data based on the spatial coordinates of the harvested area, and instantiate agricultural product digital objects. The integrated association module is used to encapsulate the three-dimensional raster data model of farmland, the digital business entities of farmers, and the digital objects of agricultural products through data inheritance relationships, and output an agricultural digital twin.

Images