Metacosm interaction system and method based on agricultural Internet of Things

Abstract

The invention discloses a meta-universe interaction system and method based on the agricultural Internet of Things, and belongs to the field of intelligent agriculture. Comprising a sensing acquisition module, a digital modeling module, an operation simulation module, a growth simulation module, a disease and pest control module, an environment regulation and control module, a man-machine interaction module, a decision support module, a data display module, a remote control module and a transaction simulation module. The system can reduce pesticide abuse, improve pest and disease control efficiency, dynamically adapt to different environments in combination with real-time monitoring data of the agricultural Internet of Things, improve adaptability, and enhance the real simulation ability of a virtual farm; and in combination with the virtual farm environment and agricultural Internet of Things data, the driving route and operation task of the agricultural machine are more accurate, the energy consumption and operation cost are reduced, the linkage of virtuality and reality is enhanced, and the operation safety of the agricultural machine is improved.

Description

Technical Field

[0001] The present invention relates to the field of smart agriculture, and in particular to a metaverse interaction system and method based on the agricultural Internet of Things. Background Art

[0002] The metaverse is the third generation of the internet, a "3D version of the internet," also known as Web 3.0 or a multi-dimensional network of mutual trust and co-creation. The metaverse is not a one-dimensional technology, but rather a multi-dimensional aggregation of technologies and an abstraction of diverse application scenarios. The continuous integration of the metaverse with the real world is gradually exhibiting the distinct characteristic of "virtual-real symbiosis." Virtual farms can be created to simulate the automated operation of various agricultural machinery and the data monitoring and processing of smart agriculture. Under the national promotion of a low-carbon economy and sustainable development, the emergence of new energy farms combines clean energy technologies with sustainable agriculture. They utilize renewable energy sources such as solar, wind, and biomass for generation and subsequent energy storage, meeting some or all of the daily operating energy needs of farm equipment and new energy agricultural machinery, thereby reducing energy dependence and carbon emissions. Virtual clean, green farms can simulate and predict the application of clean energy agricultural machinery on farms, as well as their advantages in environmental protection and other aspects. This will play a positive role in the popularization and promotion of new agricultural technologies such as clean energy agricultural machinery.

[0003] The existing metaverse interaction system and method based on the agricultural Internet of Things have low efficiency in pest control, cannot dynamically adapt to different environments, and reduce the reality simulation ability of virtual farms; in addition, the driving routes and operating tasks of agricultural machinery in the existing metaverse interaction system and method based on the agricultural Internet of Things have large deviations, high energy consumption and operating costs, and poor linkage between virtual and reality. Therefore, we propose a metaverse interaction system and method based on the agricultural Internet of Things. Summary of the Invention

[0004] The purpose of the present invention is to solve the defects in the prior art and to propose a metaverse interaction system and method based on the agricultural Internet of Things.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] A metaverse interactive system based on the agricultural Internet of Things, including a perception and acquisition module, a digital modeling module, an operation simulation module, a growth simulation module, a pest control module, an environmental control module, a human-computer interaction module, a decision support module, a data display module, a remote control module, and a transaction simulation module;

[0007] The sensing and acquisition module uses the agricultural Internet of Things sensor network to collect environmental information and transmit the data to the cloud for storage;

[0008] The digital modeling module constructs a virtual farm simulation model based on real agricultural data;

[0009] The operation simulation module is used to provide simulated operation of agricultural machinery;

[0010] The growth simulation module simulates the growth status of different crops based on environmental data and crop growth models;

[0011] The pest control module is used to combine AIoT data and computer vision to detect crop pests and diseases in real time and recommend the best pest control strategy;

[0012] The environmental control module uses meteorological data to predict future weather conditions and automatically controls agricultural equipment;

[0013] The human-computer interaction module supports users to interact with the metaverse agricultural environment and virtual NPCs;

[0014] The decision support module uses big data analysis to provide users with precise agricultural decision support;

[0015] The data display module is used to provide a visual display of agricultural production data;

[0016] The remote control module uses the metaverse environment to remotely control agricultural equipment in reality;

[0017] The transaction simulation module is used to simulate agricultural economic activities and build a complete agricultural metaverse economic ecology.

[0018] As a further solution of the present invention, the specific steps of constructing the virtual farm simulation model by the digital modeling module are as follows:

[0019] S1.1: Collect a variety of agricultural data from real farmland using IoT sensors, weather stations, and remote sensing equipment. Clean and standardize the collected agricultural data. Then, use Unreal Engine to model the terrain of a virtual farm based on the collected agricultural data. Simulate soil layers and properties in the virtual farm based on soil type, moisture, and temperature data.

[0020] S1.2: The growth simulation module establishes a crop growth model based on the collected real-world crop growth data and the interaction between crops and the environment by combining the photosynthesis model and the water-nutrient demand model to predict the growth process of crops under different environmental conditions. The constructed crop growth model is then loaded into the virtual farm. The specific construction formula of the crop growth model is as follows:

[0021]

[0022] Where G(t) represents the crop growth state at time t; G0 represents the initial crop growth state; R(t) represents the factor affecting the growth of the climate conditions at time t; L(t) represents the photosynthesis intensity at time t; S represents the growth rate constant of the crop species; W(t) represents the water requirement of the crop at time t; W0 represents the initial water requirement of the crop; P(t) represents the precipitation in the root zone of the crop at time t; E(t) represents the evaporation of the crop at time t; F represents a set of constants;

[0023] S1.3: Optimize the accuracy of the virtual farm model's simulation of agricultural production by continuously adjusting the relationship between the crop growth model and environmental factors. Utilize real-time collected actual agricultural data to verify and test the virtual farm model. Use error analysis methods to evaluate the accuracy and reliability of the virtual model, and adjust the virtual farm model in real time based on the evaluation results.

[0024] As a further solution of the present invention, the agricultural data in S1.1 specifically includes meteorological data, soil data, and crop data;

[0025] The meteorological data specifically includes temperature, humidity, and light, etc.; the soil data specifically includes soil moisture, soil temperature, and pH value, etc.; the crop data specifically includes crop growth stage, root depth, and plant spacing, etc.;

[0026] The specific calculation formula of the error analysis method described in S1.3 is as follows:

[0027]

[0028] Where, MSE represents the mean square error between the actual agricultural data and the predicted value of the virtual farm model; n represents the total number of sample data selected; y i Represents actual agricultural data; Represents the model prediction value; among them, when the MSE is higher than the preset threshold, it means that there is an abnormality in the virtual farm model, and the parameters of the virtual farm model should be readjusted.

[0029] As a further embodiment of the present invention, the specific steps of the pest control module for recommending the best pest control strategy are as follows:

[0030] S2.1: Monitor crop pest and disease conditions in real time through external agricultural networking equipment. Analyze this data using a pre-trained CNN model to identify the type, severity, and distribution of pests and diseases. Analyze the environmental conditions where pests and diseases occur using sensor data and record the analysis data.

[0031] S2.2: Based on the pest and disease detection results and the pest and disease control database, according to formula s t+1 =T(st ,a t )+ε calculates the state changes of each crop, where s t+1 represents the state of farmland pests and diseases at time t+1, T(s t ,a t ) represents the state s t Take action to prevent and control measuresa t Changes in crop status after s t represents the state of farmland pests and diseases at time t, a t represents the control measures taken at time t, ε represents random interference, and an initial pest control decision tree is established, where nodes represent farm states and edges represent possible control measures, including biological control, chemical control, physical control, and environmental control. The current farm state is used as the root node, and the current benefit value and visit count of each node are initialized.

[0032] S2.3: Calculate the UCT value of each layer of child nodes using the upper confidence interval formula, and select the child node with the highest UCT value in each layer layer by layer. Stop selecting until the selected child node is not fully expanded. Then, generate a new control strategy node based on the control measures not used in the current child node and add it to the pest control decision tree;

[0033] S2.4: Based on the currently selected child node, randomly select a control strategy and simulate the changes in crop health status under the long-term effect of the control strategy until the simulation time reaches the preset threshold. Then, the model stops and returns the simulation results to the previous node layer by layer, and updates the benefits of each node until it returns to the root node.

[0034] S2.5: Repeatedly select, expand, simulate, and backtrack until the preset iteration time is reached. Then, the final pest control decision tree is traversed and the nodes with the highest benefit values are selected layer by layer to generate a complete strategy path. At the same time, the control strategy corresponding to the path is fed back to the user. The control strategy can then be automatically executed or manually adjusted based on the user's choice.

[0035] As a further embodiment of the present invention, the specific calculation formula for identifying the type, severity and distribution of pests and diseases described in S2.1 is as follows:

[0036]

[0037] In the formula, P(C k |X) represents the probability that the crop sample X belongs to the k-type pest C; exp(f k (X)) represents the score of the crop sample X output by the CNN model belonging to category k; exp(f j (X)) represents the score of crop sample X belonging to category j; N represents the total number of pest and disease categories;

[0038] The specific calculation formula for the upper confidence interval formula described in S2.3 is as follows:

[0039]

[0040] Where Q(n) represents the cumulative revenue of node n; N(n) represents the number of times node n is visited; N(p) represents the number of times parent node p is visited; and C represents the exploration factor.

[0041] As a further solution of the present invention, the specific steps of the decision support module to provide users with precise agricultural decision support are as follows:

[0042] S3.1: Create a corresponding decision graph based on each agricultural decision problem. Each node in the graph represents a different agricultural decision state, and the paths between nodes represent the transition from one state to another. Set the initial pheromone value for all paths, initialize the population, set the population size, iteration time, and pheromone volatility coefficient, and calculate the fitness value of each agricultural decision based on the set fitness function. This is used as the heuristic information for each node.

[0043] S3.2: Take the current agricultural decision state as the starting node, and calculate the selection probability of each node connected to the starting node based on the pheromone strength and node heuristic information of each path. Then move from the current node to the node with the highest selection probability, repeating the path selection until a complete decision path is constructed. The specific calculation formula for the selection probability is as follows:

[0044]

[0045] Where, P ij represents the probability of selecting node j from node i; τ ij represents the pheromone concentration on path ij; η ij represents the heuristic information on path ij; α represents the pheromone influence factor; β represents the heuristic information influence factor; Ω represents the set of optional next path; τ ik represents the pheromone concentration on path ik; η ik Represents the heuristic information on the path ik;

[0046] S3.3: After completing the path selection, the quality of the corresponding decision path is calculated according to the agricultural production target set by the user. Based on the calculated quality, the pheromone is updated through the pheromone volatility coefficient, and the decision path construction is repeated until the preset number of iterations is reached. The search is stopped and the decision path with the highest pheromone value is selected as the optimal agricultural decision plan.

[0047] A metaverse interaction method based on the agricultural Internet of Things, the specific steps of the interaction method are as follows:

[0048] Ⅰ. Collect agricultural environmental data in real time and transmit the data to the cloud platform, while building a virtual farm based on the collected environmental data;

[0049] II. Use crop growth models to simulate the growth of different types of crops and predict the growth status and growth cycle of crops;

[0050] III. Users operate agricultural machinery through virtual farms or remote control systems, and plan the machinery's routes and tasks based on environmental data;

[0051] IV. Detect crop pests and diseases using computer vision technology and sensor data, analyze farmland images, and recommend optimal pest and disease control solutions;

[0052] V. Analyze the various data sets collected from the agricultural Internet of Things and generate real-time agricultural production status reports.

[0053] As a further solution of the present invention, the specific steps of planning the driving route and operation tasks of the agricultural machinery according to the environmental data in step III are as follows:

[0054] S4.1: Determine the number of agricultural machines built within the virtual farm based on real-time collected data, divide each field within the virtual farm into multiple operation blocks, and construct a set of particle spaces, where each particle represents a different agricultural machine path planning scheme. The particle position represents the agricultural machine's driving path and task allocation within the field, and the particle velocity represents the adjustment direction of the agricultural machine's driving path. The velocity of each particle in the particle space is initialized by random generation.

[0055] S4.2: Construct an objective function to minimize operating time, minimize fuel consumption, and maximize agricultural machinery utilization. Evaluate the fitness value of each particle based on this objective function, select the particle with the best fitness value as the global optimal position, and compare the current fitness value with the corresponding particle's own historical best position. If the current fitness value is lower than the historical best position, the historical best position remains unchanged; otherwise, it is replaced by the current particle position.

[0056] S4.3: Update the velocity of each particle based on the global optimal position and the historical best position, adjust the position of each particle based on the updated velocity, and recalculate the fitness value of each particle to update the global optimal position and the historical best position. Repeat the velocity and position updates until the fitness value of the particle corresponding to the global optimal position converges to a preset range;

[0057] S4.4: Output the agricultural machinery path planning scheme corresponding to the global optimal particle position, and visualize the driving route, operation sequence, estimated completion time, and fuel consumption evaluation in the scheme.

[0058] As a further solution of the present invention, the specific calculation formula for the velocity of each particle in the initialized particle space described in S4.1 is as follows:

[0059] V i =V min +(V max -V min )·rand()

[0060] Where V i represents the initial velocity of the i-th particle; V min Represents the lower limit of speed; V max Represents the upper limit of speed; rand() represents a random number, its range is [0, 1];

[0061] The specific calculation formula of the objective function described in S4.2 is as follows:

[0062] f(X i )=w1T(X i )+w2C(X i )-w3U(X i )

[0063] Where, f(X i ) represents the fitness value of the i-th solution X; T(X i ) represents the total operation time of the i-th solution; C(X i ) represents the fuel consumption of the ith option; U(X i ) represents the agricultural machinery utilization rate of the i-th scheme; w1, w2 and w3 represent weight coefficients respectively;

[0064] The specific calculation formula for updating the speed of each particle described in S4.3 is as follows:

[0065] V i (t+1)=ωV i (t)+c1r1(P best,i -X i )+c2r2(G best -X i )

[0066] Where V i (t+1) represents the velocity of the i-th particle in round t+1; ω represents the inertia weight; P best,i represents the historical optimal position of the i-th particle; G best represents the global optimal position among all particles; X irepresents the current position of the i-th particle; c1 and c2 represent learning factors; r1 and r2 are random numbers between 0 and 1;

[0067] The specific calculation formula for adjusting the position of each particle as described in S4.3 is as follows:

[0068] X i (t+1)=X i (t)+V i (t+1)

[0069] Where, X i (t+1) represents the new position of the i-th particle in the t+1 round, that is, the new agricultural machinery operation plan.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] 1. The present invention obtains crop pest and disease images and environmental data, and establishes a library of possible prevention and control strategies. Then, based on historical data and environmental variables, it gradually explores the short-term and long-term effects of different prevention and control measures, and then gives priority to the plans with better historical effects. At the same time, it dynamically adjusts the direction. During the simulation process, according to the newly obtained feedback information, the prevention and control plan is gradually adjusted to select the most effective pest and disease control strategy, and implementation guidance is provided. This can reduce the abuse of pesticides, improve the efficiency of pest and disease control, and combine with real-time monitoring data of the agricultural Internet of Things to dynamically adapt to different environments, improve adaptability, and enhance the reality simulation capabilities of virtual farms.

[0072] 2. The present invention generates multiple initial operation routes and task allocations based on the farmland environment and agricultural machinery performance, calculates the operation time, fuel consumption and agricultural machinery utilization rate of each plan, and assigns corresponding scores. It dynamically adjusts the agricultural machinery driving path by combining the historical best plan and the global optimal plan, and then continuously iterates the historical best plan and the global optimal plan to gradually optimize the operation sequence and driving route of the agricultural machinery. When the optimization process meets the termination conditions, the optimal agricultural machinery operation plan is output. Combined with the virtual farm environment and agricultural Internet of Things data, the driving route and operation tasks of the agricultural machinery are made more accurate, energy consumption and operation costs are reduced, and the linkage between virtual and reality is enhanced to improve the safety of agricultural machinery operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0074] Figure 1 This is a system block diagram of a metaverse interactive system based on the agricultural Internet of Things proposed by the present invention;

[0075] Figure 2This is a flowchart of a metaverse interaction method based on the agricultural Internet of Things proposed by the present invention;

[0076] Figure 3 This is a flowchart for selecting pest and disease prevention strategies based on the metaverse interaction method of the agricultural Internet of Things proposed in the present invention. DETAILED DESCRIPTION

[0077] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0078] Example 1, with reference to Figure 1 、 Figure 3 ,A metaverse interactive system based on the agricultural Internet of Things, includes a perception and acquisition module, a digital modeling module, an operation simulation module, a growth simulation module, a pest and disease control module, an environmental control module, a human-computer interaction module, a decision support module, a data display module, a remote control module, and a transaction simulation module.

[0079] The perception and collection module uses the agricultural Internet of Things sensor network to collect environmental information and transmit the data to cloud storage; the digital modeling module builds a virtual farm simulation model based on real agricultural data.

[0080] Specifically, a variety of agricultural data are collected from real farmlands through IoT sensors, weather stations and remote sensing equipment, and the collected agricultural data are cleaned and standardized. Then, based on the collected agricultural data, the Unreal Engine is used to model the terrain of the virtual farm, and based on the soil type, moisture and temperature data, the soil layer and properties are simulated in the virtual farm. The growth simulation module establishes a crop growth model based on the collected real-life crop growth data and the interaction between crops and the environment by combining the photosynthesis model and the water-nutrient demand model to predict the growth process of crops under different environmental conditions. The constructed crop growth model is loaded into the virtual farm. By continuously adjusting the relationship between the crop growth model and environmental factors, the simulation accuracy of the virtual farm model in displaying agricultural production is optimized, and the actual agricultural data collected in real time is used to verify and test the virtual farm model. At the same time, the error analysis method is used to evaluate the accuracy and reliability of the virtual model, and the virtual farm model is adjusted in real time according to the evaluation results.

[0081] In this embodiment, the specific construction formula of the crop growth model is as follows:

[0082]

[0083] Where G(t) represents the crop growth state at time t; G0 represents the initial crop growth state; R(t) represents the factor affecting the growth of the climate conditions at time t; L(t) represents the photosynthesis intensity at time t; S represents the growth rate constant of the crop species; W(t) represents the water requirement of the crop at time t; W0 represents the initial water requirement of the crop; P(t) represents the precipitation in the root zone of the crop at time t; E(t) represents the evaporation of the crop at time t; F represents a set of constants;

[0084] The specific calculation formula of the error analysis method is as follows:

[0085]

[0086] Where, MSE represents the mean square error between the actual agricultural data and the predicted value of the virtual farm model; n represents the total number of sample data selected; y i Represents actual agricultural data; Represents the model prediction value; among them, when the MSE is higher than the preset threshold, it means that there is an abnormality in the virtual farm model, and the parameters of the virtual farm model should be readjusted.

[0087] In addition, it should be noted that meteorological data specifically includes temperature, humidity, and light; soil data specifically includes soil moisture, soil temperature, and pH value; and crop data specifically includes crop growth stage, root depth, and plant spacing.

[0088] The operation simulation module is used to provide simulated operation of agricultural machinery; the growth simulation module simulates the growth status of different crops based on environmental data and crop growth models.

[0089] The pest and disease control module is used to combine AIoT data and computer vision to detect crop pests and diseases in real time and recommend the best pest and disease control strategy.

[0090] Specifically, refer to Figure 3It can be seen that the pest and disease conditions of crops are monitored in real time through various external agricultural network devices, and the real-time monitoring data is analyzed through the pre-trained CNN model to identify the types, severity and distribution of pests and diseases. The environmental conditions for the occurrence of pests and diseases are analyzed by using sensor data, and various analysis data are recorded. Based on the pest and disease detection results and the pest and disease control database, the changes in the status of each crop are calculated, and an initial pest and disease control decision tree is established, in which the nodes represent the farm status and the edges represent the control measures that can be taken, including biological control, chemical control, physical control and environmental control. The current farm status is used as the root node, and the current benefit value and visit count of each node are initialized. The UCT value of each layer of child nodes is calculated by the upper confidence interval formula, and the child node with the highest UCT value in each layer is selected layer by layer. Point, until the selected child node is not fully expanded, stop selecting, then generate a new prevention and control strategy node based on the prevention and control measures not used by the current child node, and add it to the pest control decision tree, randomly select a prevention and control strategy based on the currently selected child node, and simulate the changes in crop health status under the long-term effect of the prevention and control strategy, until the simulation time reaches the preset threshold, stop the model, and then return the simulation results to the previous node layer by layer, and update the benefit value of each node until it returns to the root node, repeat selection, expansion, simulation and backtracking until the preset iteration time is reached, then traverse the final pest control decision tree, and select nodes layer by layer to generate a complete strategy path, and at the same time, the prevention and control strategy corresponding to the path is fed back to the user, and then the prevention and control strategy is automatically executed or manually adjusted by the user.

[0091] It should be further explained that the specific calculation formula for each crop state change is as follows:

[0092] s t+1 =T(s t ,a t )+ε

[0093] Where s t+1 represents the state of farmland pests and diseases at time t+1; T(s t ,a t ) represents the state s t Take action to prevent and control measuresa t Changes in crop status after s t represents the state of farmland pests and diseases at time t; a t represents the prevention and control measures taken at time t; ε represents random interference;

[0094] The specific calculation formula for identifying the type, severity and distribution of pests and diseases is as follows:

[0095]

[0096] In the formula, P(Ck |X) represents the probability that the crop sample X belongs to the k-type pest C; exp(f k (X)) represents the score of the crop sample X output by the CNN model belonging to category k; exp(f j (X)) represents the score of crop sample X belonging to category j; N represents the total number of pest and disease categories;

[0097] The specific calculation formula for the upper confidence interval formula is as follows:

[0098]

[0099] Where Q(n) represents the cumulative revenue of node n; N(n) represents the number of times node n is visited; N(p) represents the number of times parent node p is visited; and C represents the exploration factor.

[0100] The environmental control module uses meteorological data to predict future weather conditions and automatically adjust agricultural equipment; the human-computer interaction module supports users to interact with the metaverse agricultural environment and virtual NPCs; the decision support module uses big data analysis to provide users with accurate agricultural decision support.

[0101] Specifically, a corresponding decision graph is established based on each agricultural decision problem. Each node in the graph represents a different agricultural decision state, and the path between each node represents the transition from one state to another. The initial pheromone value of all paths is set, and the population is initialized. The population size, iteration time and pheromone volatility coefficient are set. At the same time, the fitness value of each agricultural decision is calculated according to the set fitness function, and it is used as the heuristic information of each node. The current agricultural decision state is used as the starting node, and according to the pheromone intensity of each path and the node heuristic information, the selection probability of each node connected to the starting node is calculated, and the node is moved from the current node to the node with the highest selection probability. The path selection is repeated until a complete decision path is constructed. After the path selection is completed, the quality of the corresponding decision path is calculated according to the agricultural production target set by the user, and based on the calculated quality, the pheromone is updated through the pheromone volatility coefficient. The decision path construction is repeated until the preset number of iterations is reached, the search is stopped, and the decision path with the highest pheromone value is selected as the optimal agricultural decision solution.

[0102] In this embodiment, the specific calculation formula for the selection probability is as follows:

[0103]

[0104] Where, P ij represents the probability of selecting node j from node i; τ ij represents the pheromone concentration on path ij; η ijrepresents the heuristic information on path ij; α represents the pheromone influence factor; β represents the heuristic information influence factor; Ω represents the set of optional next path; τ ik represents the pheromone concentration on path ik; η ik Represents the heuristic information on the path ik.

[0105] The data display module is used to provide a visual display of agricultural production data; the remote control module uses the metaverse environment to remotely control real-world agricultural equipment; and the transaction simulation module is used to simulate agricultural economic activities and build a complete agricultural metaverse economic ecosystem.

[0106] Example 2, reference Figure 2 , a metaverse interaction method based on the agricultural Internet of Things, the specific steps of the interaction method are as follows:

[0107] Collect agricultural environmental data in real time and transmit the data to the cloud platform, while building a virtual farm based on the collected environmental data.

[0108] The crop growth model is used to simulate the growth of different types of crops and predict the growth status and growth cycle of crops.

[0109] Users operate agricultural machinery through virtual farms or remote control systems, and plan the driving routes and operating tasks of agricultural machinery based on environmental data.

[0110] Specifically, the number of agricultural machinery built in the virtual farm is determined based on real-time collected data, and each farmland in the virtual farm is divided into multiple operation blocks. A set of particle spaces is constructed, in which each particle represents a different agricultural machinery path planning scheme, the particle position represents the agricultural machinery's driving path and operation task allocation in the farmland, and the particle speed represents the adjustment direction of the agricultural machinery's driving path. By randomly generating and initializing the speed of each particle in the particle space, an objective function is constructed to minimize the operation time, minimize fuel consumption, and maximize the utilization rate of agricultural machinery. The fitness value of each particle is evaluated based on the objective function, and the particle with the best fitness value is selected as the global optimal position. The current fitness value is compared with the corresponding particle speed. The current fitness value is compared with the historical best position of the particle itself. If the current fitness value is lower than the historical best position, the historical best position remains unchanged. Otherwise, it is replaced by the current particle position. The speed of each particle is updated according to the global optimal position and the historical best position, and the position of each particle is adjusted based on the updated speed. The fitness value of each particle is recalculated to update the global optimal position and the historical best position. The speed and position updates are repeated until the fitness value of the particle corresponding to the global optimal position converges to the preset range. The agricultural machinery path planning scheme corresponding to the global optimal particle position is output, and the driving route, operation sequence, estimated completion time and fuel consumption evaluation in the scheme are visualized.

[0111] It should be further explained that the specific calculation formula for the velocity of each particle in the initialized particle space is as follows:

[0112] V i =V min +(V max -V min )·rand()

[0113] Where V i represents the initial velocity of the i-th particle; V min Represents the lower limit of speed; V max Represents the upper limit of speed; rand() represents a random number, its range is [0, 1];

[0114] The specific calculation formula of the objective function is as follows:

[0115] f(X i )=w1T(X i )+w2C(X i )-w3U(X i )

[0116] Where, f(X i ) represents the fitness value of the i-th solution X; T(X i ) represents the total operation time of the i-th solution; C(X i ) represents the fuel consumption of the ith option; U(X i ) represents the agricultural machinery utilization rate of the i-th scheme; w1, w2 and w3 represent weight coefficients respectively;

[0117] The specific calculation formula for updating the speed of each particle is as follows:

[0118] V i (t+1)=ωV i (t)+c1r1(P best,i -X i )+c2r2(G best -X i )

[0119] Where V i (t+1) represents the velocity of the i-th particle in round t+1; ω represents the inertia weight; P best,i represents the historical optimal position of the i-th particle; G best represents the global optimal position among all particles; X i represents the current position of the i-th particle; c1 and c2 represent learning factors; r1 and r2 are random numbers between 0 and 1;

[0120] The specific calculation formula for adjusting the position of each particle is as follows:

[0121] X i (t+1)=X i (t)+V i (t+1)

[0122] Where, X i (t+1) represents the new position of the i-th particle in the t+1 round, that is, the new agricultural machinery operation plan.

[0123] Detect crop pests and diseases through computer vision technology and sensor data, analyze farmland images, and recommend optimal pest and disease control solutions.

[0124] Analyze various sets of data collected from the agricultural Internet of Things and generate real-time agricultural production status reports.

Claims

1. A metaverse interactive system based on the agricultural Internet of Things, characterized by: It includes perception and acquisition module, digital modeling module, operation simulation module, growth simulation module, pest control module, environmental control module, human-computer interaction module, decision support module, data display module, remote control module and transaction simulation module; The sensing and acquisition module uses the agricultural Internet of Things sensor network to collect environmental information and transmit the data to the cloud for storage; The digital modeling module constructs a virtual farm simulation model based on real agricultural data; The operation simulation module is used to provide simulated operation of agricultural machinery; The growth simulation module simulates the growth status of different crops based on environmental data and crop growth models; The pest control module is used to combine AIoT data and computer vision to detect crop pests and diseases in real time and recommend the best pest control strategy; The environmental control module uses meteorological data to predict future weather conditions and automatically controls agricultural equipment; The human-computer interaction module supports users to interact with the metaverse agricultural environment and virtual NPCs; The decision support module uses big data analysis to provide users with precise agricultural decision support; The data display module is used to provide a visual display of agricultural production data; The remote control module uses the metaverse environment to remotely control agricultural equipment in reality; The transaction simulation module is used to simulate agricultural economic activities and build a complete agricultural metaverse economic ecology.

2. The metaverse interactive system based on the agricultural Internet of Things according to claim 1 is characterized in that: The specific steps of constructing the virtual farm simulation model by the digital modeling module are as follows: S1.1: Collect a variety of agricultural data from real farmland using IoT sensors, weather stations, and remote sensing equipment. Clean and standardize the collected agricultural data. Then, use Unreal Engine to model the terrain of a virtual farm based on the collected agricultural data. Simulate soil layers and properties in the virtual farm based on soil type, moisture, and temperature data. S1.2: The growth simulation module establishes a crop growth model based on the collected real-world crop growth data and the interaction between crops and the environment by combining the photosynthesis model and the water-nutrient demand model to predict the growth process of crops under different environmental conditions. The constructed crop growth model is then loaded into the virtual farm. The specific construction formula of the crop growth model is as follows: Where G(t) represents the crop growth state at time t; G0 represents the initial crop growth state; R(t) represents the factor affecting the growth of the climate conditions at time t; L(t) represents the photosynthesis intensity at time t; S represents the growth rate constant of the crop species; W(t) represents the water requirement of the crop at time t; W0 represents the initial water requirement of the crop; P(t) represents the precipitation in the root zone of the crop at time t; E(t) represents the evaporation of the crop at time t; F represents a set of constants; S1.3: Optimize the accuracy of the virtual farm model's simulation of agricultural production by continuously adjusting the relationship between the crop growth model and environmental factors. Utilize real-time collected actual agricultural data to verify and test the virtual farm model. Use error analysis methods to evaluate the accuracy and reliability of the virtual model, and adjust the virtual farm model in real time based on the evaluation results.

3. The metaverse interactive system based on the agricultural Internet of Things according to claim 2 is characterized in that: The specific steps for the pest control module to recommend the best pest control strategy are as follows: S2.1: Monitor crop pest and disease conditions in real time using drones, ground monitoring stations, and high-resolution cameras. Analyze this data using a pre-trained CNN model to identify the type, severity, and distribution of pests and diseases. Analyze the environmental conditions where pests and diseases occur using sensor data and record the analysis data. S2.2: Based on the pest and disease detection results and the pest and disease control database, according to formula s t+1 =T(s t ,a t )+ε calculates the state changes of each crop, where s t+1 represents the state of farmland pests and diseases at time t+1, T(s t ,a t ) represents the state s t Take action to prevent and control measuresa t Changes in crop status after s t represents the state of farmland pests and diseases at time t, a t represents the control measures taken at time t, ε represents random interference, and an initial pest control decision tree is established, where nodes represent farm states and edges represent possible control measures, including biological control, chemical control, physical control, and environmental control. The current farm state is used as the root node, and the current benefit value and visit count of each node are initialized. S2.3: Calculate the UCT value of each layer of child nodes using the upper confidence interval formula, and select the child node with the highest UCT value in each layer layer by layer. Stop selecting until the selected child node is not fully expanded. Then, generate a new control strategy node based on the control measures not used in the current child node and add it to the pest control decision tree; S2.4: Based on the currently selected child node, randomly select a control strategy and simulate the changes in crop health status under the long-term effect of the control strategy until the simulation time reaches the preset threshold. Then, the model stops and returns the simulation results to the previous node layer by layer, and updates the benefits of each node until it returns to the root node. S2.5: Repeatedly select, expand, simulate, and backtrack until the preset iteration time is reached. Then, the final pest control decision tree is traversed and the nodes with the highest benefit values are selected layer by layer to generate a complete strategy path. At the same time, the control strategy corresponding to the path is fed back to the user. The control strategy can then be automatically executed or manually adjusted based on the user's choice.

4. The metaverse interactive system based on the agricultural Internet of Things according to claim 1 is characterized in that: The specific calculation formula for identifying the type, severity, and distribution of pests and diseases as described in S2.1 is as follows: In the formula, P(C k |X) represents the probability that the crop sample X belongs to the k-type pest C; exp(f k (X)) represents the score of the crop sample X output by the CNN model belonging to category k; exp(f j (X)) represents the score of crop sample X belonging to category j; N represents the total number of pest and disease categories; The specific calculation formula for the upper confidence interval formula described in S2.3 is as follows: Where Q(n) represents the cumulative revenue of node n; N(n) represents the number of times node n is visited; N(p) represents the number of times parent node p is visited; C represents the exploration factor.

5. The metaverse interactive system based on the agricultural Internet of Things according to claim 1 is characterized in that: The specific steps of the decision support module to provide users with precise agricultural decision support are as follows: S3.1: Create a corresponding decision graph based on each agricultural decision problem. Each node in the graph represents a different agricultural decision state, and the paths between nodes represent the transition from one state to another. Set the initial pheromone value for all paths, initialize the population, set the population size, iteration time, and pheromone volatility coefficient, and calculate the fitness value of each agricultural decision based on the set fitness function. This is used as the heuristic information for each node. S3.2: Take the current agricultural decision state as the starting node, and calculate the selection probability of each node connected to the starting node based on the pheromone strength and node heuristic information of each path. Then move from the current node to the node with the highest selection probability, repeating the path selection until a complete decision path is constructed. The specific calculation formula for the selection probability is as follows: Where, P ij represents the probability of selecting node j from node i; τ ij represents the pheromone concentration on path ij; η ij represents the heuristic information on path ij; α represents the pheromone influence factor; β represents the heuristic information influence factor; Ω represents the set of optional next path; τ ik represents the pheromone concentration on path ik; η ik Represents the heuristic information on the path ik; S3.3: After completing the path selection, the quality of the corresponding decision path is calculated according to the agricultural production target set by the user. Based on the calculated quality, the pheromone is updated through the pheromone volatility coefficient, and the decision path construction is repeated until the preset number of iterations is reached. The search is stopped and the decision path with the highest pheromone value is selected as the optimal agricultural decision plan.

6. A metaverse interaction method based on the agricultural Internet of Things, used to implement the metaverse interaction system function based on the agricultural Internet of Things according to any one of claims 1 to 5, characterized in that: The specific steps of this interactive method are as follows: Ⅰ. Collect agricultural environmental data in real time and transmit the data to the cloud platform, while building a virtual farm based on the collected environmental data; II. Use crop growth models to simulate the growth of different types of crops and predict the growth status and growth cycle of crops; III. Users operate agricultural machinery through virtual farms or remote control systems, and plan the machinery's routes and tasks based on environmental data; IV. Detect crop pests and diseases using computer vision technology and sensor data, analyze farmland images, and recommend optimal pest and disease control solutions; V. Analyze the various sets of data collected from the agricultural Internet of Things and generate real-time agricultural production status reports.

7. The method for interacting with the metaverse based on the agricultural Internet of Things according to claim 6, characterized in that: The specific steps for planning the driving route and operation tasks of agricultural machinery based on environmental data in step III are as follows: S4.1: Determine the number of agricultural machines built within the virtual farm based on real-time collected data, divide each field within the virtual farm into multiple operation blocks, and construct a set of particle spaces, where each particle represents a different agricultural machine path planning scheme. The particle position represents the agricultural machine's driving path and task allocation within the field, and the particle velocity represents the adjustment direction of the agricultural machine's driving path. The velocity of each particle in the particle space is initialized by random generation. S4.2: Construct an objective function to minimize operating time, minimize fuel consumption, and maximize agricultural machinery utilization. Evaluate the fitness value of each particle based on this objective function, select the particle with the best fitness value as the global optimal position, and compare the current fitness value with the corresponding particle's own historical best position. If the current fitness value is lower than the historical best position, the historical best position remains unchanged; otherwise, it is replaced by the current particle position. S4.3: Update the velocity of each particle based on the global optimal position and the historical best position, adjust the position of each particle based on the updated velocity, and recalculate the fitness value of each particle to update the global optimal position and the historical best position. Repeat the velocity and position updates until the fitness value of the particle corresponding to the global optimal position converges to a preset range; S4.4: Output the agricultural machinery path planning scheme corresponding to the global optimal particle position, and visualize the driving route, operation sequence, estimated completion time, and fuel consumption evaluation in the scheme.

8. The method for interacting with the metaverse based on the agricultural Internet of Things according to claim 7, characterized in that: The specific calculation formula for the velocity of each particle in the initialized particle space described in S4.1 is as follows: V i =V min +(V max -V min )·rand() Where V i represents the initial velocity of the i-th particle; V min Represents the lower limit of speed; V max Represents the upper limit of speed; rand() represents a random number, its range is [0, 1]; The specific calculation formula of the objective function described in S4.2 is as follows: f(X i )=w1T(X i )+w2C(X i )-w3U(X i ) Where, f(X i ) represents the fitness value of the i-th solution X; T(X i ) represents the total operation time of the i-th solution; C(X i ) represents the fuel consumption of the ith option; U(X i ) represents the agricultural machinery utilization rate of the i-th scheme; w1, w2 and w3 represent weight coefficients respectively; The specific calculation formula for updating the speed of each particle described in S4.3 is as follows: V i (t+1)=ωV i (t)+c1r1(P best,i -X i )+c2r2(G best -X i ) Where V i (t+1) represents the velocity of the i-th particle in round t+1; ω represents the inertia weight; P best,i represents the historical optimal position of the i-th particle; G best represents the global optimal position among all particles; X i represents the current position of the i-th particle; c1 and c2 represent learning factors; r1 and r2 are random numbers between 0 and 1; The specific calculation formula for adjusting the position of each particle as described in S4.3 is as follows: X i (t+1)=X i (t)+V i (t+1) Where, X i (t+1) represents the new position of the i-th particle in the t+1 round, that is, the new agricultural machinery operation plan.

Images