Intelligent agricultural condition monitoring and early warning method and system

Abstract

The invention discloses an intelligent agricultural condition monitoring and early warning method and system, and the method comprises the steps: constructing a three-dimensional geographic model of a monitored farmland, and carrying out the grid division to generate a plurality of agricultural monitoring grids; obtaining multi-source agricultural condition data, predicting the agricultural condition risk level of each grid through a machine learning model, screening the grid with the high level as a target early warning area, and identifying a risk source and an initial monitoring point; establishing a multi-objective decision model according to the space coverage rate, the data acquisition precision and the deployment cost of the initial monitoring point, and solving the model by adopting a genetic algorithm to determine an optimized monitoring point; and simulating agricultural condition evolution of the optimized monitoring points in an extreme scene, screening the optimized monitoring points of which the stability indexes are higher than a preset value as final agricultural condition early warning nodes, and generating a dynamic monitoring network. According to the invention, problems of one-sided agricultural condition evaluation and inaccurate risk area positioning in the prior art are solved. And the agricultural condition change can be responded more timely and accurately, and the reliability and the stability of the monitoring system under complex and extreme conditions are remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural condition monitoring, and in particular to an intelligent agricultural condition monitoring and early warning method and system. Background Art

[0002] Accurate and efficient agricultural monitoring and early warning are crucial in agricultural production. However, existing agricultural monitoring and early warning technologies have numerous shortcomings. Traditional agricultural monitoring methods often rely on manual field surveys, which are not only inefficient but also difficult to achieve comprehensive, real-time monitoring of large areas of farmland. For example, plant protection personnel obtain information on crop pests and diseases through manual inspections and visual assessments. This is not only labor-intensive, but also highly dependent on individual experience and accuracy, and lacks automation and intelligence.

[0003] While some monitoring systems that utilize sensors and other technologies have achieved a certain degree of automation in data collection, they suffer from significant shortcomings in data processing and analysis. On the one hand, data sources are relatively limited, failing to fully capture key information from the crop growth stages and making it difficult to conduct a comprehensive and accurate assessment of agricultural conditions. On the other hand, existing systems lack precision and foresight in risk prediction and early warning, making it impossible to effectively identify high-risk areas and formulate response strategies in advance. Furthermore, existing technologies often lack scientific optimization methods for the layout of monitoring points, resulting in low overall efficiency and stability of the monitoring network, making it difficult to ensure accurate acquisition of agricultural information and timely early warning in extreme scenarios. Summary of the Invention

[0004] In order to solve at least one of the technical problems mentioned above, the present invention provides an intelligent agricultural monitoring and early warning method and system.

[0005] In a first aspect, the present invention provides an intelligent agricultural monitoring and early warning method, the method comprising:

[0006] Construct a three-dimensional geographic model of the monitored farmland, divide the model into grids based on soil type, crop distribution, and terrain data, and generate multiple agricultural monitoring grids;

[0007] Acquire multi-source agricultural data, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. Use machine learning models to predict the agricultural risk level of each grid. Select agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identify risk sources and initial monitoring points within the target warning areas.

[0008] A multi-objective decision-making model was established, with the spatial coverage of initial monitoring points, data collection accuracy, and deployment cost as optimization objectives. The constraints included the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. An improved genetic algorithm was used to solve the model and determine the optimal monitoring points.

[0009] Simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate stability indicators based on early warning accuracy, response time, and cost volatility, select optimized monitoring points with stability indicators higher than preset values ​​as the final agricultural condition early warning nodes, and generate a dynamic monitoring network.

[0010] Preferably, the predicting of the agricultural risk level of each grid by a machine learning model includes:

[0011] Constructing a feature set, the feature set including soil pH value, nitrogen, phosphorus and potassium content, cumulative precipitation, temperature fluctuation rate, crop disease resistance index and historical disaster frequency;

[0012] The risk prediction model was trained using a random forest algorithm, and model parameters were optimized through cross-validation to output the probability distribution of drought risk, waterlogging risk, and pest and disease risk for each grid.

[0013] A comprehensive risk level is generated based on the weighted summation of risk probabilities, and high, medium and low risk areas are divided, and high-risk areas are mapped to target warning areas.

[0014] Preferably, the construction of the multi-objective decision model includes defining a single objective function, including:

[0015] Maximize spatial coverage:

[0016] ;

[0017] Where, is the spatial coverage, For the The effective coverage area of ​​each monitoring point reflects the physical range within which the monitoring point can collect data. Monitoring points, 、 is the terrain attenuation coefficient, which is used to quantify the weakening effect of terrain complexity (such as slope and vegetation density) on the coverage effect; The base area threshold is used to adjust the sensitive interval of the logic function to ensure the nonlinear mapping characteristics of the coverage rate.

[0018] Maximize data accuracy:

[0019] ;

[0020] Where, is the data accuracy, is the sensor resolution; is the Euclidean distance between the monitoring point and the risk source; is the data calibration frequency, 、 is the weight factor;

[0021] Minimize deployment costs:

[0022] ;

[0023] Where, For deployment cost, For hardware costs, is the communication link cost, For maintenance costs, is the cost coefficient;

[0024] The above single objective function is integrated into a multi-objective decision model through fuzzy analytic hierarchy process, and constraints are added:

[0025] ;

[0026] ;

[0027] ;

[0028] Where, For the warning response time, The upper limit of the warning response time, is the total cost threshold, This is the minimum data accuracy requirement.

[0029] Preferably, the extreme scenarios include:

[0030] Meteorological extreme events: spatiotemporal distribution patterns of persistent drought, heavy rain and flooding, and low temperature and freezing damage;

[0031] Biological disaster events: the coupling relationship between pest and disease spread paths, egg hatching rates, and crop susceptible periods;

[0032] Human intervention events: simulation of soil salinization caused by irrigation interruption and excessive fertilization.

[0033] Preferably, the simulation and optimization of agricultural condition evolution at monitoring points under extreme scenarios includes:

[0034] Construct a multidimensional scenario library based on extreme scenarios;

[0035] Based on computational fluid dynamics and agent-based models, the propagation trajectory of risk sources in various scenarios is dynamically simulated, and the data capture efficiency and warning delay of monitoring nodes are evaluated to generate simulation results.

[0036] Calculate the stability index based on the simulation results:

[0037] ;

[0038] Where, For stability score, For the early warning accuracy, is the average response time, is the cost volatility; is the normalized weight.

[0039] Preferably, the method further comprises:

[0040] Through the coordinated inspection of drone clusters and ground sensors, the monitoring network topology is updated in real time;

[0041] When a jump in the local risk level is detected, the adaptive adjustment algorithm of the monitoring point density is automatically triggered, and mobile monitoring nodes are added to the target area. At the same time, redundant nodes in low-risk areas are shut down to reduce energy consumption.

[0042] In a second aspect, the present invention further provides an intelligent agricultural monitoring and early warning system, the system comprising:

[0043] The gridding unit is used to construct a three-dimensional geographic model of the monitored farmland. The model is gridded based on soil type, crop distribution and terrain data to generate multiple agricultural monitoring grids.

[0044] The monitoring point screening unit is used to obtain agricultural data from multiple sources, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. It uses machine learning models to predict the agricultural risk level of each grid, selects agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identifies risk sources and initial monitoring points within the target warning areas.

[0045] The monitoring point optimization unit is used to establish a multi-objective decision-making model with the spatial coverage, data collection accuracy, and deployment cost of the initial monitoring points as optimization targets. The constraints include the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. An improved genetic algorithm is used to solve the model to determine the optimal monitoring points.

[0046] The early warning node determination unit is used to simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate the stability index based on the early warning accuracy, response time and cost volatility, select the optimized monitoring points with stability index higher than the preset value as the final agricultural condition early warning nodes, and generate a dynamic monitoring network.

[0047] In a third aspect, the present invention also provides an electronic device comprising a processor and a memory, wherein the memory is used to store computer program code, and the computer program code comprises computer instructions. When the processor executes the computer instructions, the electronic device executes the method as described in the first aspect above and any possible implementation thereof.

[0048] In a fourth aspect, the present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions, and when the program instructions are executed by a processor of an electronic device, the processor executes the method as described in the first aspect above and any possible implementation method thereof.

[0049] Compared with the prior art, the present invention has the following beneficial effects:

[0050] The present invention constructs a three-dimensional geographic model of the monitored farmland and performs grid division based on soil type, crop distribution and terrain data. It can accurately subdivide the farmland into multiple agricultural monitoring grids, laying the foundation for subsequent refined monitoring. After obtaining multi-source agricultural data, the agricultural risk level of each grid is predicted with the help of a machine learning model. It can efficiently screen out agricultural monitoring grids with risk levels exceeding a preset threshold as target early warning areas, and accurately identify risk sources and initial monitoring points, thus changing the previous problems of one-sided agricultural assessments and inaccurate positioning of risk areas. By establishing a multi-objective decision-making model and using an improved genetic algorithm to solve it, the optimized monitoring points are determined. This process fully considers key factors such as the spatial coverage of the monitoring points, data acquisition accuracy and deployment costs. Compared with the traditional arbitrary or unreasonable layout of monitoring points, it greatly improves the scientificity and rationality of the monitoring network, effectively controlling costs while ensuring data quality. By simulating the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, and calculating stability indicators based on early warning accuracy, response time, and cost volatility, we screened out optimized monitoring points with stability indicators higher than preset values ​​as the final agricultural condition early warning nodes and generated a dynamic monitoring network. This significantly enhanced the reliability and stability of the monitoring system under complex and extreme conditions, and enabled it to respond to changes in agricultural conditions more promptly and accurately, providing strong guarantees for agricultural production, and realizing the transformation from traditional extensive agricultural condition monitoring and early warning to a precise, efficient, intelligent, and stable modern agricultural condition monitoring and early warning model.

[0051] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0053] The accompanying drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and, together with the specification, are used to explain the technical solutions of the present disclosure.

[0054] Figure 1 A flow chart of an intelligent agricultural monitoring and early warning method provided by an embodiment of the present invention;

[0055] Figure 2 for Figure 1 Schematic diagram of the flow of sub-steps of step S20;

[0056] Figure 3 A schematic structural diagram of an intelligent agricultural monitoring and early warning system provided by an embodiment of the present invention;

[0057] Figure 4 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0058] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0059] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0060] See also Figure 1 , Figure 1 The following is a flow chart of an intelligent agricultural monitoring and early warning method provided by an embodiment of the present invention. Figure 1 As shown, the method includes:

[0061] S10, constructing a three-dimensional geographic model of the monitored farmland, dividing the model into grids based on soil type, crop distribution, and terrain data, and generating multiple agricultural monitoring grids;

[0062] Traditional agricultural monitoring relies on manual sampling, which is difficult to cover in complex terrain and diverse planting structures. Three-dimensional models can intuitively reflect the three-dimensional structure of farmland, and gridding can subdivide the monitoring area into controllable units, improving monitoring accuracy.

[0063] Terrain elevation data is acquired through drone LiDAR, combined with satellite remote sensing imagery to extract crop distribution. Soil sampling points are used to construct a spatial distribution model of soil types. Kriging interpolation is used to generate a continuous terrain surface. Combined with ArcGIS grid analysis tools, grids are created at a resolution of 50m x 50m or 100m x 100m. Each grid is assigned attributes such as soil texture, fertility level, and crop variety. This creates standardized monitoring units, providing a unified spatial framework for subsequent multi-source data fusion and improving monitoring coverage compared to traditional methods.

[0064] S20. Obtain agricultural data from multiple sources, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. Use machine learning models to predict the agricultural risk level of each grid. Select agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identify risk sources and initial monitoring points within the target warning areas.

[0065] A single data source cannot fully reflect changes in agricultural conditions. Machine learning can explore implicit patterns in data and enable early identification of risks.

[0066] In this step, data acquisition includes: 1) Soil moisture: deploying a distributed TDR sensor network; 2) Meteorological indicators: accessing data from the Meteorological Bureau's micro-meteorological stations; 3) Crop growth: extracting the NDVI index using drone multispectral cameras; and 4) Pest and disease data: integrating historical pest monitoring data with satellite pest and disease remote sensing inversion results. A random forest algorithm is used to train a risk prediction model using historical disaster data as labels. This input is a 12-dimensional feature vector, which outputs a risk score on a scale of 0-10. A risk threshold of 7 or higher is set as a high-risk area. Spatial overlay analysis is used to identify risk sources (e.g., continuously cropped fields) and initial monitoring points (the center of risk). This improves early warning accuracy and enables earlier warnings compared to traditional manual forecasting.

[0067] See also Figure 2 In one embodiment, the method of predicting the agricultural risk level of each grid using a machine learning model includes:

[0068] S201, constructing a feature set, wherein the feature set includes soil pH value, nitrogen, phosphorus and potassium content, cumulative precipitation, temperature fluctuation rate, crop disease resistance index and historical disaster frequency;

[0069] Multi-source data integration:

[0070] Soil data: Real-time collection of pH, nitrogen, phosphorus, and potassium content (EC value inversion) through a distributed sensor network;

[0071] Meteorological data: Obtain the cumulative precipitation and daily temperature difference data for the past 30 days from the Meteorological Bureau API;

[0072] Crop data: Extracting chlorophyll content based on drone hyperspectral images and combining it with a variety database to generate a disease resistance index;

[0073] Historical data: Establish a GIS spatial database to record the location, intensity, and frequency of disasters in the past five years;

[0074] Continuous features were normalized using the Z-score method. The contribution of each feature was calculated using the Gini index of the random forest. Redundant features (with importance less than 0.05) were removed, ultimately retaining 12 core features. Feature dimension compression reduces computational complexity and improves model training speed.

[0075] S202. Use the random forest algorithm to train the risk prediction model, optimize the model parameters through cross-validation, and output the probability distribution of drought risk, waterlogging risk, and pest and disease risk for each grid;

[0076] The training and test sets were split in a 7:3 ratio. The number of decision trees was set to 200, with a maximum depth of 8 layers. Finally, 5-fold cross-validation was used, with evaluation metrics including the area under the receiver operating characteristic (ROC) curve, precision, and recall. A soft vote was then performed on the output of each tree to generate the final probability distribution.

[0077] S203. Generate a comprehensive risk level based on the weighted sum of risk probabilities, divide the risk areas into high, medium and low risk areas, and map the high risk areas to the target warning areas.

[0078] A comprehensive risk level is generated based on the weighted summation of risk probabilities: R ≥ 0.7 for high-risk areas, 0.4 ≤ R < 0.7 for medium-risk areas, and R < 0.4 for low-risk areas. The probability grid is converted into a vector polygon and spatially overlaid with the grid data in step S10 for analysis. This allows for a quantitative expression of risk, improving accuracy compared to traditional qualitative assessments. The generated risk heat map can intuitively display the spatial distribution of risk, providing a basis for precise prevention and control.

[0079] This example introduces "temperature difference fluctuation rate" as a sensitive indicator for crop stress warning, constructs the "historical disaster frequency-spatial autocorrelation" feature to capture the risk aggregation effect; uses ensemble learning to reduce the overfitting risk of a single decision tree; and achieves parameter self-optimization through cross-validation to reduce the cost of manual parameter adjustment.

[0080] S30. Taking the spatial coverage of the initial monitoring points, data collection accuracy, and deployment cost as optimization objectives, a multi-objective decision-making model is established, where the constraints include the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. An improved genetic algorithm is used to solve the model to determine the optimal monitoring points;

[0081] Traditional monitoring point deployment relies on experience, which can easily lead to low coverage or waste of costs. Multi-objective optimization can balance technical indicators and economic constraints.

[0082] In one embodiment, the construction of the multi-objective decision model includes defining a single objective function, including:

[0083] Maximize spatial coverage:

[0084] ;

[0085] Where, is the spatial coverage, For the The effective coverage area of ​​each monitoring point reflects the physical range within which the monitoring point can collect data. Monitoring points, 、 is the terrain attenuation coefficient, which is used to quantify the weakening effect of terrain complexity (such as slope and vegetation density) on the coverage effect; The base area threshold is used to adjust the sensitive interval of the logic function to ensure the nonlinear mapping characteristics of the coverage rate.

[0086] Maximize data accuracy:

[0087] ;

[0088] Where, is the data accuracy, is the sensor resolution; is the Euclidean distance between the monitoring point and the risk source; is the data calibration frequency, 、 is the weight factor;

[0089] Minimize deployment costs:

[0090] ;

[0091] Where, For deployment cost, For hardware costs, is the communication link cost, For maintenance costs, is the cost coefficient;

[0092] The above single objective function is integrated into a multi-objective decision model through fuzzy analytic hierarchy process, and constraints are added:

[0093] ;

[0094] ;

[0095] ;

[0096] Where, For the warning response time, The upper limit of the warning response time, is the total cost threshold, This is the minimum data accuracy requirement.

[0097] The spatial coverage function adopts the S-type logistic growth model , simulating the terrain attenuation effect. When the effective area of ​​the monitoring point Exceeding the threshold When , the coverage efficiency shows an accelerated improvement characteristic, which is in line with the increasing marginal benefit law of actual monitoring equipment coverage. The data accuracy function introduces the inverse square of the distance , reflecting the physical law that the monitoring resolution decays with distance, and at the same time, through the calibration frequency To ensure long-term data reliability, the cost model uses a linear combination to achieve the quantitative superposition of factors such as hardware, communication, and maintenance.

[0098] Triangular fuzzy numbers are used to handle the subjective ambiguity of expert judgment. A fuzzy judgment matrix is ​​constructed to determine the weights of each objective. The fuzzy weights are converted into deterministic weight coefficients using the λ-cut method. The weighted summation method is used to transform the multi-objective optimization problem into a single-objective optimization problem, which is then determined by the analytic hierarchy process.

[0099] In the specific implementation of the solution process:

[0100] 1) Parameter calibration stage:

[0101] Terrain attenuation coefficient 、 , calculate the average slope using DEM elevation data , establish a mapping relationship ;

[0102] Base area , take 60% of the nominal coverage area of ​​the monitoring equipment as the initial value;

[0103] Weighting Factor 、 , calculated using the entropy weight method, for example .

[0104] 2) Model solving process:

[0105] 2.1) Data collection: GIS geographic data, equipment parameter database, cost database;

[0106] 2.2) Constraint processing: Use penalty function method to Equal constraints are integrated into the objective function;

[0107] 2.3) Optimization solution: NSGA-II multi-objective genetic algorithm was used with a population size of 100 and 500 iterations.

[0108] 2.4) Pareto frontier analysis: Output a set of non-inferior solutions for decision makers to choose.

[0109] 3) Deployment verification:

[0110] Monte Carlo simulation: randomly generate 1,000 terrain-cost combinations to verify the robustness of the model;

[0111] On-site debugging: After deployment, collect actual coverage data and provide feedback for correction 、 parameter.

[0112] Based on the Pareto optimality theory, this model integrates three conflicting objective functions through the fuzzy analytic hierarchy process to form a unified evaluation system. The core idea is to find a set of non-dominated solutions (Pareto optimal solution set) that optimizes all three objective functions while satisfying the constraints.

[0113] The terrain attenuation effect is described by the logistic function, which is more in line with actual monitoring scenarios than the traditional linear model. The inverse square law in physics is introduced to quantify the impact of the distance between the monitoring point and the risk source on data accuracy. The fuzzy hierarchical analysis method is used to dynamically determine the weight of each target to adapt to the needs of different agricultural scenarios.

[0114] S40. Simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate stability indicators based on early warning accuracy, response time, and cost volatility, select optimized monitoring points with stability indicators higher than preset values ​​as final agricultural condition early warning nodes, and generate a dynamic monitoring network.

[0115] Conventional monitoring networks are prone to failure in extreme weather conditions; stability assessment can enhance system robustness.

[0116] Preferably, the extreme scenarios include:

[0117] Meteorological extreme events: spatiotemporal distribution patterns of persistent drought, heavy rain and flooding, and low temperature and freezing damage;

[0118] Biological disaster events: the coupling relationship between pest and disease spread paths, egg hatching rates, and crop susceptible periods;

[0119] Human intervention events: simulation of soil salinization caused by irrigation interruption and excessive fertilization.

[0120] In one embodiment, the simulation and optimization of agricultural condition evolution at monitoring points under extreme scenarios includes:

[0121] Construct a multidimensional scenario library based on extreme scenarios;

[0122] Based on computational fluid dynamics and agent-based models, the propagation trajectory of risk sources in various scenarios is dynamically simulated, and the data capture efficiency and warning delay of monitoring nodes are evaluated to generate simulation results.

[0123] Calculate the stability index based on the simulation results:

[0124] ;

[0125] Where, For stability score, For the early warning accuracy, is the average response time, is the cost volatility; is the normalized weight.

[0126] This solution simulates the evolution of agricultural conditions under extreme scenarios, combines multi-dimensional dynamic assessment with an optimized screening mechanism, and builds a highly stable agricultural monitoring and early warning network. Its core principles are:

[0127] Extreme scenario rehearsal: Use a multi-dimensional scenario library to simulate the impact of three types of extreme events, meteorological, biological, and human, on agricultural systems, and quantify the risk transmission path and the response capacity of monitoring nodes.

[0128] Dynamic stability assessment: based on early warning accuracy , average response time Cost volatility Constructing composite indicators , screen out monitoring points with strong anti-interference ability.

[0129] Network adaptive optimization: Through dynamic monitoring of network generation mechanisms, it ensures that early warning nodes automatically adjust with environmental changes, thereby improving the overall robustness of the system.

[0130] Specifically, the implementation includes

[0131] 1) Construction of multi-dimensional scenario library:

[0132] Data fusion: Integrate meteorological satellite data, soil sensor records, historical outbreak patterns of pests and diseases, and human intervention logs to build an extreme event database covering time and space.

[0133] Scenario parameterization includes the following three types:

[0134] Meteorological events: define drought index, rainfall intensity-duration-frequency curve, and low temperature duration threshold.

[0135] Biological disasters: Establish a correlation matrix between the spread rate of pests and diseases and temperature, humidity, and crop growth stage.

[0136] Human-induced events: simulation of irrigation interruption cycles, and fertilizer application-soil electrical conductivity relationship model.

[0137] 2) Dynamic deduction of risk communication:

[0138] Computational Fluid Dynamics (CFD) Modeling:

[0139] Simulate the spatial spread of meteorological hazards (such as floods) and output the risk coverage and propagation speed. Example: Use the Navier-Stokes equations to simulate the path of rainstorm runoff and predict the inundation area of ​​farmland.

[0140] Agent-Based Model (ABM):

[0141] Build a pest and disease transmission agent and define migration rules (such as wind propagation probability and the influence of host crop density). Example: Simulate an exponential outbreak caused by the overlap of insect egg hatching rate and crop susceptible period.

[0142] 3) Monitoring node evaluation and screening:

[0143] Data capture efficiency calculation: Calculate the probability that a monitoring node successfully identifies a risk event in the simulation. = Number of correct warnings / total number of risk events.

[0144] Response time quantification: records the average delay from the time the risk reaches the monitoring range to the time the warning is triggered .

[0145] Cost fluctuation analysis: Calculate the coefficient of variation of node maintenance costs under extreme scenarios. , is the standard deviation, is the mean).

[0146] Stability index calculation: According to Formula calculation, weight distribution can be They are 0.5, 0.3 and 0.2 respectively.

[0147] 4) Dynamic monitoring network generation:

[0148] Node selection: Filter Monitoring points with a value greater than a threshold (e.g., 0.8) form a backbone network.

[0149] Redundant design: Deploy backup nodes in key areas and automatically switch when backbone nodes fail.

[0150] Real-time update: Dynamically adjust node weights based on Internet of Things (IoT) data streams to adapt to sudden environmental changes.

[0151] In summary, in this embodiment, the warning accuracy will be improved and the average response time will be shortened in the simulation of continuous drought. Optimization has reduced node maintenance cost volatility by 40%, significantly reducing the risk of budget overruns in extreme scenarios. In heavy rain and flooding scenarios, the dynamic network maintained over 85% of nodes operating effectively. Monitoring network coverage has been significantly improved during coupled pest and disease events. The resulting risk heat maps and node stability maps can guide agricultural insurance pricing and emergency resource allocation. This closed-loop "simulation-optimization-verification" solution provides a comprehensive theoretical and practical solution for agricultural disaster early warning, particularly applicable to the demands of precision agriculture management in the context of intensified climate change.

[0152] Preferably, in one embodiment, the method further comprises:

[0153] Through the coordinated inspection of drone clusters and ground sensors, the monitoring network topology is updated in real time;

[0154] When a jump in the local risk level is detected, the adaptive adjustment algorithm of the monitoring point density is automatically triggered, and mobile monitoring nodes are added to the target area. At the same time, redundant nodes in low-risk areas are shut down to reduce energy consumption.

[0155] Ground sensors are deployed extensively across the monitoring area to collect environmental data such as temperature, humidity, gas concentration, and vibration. Simultaneously, a drone swarm is assembled, each equipped with high-precision sensors, communication modules, and positioning systems. The drone swarm communicates with the ground sensors via wireless networks (such as 5G or ad hoc networks), forming a monitoring network.

[0156] Ground sensors continuously collect data and transmit it to the control center in real time. Drone swarms conduct regular inspections along pre-set routes. During these inspections, drones use their onboard sensors to collect data from areas difficult to cover or key monitoring areas for ground sensors. This data is also transmitted back to the control center. Based on the data received from drones and ground sensors, the control center uses graph theory algorithms (such as the minimum spanning tree algorithm and Dijkstra's algorithm) to construct and update the monitoring network topology in real time. This topology intuitively displays the location, connectivity, and data transmission paths of each monitoring node, providing basic data support for subsequent risk monitoring and node adjustments.

[0157] The control center pre-defines a risk assessment model that considers factors such as the type of monitoring data, numerical trends, and historical data comparisons to categorize monitoring areas into risk levels (e.g., low risk, medium risk, high risk). The control center continuously analyzes data transmitted from each monitoring node and assesses the risk level of each area in real time.

[0158] When the control center detects an increase in the local risk level in a certain area (for example, from low risk to high risk), it automatically triggers an adaptive adjustment algorithm for the monitoring point density. This algorithm first calculates the number and locations of additional mobile monitoring nodes required in the target area based on the scope of the risk area, the risk type, and the distribution of existing monitoring nodes. The control center then sends instructions to the drone swarm, dispatching some drones as mobile monitoring nodes to fly to the target area, hover at designated locations, or fly along specific trajectories to collect data in real time, thereby increasing the monitoring density of the target area. Simultaneously, the control center evaluates monitoring nodes in low-risk areas and shuts down redundant nodes (i.e., ground sensor nodes with high data similarity and minimal impact on overall monitoring) to reduce unnecessary energy consumption. During the node adjustment process, the control center continuously monitors the network status to ensure the stability and integrity of data transmission.

[0159] The coordinated inspections of drone swarms and ground sensors achieve comprehensive, all-encompassing coverage of the monitored area. Drones can quickly reach areas difficult for ground sensors to reach (such as mountainous areas, rooftops of tall buildings, and hazardous areas) for data collection, thus filling in the gaps in ground sensors' monitoring. Furthermore, the real-time monitoring network topology accurately reflects the operating status and data transmission paths of each monitoring node. When the local risk level increases, node adjustments are swiftly triggered, and mobile monitoring nodes are added to the target area. This makes monitoring data more intensive and accurate, enabling timely capture of subtle changes in risk areas. This provides reliable data support for risk warning and response, effectively improving the accuracy and timeliness of monitoring. Furthermore, this reduces equipment operating time, maintenance costs, and replacement frequency, further reducing operating costs and improving the cost-effectiveness and sustainability of the monitoring system.

[0160] In summary, the present invention constructs a three-dimensional geographic model of the monitored farmland and performs grid division based on soil type, crop distribution and terrain data, which can accurately subdivide the farmland into multiple agricultural monitoring grids, laying the foundation for subsequent refined monitoring. After obtaining multi-source agricultural data, the agricultural risk level of each grid is predicted with the help of a machine learning model, and the agricultural monitoring grids with risk levels exceeding a preset threshold can be efficiently screened out as target early warning areas, and the risk sources and initial monitoring points can be accurately identified, thus changing the previous problems of one-sided agricultural assessments and inaccurate positioning of risk areas. By establishing a multi-objective decision-making model and using an improved genetic algorithm to solve it, the optimized monitoring points are determined. This process fully considers key factors such as the spatial coverage of the monitoring points, data acquisition accuracy and deployment cost. Compared with the traditional arbitrary or unreasonable monitoring point layout, it greatly improves the scientificity and rationality of the monitoring network, effectively controlling costs while ensuring data quality. By simulating the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, and calculating stability indicators based on early warning accuracy, response time, and cost volatility, we screened out optimized monitoring points with stability indicators higher than preset values ​​as the final agricultural condition early warning nodes and generated a dynamic monitoring network. This significantly enhanced the reliability and stability of the monitoring system under complex and extreme conditions, and enabled it to respond to changes in agricultural conditions more promptly and accurately, providing strong guarantees for agricultural production, and realizing the transformation from traditional extensive agricultural condition monitoring and early warning to a precise, efficient, intelligent, and stable modern agricultural condition monitoring and early warning model.

[0161] See also Figure 3 In one embodiment, the present invention further provides an intelligent agricultural monitoring and early warning system, the system comprising:

[0162] The grid division unit 100 is used to construct a three-dimensional geographic model of the monitored farmland, divide the model into grids based on soil type, crop distribution and terrain data, and generate multiple agricultural monitoring grids;

[0163] The monitoring point screening unit 200 is used to obtain agricultural data from multiple sources, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. It uses a machine learning model to predict the agricultural risk level of each grid, selects agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identifies risk sources and initial monitoring points within the target warning areas.

[0164] The monitoring point optimization unit 300 is used to establish a multi-objective decision-making model with the spatial coverage, data collection accuracy, and deployment cost of the initial monitoring points as optimization objectives. The constraints include the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. The improved genetic algorithm is used to solve the model to determine the optimized monitoring points.

[0165] The early warning node determination unit 400 is used to simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate stability indicators based on early warning accuracy, response time and cost volatility, select optimized monitoring points with stability indicators higher than preset values ​​as final agricultural condition early warning nodes, and generate a dynamic monitoring network.

[0166] It can be understood that the functions or modules included in the system provided in this embodiment can be used to execute the method described in the above method embodiment. Its specific implementation can refer to the description of the above method embodiment. For the sake of brevity, it will not be repeated here.

[0167] The present invention also provides an electronic device, including a processor and a memory, wherein the memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device executes a method as described in any one of the possible implementation modes.

[0168] The present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions, and when the program instructions are executed by a processor of an electronic device, the processor executes a method as described in any one of the possible implementation methods described above.

[0169] See also Figure 4 , Figure 4 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention.

[0170] The electronic device 2 includes a processor 21, a memory 22, an input device 23, and an output device 24. The processor 21, memory 22, input device 23, and output device 24 are coupled via a connector, which may include various interfaces, transmission lines, or buses, etc., although this is not limited in the present embodiment. It should be understood that in various embodiments of the present invention, coupling refers to interconnection in a specific manner, including direct connection or indirect connection through other devices, such as various interfaces, transmission lines, buses, etc.

[0171] The processor 21 may be one or more graphics processing units (GPUs). If the processor 21 is a GPU, the GPU may be a single-core GPU or a multi-core GPU. Alternatively, the processor 21 may be a processor group consisting of multiple GPUs, with the multiple processors coupled to each other via one or more buses. Alternatively, the processor may be another type of processor, and this is not limited in this embodiment of the present invention.

[0172] The memory 22 can be used to store computer program instructions and various computer program codes, including program codes for executing the embodiments of the present invention. Optionally, the memory includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and is used for related instructions and data.

[0173] The input device 23 is used to input data and / or signals, and the output device 24 is used to output data and / or signals. The input device 23 and the output device 24 can be independent devices or an integrated device.

[0174] It is understandable that in the embodiment of the present invention, the memory 22 is not only used to store relevant instructions, and the embodiment of the present invention does not limit the specific data stored in the memory.

[0175] It is understandable that Figure 4 Only a simplified design of an electronic device is shown. In actual applications, the electronic device may further include other necessary components, including but not limited to any number of input / output devices, processors, memories, etc., and all video analysis devices that can implement the embodiments of the present invention are within the scope of protection of the present invention.

[0176] Those skilled in the art will appreciate that the units and algorithm steps of each example 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 performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0177] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here. Those skilled in the art will also clearly understand that the descriptions of the various embodiments of the present invention have different focuses. For the convenience and brevity of description, the same or similar parts may not be repeated in different embodiments. Therefore, for parts not described or not described in detail in one embodiment, reference can be made to the descriptions of other embodiments.

[0178] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0179] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0180] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0181] In the above embodiments, all or part of the embodiments can be implemented using software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments can be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are fully or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted via the computer-readable storage medium. The computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible by a computer or a data storage device such as a server or data center that integrates one or more available media. The available medium may be a magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a digital versatile disc (DVD)), or a semiconductor medium (eg, a solid state disk (SSD)).

[0182] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. An intelligent agricultural monitoring and early warning method, characterized in that: The method comprises: Construct a three-dimensional geographic model of the monitored farmland, divide the model into grids based on soil type, crop distribution, and terrain data, and generate multiple agricultural monitoring grids; Acquire multi-source agricultural data, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. Use machine learning models to predict the agricultural risk level of each grid. Select agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identify risk sources and initial monitoring points within the target warning areas. A multi-objective decision-making model was established, with the spatial coverage of initial monitoring points, data collection accuracy, and deployment cost as optimization objectives. The constraints included the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. An improved genetic algorithm was used to solve the model and determine the optimal monitoring points. Simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate stability indicators based on early warning accuracy, response time, and cost volatility, select optimized monitoring points with stability indicators higher than preset values ​​as the final agricultural condition early warning nodes, and generate a dynamic monitoring network.

2. The intelligent agricultural monitoring and early warning method according to claim 1 is characterized in that: The machine learning model is used to predict the agricultural risk level of each grid, including: Constructing a feature set, the feature set including soil pH value, nitrogen, phosphorus and potassium content, cumulative precipitation, temperature fluctuation rate, crop disease resistance index and historical disaster frequency; The risk prediction model was trained using a random forest algorithm, and model parameters were optimized through cross-validation to output the probability distribution of drought risk, waterlogging risk, and pest and disease risk for each grid. A comprehensive risk level is generated based on the weighted summation of risk probabilities, and high, medium and low risk areas are divided, and high-risk areas are mapped to target warning areas.

3. The intelligent agricultural monitoring and early warning method according to claim 1 is characterized in that: The construction of the multi-objective decision-making model includes defining a single objective function, including: Maximize spatial coverage: ； Where, is the spatial coverage, For the The effective coverage area of ​​each monitoring point reflects the physical range within which the monitoring point can collect data. Monitoring points, 、 is the terrain attenuation coefficient, which is used to quantify the weakening effect of terrain complexity (such as slope and vegetation density) on the coverage effect; The base area threshold is used to adjust the sensitive interval of the logic function to ensure the nonlinear mapping characteristics of the coverage rate. Maximize data accuracy: ； Where, is the data accuracy, is the sensor resolution; is the Euclidean distance between the monitoring point and the risk source; is the data calibration frequency, 、 is the weight factor; Minimize deployment costs: ； Where, For deployment cost, For hardware costs, is the communication link cost, For maintenance costs, is the cost coefficient; The above single objective function is integrated into a multi-objective decision model through fuzzy analytic hierarchy process, and constraints are added: ； ； ； Where, For the warning response time, The upper limit of the warning response time, is the total cost threshold, This is the minimum data accuracy requirement.

4. The intelligent agricultural monitoring and early warning method according to claim 1 is characterized in that: The extreme scenarios include: Meteorological extreme events: spatiotemporal distribution patterns of persistent drought, heavy rain and flooding, and low temperature and freezing damage; Biological disaster events: the coupling relationship between pest and disease spread paths, egg hatching rates, and crop susceptible periods; Human intervention events: simulation of soil salinization caused by irrigation interruption and excessive fertilization.

5. The intelligent agricultural monitoring and early warning method according to claim 4 is characterized in that: The simulation optimization of agricultural conditions at monitoring points under extreme scenarios includes: Construct a multidimensional scenario library based on extreme scenarios; Based on computational fluid dynamics and agent-based models, the propagation trajectory of risk sources in various scenarios is dynamically simulated, and the data capture efficiency and warning delay of monitoring nodes are evaluated to generate simulation results. Calculate the stability index based on the simulation results: ； Where, For stability score, For the early warning accuracy, is the average response time, is the cost volatility; is the normalized weight.

6. The intelligent agricultural monitoring and early warning method according to claim 1 is characterized in that: The method further comprises: Through the coordinated inspection of drone clusters and ground sensors, the monitoring network topology is updated in real time; When a jump in the local risk level is detected, the adaptive adjustment algorithm of the monitoring point density is automatically triggered, and mobile monitoring nodes are added to the target area. At the same time, redundant nodes in low-risk areas are shut down to reduce energy consumption.

7. An intelligent agricultural monitoring and early warning system, characterized in that: The system comprises: The gridding unit is used to construct a three-dimensional geographic model of the monitored farmland. The model is gridded based on soil type, crop distribution and terrain data to generate multiple agricultural monitoring grids. The monitoring point screening unit is used to obtain agricultural data from multiple sources, including soil moisture, meteorological indicators, crop growth stages, and historical pest and disease data. It uses machine learning models to predict the agricultural risk level of each grid, selects agricultural monitoring grids with risk levels exceeding a preset threshold as target warning areas, and identifies risk sources and initial monitoring points within the target warning areas. The monitoring point optimization unit is used to establish a multi-objective decision-making model with the spatial coverage, data collection accuracy, and deployment cost of the initial monitoring points as optimization targets. The constraints include the maximum response distance between the monitoring points and the risk source, the minimum data update frequency, and the cost upper limit. An improved genetic algorithm is used to solve the model to determine the optimal monitoring points. The early warning node determination unit is used to simulate the evolution of agricultural conditions at optimized monitoring points under extreme scenarios, calculate the stability index based on the early warning accuracy, response time and cost volatility, select the optimized monitoring points with stability index higher than the preset value as the final agricultural condition early warning nodes, and generate a dynamic monitoring network.

8. An electronic device, characterized in that: include: A processor and a memory, the memory being used to store computer program code, the computer program code comprising computer instructions, and when the processor executes the computer instructions, the electronic device executes the intelligent agricultural monitoring and early warning method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes the intelligent agricultural monitoring and early warning method according to any one of claims 1 to 6.

Images