Agricultural pest diagnosis and early warning method and system based on intelligent algorithm

Abstract

The application provides a kind of agricultural disease and insect pest diagnosis early warning method and system based on intelligent algorithm, system includes: spore collection and detection module, credible evaluation module, wind field inversion and source area identification module, incubation period modeling module, risk determination module, prevention and control decision and scheduling module and output interface module.Said spore collection and detection module obtains spore living body load value in real time by air sampling and quantitative PCR detection technology;The credible evaluation module realizes the effective triggering state identification of pathogen by multidimensional threshold determination;Wind field inversion and source area identification module determines upstream source area and propagation path using particle trajectory inversion and clustering algorithm;Incubation period modeling module calculates incubation period based on environmental parameters and Arrhenius type model and deduces potential disease time window;The application realizes the quantitative modeling and intelligent early warning of the whole process of disease and insect pest transmission, can improve the timeliness and accuracy of prevention and control, and reduce the prevention and control cost.

Description

Technical Field

[0001] This invention relates to the field of intelligent algorithm-based pest and disease diagnosis technology, and in particular to an agricultural pest and disease diagnosis and early warning method and system based on intelligent algorithms. Background Technology

[0002] Agricultural pests and diseases are a major factor restricting grain yield and crop quality, characterized by rapid spread, insidious incubation periods, strong regionality, and irreversible outbreaks. Currently, agricultural pest and disease control mainly employs two methods:

[0003] The first method involves manual field inspections, experience-based judgment, and post-disaster spraying. However, this method struggles to capture early signs of pest and disease transmission in a timely manner, exhibiting significant lag and uncertainty. The second method utilizes equipment such as spore samplers for pathogen monitoring. However, this approach largely remains at the detection stage, failing to achieve a systematic closed-loop management system encompassing pathogen collection, transmission prediction, and control decisions. Therefore, we propose an agricultural pest and disease diagnosis and early warning method and system based on intelligent algorithms.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a method and system for diagnosing and warning agricultural pests and diseases based on intelligent algorithms, thereby solving the technical problems mentioned in the background section.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for diagnosing and warning agricultural pests and diseases based on intelligent algorithms includes the following steps:

[0008] S1. Collect pathogen spore samples from the air in real time; input the collected spore samples into the molecular detection unit, obtain the corresponding cycle threshold through quantitative PCR reaction, and obtain the spore live load value by matching the cycle threshold with the preset calibration curve.

[0009] S2. Using a multidimensional threshold determination model, the confidence level of pathogen spore activity in the sample is calculated to obtain the spore activity level and confidence interval. When the live load value exceeds the preset threshold and the confidence interval converges, the effective triggering state of pathogen transmission is determined.

[0010] S3. Call high-resolution wind field data to perform particle trajectory inversion calculation on the upstream spatial region of the target farmland to obtain candidate source regions of spores, and perform spatiotemporal reverse correlation on the candidate source regions to form a propagation path;

[0011] S4. Combine the propagation path of the candidate source area with the physiological response parameters of the target crop to construct a pathogen incubation period model. The pathogen incubation period model calculates the incubation period length based on multiple factors such as crop variety, temperature and humidity, and leaf surface moisture, and extrapolates the potential disease outbreak time window after the pathogen arrives at the target plot.

[0012] S5. Combining the spore live load level and the intensity of the transmission path, the pathogen transmission risk index is calculated using an intelligent algorithm; when the transmission risk index exceeds the warning threshold, high-risk plots of target farmland and warning levels are identified, and a spatiotemporal distribution matrix is ​​generated.

[0013] S6. Based on high-risk plots and early warning levels, calculate the optimal timing of prevention and control measures and determine the application window when the pathogen is controllable; based on the transmission path and risk level, output the corresponding minimum control zone and operation scheduling information to achieve early and precise prevention and control of pests and diseases.

[0014] S1 specifically includes:

[0015] Spore-capturing devices were set up in the target farmland area to continuously collect samples of pathogenic spore particles from the air at a preset sampling frequency;

[0016] The collected spore samples are transferred to the molecular detection unit and pre-processed under temperature-controlled conditions.

[0017] Quantitative PCR was performed on the pretreated spore samples to obtain the cycle threshold for each test sample;

[0018] The cyclic threshold is compared with the preset spore live load calibration curve to obtain the corresponding spore live load value;

[0019] The spore viability load value is sent to the subsequent reliable evaluation module as a valid detection input.

[0020] S2 specifically includes:

[0021] The spore live load value from S1 is received and matched with a pre-set multidimensional threshold determination model;

[0022] Based on the spore live load value, the activity level is calculated and the corresponding confidence interval is generated;

[0023] Determine whether the in vivo load value exceeds the activity threshold and whether the confidence interval converges;

[0024] When the live load value exceeds the activity threshold and the confidence interval meets the convergence condition, it is marked as an effective triggering state for pathogen transmission.

[0025] The effective triggering state and the corresponding live load value are output to the wind field inversion module as a trigger signal for retrospective analysis.

[0026] S3 specifically includes:

[0027] Acquire high-resolution wind field meteorological data corresponding to the target farmland area, including wind speed, wind direction, and turbulence characteristics;

[0028] Under pathogen transmission triggering conditions, the particle trajectory inversion algorithm is invoked to establish a spatiotemporal reverse trajectory model of the target area;

[0029] Based on the results of reverse trajectory calculation, candidate upstream source regions of spore origin are identified;

[0030] Candidate upstream source regions are matched with wind field temporal characteristics to form continuous propagation paths;

[0031] The propagation path and source region location are output to the incubation period modeling module as the basis for extrapolating the propagation time window.

[0032] S4 specifically includes:

[0033] Obtain physiological response parameters corresponding to the target crop, including variety, temperature and humidity range, and leaf surface moisture content;

[0034] By combining the transmission pathways of candidate upstream source regions with the physiological response parameters of target crops, a pathogen incubation period model was established.

[0035] Based on the incubation period model, the incubation period from the arrival of the pathogen at the target site to the onset of disease is calculated;

[0036] By superimposing the incubation period length with the wind field inversion time, the potential time window for disease onset after the pathogen arrives can be estimated.

[0037] The potential onset time window is output to the transmission risk determination module as a time constraint for risk assessment.

[0038] S5 specifically includes:

[0039] Receive potential disease time window, spore viable load level and transmission route intensity;

[0040] Input the data into the intelligent risk calculation model to comprehensively assess the pathogen transmission risk index;

[0041] The risk level is determined by comparing the transmission risk index with the early warning threshold.

[0042] When the risk index exceeds the warning threshold, high-risk plots of the target farmland are marked, and a spatiotemporal distribution matrix is ​​generated;

[0043] The risk level and information on high-risk plots are output to the prevention and control decision-making module as input for calculating the pesticide application window.

[0044] S6 specifically includes:

[0045] Based on the location of high-risk sites and the transmission path, calculate the optimal time window for prevention and control measures;

[0046] Determine the priority order of pesticide application based on risk level and transmission intensity;

[0047] By using the spatiotemporal propagation matrix, the boundary of the prevention and control area is optimized to obtain the minimum prevention and control zone range;

[0048] Generate control instructions that include the scope of the control zone, the application time window, and the operation scheduling plan;

[0049] The control instructions are pushed to the execution terminal to achieve early and precise control of pests and diseases.

[0050] An agricultural pest and disease diagnosis and early warning system based on intelligent algorithms includes:

[0051] The spore collection and detection module is used to deploy air spore sampling devices in the target farmland area, collect pathogenic spore particle samples in the air in real time according to the preset sampling frequency, and perform sample pretreatment and quantitative PCR detection under temperature control conditions.

[0052] The confidence assessment module is used to receive the spore live load value and calculate the activity level and confidence interval based on the multidimensional threshold judgment model. When the live load value exceeds the threshold and the confidence interval converges, an effective triggering state for pathogen transmission is generated.

[0053] The wind field inversion and source region identification module is used to call high-resolution wind field data when triggered, calculate candidate upstream source regions of spore origin based on particle trajectory inversion algorithm, identify dominant source regions using clustering algorithm, and output propagation path and source region location information;

[0054] The incubation period modeling module is used to collect physiological response parameters of temperature, humidity, and leaf moisture of the target crop, construct an Arrhenius-type incubation period model, calculate the incubation period length based on the transmission path in the source area, and extrapolate the potential disease outbreak time window.

[0055] The risk assessment module receives spore live load value, transmission path length, and potential disease time window input. It uses intelligent algorithms to calculate the transmission risk index, compares the transmission risk index with the warning threshold, and determines high-risk plots and risk levels when the threshold is exceeded, and generates a spatiotemporal distribution matrix.

[0056] The prevention and control decision-making and scheduling module is used to calculate the optimal application time window for prevention and control measures based on high-risk plots and risk levels, perform meteorological correction on the application window, determine the minimum prevention and control zone, and perform capacity matching and resource scheduling in combination with equipment capabilities.

[0057] The output interface module is used to output the application time window, prevention and control area, scheduling information, and risk level data to the operation execution terminal in a structured format.

[0058] The beneficial effects of this invention are as follows:

[0059] This invention utilizes a cyclone-type air sampling device to continuously collect pathogenic spores from the air. Combined with quantitative PCR and Ct value calibration mapping technology, it obtains spore concentration data on an hourly timescale. It features high sensitivity, fast response, and high degree of automation, enabling the early detection of pathogen diffusion signals that are difficult to identify using traditional trapping techniques, thereby facilitating proactive deployment of pest and disease control.

[0060] This invention utilizes high-resolution meteorological wind field data and particle trajectory inversion algorithm, combined with DBSCAN clustering technology, to effectively identify the dominant upstream source area of ​​pathogens. By coupling analysis of trajectory density and wind field characteristics, high-precision source tracing can be achieved under complex terrain and variable wind direction conditions. This not only improves spatial positioning accuracy but also provides a scientific basis for regional prevention and control, supporting precise strike-style prevention and control strategies.

[0061] This invention constructs an Arrhenius-type incubation period model based on key farmland environmental parameters such as temperature, humidity, and leaf surface moisture, enabling real-time calculation and dynamic correction of the incubation period. By coupling with transmission path information, it can accurately extrapolate potential disease outbreak time windows, forming a high-precision early warning in the time dimension. Compared with static empirical models, this method can better cope with variable weather conditions and improve the reliability and timeliness of predictions.

[0062] This invention determines the weights of multiple factors using the entropy weight method, and combines ROC curve analysis and quantile threshold setting to automatically calculate and classify the transmission risk index R (L1–L4). This risk index is not only objective and repeatable, but also has good generalization ability, adapting to different pathogen types and planting environments in multiple regions. Through standardized classification, it is possible to achieve quantitative comparison of risks and coordinated prevention and control between regions.

[0063] This invention utilizes potential disease outbreak time windows, meteorological correction functions, and propagation radius models to automatically calculate the optimal application window and minimum control radius, and combines this with the capacity of spraying equipment for capacity matching and scheduling. Compared with traditional experience-based application, it can significantly reduce application timing deviations, reduce the area covered by pesticides, and simultaneously improve control coverage and effectiveness, effectively reducing costs and environmental impact. It establishes a complete technical path encompassing spore detection, wind field tracing, incubation period modeling, risk assessment, application decision-making, and scheduling execution. Through a unified data interface and modular architecture, it enables pest and disease monitoring, prediction, early warning, and control on a single platform, achieving minute-level response from "discovery" to "control." Attached Figure Description

[0064] Figure 1 This is a schematic diagram of an agricultural pest and disease diagnosis and early warning method based on intelligent algorithms according to the present invention.

[0065] Figure 2 This is a schematic diagram of the framework of an agricultural pest and disease diagnosis and early warning system based on intelligent algorithms according to the present invention. Detailed Implementation

[0066] The technical solutions of 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.

[0067] Example 1: As Figure 1 As shown in the figure, this embodiment provides a method for diagnosing and warning of agricultural pests and diseases based on intelligent algorithms, including the following steps:

[0068] S1. Collect pathogen spore samples from the air in real time; input the collected spore samples into the molecular detection unit, obtain the corresponding cycle threshold (Ct value) through quantitative PCR reaction, and obtain the spore live load value by matching the cycle threshold with the preset calibration curve.

[0069] S2. Using a multidimensional threshold determination model, the confidence level of pathogen spore activity in the sample is calculated to obtain the spore activity level and confidence interval. When the live load value exceeds the preset threshold and the confidence interval converges, the effective triggering state of pathogen transmission is determined.

[0070] S3. Call high-resolution wind field data to perform particle trajectory inversion calculation on the upstream spatial region of the target farmland to obtain candidate source regions of spores, and perform spatiotemporal reverse correlation on the candidate source regions to form a propagation path;

[0071] S4. Combine the propagation path of the candidate source area with the physiological response parameters of the target crop to construct a pathogen incubation period model. The pathogen incubation period model calculates the incubation period length based on multiple factors such as crop variety, temperature and humidity, and leaf surface moisture, and extrapolates the potential disease outbreak time window after the pathogen arrives at the target plot.

[0072] S5. Combining the spore live load level and the intensity of the transmission path, the pathogen transmission risk index is calculated using an intelligent algorithm; when the transmission risk index exceeds the warning threshold, high-risk plots of target farmland and warning levels are identified, and a spatiotemporal distribution matrix is ​​generated.

[0073] S6. Based on high-risk plots and early warning levels, calculate the optimal timing of prevention and control measures and determine the application window when the pathogen is controllable; based on the transmission path and risk level, output the corresponding minimum control zone and operation scheduling information to achieve early and precise prevention and control of pests and diseases.

[0074] S1 specifically includes the following sub-steps:

[0075] S110. Spore Sampling and Environmental Parameter Setting: Airborne spore sampling devices were deployed in the target farmland area. Cyclone-type air samplers were used, with a sampling height set at 1.5m, a sampling particle size range of 2–50μm, and a sampling frequency of 10min / time. The sampling device's outer casing had an IP54 protection rating, suitable for high humidity and moderate wind speed environments in the field. Basic environmental parameters such as temperature, humidity, and wind speed were simultaneously collected during sampling to ensure the stability of the spore detection results.

[0076] S120. Sample Pretreatment: Collected spore samples were sent to the molecular detection unit under a constant temperature of 25±2℃. The samples were lysed with standard lysis buffer for 10 min, and DNA was extracted using a silica gel membrane purification column. The entire sample processing took 15±5 min to ensure the stability and consistency of DNA extraction. The sample was immediately subjected to amplification and detection after processing to avoid Ct value shifts caused by degradation.

[0077] S130, Ct value detection and repeat assays: Quantitative amplification was performed using a commercial qPCR detection system (e.g., ABI 7500). The amplification system was prepared in standard 20 μL form, and 40 amplification cycles were performed. Each sample was tested three times, and the arithmetic mean of the Ct values ​​was taken as the final result, with the standard deviation of error controlled within ±0.5Ct. The detection sensitivity was not less than 10² CFU / m³, and the Ct detection range was 15–40.

[0078] S140, Ct calibration and live load mapping: Ct-spore live load calibration curves were established through serial dilution experiments of standard strains.

[0079] ;

[0080] in The value represents the spore live load, and a and b are calibration constants. Parameters a and b are obtained by fitting the standard sample concentration with the Ct value. This calibration process is a routine detection step in existing technology. The above formula accurately maps the detection signal to the actual spore load, providing quantifiable input parameters for subsequent risk assessment.

[0081] Typical calibration intervals are shown in the table below:

[0082]

[0083] The linear fit R of the calibration curve in the Ct=15~40 range 2 ≥0.98 ensures the repeatability and accuracy of Ct values ​​and in vivo load.

[0084] S150, Data Output and Module Connection: The calculated spore live load value is output to the reliability assessment module as the input for judging the effective triggering state of subsequent pathogen transmission; the data output includes the live load value, Ct detection interval, detection timestamp and corresponding sampling point information, to ensure that the data reference between the front and back modules is consistent, traceable and does not cause semantic jumps.

[0085] S2 specifically includes the following sub-steps:

[0086] S210, Spore live load threshold matching: The system receives the spore live load value Load from S150 and compares it with the preset multidimensional threshold model. Matching is performed. The threshold is determined based on 3–5 years of historical field monitoring data. The optimal cutoff value is determined through statistical analysis and ROC curve method to ensure a balance between sensitivity and specificity.

[0087] Taking wheat scab spore monitoring as an example, when the spore viable load exceeds 1.2 × 10⁻⁶... 3 CFU / m 3 If the value is higher than this for two consecutive sampling periods, the system will initially determine it as a "high-risk load zone".

[0088] S220. Confidence Interval Calculation and Convergence Judgment: Repeat the detection of spore viability load values ​​for each sampling point within a 10-minute sampling period (n=3), and calculate the mean. 1. Population standard deviation σ, and use the confidence interval formula:

[0089] ;

[0090] The significance level , It corresponds to the confidence level. The standard normal distribution quantile (e.g., 1.96 at 95% confidence level), where n is the sample size.

[0091] For example, when the results of three repeated tests are 1150, 1200, and 1180 CFU / m³, respectively. 3 ,but:

[0092] ;

[0093] The confidence interval is obtained as [1148.2, 1205.2] CFU / m 3 , with a threshold of 1.2 × 103 It converges close to but stably, satisfying the triggering condition.

[0094] S230, Threshold Determination Logic: Under the premise that the confidence interval converges, the system determines whether the live load value exceeds the trigger threshold; if... and When the trigger threshold is reached, the system enters a valid trigger state.

[0095] Threshold models allow for dynamic adjustments based on crop variety and climate season:

[0096] Spring wheat planting area: Trigger threshold 1.0×10 3 CFU / m 3 ;

[0097] Summer maize area: Trigger threshold 1.4 × 10 3 CFU / m 3 ;

[0098] Rice paddy area: Trigger threshold 0.9×10 3 CFU / m 3 .

[0099] This multi-threshold design improves the algorithm's adaptability to different ecological zones.

[0100] S240. Trigger State Generation: When the live load value exceeds the threshold and the confidence interval converges, the system generates an effective trigger state for pathogen transmission. To ensure real-time performance, the system's judgment module automatically runs once every 10 minutes to refresh the signal, guaranteeing both temporal continuity and dynamism. Trigger Status This includes the trigger timestamp, sampling point coordinates, mean live load, and upper and lower limits of the confidence interval.

[0101] S250, Data Output and Subsequent Module Interaction: The system will... The information and its accompanying data are output to the wind field inversion module for upstream source region location and trajectory analysis.

[0102] The output data structure is uniformly set as follows:

[0103] ;

[0104] TimeStamp represents the timestamp of the data collection and processing for this test, and SampleID represents the sampling point number or the unique identifier of the device. This indicates the lower limit of the confidence interval for spore detection; This indicates the upper limit of the confidence interval; this structure ensures complete consistency with the calling parameters of S310–S350, avoiding semantic jumps or data ambiguity.

[0105] S3 specifically includes the following sub-steps:

[0106] S310. Wind Field Data Acquisition and Preprocessing: The system collects near-surface wind field data from national and local meteorological monitoring networks. Data sources include:

[0107] Actual measured data from ground-based automatic weather stations;

[0108] Numerical weather forecasting models (such as WRF or ERA5);

[0109] If available, supplement with LIDAR or SODAR wind profiler data for vertical correction.

[0110] The collected wind field data was set to a time resolution of 30 minutes and a spatial resolution of 1 km, including wind speed. ,wind direction turbulence intensity With atmospheric stability S.

[0111] Before entering trajectory inversion, the wind field data will be resampled through three-dimensional grid interpolation and time linear interpolation to ensure the continuity of trajectory calculation.

[0112] S320, Lagrange particle trajectory inversion: in the triggered state Below, the system uses the center point of the target farmland as the particle release point and randomly and uniformly releases N particles. p =1000 particles, and perform inverse integration on the evolution of their positions over time.

[0113] The formula for calculating the particle trajectory position is:

[0114] ;

[0115] in This represents the spatial position vector of the particle at the trigger time t, which is the location where the spore was detected at the time of detection, usually the monitoring center point of the target farmland; Indicates before the trigger time point The particle position in time, that is, the possible source location of the spore, is the upstream source region calculated through wind field inversion. This is the inversion time length, used to determine the integration interval, which is the time window for backtracking upstream. For example, when The hour indicates the airflow path two days in advance. It is a time variable in the integration process, used to describe the continuous change of the wind field throughout the entire retrospective time period, and is an intermediate variable in a mathematical integral.

[0116] Represents the relationship between spatial location x and time. The wind speed vector below contains wind speed components in the horizontal and vertical directions (e.g., u, v, w), and can be derived from automatic weather station observation data or meteorological numerical forecast data.

[0117] To suppress high-frequency jitter during the inversion process, a turbulent random disturbance term is added to the trajectory calculation:

[0118] ;

[0119] in The standard deviation of turbulence (measured turbulence intensity 5%–15%). This represents the corrected wind field velocity, which is based on the original wind field with a random disturbance term superimposed to simulate the turbulence uncertainty in the actual wind field. It has a mean of zero and a standard deviation of Gaussian white noise is used to represent random fluctuations in wind speed caused by turbulence.

[0120] The integration process employs a fourth-order Runge-Kutta numerical integration method to solve the above formula, gradually approximating the integration result by discretizing the time interval into smaller step sizes. In this embodiment, the integration step size is set to 5 minutes to ensure the accuracy of trajectory calculation.

[0121] S330, Upstream Source Region Clustering and Identification: N obtained through inversion p Set of trajectory endpoint distributions of individual particles As input for clustering.

[0122] The system uses the DBSCAN clustering algorithm to perform cluster analysis on the endpoint distribution, with neighborhood radius... Minimum number of samples When the trajectory density within a certain area At that time, it was determined to be a candidate upstream source region. If the clustering results contain multiple source regions, the dominant source region is selected based on the trajectory weight and time proximity. This step ensures that the source region identification has stability and robustness.

[0123] S340, Propagation Path Construction and Smoothing: The system will select candidate upstream source regions. Reconstruct the time-series trajectories obtained from particle inversion to generate a propagation path set. ;

[0124] To reduce path jitter caused by instantaneous changes in the wind field, the system performs Bezier curve smoothing and least squares fitting on the trajectory path.

[0125] ;

[0126] Where p(t) represents the propagation path function, which describes the theoretical propagation trajectory of particles or spores at time t. This function is obtained by smoothing the spatial distribution of the inverted particle swarm and reflects the average movement path of the pathogen under the influence of the airflow field. This represents the spatial coordinates of the i-th trajectory point, which is the actual location obtained in the particle inversion calculation. Each trajectory point corresponds to a time node. , indicating the particle's time The position at any given moment. This means that by optimizing the path function p(t), the sum of squared errors between all trajectory points and the smoothed path is minimized, thus obtaining the optimal smoothed path. This least-squares smoothing fitting effectively eliminates path jitter caused by local wind field disturbances and inversion noise, making the propagation path more spatially continuous and stable. When the root mean square residual (RMSE) of the smoothed path is less than 0.5 km, the path is determined to be the optimal propagation path. .

[0127] S350, Data Output and Subsequent Calls: The system will output data from the upstream source region. With the optimal propagation path As structured data output to the latent period modeling module (S410–S450), the data format includes:

[0128] ;

[0129] Wherein: TrajectoryDensity represents trajectory density, ConfidenceScore represents source region identification confidence (calculated based on clustering score and path stability), and TimeWindow represents propagation time interval; the output path and source region data correspond completely to the incubation period model input, avoiding semantic jumps or interface ambiguities.

[0130] S4 specifically includes the following sub-steps:

[0131] S410. Physiological Environmental Parameter Acquisition: Deploy a micro-meteorological sensing network in the target farmland area, including temperature sensors, humidity sensors, and leaf surface humidity sensors. The sensor sampling frequency is set to 10 min / time, and the spatial resolution is 1 sensor node / 100 m².

[0132] The collected environmental parameters include:

[0133] Air temperature T(t) (unit: °C, range 10–35 °C); relative humidity RH(t) (unit: %, range 60–100%); leaf surface humidity L w (t) (unit: 0–1 represents the proportion of wetness), indicating the duration of time the leaf surface remains wet, which is crucial for the germination of pathogen spores.

[0134] The data undergoes real-time preprocessing and anomaly removal via local edge computing nodes. The 3σ rule is used to remove sensor anomalies, ensuring the stability and continuity of physiological environmental data.

[0135] S420. Latent period model construction: The latent period model adopts a combination of Arrhenius type and empirical correction, comprehensively considering environmental factors and crop variety correction coefficient k. v The formula for calculating the incubation period length τ is as follows:

[0136] ;

[0137] in It is the incubation period length (unit: h), which represents the time required for a pathogen to complete its latent development under specific environmental conditions; It is the variety correction factor (0.8–1.3), which is determined by field measurements. It is an empirical correction value used to reflect the differences in resistance of different crop varieties to the same pathogen. , , These are empirical fitting coefficients, corresponding to the weights of the three environmental factors—temperature, relative humidity, and leaf surface wetting time—in the incubation period model, with preferred values ​​of 0.015, 0.002, and 0.1, respectively. R is the empirical activation energy constant (unit: J / mol), used to characterize the sensitivity of pathogen physiological responses to temperature changes; R is the gas constant (8.314 J / mol·K); T(t) is the temperature (°C), converted to K as an exponential term; the above formula ensures that the incubation period length can dynamically respond to environmental stress when the temperature and humidity change, and has continuity and physical meaning.

[0138] S430. Calculation of incubation period length: The system calculates the incubation period length based on the 24–48 h rolling average of collected temperature, humidity, and leaf surface moisture. .

[0139] Typical parameter range examples are as follows:

[0140]

[0141] When environmental conditions are high temperature and high humidity, the incubation period can be shortened to less than 24 hours; when the environment is cold or the humidity is insufficient, the incubation period can be extended to more than 96 hours.

[0142] S440. Potential onset time window projection: Pathogen arrival time calculated from the transmission path. With incubation period length Superimposed to determine the potential onset time window W:

[0143] ;

[0144] in This indicates time uncertainty, calculated from the standard deviation of environmental fluctuations:

[0145] ;

[0146] in The short-term standard deviation of temperature fluctuations is used to quantify the contribution of temperature uncertainty to the onset time. This represents the short-term standard deviation of relative humidity fluctuations, used to quantify the impact of humidity fluctuations. This represents the short-term standard deviation of the fluctuation in leaf surface moisture time. Leaf surface moisture is a key condition for pathogen germination.

[0147] For example, when hour: This indicates that field symptoms may appear approximately 2–3 days after the pathogen arrives.

[0148] S450, Data Output and Subsequent Retrieval: The system outputs the incubation period length and the potential onset time window W to the transmission risk assessment module (S510–S550). The output structure includes:

[0149] ;

[0150] in It is the average incubation period length; It is a matter of time uncertainty; , These are the upper and lower boundaries of the time window; It is the crop variety correction factor; It is the environmental time window used to calculate the incubation period; this output data is directly used as the time input variable of the risk assessment model, realizing seamless integration with the early warning decision module.

[0151] S5 specifically includes the following sub-steps:

[0152] S510, Risk Index Calculation Input: The propagation risk assessment module receives the following inputs: Spore live load value (Load); propagation path length. Potential onset time window (from S450); historical incidence rate and control effectiveness data H (used for weight and threshold determination).

[0153] Load reflects the quantity and intensity of pathogens. The inputs reflect the pathogen's potential for transmission, while W reflects the duration of the pathogen's latent threat in the target area. These inputs together determine the actual risk level of pests and diseases occurring in the target area.

[0154] S520. Risk Function Calculation and Weight Determination: The risk index R is calculated using the following formula:

[0155] ;

[0156] in: ; , These are the maximum observed values ​​in the historical monitoring data; , , The weighting coefficients are determined using the entropy weighting method: the relative impact weight of each input on the risk is determined by the variance contribution of historical H data.

[0157] For example: contribution of pathogen concentration variance Path length variance contribution Time window variance contribution This method of determining weights is objective, can be repeatedly calculated, and avoids being questioned for being "artificially set".

[0158] S530. Risk Threshold Determination and Classification: Risk Threshold R th The determination is made using a two-step method:

[0159] ROC curve analysis: historical disease records are used as the true label; the calculated R is used as the predictor variable; the optimal threshold point is determined by the YoudenIndex maximization method to ensure a balance between sensitivity and specificity.

[0160] P90 quantile method supplementary calibration: The 90th percentile of the R distribution during historical high-risk periods is taken as the upper limit of the threshold for verification, ensuring responsiveness to extreme risks.

[0161] The risk classification standards (L1–L4) are as follows:

[0162]

[0163] For example, the detection value is:

[0164] ;

[0165] ;

[0166] but: ;

[0167] R=0.605 → Corresponds to L3 level risk → The system triggers an early warning and enters the precise prevention and control phase.

[0168] S540, High-Risk Land Parcel Marking and Aggregation: When R≥R th At that time, the system marks high-risk plots in the target farmland area.

[0169] Spatial proximity aggregation rules are adopted: if the distance between the center points of two high-risk grids is ≤50m, they are automatically merged into the same high-risk cluster; if a plot of land is consecutively located in more than 3 grids (Level L4), a key area marking is triggered. The high-risk plot marking matrix includes fields such as risk level, coordinates, area, and cluster ID.

[0170] S550, Early Warning Output and Interface Format: Risk assessment results are output to the prevention and control decision-making module in a structured format. The unified interface format is as follows:

[0171] ;

[0172] Where: GridID: Grid number;

[0173] R: Transmission Risk Index;

[0174] Risk Level: L1–L4;

[0175] ClusterID: Cluster ID for aggregated risk clusters;

[0176] Coordinates: Grid center coordinates;

[0177] Area: Corresponding farmland area;

[0178] Timestamp: The time of determination.

[0179] S6 specifically includes the following sub-steps:

[0180] S610. Determining the application time window: Application time window Based on the potential onset time window W shifted forward by a preventative time difference. Sure:

[0181] ;

[0182] ;

[0183] Among them: W start W end These are the upper and lower bounds of the potential onset time window derived from S440; This is a preventative offset, determined based on statistical analysis of previous disease incidence curves, generally taken as 12–24 hours; the standard for this value is: when the pathogen incubation period is <48 hours. Use the upper limit of 24 hours; when the incubation period is long, use 12 hours; this ensures that the application window falls in the middle to late stage of the pathogen spore incubation period but before the disease threshold, maximizing the control efficiency.

[0184] S620, Application Window Weather Correction: Because weather conditions have a significant impact on application effectiveness, the system adjusts the initial application window... Make corrections:

[0185] in It is a meteorological correction function, which is a logical control process that automatically determines whether to postpone or maintain the application time based on meteorological forecast data. For example, when the wind speed is >5m / s or the probability of precipitation is >50%, the window is postponed by 12 hours; when the temperature is <15℃ and the humidity is <60%, the window is postponed by 6 hours; when the meteorological conditions meet the ideal spraying requirements (wind speed <3m / s, no precipitation), no correction is made.

[0186] The final application window is formed after revision. ,in ; This indicates the revised start point of the application window. If meteorological conditions do not meet the application requirements, this start point may be postponed. This indicates the revised end time of the application window, calculated based on the time difference adjusted from the starting point.

[0187] S630, Determination of Minimum Control Zone: Radius of Minimum Control Zone Based on the length of the pathogen transmission path Risk aggregation cluster radius and wind speed conditions calculate:

[0188] ;

[0189] in Indicates the geometric radius of the cluster (obtained from the S540 clustering results); Indicates the length of the pathogen transmission path; Indicates current or predicted wind speed; and These are the path and wind speed weighting coefficients (measured calibration values, typical values:). ).

[0190] For example: ; .

[0191] S640, Application Capacity Matching and Resource Scheduling: Based on the minimum control zone radius Calculate the required area for prevention and control:

[0192] ;

[0193] Combined with the unit application capacity of the spraying equipment (Unit: m² / h), calculate the total application time:

[0194] ;

[0195] Where n is the number of available spraying equipment; the system automatically matches n based on equipment availability, transportation time and spraying window duration. This indicates the unit spraying capacity of a single spraying device (i.e., the area that can be sprayed per unit time), which depends on the specifications of the spraying system, the spray width, and the travel speed; ensuring pest control operations are carried out in a timely manner. Completed within the specified time.

[0196] For example: Single unit capacity: 25,000 m² / h; 3 units available: This time is less than the application window width (12h), which can meet the requirements for timely prevention and control.

[0197] S650, Data Output and Execution Interface: The application window and control zone information are output to the agricultural operation scheduling system in a structured format. The unified interface format is as follows:

[0198] ;

[0199] Field description:

[0200] ClusterID: High-risk cluster ID; , : Start and end times of the application window; Minimum prevention zone radius; Minimum prevention zone area; : Time required for application; n: Number of devices to be dispatched; Coordinates: Coordinates of the center of the control zone; RiskLevel: Corresponding risk level (L1–L4); Timestamp: Generation timestamp.

[0201] These data serve as direct input to the downstream "operation scheduling system," enabling intelligent spraying scheduling, path planning, and personnel dispatch.

[0202] Example 2: Figure 2 As shown, this embodiment provides an agricultural pest and disease diagnosis and early warning system based on intelligent algorithms, including:

[0203] The spore collection and detection module is used to deploy airborne spore sampling devices in the target farmland area, collect pathogenic spore particle samples in the air in real time according to a preset sampling frequency, and perform sample pretreatment and quantitative PCR detection under temperature control conditions; the spore collection and detection module maps the cycle threshold (Ct value) to the spore viable load value through a calibration curve, and outputs the load value, detection timestamp and sampling point information;

[0204] The confidence assessment module is used to receive the spore live load value, calculate the activity level and confidence interval based on the multidimensional threshold judgment model, and generate an effective pathogen transmission trigger state when the live load value exceeds the threshold and the confidence interval converges, and output the trigger state and load value to the wind field inversion module.

[0205] The wind field inversion and source region identification module is used to call high-resolution wind field data in the triggered state, calculate the candidate upstream source region of spore origin based on the particle trajectory inversion algorithm, and identify the dominant source region using the clustering algorithm, and output the propagation path and source region location information.

[0206] The incubation period modeling module is used to collect physiological response parameters of the target crop, such as temperature, humidity, and leaf surface moisture, to construct an Arrhenius-type incubation period model, calculate the incubation period length based on the transmission path in the source area, and extrapolate the potential disease outbreak time window.

[0207] The risk assessment module receives inputs such as spore live load value, transmission path length, and potential disease time window. It uses intelligent algorithms to calculate the transmission risk index, compares the transmission risk index with the warning threshold, and determines high-risk plots and risk levels when the threshold is exceeded, and generates a spatiotemporal distribution matrix.

[0208] The prevention and control decision and scheduling module is used to calculate the optimal application time window for prevention and control measures based on high-risk plots and risk levels, perform meteorological correction on the application window, determine the minimum prevention and control zone, and perform capacity matching and resource scheduling in combination with equipment capabilities.

[0209] The output interface module is used to output data such as application time window, control zone range, scheduling information, and risk level in a structured format to the operation execution terminal, so as to achieve early and precise prevention and control of pests and diseases.

[0210] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0211] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are 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 from one computer-readable storage medium to another. For example, 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., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state drive.

[0212] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0213] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0214] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0215] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0216] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0217] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0218] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0219] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart algorithm-based agricultural pest and disease diagnosis and early warning method, characterized in that, Includes the following steps: S1. Collect pathogen spore samples from the air in real time; input the collected spore samples into the molecular detection unit, obtain the corresponding cycle threshold through quantitative PCR reaction, and obtain the spore live load value by matching the cycle threshold with the preset calibration curve. S2. Using a multidimensional threshold determination model, the confidence level of pathogenic spore activity in the sample is calculated to obtain the spore activity level and confidence interval. When the live load value exceeds the preset threshold and the confidence interval converges, the effective triggering state of pathogen transmission is determined; S3. Call high-resolution wind field data to perform particle trajectory inversion calculation on the upstream spatial region of the target farmland to obtain candidate source regions of spores, and perform spatiotemporal reverse correlation on the candidate source regions to form a propagation path; S4. Combine the propagation path of the candidate source area with the physiological response parameters of the target crop to construct a pathogen incubation period model. The pathogen incubation period model calculates the incubation period length based on multiple factors such as crop variety, temperature and humidity, and leaf surface moisture, and extrapolates the potential disease outbreak time window after the pathogen arrives at the target plot. S5. Combining the spore live load level and the intensity of the transmission route, the pathogen transmission risk index is calculated using an intelligent algorithm; When the transmission risk index exceeds the warning threshold, high-risk plots of target farmland and warning levels are identified, and a spatiotemporal distribution matrix is ​​generated.

2. The intelligent algorithm-based agricultural disease and pest diagnosis and early warning method according to claim 1, characterized in that, It also includes S6, which calculates the optimal timing of prevention and control measures based on high-risk plots and warning levels, and determines the application window when the pathogen is controllable; Based on the transmission path and risk level, the corresponding minimum control zone and operation scheduling information are output to achieve early and precise prevention and control of pests and diseases.

3. The method of claim 1, wherein the method is characterized by, S1 specifically includes: Spore-capturing devices were set up in the target farmland area to continuously collect samples of pathogenic spore particles from the air at a preset sampling frequency; The collected spore samples are transferred to the molecular detection unit and pre-processed under temperature-controlled conditions. Quantitative PCR was performed on the pretreated spore samples to obtain the cycle threshold for each test sample; The cyclic threshold is compared with the preset spore live load calibration curve to obtain the corresponding spore live load value; The spore viability load value is sent to the subsequent reliable evaluation module as a valid detection input.

4. The intelligent algorithm-based agricultural disease and pest diagnosis and early warning method according to claim 1, characterized in that, S2 specifically includes: The spore live load value from S1 is received and matched with a pre-set multidimensional threshold determination model; Based on the spore live load value, the activity level is calculated and the corresponding confidence interval is generated; Determine whether the in vivo load value exceeds the activity threshold and whether the confidence interval converges; When the live load value exceeds the activity threshold and the confidence interval meets the convergence condition, it is marked as an effective triggering state for pathogen transmission. The effective triggering state and the corresponding live load value are output to the wind field inversion module as a trigger signal for retrospective analysis.

5. The intelligent algorithm-based agricultural pest and disease diagnosis and early warning method according to claim 1, characterized in that, S3 specifically includes: Acquire high-resolution wind field meteorological data corresponding to the target farmland area, including wind speed, wind direction, and turbulence characteristics; Under pathogen transmission triggering conditions, the particle trajectory inversion algorithm is invoked to establish a spatiotemporal reverse trajectory model of the target area; Based on the results of reverse trajectory calculation, candidate upstream source regions of spore origin are identified; Candidate upstream source regions are matched with wind field temporal characteristics to form continuous propagation paths; The propagation path and source region location are output to the incubation period modeling module as the basis for extrapolating the propagation time window.

6. The intelligent algorithm-based agricultural disease and pest diagnosis and early warning method according to claim 1, characterized in that, S4 specifically includes: Obtain physiological response parameters corresponding to the target crop, including variety, temperature and humidity range, and leaf surface moisture content; By combining the transmission pathways of candidate upstream source regions with the physiological response parameters of target crops, a pathogen incubation period model was established. Based on the incubation period model, the incubation period from the arrival of the pathogen at the target site to the onset of disease is calculated; By superimposing the incubation period length with the wind field inversion time, the potential time window for disease onset after the pathogen arrives can be estimated. The potential onset time window is output to the transmission risk determination module as a time constraint for risk assessment.

7. The method for diagnosing and warning agricultural pests and diseases based on intelligent algorithms according to claim 1, characterized in that, S5 specifically includes: Receive potential disease time window, spore viable load level and transmission route intensity; Input the data into the intelligent risk calculation model to comprehensively assess the pathogen transmission risk index; The risk level is determined by comparing the transmission risk index with the early warning threshold. When the risk index exceeds the warning threshold, high-risk plots of the target farmland are marked, and a spatiotemporal distribution matrix is ​​generated; The risk level and information on high-risk plots are output to the prevention and control decision-making module as input for calculating the pesticide application window.

8. The method for diagnosing and warning agricultural pests and diseases based on intelligent algorithms according to claim 2, characterized in that, S6 specifically includes: Based on the location of high-risk sites and the transmission path, calculate the optimal time window for prevention and control measures; Determine the priority order of pesticide application based on risk level and transmission intensity; By using the spatiotemporal propagation matrix, the boundary of the prevention and control area is optimized to obtain the minimum prevention and control zone range; Generate control instructions that include the scope of the control zone, the application time window, and the operation scheduling plan; The control instructions are pushed to the execution terminal to achieve early and precise control of pests and diseases.

9. An agricultural pest and disease diagnosis and early warning system based on intelligent algorithms, employing the agricultural pest and disease diagnosis and early warning method based on intelligent algorithms as described in any one of claims 1-8, characterized in that, include: The spore collection and detection module is used to deploy air spore sampling devices in the target farmland area, collect pathogenic spore particle samples in the air in real time according to the preset sampling frequency, and perform sample pretreatment and quantitative PCR detection under temperature control conditions. The confidence assessment module is used to receive the spore live load value and calculate the activity level and confidence interval based on the multidimensional threshold judgment model. When the live load value exceeds the threshold and the confidence interval converges, an effective triggering state for pathogen transmission is generated. The wind field inversion and source region identification module is used to call high-resolution wind field data when triggered, calculate candidate upstream source regions of spore origin based on particle trajectory inversion algorithm, identify dominant source regions using clustering algorithm, and output propagation path and source region location information; The incubation period modeling module is used to collect physiological response parameters of temperature, humidity, and leaf moisture of the target crop, construct an Arrhenius-type incubation period model, calculate the incubation period length based on the transmission path in the source area, and extrapolate the potential disease outbreak time window. The risk assessment module receives spore live load value, transmission path length, and potential disease time window inputs, uses intelligent algorithms to calculate the transmission risk index, compares the transmission risk index with the warning threshold, determines high-risk plots and risk levels when the threshold is exceeded, and generates a spatiotemporal distribution matrix. The prevention and control decision-making and scheduling module is used to calculate the optimal application time window for prevention and control measures based on high-risk plots and risk levels, perform meteorological correction on the application window, determine the minimum prevention and control zone, and perform capacity matching and resource scheduling in combination with equipment capabilities. The output interface module is used to output the application time window, prevention and control area, scheduling information, and risk level data to the operation execution terminal in a structured format.

Images