Multi-factor-based intelligent pest and disease warning system for rice and citrus

Abstract

The present application relates to the field of agricultural pest warning technology, and discloses a multi-factor-based intelligent pest warning system for rice and citrus, comprising a regional weather station interface, a data initialization module, a micro-meteorological calculation module, a thermodynamic deduction module, a hierarchical scheduling module, a hardware interlocking module and a plant protection unmanned aerial vehicle execution terminal. The system combines macro-weather flow data with a digital elevation model to deduce micro-environment meteorological sequences; calculates canopy wet retention time and evaporation decay slope to generate warning signals and scheduling instructions; the hardware interlocking module sends control intervention parameters to the plant protection unmanned aerial vehicle execution terminal for operation; and the micro-meteorological calculation module uses attitude telemetry data returned by the unmanned aerial vehicle to extract wind resistance angle true values and update wind speed decay factors. The present application solves the problems of insufficient representativeness of macro-weather data and the disconnection between the warning system and pesticide application equipment, and realizes dynamic calibration of the weather deduction model.

Description

Technical Field

[0001] This invention relates to the field of agricultural pest and disease early warning technology, specifically to a multi-factor-based intelligent early warning system for rice and citrus pests and diseases. Background Technology

[0002] The occurrence of pests and diseases in rice and citrus is related to environmental meteorological conditions. Existing early warning systems mostly rely on data provided by regional meteorological stations to determine the conditions that induce pests and diseases and guide subsequent pesticide application. However, meteorological stations typically obtain macro-meteorological data, while rice and citrus growing areas have undulating terrain, and macro-meteorological data cannot reflect the changes in the micro-meteorological environment within complex local plots. Due to the lack of downscaling processing for topographic spatial features, directly applying macro-meteorological data leads to insufficient representativeness of local micro-environment meteorological parameters.

[0003] Meanwhile, existing agricultural early warning systems only output risk alerts to terminals, and the early warning platform lacks data linkage with the underlying plant protection and pesticide application equipment. When a disease warning occurs, the warning data cannot be directly converted into control commands for the hardware. In undulating terrain and when the crop canopy surface is moist, pesticides applied by agricultural drones using fixed parameters can physically run off, affecting the actual control effectiveness.

[0004] Furthermore, existing meteorological forecasting models mostly employ static environmental attenuation parameters, failing to utilize the physical state data of agricultural drones in actual operating environments, and the system lacks data collection channels for the real wind field environment near the ground. This results in a lack of dynamic data feedback mechanisms in the meteorological forecasting algorithms, making it impossible to perform reverse iteration and closed-loop calibration of parameters such as wind speed attenuation during the forecasting process. The resulting prediction bias reduces the accuracy of the early warning system. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a multi-factor-based intelligent early warning system for rice and citrus pests and diseases. It solves the problems of insufficient representativeness of large-area meteorological data in complex field applications, the disconnect between the early warning system and the application equipment, and the lack of dynamic calibration of meteorological simulation models, which leads to low early warning accuracy.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solution: a multi-factor-based intelligent early warning system for rice and citrus pests and diseases, including a regional meteorological station interface, a data initialization module deployed in a cloud processing center, a micro-meteorological calculation module, a thermodynamic deduction module, a hierarchical scheduling module, a hardware interlock module, and a plant protection drone execution terminal.

[0007] The regional meteorological station interface is used to acquire macro-meteorological flow data for the target area.

[0008] The data initialization module is used to receive macroscopic meteorological flow data and extract and generate terrain elevation parameters by combining them with digital elevation models.

[0009] The micro-meteorological calculation module is used to extrapolate and generate micro-environmental meteorological sequences based on topographic elevation parameters and macro-meteorological flow data.

[0010] The thermodynamic deduction module is used to calculate and output the canopy wet retention time and evaporation attenuation slope based on the difference of the microenvironment meteorological sequence.

[0011] The hierarchical scheduling module is used to compare the canopy wet retention time and evaporation attenuation slope with thresholds to generate multi-factor disease early warning signals and hierarchical scheduling instructions.

[0012] The hardware interlock module is used to issue control intervention parameters through the plant protection drone cloud platform based on multi-factor disease early warning signals and hierarchical scheduling instructions.

[0013] The agricultural drone execution terminal is used to operate over the target plot based on control intervention parameters and transmit attitude telemetry data streams back.

[0014] The micro-meteorological calculation module is also used to extract the true value of the fuselage wind resistance tilt angle using attitude telemetry data stream and update the wind speed attenuation factor.

[0015] Preferably, during the data initialization phase, the data initialization module extracts the local elevation matrix corresponding to the location of the target plot from the digital elevation model.

[0016] Perform spatial surface fitting operation on the local elevation matrix to generate topographic elevation parameters that include the geometric center slope and the base aspect angle of the target plot's center point.

[0017] Historical interaction logs stored on farmers' mobile terminals associated with target plots corresponding to terrain elevation parameters were extracted, and historical prevention and control response was calculated using the historical response evaluation formula.

[0018] The above process extracts parameters that characterize the undulation features of the plots and quantifies the execution efficiency of farmers, providing a digital benchmark for subsequent simulation and scheduling.

[0019] Preferably, when generating the microenvironmental meteorological sequence, the micrometeorological calculation module extracts the macro wind direction angle from the macro meteorological flow data and the basic slope aspect angle from the topographic elevation parameters, and calculates the initial wind speed attenuation factor using the initial wind speed attenuation factor calculation formula.

[0020] The macro-average wind speed is extracted from the macro-meteorological flow data, and the local micro-wind speed is calculated using the local micro-wind speed extrapolation formula in combination with the initial wind speed attenuation factor.

[0021] The solar radiation reduction ratio of the target plot is determined based on the solar altitude angle at the current moment, and the macro air temperature and macro air relative humidity in the macro meteorological flow data are offset compensated to output the calibration temperature and calibration humidity.

[0022] The local micro-wind speed, calibrated temperature, calibrated humidity, and cumulative solar radiation from macro-meteorological flow data are aligned and encapsulated according to time-domain characteristics to generate a micro-environment meteorological sequence.

[0023] The above process downscales macroscopic meteorological factors to micro-topographic environments, compensating for the impact of topographic undulations on wind fields and sunshine.

[0024] Preferably, in the thermodynamic deduction stage, the thermodynamic deduction module analyzes the calibration temperature and calibration humidity in the microenvironment meteorological sequence and calculates the surface dew point temperature using the surface dew point temperature conversion formula.

[0025] The surface dew point temperature is compared with the real-time calibration temperature using logic.

[0026] When the calibration temperature is determined to be lower than or equal to the surface dew point temperature, the crop canopy of the target plot is determined to have entered the dew formation stage, and time-domain integration is started simultaneously to record the dew formation duration and intensity.

[0027] Subsequently, the thermodynamic evaporation rate was calculated using the thermodynamic evaporation rate calculation formula.

[0028] The initial liquid water thickness accumulated during the condensation stage is used as the starting value, and the loss depth obtained by converting the thermodynamic evaporation rate is deducted hourly according to the preset time step.

[0029] When the remaining liquid water thickness decreases to zero, the time count stops and the canopy wet residence time is output. The evaporation decay slope is calculated using the evaporation decay slope calculation formula to obtain the canopy wet residence time and evaporation decay slope.

[0030] The above process is based on the principle of phase transition and energy balance to complete differential deduction and obtain the quantitative transformation relationship between meteorological environment and physical conditions for pathogen induction.

[0031] Preferably, during the hierarchical scheduling phase, the hierarchical scheduling module generates a multi-factor disease early warning signal when the acquired canopy wet retention time is greater than or equal to a preset retention time threshold and the evaporation attenuation slope is less than or equal to a preset evaporation slope upper limit.

[0032] When the historical control response is determined to be less than or equal to the preset response grading benchmark, a standardized plant protection operation order containing specific operation parameters is generated based on the geographic coordinate information of the target plot and the category of multi-factor disease early warning signal.

[0033] The system encapsulates and transmits multifactor disease early warning signals and instructions containing standardized plant protection operation orders, generating multifactor disease early warning signals and hierarchical scheduling instructions.

[0034] The preset residence time threshold and preset evaporation slope upper limit are pre-calibrated based on the biological germination experimental data of the target pathogen.

[0035] When the historical response rate is determined to be greater than the response rate grading benchmark, an agricultural efficacy guidance suggestion containing disease type, risk level and recommended application window period is generated and pushed to the corresponding linked farmer's mobile terminal.

[0036] The above process filters out the conditions that induce disease and classifies and distributes intervention resources based on the historical behavior data of the operators.

[0037] Preferably, in parameter control and interlock execution, the hardware interlock module extracts the geometric center slope from the terrain elevation parameters and the canopy wet retention time from the evaporation attenuation slope, normalizes them, and then uses preset weighting coefficients to perform weighted summation to calculate the loss risk index.

[0038] When the risk index of leakage is determined to be greater than or equal to the preset safety boundary, a low-level electrical limiting command is generated, which includes a forced overwrite value for the upper limit of the centrifugal atomizing disc rotation speed and a cutoff parameter for the water pump duty cycle.

[0039] The underlying electrical limiting command is encapsulated into control intervention parameters that include underlying device limiting information.

[0040] The above process constrains the output characteristics of the underlying hardware devices based on the risk assessment results in the cloud, reducing the loss of pesticide solution caused by conventional spraying in wet and steep terrain conditions.

[0041] Preferably, during the data closed-loop calibration phase, the agricultural drone execution terminal receives control intervention parameters and starts the pesticide application operation on the target plot under limited physical output conditions, and collects the flight attitude and position information of the airborne sensors in real time to generate attitude telemetry data streams to be transmitted back to the cloud processing center.

[0042] The micro-meteorological calculation module receives attitude telemetry data streams and performs data cleaning, removing data from the maneuvering disturbance range where the rate of change of acceleration exceeds the preset range, and extracting wind steady-state data that meets the dynamic equilibrium conditions.

[0043] The true value of the fuselage wind resistance tilt angle is extracted from the wind measurement steady-state data, and the true value of the measured micro-wind speed at the corresponding moment is extracted using the wind resistance tilt angle mapping micro-wind speed formula.

[0044] The macro-average wind speed in the macro-meteorological flow data is called up, and the initial wind speed attenuation factor in the extrapolation process is updated in a closed loop using the measured micro-wind speed true value and the inverse iterative formula of the wind speed model.

[0045] The above process uses the dynamic characteristics of UAV operation to infer the actual near-surface micro-wind field, performs reverse iteration on the static attenuation parameters of the meteorological inference algorithm, and completes the calculation and calibration of prediction bias.

[0046] This invention provides an intelligent early warning system for rice and citrus diseases and pests based on multiple factors. It has the following beneficial effects:

[0047] 1. This invention extracts terrain elevation parameters through a data initialization module combined with a digital elevation model, and uses a micro-meteorological calculation module to downscale macro-meteorological flow data to calculate a micro-environmental meteorological sequence including local micro-wind speed, calibrated temperature, and calibrated humidity. This structure transforms large-area meteorological data into local meteorological parameters that conform to the undulation state of the target plot, solving the problem of insufficient representativeness of macro-meteorological data when directly applied in complex terrain, and improving the reliability of environmental data extraction.

[0048] 2. This invention generates standardized plant protection operation orders based on canopy moisture retention time and historical control response through a hierarchical scheduling module and a hardware interlock module. It also generates control intervention parameters, including centrifugal atomizing disc rotation speed and water pump duty cycle limit information, based on the loss risk index, and directly sends these parameters to the plant protection drone execution terminal. This method establishes a direct link between cloud-based early warning commands and the hardware output of the underlying pesticide application equipment, solving the problem of disconnect between the early warning system and the execution end, and reducing the physical loss of pesticide solution in complex terrain and moist canopies.

[0049] 3. This invention utilizes the attitude telemetry data stream transmitted back from the agricultural drone's execution terminal during operation. The micro-meteorological calculation module extracts the true value of the fuselage's wind resistance angle, which meets the dynamic equilibrium conditions, and converts it into the measured true value of the micro-wind speed. This is then used to perform a closed-loop update of the initial wind speed attenuation factor during the simulation process. This process leverages the actual physical characteristics of the spraying equipment during operation to perform inverse iteration and calibration of the static parameters of the meteorological simulation model, solving the problem of warning deviations caused by the lack of dynamic feedback in meteorological simulation algorithms. Attached Figure Description

[0050] Figure 1 This is an architecture diagram of the intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to the present invention.

[0051] Figure 2 This is a flowchart illustrating the workflow of the intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to the present invention.

[0052] Figure 3This is a comparison chart of the calibration effects of the wind speed model of the present invention;

[0053] Figure 4 This is a graph showing the relationship between drug application risk and duty cycle limit in this invention. Detailed Implementation

[0054] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] See attached document Figure 1 This invention provides a multi-factor-based intelligent early warning system for rice and citrus pests and diseases, including a cloud processing center, a regional meteorological station interface, a plant protection drone cloud platform, a plant protection drone execution terminal, and a farmer's mobile terminal.

[0056] The cloud processing center establishes data transmission connections with regional meteorological stations, agricultural drone cloud platforms, and farmers' mobile terminals via communication networks. Internally, the cloud processing center is equipped with data initialization, micro-meteorological calculation, thermodynamics simulation, hierarchical scheduling, and hardware interlocking modules.

[0057] The regional meteorological station interface communicates with the cloud processing center and is used to acquire macro-meteorological flow data of the target area.

[0058] The data initialization module receives macro-meteorological flow data and extracts the elevation characteristics of rice and citrus target plots using a digital elevation model to generate topographic elevation parameters.

[0059] The micro-meteorological calculation module receives topographic elevation parameters and macro-meteorological flow data, performs spatial interpolation and extrapolation, and generates micro-environmental meteorological sequences at specific altitudes.

[0060] The thermodynamic extrapolation module acquires the microenvironment meteorological sequence, and performs surface dew point temperature calculation and time-domain integration based on it, extrapolating and outputting the canopy wet retention time and evaporation attenuation slope.

[0061] The hierarchical scheduling module compares the physical thresholds of pests and diseases based on the canopy moisture retention time and evaporation attenuation slope, and generates multi-factor disease early warning signals and hierarchical scheduling instructions.

[0062] Based on multi-factor disease early warning signals and hierarchical scheduling instructions, the hardware interlock module sends control intervention parameters containing underlying equipment limitation information to the agricultural drone execution terminal via the agricultural drone cloud platform.

[0063] The agricultural drone's execution terminal performs spraying operations over the target plot based on control intervention parameters, and transmits attitude telemetry data streams back to the cloud processing center during the operation. The micro-meteorological calculation module uses the attitude telemetry data stream to extract the true value of the fuselage's wind resistance tilt angle and performs reverse iterative updates to the wind speed attenuation factor during the simulation process.

[0064] See attached document Figure 2 This invention provides a multi-factor-based intelligent early warning method for rice and citrus pests and diseases, comprising the following steps:

[0065] S1, Multi-source data initialization, the cloud processing center calls the regional meteorological station interface to obtain macro-meteorological flow data over a large area, and combines the elevation model to extract the topographic elevation parameters of the target plots of rice and citrus.

[0066] S2, Micro-meteorological forward extrapolation: The micro-meteorological calculation module receives macro-meteorological flow data and topographic elevation parameters, performs spatial interpolation and offset calculation, and generates micro-environmental meteorological sequences within the plot area;

[0067] S3, thermodynamic differential calculation, the thermodynamic deduction module calculates the dew point based on the microenvironment meteorological sequence and performs time-domain accumulation, and outputs the canopy wet retention time and evaporation attenuation slope;

[0068] S4, Early Warning Triggering and Scheduling: The hierarchical scheduling module logically compares the canopy moisture retention time and evaporation attenuation slope with the pathogen germination conditions to generate multi-factor disease early warning signals and standardized plant protection operation orders.

[0069] S5, parameter interlocking and calibration closed loop, the hardware interlocking module sends the underlying electrical limiting command to the plant protection drone execution terminal based on the multi-factor disease early warning signal and the standardized plant protection operation work. The plant protection drone execution terminal operates in the limited state and sends back the attitude telemetry data stream. The micro-meteorological calculation module uses the attitude telemetry data stream to extract the wind steady state and perform reverse calibration of the wind speed model.

[0070] The following section will provide a detailed explanation of each core step in the above workflow and its implementation principles, taking into account specific application scenarios.

[0071] In this embodiment, a multi-source data initialization operation is performed by a data initialization module deployed in a cloud processing center, aiming to establish the digital benchmark required for subsequent simulations. The specific execution process of step S1 is as follows:

[0072] In step S101, the data acquisition phase, the cloud processing center calls the regional meteorological station interface to obtain macroscopic meteorological flow data within the target area. This macroscopic meteorological flow data specifically covers the macroscopic average wind speed, macroscopic wind direction angle, macroscopic air temperature, macroscopic relative humidity, and cumulative solar radiation within the region. The data initialization module aligns the received meteorological factors based on a unified time reference and performs outlier removal. The specific implementation of using standard communication protocols to call the meteorological platform interface to obtain external environmental characteristic data can be achieved by those skilled in the art using conventional network communication technologies, and will not be elaborated upon here.

[0073] S102, after data alignment is completed, the data initialization module imports the digital elevation model (DEM) data and performs a spatial masking operation using a pre-stored boundary coordinate set to extract the local elevation matrix corresponding to the target plot's location from the DEM. The data initialization module then generates terrain elevation parameters characterizing the plot's geometric undulations by performing spatial surface fitting on the local elevation matrix. Preferably, these terrain elevation parameters include the geometric center slope and the base aspect angle at the target plot's center point. These parameters assist the subsequent micrometeorological calculation module in correcting for wind speed and sunshine duration based on terrain undulations during simulations.

[0074] S103, the data initialization module retrieves the crop attribute categories associated with the target plot and configures the lower boundary conditions for meteorological simulation based on a preset mapping relationship. In this embodiment, the crop attribute categories are distinguished as rice and citrus. When the crop attribute category is identified as rice, the data initialization module assigns an initial aerodynamic roughness value to the plain depression of the plot; when the crop attribute category is identified as citrus, the data initialization module assigns an initial aerodynamic roughness value to the hilly slope of the plot. By converting the morphological characteristics of different crops into corresponding roughness values, physical support is provided for the generation of microenvironment meteorological sequences.

[0075] S104, To quantify farmers' execution efficiency, the data initialization module extracts historical interaction logs from the cloud processing center of farmers' mobile terminals associated with the target plots corresponding to the terrain elevation parameters, and then calculates the historical prevention and control response rate. The historical prevention and control response rate is calculated using a historical response rate evaluation formula, which is:

[0076] ;

[0077] In the formula, Indicates the historical response rate to prevention and control; This represents the total number of historical warning events, and its value is an integer greater than or equal to 1. Indicates the first The timestamp for each historical pesticide scan and release is in seconds. Indicates the first The timestamp of each historical early warning issuance is in seconds; Indicates the first The single-response delay time is measured in seconds. Indicates the summation symbol; This indicates the cumulative response delay time, measured in seconds.

[0078] In this embodiment, if the denominator of the calculated cumulative response lag time is zero, the data initialization module will set the historical prevention and control response rate to a system-preset upper limit constant. The calculated historical prevention and control response rate will be called by the hierarchical scheduling module as the decision-making basis for issuing technical recommendations or generating work orders.

[0079] Through the implementation of step S1 above, the system has completed the multi-dimensional data integration of the physical environment, meteorological fluids, and management entities, providing the necessary input parameters for the subsequent forward extrapolation of meteorological offsets by the micro-meteorological calculation module.

[0080] The following section will further explain in detail how the micro-meteorological calculation module uses the above parameters to perform spatial interpolation and inverse calibration, in conjunction with the physical kinematic model.

[0081] In this embodiment, the micro-meteorological computing module deployed in the cloud processing center performs forward extrapolation of micro-topographic meteorological offsets. By establishing a downscaling mapping relationship between topographic physical features and meteorological fluids, a micro-environmental meteorological sequence with plot-scale characteristics is generated. The specific execution logic of step S2 is described as follows:

[0082] S201, for scenarios where the spatial wind field is constrained by terrain, the micro-meteorological calculation module receives macro-meteorological flow data and terrain elevation parameters transmitted by the data initialization module, and constructs a local wind speed attenuation model accordingly. In this model, the degree of attenuation of macro-airflow after entering the canopy of the target plot is determined by the spatial geometric angle between the wind direction and the slope aspect. The micro-meteorological calculation module extracts the macro-wind direction angle from the macro-meteorological flow data and the basic slope aspect angle from the terrain elevation parameters, and calculates the initial wind speed attenuation factor using the initial wind speed attenuation factor calculation formula. The module's initial wind speed attenuation factor calculation formula is as follows:

[0083] ;

[0084] In the formula, This represents the initial wind speed attenuation factor; This indicates the macroscopic wind direction angle, and its unit is degrees. This indicates the slope angle of the foundation, and its unit is degrees; The terrain closure coefficient is a constant determined by the ratio of the extreme elevation difference of the digital elevation model within a preset range around the target plot to the horizontal distance, and its value ranges from 0 to 1. The absolute value operator is represented by the symbol. Indicates the symbol for cosine trigonometric functions; This represents the terrain transparency coefficient. The initial wind speed attenuation factor, through its cosine projection component, characterizes the degree to which the slope orientation mechanically impedes the windward airflow.

[0085] S202, after obtaining the physical constraints of the spatial distribution of wind speed, the micro-meteorological calculation module performs a localized migration transformation on the regional-scale wind speed. The micro-meteorological calculation module extracts the macro-average wind speed from the macro-meteorological flow data and, combined with the initial wind speed attenuation factor calculated above, calculates the local micro-wind speed using a local micro-wind speed extrapolation formula. The module's local micro-wind speed extrapolation formula is as follows:

[0086] ;

[0087] In the formula, This indicates the local micro-wind speed, and its unit is meters per second; This represents the macroscopic average wind speed, and its unit is meters per second. This represents the initial wind speed attenuation factor. The calculated local micro-wind speed reflects the actual wind dynamic intensity at the top of the crop canopy in the target plot, serving as the kinetic input for subsequent thermodynamic moisture evaporation calculations.

[0088] S203, the micro-meteorological calculation module further performs spatial interpolation offset calculations on temperature and humidity parameters to address the differences in radiation flux caused by terrain shading. In this embodiment, the micro-meteorological calculation module extracts the geometric center slope from the terrain elevation parameters and determines the solar radiation reduction ratio of the target plot based on the current solar altitude angle. The micro-meteorological calculation module reduces the cumulative solar radiation in the macro-meteorological flow data and uses the reduced radiation energy change value as a correction weight to perform offset compensation on macro-air temperature and macro-air relative humidity through a linear interpolation algorithm, thereby outputting calibrated temperature and calibrated humidity corresponding to the terrain distribution of the target plot.

[0089] In this embodiment, the micro-meteorological calculation module aligns and encapsulates the generated local micro-wind speed, calibration temperature, calibration humidity, and received cumulative solar radiation according to time-domain characteristics, generating a micro-environment meteorological sequence covering multi-dimensional meteorological features. This micro-environment meteorological sequence is transmitted to the thermodynamic deduction module, providing an environmental background benchmark for subsequent analysis of crop leaf surface moisture retention characteristics.

[0090] The implementation of step S2, through forward extrapolation logic, realizes the downward mapping of macro-meteorological factors to micro-topographic environment, solves the problem of insufficient representativeness of meteorological station observation data in complex hilly or plot environments, and lays a data foundation for the quantitative extrapolation of disease risk.

[0091] The following section will further explain in detail the implementation principle of how the thermodynamic deduction module uses microenvironmental meteorological sequences to calculate the canopy moisture retention time, taking into account the physical characteristics of plant pathology.

[0092] In this embodiment, the thermodynamic extrapolation module deployed in the cloud processing center receives the microenvironmental meteorological sequence output by the micrometeorological calculation module and performs differential extrapolation based on the principle of water phase change and the energy balance equation. The specific execution process of step S3 is as follows:

[0093] S301, as a prerequisite for physical state determination, the thermodynamic deduction module analyzes the calibration temperature and calibration humidity in the microenvironment meteorological sequence. Considering that water vapor condensation in the air is constrained by the thermal state of the Earth's surface, the thermodynamic deduction module uses the surface dew point temperature conversion formula to calculate the surface dew point temperature. The module's surface dew point temperature conversion formula is:

[0094] ;

[0095] In the formula, This indicates the surface dew point temperature, and its unit is degrees Celsius. This indicates the calibrated humidity, expressed as a percentage. This indicates the calibration temperature, and its unit is degrees Celsius. This represents the first Magnus constant, which is 17.27 in this embodiment; This represents the second Magnus constant, which in this embodiment is 237.7; This represents the natural logarithm operator.

[0096] The thermodynamics deduction module performs a logical comparison between the calculated surface dew point temperature and the real-time calibration temperature. When the calibration temperature is determined to be lower than or equal to the surface dew point temperature, the thermodynamics deduction module determines that the crop canopy of the target plot has entered the dew formation stage and simultaneously starts time-domain integration to record the dew formation duration and intensity.

[0097] In S302, during the stage where a liquid water film exists on the crop surface, the thermodynamic simulation module, combining wind dynamics and radiation energy data from the microenvironment meteorological sequence, determines the dissipation rate by simulating the mass transfer process of water from the canopy to the atmosphere. The thermodynamic simulation module uses the thermodynamic evaporation rate calculation formula to calculate the thermodynamic evaporation rate. The module's thermodynamic evaporation rate calculation formula is as follows:

[0098] ;

[0099] In the formula, It represents the thermodynamic evaporation rate, and its unit is millimeters per hour; This indicates the local micro-wind speed, and its unit is meters per second; This represents the saturated vapor pressure, and its unit is kilopascal (kPa). This indicates the actual water vapor pressure, and its unit is kilopascal (kPa). This represents the saturated water vapor pressure difference, and its unit is kilopascal (kPa). This represents the cumulative solar radiation, and its unit is watts per square meter. This represents the wind speed sensing coefficient, which is a preset empirical constant. This represents the radiation conversion factor, used to characterize the contribution rate of a unit of radiation energy to water evaporation.

[0100] In this step, the saturated vapor pressure is derived by the thermodynamic derivation module through the Clausius-Clapeyron equation based on the calibration temperature; the actual vapor pressure is determined by the product of the saturated vapor pressure and the calibration humidity.

[0101] S303, as a closed-loop simulation of water dynamic balance, uses a thermodynamic simulation module to perform time-domain difference calculations to determine the lifespan of liquid water on the leaf surface. The module uses the initial liquid water thickness accumulated during the condensation stage as a starting point and subtracts the loss depth obtained from the thermodynamic evaporation rate hourly according to a preset time step. When the remaining liquid water thickness decreases to zero, the module stops time counting and outputs the canopy wetness retention time. It then calculates the evaporation decay slope using the evaporation decay slope calculation formula to obtain the canopy wetness retention time and the evaporation decay slope.

[0102] The evaporation attenuation slope is calculated using the evaporation attenuation slope calculation formula. The formula for calculating the module evaporation attenuation slope is as follows:

[0103] ;

[0104] In the formula, This represents the evaporation decay slope, used to characterize the rate at which liquid water in the canopy disappears; This represents the thermodynamic evaporation rate at the current moment, and its unit is millimeters per hour. This indicates the thermodynamic evaporation rate at the previous moment, and its unit is millimeters per hour. This indicates the time interval for calculating the step size, and its unit is hours.

[0105] In this embodiment, the canopy wetness retention time and evaporation attenuation slope generated by the thermodynamic deduction module are transmitted to the hierarchical scheduling module. This set of parameters, by describing the actual physical wetness state of the crop receptor surface, avoids the ambiguity of macroscopic temperature and humidity thresholds, and provides a direct basis for determining the risk of pathogen spore germination.

[0106] By executing step S3, the system achieves a quantitative transformation from microenvironmental meteorological data to the physical conditions for pathogen induction. This differential calculation logic based on energy balance can reflect the differentiated characteristics of crop canopy humid environments under different topographical and meteorological backgrounds.

[0107] After acquiring the physical characteristic parameters, the hierarchical scheduling module will conduct multi-factor logical comparisons and early warning decisions based on these parameters. The following section will explain in detail the specific implementation principles of early warning triggering and hierarchical scheduling, in conjunction with specific business logic.

[0108] In this embodiment, the hierarchical scheduling module deployed in the cloud processing center performs multi-factor logical comparison and prevention resource allocation, transforming the physical state parameters inferred from the front end into agricultural production guidance actions. The specific execution process of step S4 is as follows:

[0109] S401, as the core step in early warning determination, the hierarchical scheduling module receives the canopy moisture retention time and evaporation attenuation slope transmitted by the thermodynamic deduction module. Considering the different sensitivities of various pathogens to the microenvironment, the hierarchical scheduling module retrieves pathogen germination conditions pre-stored in the system database. These pathogen germination conditions specifically include a preset retention time threshold and a preset evaporation slope upper limit. In this embodiment, the aforementioned preset retention time threshold and preset evaporation slope upper limit are pre-calibrated and configured based on biological germination experimental data of target pathogens of rice or citrus (e.g., rice blast fungus or citrus canker fungus).

[0110] The hierarchical scheduling module performs a two-factor logical evaluation calculation based on this: when the acquired canopy wet retention time is greater than or equal to a preset retention time threshold, and the acquired evaporation attenuation slope is less than or equal to a preset evaporation slope upper limit, the canopy microenvironment of the target plot is determined to meet the physical window conditions for disease induction, and a corresponding multi-factor disease early warning signal is generated accordingly. If the acquired canopy wet retention time is less than the preset retention time threshold, or the acquired evaporation attenuation slope is greater than the preset evaporation slope upper limit, the hierarchical scheduling module determines that the current environment does not meet the conditions for disease induction, and the system maintains a normalized environmental monitoring state.

[0111] S402, based on generating multi-factor disease early warning signals, the system needs to further plan the entities responsible for implementing control actions. The hierarchical scheduling module retrieves the historical control response rate of the farmers corresponding to the target plots. This historical control response rate is an evaluation index derived by the system through normalized statistics based on the time difference between when the farmer actually started plant protection operations after receiving control guidance and suggestions in the past. It is used to characterize the farmer's initiative in independent intervention.

[0112] By comparing the historical prevention and control response rate with a preset response rate grading benchmark, the grading and scheduling module determines the guidance and recommendation issuance strategy for the target plot. In this embodiment, the preset response rate grading benchmark is a pre-set constant, and its value ranges from 0.6 to 0.85.

[0113] As a preferred approach, if the historical response rate to pest control is determined to be greater than the preset response rate grading benchmark, it indicates that the corresponding farmer has good self-control implementation capability and timeliness. The grading and scheduling module generates agricultural efficacy guidance suggestions including disease type, risk level, and recommended application window period, and pushes these suggestions to the corresponding bound farmer's mobile terminal through the communication network of the cloud processing center. The specific process of the server pushing data packets to the mobile terminal can be implemented by those skilled in the art using conventional mobile terminal information push protocols, which are well-known technologies in the field and will not be elaborated upon here.

[0114] S403 addresses scenarios where there is a risk of delayed historical interventions by farmers. In such cases, the system will initiate an automated resource scheduling process. When the historical response rate is determined to be less than or equal to a preset response rate grading benchmark, the grading scheduling module generates a standardized plant protection operation order containing specific operational parameters based on the geographic coordinates of the target plot and the category of multi-factor disease early warning signals.

[0115] In this embodiment, the standardized plant protection operation order includes the target plot boundary coordinate set, the recommended pesticide type for the current disease, the preset operating speed, and the initial spraying height determined based on the current micro-meteorological conditions. This standardized plant protection operation order is encapsulated and output as a specific carrier of hierarchical scheduling instructions. Subsequently, the hierarchical scheduling module encapsulates and transmits the multi-factor disease early warning signal and the instructions containing the standardized plant protection operation order, generating a multi-factor disease early warning signal and hierarchical scheduling instructions, and transmits them to the hardware interlock module to trigger subsequent parameter control of the plant protection drone execution terminal.

[0116] By executing step S4, the system achieves logical filtering of disease microenvironment inducing conditions and distributes prevention and control strategies in a differentiated manner based on the historical behavior data of the operator, thus completing the closed loop of agricultural prevention and control effectiveness.

[0117] After completing the above-mentioned early warning and scheduling logic, the following will further explain in detail the specific implementation principle of how the hardware interlock module performs the underlying electrical limiting and wind field reverse calibration of the UAV based on the above-mentioned work order and early warning signal, in conjunction with the control principle of physical equipment.

[0118] In this embodiment, through data interaction between the cloud and the execution terminal, the system transforms early warning decisions into physical control constraints on the terminal devices, and utilizes operational feedback from the aircraft to optimize the parameters of the front-end meteorological projection model. The specific execution logic of step S5 is described as follows:

[0119] After receiving the operation instructions from the cloud, the S501 hardware interlock module performs a pesticide application risk assessment based on the multi-factor disease early warning signals and hierarchical scheduling instructions containing standardized plant protection operation work orders generated above.

[0120] As a specific implementation method, the hardware interlock module extracts the geometric center slope and canopy wet retention time from the terrain elevation parameters, and the canopy wet retention time from the evaporation attenuation slope. Considering that in steep terrain and environments with a large amount of liquid water remaining on the leaves, conventional pesticide spraying can easily lead to pesticide runoff and loss, the hardware interlock module normalizes the geometric center slope and canopy wet retention time, and then uses a preset weighting coefficient to perform a weighted summation to calculate the loss risk index. When the loss risk index is determined to be greater than or equal to a preset safety boundary, the system determines that the physical output characteristics of the pesticide application equipment need to be intervened and limited; conversely, if the loss risk index is less than the preset safety boundary, the original operating parameters of the work order are maintained without intervention.

[0121] S502, based on the aforementioned risk assessment results requiring intervention, the hardware interlock module generates a low-level electrical limiting command for the agricultural drone spraying system. In this embodiment, the low-level electrical limiting command includes a forced overwrite value for the upper limit of the centrifugal atomizing disc rotation speed and a cutoff parameter for the water pump duty cycle. These settings, by reducing droplet size and limiting instantaneous spray flow, adapt to the water-holding capacity of the wet canopy. Subsequently, the hardware interlock module encapsulates this low-level electrical limiting command into control intervention parameters containing low-level device limitation information, providing a data carrier for subsequent transmission to the drone.

[0122] In S503, to achieve a closed-loop interlock between hardware and software control, the hardware interlock module sends module control intervention parameters to the agricultural drone execution terminal via the agricultural drone cloud platform. In the terminal control logic, upon receiving these parameters, the flight control system of the agricultural drone execution terminal performs a network handshake and data packet verification mechanism. If the verification fails or communication is interrupted, the flight control system triggers a safety protection mechanism, achieving physical interlock by cutting off the ESC control signal line of the water pump, thus restricting pesticide spraying. After successful verification and obtaining de-control permission, the agricultural drone execution terminal receives the control intervention parameters and initiates pesticide application operations on the target plot under restricted physical output conditions.

[0123] S504, during operation, the agricultural drone's execution terminal collects flight attitude and position information from onboard sensors in real time to generate an attitude telemetry data stream to be transmitted back to the cloud processing center. After this data stream is transmitted to the cloud, due to the drone's acceleration and deceleration maneuvers during operation, the original attitude data contains tilt angle changes caused by maneuvering acceleration. The micro-meteorological calculation module performs data cleaning on this attitude telemetry data stream. The micro-meteorological calculation module removes data from maneuvering disturbance ranges where the rate of acceleration change exceeds a preset range, which is determined by the factory calibration threshold of the flight control system's dynamic response characteristics. After cleaning, the system extracts wind measurement steady-state data that meets the dynamic equilibrium conditions. This completes the wind measurement steady-state extraction process.

[0124] S505, after acquiring the aforementioned steady-state data, the micro-meteorological calculation module uses the extracted data to perform inverse calibration of the wind speed model. In physical terms, when the UAV is in stable flight, the horizontal thrust component generated by its fuselage tilt reaches equilibrium with the horizontal wind resistance experienced by the fuselage. The micro-meteorological calculation module extracts the true value of the fuselage's wind-resistant tilt angle from this data and uses the wind-resistant tilt angle mapping micro-wind speed formula to extract the corresponding measured true value of the micro-wind speed. The module's wind-resistant tilt angle mapping micro-wind speed formula is:

[0125] ;

[0126] In the formula, This represents the true value of the measured micro-wind speed, and its unit is meters per second. This indicates the takeoff mass of the drone, expressed in kilograms. This represents the gravitational acceleration constant, and its unit is meters per second squared. This represents the true value of the fuselage's wind resistance angle, and its unit is degrees; Indicates the symbol for the tangent trigonometric function; This indicates air density, and its unit is kilograms per cubic meter. This indicates the windward reference area of ​​the drone, with the unit being square meters. This represents the drag coefficient, which is the preset calibration constant for the corresponding model; This represents the square root operator.

[0127] After obtaining the above measured true values, the micro-meteorological calculation module performs error convergence on the wind speed spatial mapping relationship in the forward extrapolation process. In this embodiment, the micro-meteorological calculation module calls the macro-average wind speed from the macro-meteorological flow data, and uses the measured micro-wind speed true values ​​to perform closed-loop updates on the initial wind speed attenuation factor in the extrapolation process using the wind speed model inverse iteration formula. The module's wind speed model inverse iteration formula is as follows:

[0128] ;

[0129] In the formula, This indicates the updated wind speed attenuation factor; This represents the initial wind speed attenuation factor; This represents the inverse iterative learning rate, which ranges from 0 to 1 and is used to control the weight step size for a single calibration. This represents the macroscopic average wind speed, and its unit is meters per second.

[0130] Through the closed-loop implementation of step S5, the system uses UAV flight attitude data with spatial positioning to infer the actual near-surface micro-wind field. This mechanism not only ensures the standardized execution of plant protection operations within safe thresholds, but also improves the environmental adaptability of the meteorological inference algorithm through model iteration updates, effectively reducing the calculation deviation of a single static terrain wind speed model.

[0131] To aid in understanding the present invention, a specific application example is provided below, taking into account the actual business scenario of a hilly citrus orchard. Related experimental verification and effect comparison data are also provided to detail the execution process and technical effects of the system's underlying control parameter interlocking and wind measurement model reverse self-calibration closed loop.

[0132] In this specific application scenario, the target area was selected as a contiguous citrus orchard in a hilly terrain. Due to the significant geometric undulations of the area and the presence of noticeable dew on the canopy leaves in the early morning hours of the current season, the system's cloud processing center, after acquiring macro-meteorological flow data, derived a forward-looking analysis that the target area was in a microenvironment characterized by high humidity and some wind disturbance. Based on the analysis results, the hierarchical scheduling module generated multi-factor disease early warning signals and issued hierarchical scheduling instructions containing standardized plant protection operation orders.

[0133] Upon receiving the aforementioned instruction, the hardware interlock module extracts the geometric center slope and canopy moisture retention time of the citrus orchard plot. The system normalizes and weights these two parameters characterizing the physical state, calculating a loss risk index of 0.82. Since this value exceeds the system's preset safety boundary value of 0.70, the hardware interlock module determines that under the current moist, steep slope conditions, conventional high-flow-rate spraying would cause the pesticide solution to slide down the dew-covered citrus leaves and run into the soil, failing to form effective adhesion.

[0134] Based on this assessment, the hardware interlock module generated a low-level electrical limiting command, forcibly locking the duty cycle cutoff limit of the agricultural drone's water pump to 40% and overwriting the upper limit of the centrifugal atomizing disc's rotation speed to a high-speed mode, thereby forcibly adjusting the operation to a low-flow fine-droplet spraying mode. This low-level electrical limiting command was encapsulated as a control intervention parameter and sent to the agricultural drone's execution terminal via the agricultural drone cloud platform. After completing network handshake and verification, the terminal flight control system obtained decontrol permission and started the pesticide application operation in the citrus orchard under restricted physical output conditions.

[0135] During the drone spraying operation, the drone's terminal continuously transmits attitude telemetry data streams back to the cloud processing center. The micro-meteorological computing module receives this data stream and performs data cleaning, filtering out acceleration and deceleration maneuver data when the drone changes lanes at the edge of the plot, and extracting the wind stability data during constant speed and level flight on a straight flight path. Based on telemetry feedback, the true value of the extracted fuselage wind resistance angle is 3 degrees.

[0136] To correct the initial static terrain wind speed model, the micro-meteorological calculation module performs inverse calibration of the wind speed model using the aforementioned steady-state wind measurement data. The module uses the wind resistance tilt angle mapping micro-wind speed formula, substituting it into the current UAV's physical parameters, to calculate the actual wind field intensity at the operating altitude. The module's wind resistance tilt angle mapping micro-wind speed formula is as follows:

[0137] ;

[0138] In the formula, This represents the true value of the measured micro-wind speed, calculated to be 3.73, with units of meters per second. This indicates the takeoff mass of the drone; substitute the specific value 30, where the unit is kilograms. This represents the gravitational acceleration constant. Substituting the specific value 9.8, its unit is meters per second squared. This represents the true value of the fuselage's wind resistance angle. Substitute the specific value 3 into it; its unit is degrees. Indicates the symbol for the tangent trigonometric function; This represents air density; substitute the specific value 1.225, and its unit is kilograms per cubic meter. This represents the reference area of ​​the drone facing the wind. Substitute the model's fixed value of 1.5, and its unit is square meters. This represents the drag coefficient, substituted into the model's fixed constant of 1.2; This represents the square root operator.

[0139] After obtaining the true value of the measured micro-wind speed, the micro-meteorological calculation module calls the macro-average wind speed (value 5.0 m / s) provided by the weather station interface at that moment, and extracts the initial wind speed attenuation factor (value 0.5) used in the forward extrapolation stage. The micro-meteorological calculation module uses the inverse iteration formula of the wind speed model to update this attenuation factor in a closed loop. The inverse iteration formula of the module's wind speed model is as follows:

[0140] ;

[0141] In the formula, This indicates the updated wind speed attenuation factor, which is calculated to be 0.549. This represents the initial wind speed attenuation factor; substitute the specific value 0.5. This represents the reverse iterative learning rate; substitute the set value of 0.2. This represents the true value of the measured micro-wind speed. Substitute the specific value 3.73 into the equation. The unit is meters per second. This represents the macroscopic average wind speed, substituted with the specific value 5.0, with units of meters per second. The updated wind speed attenuation factor is written back to the system database for meteorological projections in subsequent batches of tasks for this plot, achieving autonomous learning and convergence of algorithm parameters.

[0142] To verify the technical effectiveness of this solution, multiple high-precision ultrasonic anemometers were deployed in the target citrus orchard as ground truth references, and comparative experiments were conducted on 20 consecutive operational batches. The experiments recorded the micro-meteorological prediction errors of the system without introducing a reverse calibration mechanism (i.e., relying solely on the initial digital elevation model) and with the introduction of a closed-loop iterative mechanism. Loss risk index and terminal response data were also collected for each batch.

[0143] See attached document Figure 3 ,Should Figure 3 The system intuitively demonstrates the improved prediction accuracy achieved by introducing a drone-based wind measurement model for inverse self-calibration closed-loop. Figure 3 The solid black line with circles represents the actual measured micro-wind speed, the dark gray dashed line represents the wind speed extrapolated from the initial model, and the light gray dotted line represents the wind speed extrapolated after reverse calibration. Due to the lack of local measured data, the dark gray dashed line generated based on static terrain parameters has a systematic deviation from the actual physical environment and cannot reflect the true wind field fluctuations. However, by implementing the parameter interlocking and calibration closed-loop mechanism constructed in this invention, the system uses the attitude telemetry data transmitted back by the agricultural drone during operation to extrapolate the actual wind speed and dynamically updates the wind speed attenuation factor using the wind speed model's reverse iterative formula. As the number of operation batches increases, the light gray dotted line shows a gradual convergence towards the solid black line with circles. The above process uses the drone as a dynamic data acquisition end, overcoming the technical shortcomings of traditional extrapolation algorithms in environmentally unsuitable conditions under complex terrain, improving the system's accuracy in extrapolating micro-meteorological conditions, and meeting the data requirements of pesticide application operations.

[0144] See attached document Figure 4 ,Should Figure 4 It demonstrates the triggering logic and physical constraint behavior of software and hardware collaboration in defense and control. Figure 4 The solid black line with squares represents the loss risk index, and the dark gray dotted line represents the pump duty cycle cutoff limit. Combined with the technical solution of this invention, the system overcomes the bottleneck of previous disconnect between cloud-based decision-making and terminal execution. Figure 4Within a specific batch range, when the target plot's black solid line crosses the preset safety boundary due to significant elevation fluctuations and canopy condensation, the hardware interlock module, based on the operational procedure locking mechanism, directly issues a low-level electrical limiting command. The dark gray dotted line within this range exhibits a step-down to a 40% forced overwrite state, indicating that the flight control system, upon receiving intervention parameters, cuts off full-power output permission, limiting the instantaneous spray flow. This data demonstrates that the forced overwrite and interlock de-control mechanism disclosed in this invention can constrain the underlying hardware equipment in real-time and stably based on cloud-based risk assessment results, preventing pesticide waste and loss caused by improper application, and achieving a complete technical closed loop from information early warning to physical intervention.

Claims

1. A multi-factor-based intelligent early warning system for rice and citrus diseases and pests, characterized in that, include: The regional meteorological station interface is used to acquire macro-meteorological flow data for the target area; The data initialization module deployed in the cloud processing center is used to receive the macro-meteorological flow data and extract and generate terrain elevation parameters in combination with the digital elevation model; The micro-meteorological calculation module is used to deduce and generate micro-environmental meteorological sequences based on the topographic elevation parameters and the macro-meteorological flow data. Thermodynamic deduction module is used to calculate and output the canopy wet retention time and evaporation attenuation slope based on the microenvironment meteorological sequence difference; The hierarchical scheduling module is used to compare the canopy wet retention time and evaporation attenuation slope with thresholds to generate multi-factor disease early warning signals and hierarchical scheduling instructions. The hardware interlock module is used to send control intervention parameters through the plant protection drone cloud platform based on the multi-factor disease early warning signal and hierarchical scheduling instructions. The agricultural drone execution terminal is used to operate over the target plot based on the control intervention parameters and transmit attitude telemetry data streams back. The micro-meteorological calculation module is also used to extract the true value of the fuselage wind resistance tilt angle using the attitude telemetry data stream and update the wind speed attenuation factor.

2. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 1, characterized in that, The data initialization module is specifically used for: Extract the local elevation matrix corresponding to the location of the target plot from the digital elevation model; Perform a spatial surface fitting operation on the local elevation matrix to generate the topographic elevation parameters, which include the geometric center slope and the basic slope aspect angle of the target plot's center point. Historical interaction logs stored on farmers' mobile terminals associated with the target plots corresponding to the terrain elevation parameters are extracted, and historical prevention and control response is calculated using the historical response evaluation formula.

3. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 2, characterized in that, The micro-meteorological calculation module is specifically used to generate the micro-environmental meteorological sequence based on the topographic elevation parameters and the macro-meteorological flow data: Extract the macro wind direction angle from the macro meteorological flow data and the basic slope aspect angle from the topographic elevation parameters, and calculate the initial wind speed attenuation factor using the initial wind speed attenuation factor calculation formula. Extract the macro average wind speed from the macro meteorological flow data, combine it with the initial wind speed attenuation factor to calculate the local micro wind speed using the local micro wind speed extrapolation formula, and determine the solar radiation reduction ratio of the target plot based on the solar altitude angle at the current moment. Perform offset compensation on the macro air temperature and macro air relative humidity in the macro meteorological flow data to output the calibration temperature and calibration humidity. The local micro-wind speed, the calibrated temperature, the calibrated humidity, and the cumulative solar radiation in the macro-meteorological flow data are aligned and encapsulated according to time-domain characteristics to generate the micro-environment meteorological sequence.

4. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 3, characterized in that, The thermodynamic deduction module is specifically used for calculating the surface dew point temperature based on the microenvironment meteorological sequence as follows: The calibration temperature and calibration humidity in the microenvironment meteorological sequence are analyzed, and the surface dew point temperature is calculated using the surface dew point temperature conversion formula. The calculated surface dew point temperature is compared with the real-time calibration temperature using logic. When the calibrated temperature is determined to be lower than or equal to the surface dew point temperature, the crop canopy of the target plot is determined to have entered the dew formation stage, and time-domain integration is started simultaneously to record the dew formation duration and intensity.

5. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 4, characterized in that, The thermodynamic deduction module is specifically used in the differential calculation and output of the canopy wettability retention time and evaporation attenuation slope for: The thermodynamic evaporation rate is calculated using the thermodynamic evaporation rate calculation formula; The initial liquid water thickness accumulated during the condensation stage is used as the starting value, and the loss depth obtained by the thermodynamic evaporation rate is deducted hourly according to the preset time step. When the remaining liquid water thickness decreases to zero, the time count stops and the canopy wet residence time is output. The evaporation decay slope is calculated using the evaporation decay slope calculation formula to obtain the canopy wet residence time and evaporation decay slope.

6. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 5, characterized in that, The hierarchical scheduling module, when generating the multi-factor disease early warning signal and hierarchical scheduling instruction by performing threshold comparison between the canopy wet retention time and the evaporation attenuation slope, is specifically used for: When the canopy wet retention time is greater than or equal to a preset retention time threshold and the evaporation attenuation slope is less than or equal to a preset evaporation slope upper limit, a multi-factor disease early warning signal is generated. When the historical prevention and control response is determined to be less than or equal to the preset response grading benchmark, a standardized plant protection operation order containing specific operation parameters is generated based on the geographic coordinate information of the target plot and the category of the multi-factor disease early warning signal. The multi-factor disease early warning signal and the instructions containing the standardized plant protection operation order are encapsulated and transmitted to generate the multi-factor disease early warning signal and hierarchical scheduling instructions. Among them, the preset retention time threshold and the preset evaporation slope upper limit are pre-calibrated based on the biological germination experimental data of the target pathogens in rice or citrus. The responsiveness grading benchmark is a constant ranging from 0.6 to 0.

85.

7. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 6, characterized in that, The hierarchical scheduling module is also used for: When the historical control response rate is determined to be greater than the response rate grading benchmark, an agricultural control efficacy guidance suggestion containing disease type, risk level and recommended application window period is generated and pushed to the corresponding bound farmer's mobile terminal.

8. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 6, characterized in that, When the hardware interlock module issues the control intervention parameters based on the multi-factor disease early warning signal and the hierarchical scheduling instruction, it is specifically used for: After normalizing the geometric center slope extracted from the topographic elevation parameters and the canopy wet retention time and evaporation attenuation slope, the loss risk index is calculated by weighting and summing the results using preset weighting coefficients. When the loss risk index is determined to be greater than or equal to the preset safety boundary, a low-level electrical limiting command is generated, which includes a forced overwrite value for the upper limit of the centrifugal atomizing disc rotation speed and a cutoff parameter for the water pump duty cycle. The underlying electrical limiting command is encapsulated into control intervention parameters that include underlying device limiting information.

9. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 8, characterized in that, The plant protection drone's execution terminal receives the control intervention parameters and starts the pesticide application operation on the target plot under limited physical output conditions. It also collects the flight attitude and position information of the onboard sensors in real time to generate an attitude telemetry data stream to be transmitted back to the cloud processing center. The micro-meteorological calculation module receives the attitude telemetry data stream and performs data cleaning, removing data from the maneuvering disturbance range where the rate of change of acceleration exceeds a preset range, and extracting wind measurement steady-state data that meets the dynamic equilibrium conditions.

10. The intelligent early warning system for rice and citrus diseases and pests based on multiple factors according to claim 9, characterized in that, The micro-meteorological calculation module is specifically used to extract the true value of the fuselage wind resistance tilt angle and update the wind speed attenuation factor using the attitude telemetry data stream: The true value of the fuselage wind resistance tilt angle is extracted from the extracted wind measurement steady-state data, and the true value of the measured micro-wind speed at the corresponding moment is extracted using the wind resistance tilt angle mapping micro-wind speed formula. The macro-average wind speed in the macro-meteorological flow data is called, and the initial wind speed attenuation factor in the extrapolation process is updated in a closed loop using the measured micro-wind speed true value and the inverse iterative formula of the wind speed model.

Images