Crop disease and pest early warning method based on artificial intelligence

Abstract

The invention discloses a crop disease and insect pest early warning method based on artificial intelligence, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the multi-modal feature fusion analysis of a standardized training data set through an intelligent analysis terminal, and obtaining a meteorological sensitivity coefficient, a crop resistance coefficient and a disease and insect pest outbreak probability coefficient; inputting the three coefficients into a pest and disease risk quantification model, and outputting a pest and disease risk index; the early warning decision terminal triggers a multi-stage early warning rule according to the pest risk index, and generates a customized early warning scheme containing emergency measure suggestions; and the feedback verification terminal receives feedback data after the user implements the prevention and control measures, and dynamically updates the model parameters. According to the invention, the early warning decision terminal triggers the multi-stage early warning rule according to the disease and insect pest risk index, and generates the customized early warning scheme containing the emergency measure suggestion, thereby achieving the graded response and precise implementation of disease and insect pest early warning, and improving the pertinence and effectiveness of prevention and control measures.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and particularly relates to a method for warning of crop diseases and insect pests based on artificial intelligence. Background Art

[0002] With the growth of the global population and the acceleration of the agricultural modernization process, the prevention and control of crop diseases and insect pests have become increasingly important for ensuring food security and the sustainable development of agriculture. Factors such as intensified climate change and the complexity of planting structures have led to a significant increase in the outbreak frequency and damage degree of diseases and insect pests. The traditional prevention and control mode relying on manual inspections and experience judgments can no longer meet the requirements of precise and intelligent management.

[0003] At present, the technology for monitoring and warning of crop diseases and insect pests is gradually developing from single-parameter monitoring to the integration of multi-source data. Some studies have tried to collect environmental data such as temperature, humidity, and light using Internet of Things sensors and combine them with statistical models for risk prediction. However, the existing technologies generally have problems such as single data dimension and insufficient depth of multi-modal feature fusion, especially the lack of real-time dynamic monitoring of crop physiological states and disease and insect pest characteristics, resulting in poor adaptability of the models to complex field environments. At the same time, traditional warning methods are mostly based on fixed threshold rules, making it difficult to dynamically reflect the coupling effects of meteorological conditions, crop resistance, and the population dynamics of diseases and insect pests, and the accuracy and timeliness of warning results are limited.

[0004] The main defects of the existing technologies are mainly reflected in: First, at the data collection level, there is a lack of systematic monitoring of key physiological and pathological indicators such as crop chlorophyll content, stem water potential, and pathogen spore concentration, and a complete map of disaster-causing factors for diseases and insect pests cannot be constructed; second, at the model analysis level, a quantitative correlation mechanism between meteorological sensitivity, crop resistance, and the outbreak probability of diseases and insect pests has not been established, making it difficult to accurately quantify the risks under the synergistic action of multiple factors; third, at the system closed-loop level, there is a lack of a model iterative optimization mechanism driven by the feedback of prevention and control effects, resulting in the inability to dynamically adjust the warning plan according to the actual prevention and control effects, reducing the practicability and sustainability of the technology.

[0005] Therefore, it is urgent and necessary to invent a method for warning of crop diseases and insect pests based on artificial intelligence to solve the above problems. Summary of the Invention

[0006] The purpose of the present invention is to provide a method for warning of crop diseases and insect pests based on artificial intelligence to solve the problems raised in the above background art.

[0007] To achieve the above purpose, the present invention provides the following technical solution: A method for warning of crop diseases and insect pests based on artificial intelligence, including an intelligent monitoring terminal, a data processing terminal, an intelligent analysis terminal, a warning decision terminal, and a feedback verification terminal, specifically including the following steps: S1. The intelligent monitoring terminal regularly collects multi-dimensional environmental data through the Internet of Things sensor array deployed in the farmland during the crop growth cycle to form a multi-source fusion data set; S2. The data processing terminal performs outlier removal and normalization processing on the multi-source fusion data set to generate a standardized training data set; S3. The intelligent analysis terminal performs multi-modal feature fusion analysis on the standardized training data set to obtain the meteorological sensitivity coefficient, crop resistance coefficient, and pest and disease outbreak probability coefficient, and inputs the three coefficients into the pest and disease risk quantification model to output the pest and disease risk index; S4. The early warning decision-making terminal triggers multi-level early warning rules according to the pest and disease risk index to generate a customized early warning plan including emergency measure suggestions; S5. The feedback verification terminal receives the feedback data after the user implements the prevention and control measures and dynamically updates the model parameters.

[0008] Preferably, the multi-source fusion data set includes a meteorological environment data set, a crop physiology data set, and a pest and disease characteristic data set; the meteorological environment data set includes the three-dimensional coordinates of the farmland, air temperature, relative humidity, light intensity, and precipitation; the crop physiology data set includes the relative chlorophyll content, initial fluorescence, maximum fluorescence, and stem water potential; the pest and disease characteristic data set includes the pest population density, spore concentration of the pathogen, and the proportion of the area of typical disease spots.

[0009] Preferably, the meteorological sensitivity coefficient is specifically: , where t0 is the starting time of data collection, t is the current time, k is the temperature response decay coefficient, T air (t) is the air temperature at time t, T opt is the optimum temperature for crop growth, M(t) is the air humidity at time t, M base is the optimum humidity for crop growth, I light (t) is the light intensity at time t, I thresh is the light intensity threshold, P rain，max is the single maximum precipitation, e is the natural constant, and ln is the logarithm with base e.

[0010] Preferably, the crop resistance coefficient is specifically: , where α, β, γ, and δ are weight factors, α + β + γ + δ = 1 and α, β, γ, and δ ∈ [0,1], Chl is the relative chlorophyll content, Ψ stem is the stem water potential, Ψ wilt is the wilting water potential critical value, F min is the minimum fluorescence in the dark adaptation, A lesionis the proportion of the area of typical lesions, F v is the variable fluorescence, and the calculation formula is F v =F m -F o , where F m is the maximum fluorescence, and F o is the initial fluorescence.

[0011] Preferably, the pest and disease outbreak probability coefficient is specifically: , where λ, η, and ρ are adjustment factors, D pest is the pest population density, S spores is the concentration of pathogen spores, A lesion is the proportion of the area of typical lesions, Chl is the relative chlorophyll content, and Chl crit is the chlorophyll critical value, T air is the air temperature, T pest is the optimal temperature for pest activities, and e is the natural constant.

[0012] Preferably, the pest and disease risk quantification model is specifically: , where , ω1, ω2, and ω3 are dynamic weight coefficients, C is the meteorological sensitivity coefficient, R is the crop resistance coefficient, P is the pest and disease outbreak probability coefficient, τ is the historical data correction factor, and Q hist is the pest and disease risk index in the same historical period, and e is the natural constant.

[0013] Preferably, the multi-level warning rules include: If the pest and disease risk index Q≥Q1, trigger a first-level warning and take chemical emergency measures; If Q2≤Q<Q1, trigger a second-level warning and implement biological control; If Q3≤Q<Q2, trigger a third-level warning and start enhanced patrols; If 0≤Q<Q3, trigger a fourth-level warning and recommend routine monitoring.

[0014] Preferably, the customized warning plan includes: Match a list of recommended pesticides for different pest and disease types; Generate a drone spraying path plan based on the three-dimensional coordinates of the farmland; Provide the best prevention and control window period in combination with weather forecast data.

[0015] Preferably, the specific implementation method of the feedback verification terminal is: A1. Obtain the lesion regression rate and pest population survival rate after pesticide application through intelligent monitoring terminals and data processing terminals; A2. The feedback verification terminal inversely corrects the model parameters according to the lesion regression rate and pest population survival rate, and iteratively optimizes the meteorological sensitivity coefficient, crop resistance coefficient, and pest and disease outbreak probability coefficient; A3. Reallocate the weights of the meteorological sensitivity coefficient, crop resistance coefficient, and pest and disease outbreak probability coefficient according to the correction results.

[0016] Preferably, the rules for iterative optimization include: Correction of the meteorological sensitivity coefficient: C new =C old ×(1 - η C ×S surv ) where η C is the adjustment factor, S surv is the pest survival rate, C old is the meteorological sensitivity coefficient before iterative optimization, and C new is the meteorological sensitivity coefficient after iterative optimization; Correction of the crop resistance coefficient: R new =R old ×(1 + η R ×(1 - S surv )) where η R is the adjustment factor, S surv is the pest survival rate, R old is the crop resistance coefficient before iterative optimization, and R new is the crop resistance coefficient after iterative optimization; Correction of the pest and disease outbreak probability coefficient: P new =P old ×(1 + η P ×(1 - R rec )) where η P is the adjustment factor, R rec is the lesion regression rate, P old is the pest and disease outbreak probability coefficient before iterative optimization, and P new is the pest and disease outbreak probability coefficient after iterative optimization; The implementation method of step A3 is: Compare the corrected model value with the model prediction value, and calculate the error: , where Q predicted is the model prediction value, Qactual is the corrected model value; Combining the error and the corrected coefficient value, update the weight coefficient: , , , where E total =(C new +R new +P new ) × (1 + E).

[0017] Technical effects and advantages of the present invention: Through the intelligent analysis terminal of the present invention, multi-modal feature fusion analysis is performed on the standardized training data set to obtain the meteorological sensitivity coefficient, crop resistance coefficient, and pest and disease outbreak probability coefficient, and input them into the pest and disease risk quantification model to output the risk index, realizing multi-dimensional and quantitative analysis of the pest and disease risk, and accurately evaluating the occurrence risk of pests and diseases; Through the early warning decision terminal of the present invention, multi-level early warning rules are triggered based on the pest and disease risk index, and a customized early warning plan including emergency measure suggestions is generated, realizing hierarchical response and precise implementation of pest and disease early warning, and improving the pertinence and effectiveness of prevention and control measures; Through the feedback verification terminal of the present invention, the feedback data after the user implements the prevention and control measures is received, the model parameters are dynamically updated, forming a self-optimizing closed loop of the system, continuously improving the prediction accuracy and early warning ability of the model, and adapting to the changes in different farmland environments and prevention and control requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 is the schematic diagram of the device connection of the present invention.

[0019] Figure 2 is the flowchart of the method steps of the present invention.

[0020] Figure 3 is the specific implementation flowchart of the feedback verification terminal of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] The present invention provides as Figure 1An artificial intelligence-based crop pest and disease early warning method is shown, comprising an intelligent monitoring terminal, a data processing terminal, an intelligent analysis terminal, an early warning decision terminal and a feedback verification terminal; It should be noted that the core function of the intelligent monitoring terminal is to continuously collect multi-dimensional environmental, physiological and pest and disease characteristic data throughout the entire crop growth cycle through a diverse array of IoT sensors deployed in farmland, including meteorological sensors, crop physiological sensors such as chlorophyll meters, fluorescence meters, and stem water potential probes, as well as pest and disease monitoring equipment such as insect monitoring lights, spore capture devices, and image recognition cameras, and finally integrate them to form a multi-source fusion data set including meteorological environment, crop physiological and pest and disease characteristics. The terminal mainly relies on various dedicated sensor nodes deployed on the farmland in terms of physical implementation, and performs preliminary data aggregation through edge computing gateways or direct connections to the cloud platform. The data processing terminal is responsible for professional data preprocessing of the original multi-source fusion data set from the intelligent monitoring terminal. The core tasks include removing outliers and performing normalization processing to generate a standardized data set suitable for model training. Its physical carrier can be an edge computing device for preliminary processing, but it mainly relies on a cloud-based big data processing platform, such as a Spark or Hadoop framework or a real-time data stream processing engine, such as Apache Flink, to complete data cleaning, conversion and standardization using these powerful computing resources. The intelligent analysis terminal is highly dependent on powerful computing capabilities and is usually deployed on a GPU server cluster in the cloud to perform complex model training and reasoning, or uses edge AI reasoning chips such as NVIDIA Jetson to run lightweight models for real-time analysis; The early warning decision terminal automatically triggers the corresponding level of early warning based on the pest and disease risk index output by the intelligent analysis terminal and the preset multi-level early warning rules, and generates a highly customized early warning plan containing specific action suggestions; its implementation relies on the business logic engine deployed in the cloud, such as the Drools rule engine, the geographic information system platform, and the agricultural knowledge graph database, which can accurately match the recommended list of pesticides, generate drone pesticide application path planning based on the three-dimensional coordinates of the farmland, and provide the best prevention and control window period in combination with the weather forecast, and finally push the early warning information and plan to the user through the Web console; The feedback verification terminal forms a self-optimizing closed loop of the system. By the intelligent monitoring terminal, the effect data after the user implements the prevention and control measures is obtained again, such as the lesion regression rate and the pest population survival rate after pesticide application. These feedback data are used to drive the dynamic update of the model parameters. The specific process includes reversely correcting the calculation formulas of the meteorological sensitivity coefficient, the crop resistance coefficient, and the pest and disease outbreak probability coefficient according to the feedback indicators, and redistributing the dynamic weight coefficients of these three coefficients in the risk quantification model in combination with the prediction error. Its physical implementation mainly relies on the machine learning operation and maintenance platform deployed in the cloud, such as MLflow, to manage the iterative training and version update of the model, and uses online learning or incremental learning technology to continuously optimize the model performance and improve the accuracy of subsequent early warnings.

[0023] The present invention provides a method for warning of crop pests and diseases based on artificial intelligence as Figure 2 shown, which specifically includes the following steps: S1. The intelligent monitoring terminal regularly collects multi-dimensional environmental data through the Internet of Things sensor array deployed in the farmland during the crop growth cycle to form a multi-source fusion data set; Furthermore, in the above technical solution, the multi-source fusion data set includes a meteorological environment data set, a crop physiological data set, and a pest and disease characteristic data set; the meteorological environment data set includes the three-dimensional coordinates of the farmland, air temperature, relative humidity, light intensity, and precipitation; the crop physiological data set includes the relative chlorophyll content, initial fluorescence, maximum fluorescence, and stem water potential; the pest and disease characteristic data set includes the pest population density, the concentration of pathogen spores, and the proportion of the area of typical lesions.

[0024] It should be noted that the three-dimensional coordinates of the farmland are obtained through a global navigation satellite system, such as GPS and Beidou, combined with differential positioning technology, to obtain the longitude, latitude, and altitude of the farmland plot, and realize the three-dimensional coordinate acquisition with centimeter-level to meter-level accuracy; The air temperature is regularly collected through a protective temperature sensor deployed about 1.5 meters above the ground in the farmland and equipped with a radiation shield, such as a platinum resistance PT100, and the data is recorded every 5-10 minutes to accurately reflect the temperature environment near the crop canopy; The relative humidity is measured by a capacitive humidity sensor, which works based on the principle of the change of the dielectric constant after the polymer film absorbs moisture. To ensure the accuracy of the data, it is necessary to calibrate it regularly with a saturated salt solution, such as 75% NaCl; The light intensity is collected by a photosynthetic photon sensor, such as a PAR sensor. This device is specifically used to measure the photosynthetically active radiation in the 400-700nm band. When installed, it needs to be placed horizontally above the crop canopy to avoid shadow interference with the readings; The precipitation is monitored using a tipping bucket rain gauge. Its working principle is that every 0.2 mm of accumulated rainfall triggers the tipping bucket to flip once, and the total precipitation is calculated by recording the number of flips through a reed switch. The device is usually equipped with anti-bird needles and filters to prevent debris blockage from affecting the accuracy; The relative chlorophyll content is measured using a handheld chlorophyll meter, such as the SPAD-502. This instrument calculates the relative chlorophyll index by computing the transmission ratio of 650 nm red light to 940 nm infrared light by the leaf. When sampling, healthy leaves that are fully expanded at the upper part of the plant should be selected, avoiding the main vein, and multiple measurements should be taken at different points and the average value should be obtained; The initial fluorescence and maximum fluorescence are obtained using a modulated chlorophyll fluorometer, such as the PAM-2500. Before measurement, the leaf needs to be dark adapted for 20 minutes, and then the initial fluorescence is excited under weak measuring light. The intensity of this weak measuring light is usually lower than 0.1 μmol photons m -2 s -1 , and then a saturating pulse light is applied. The intensity of this pulse light usually reaches about 3000 μmol photons m -2 s -1 , which is used to excite the leaf to produce the maximum fluorescence; The stem water potential is measured using a pressure chamber water potential meter, such as the PMS Model 1505D. During the operation, branches with leaves are cut before sunrise and quickly sealed in the pressure chamber. Pressure is gradually applied until sap appears at the cut of the xylem vessel. At this time, the corresponding pressure value is the stem water potential. Sampling in the early morning can avoid the interference of transpiration on the results to the greatest extent; The pest population density is mainly measured by an automatic insect situation forecasting lamp, such as using an insect-trapping lamp with a specific wavelength band, or a sex pheromone trap to capture insects. Combined with the built-in image recognition system, the captured insects are automatically counted and classified. Finally, the captured quantity is converted into the pest population density per unit area, such as per hectare; The concentration of pathogenic spores is collected using a spore trap, such as the TPBZ3 type. This device sucks the spores in the air through air flow and impacts and adheres them onto a glass slide. Subsequently, a microscopic imaging system is used to obtain an image, and then an AI recognition model based on deep learning, such as YOLOv5, is applied to automatically identify and count the spores in the image. The final result is expressed as the number of spores per cubic meter of air; The proportion of the area of typical disease spots is determined by using a high-definition camera deployed in the field, especially a multispectral camera such as the RedEdge-MX, to take images of crop leaves. Through image processing techniques, such as segmentation based on the HSV color space, the disease spot areas on the leaves are identified and segmented, and finally the percentage of the pixel area of the disease spots in the total pixel area of the leaves is calculated.

[0025] S2, the data processing terminal performs outlier removal and normalization processing on the multi-source fusion data set to generate a standardized training data set; It should be noted that the specific implementation process of the data processing terminal is to first remove outliers from the original multi-source fusion data set from the intelligent monitoring terminal, strictly check according to the preset data verification rules, and use physical reasonable range filtering for the meteorological environment data set, such as excluding abnormal temperature values exceeding -10°C to 50°C. For the crop physiological data set, a dynamic threshold algorithm is used to identify and remove data points that significantly deviate from normal fluctuations, such as chlorophyll readings that exceed 120% of the historical highest value. For the pest and disease characteristic data set, an isolation forest algorithm is used to detect statistical outliers; for the data gaps generated after removing the outliers, linear interpolation is used to fill the missing values in the continuous time series data. After completing the abnormal processing, data normalization is performed immediately to eliminate the dimension differences of different sensors. For numerical data, such as meteorological environment data sets, crop physiological data sets, and pathogen spore concentrations, Min-Max normalization is used to linearly map them to the [0, 1] interval. For pest and disease data with high skewness characteristics, such as pest population density and typical lesion area ratio, logarithmic transformation is applied to reduce the data distribution inclination, and finally a standardized training data set is obtained.

[0026] S3. The intelligent analysis terminal performs multimodal feature fusion analysis on the standardized training data set to obtain the meteorological sensitivity coefficient, crop resistance coefficient and pest outbreak probability coefficient, and inputs the three coefficients into the pest risk quantification model to output the pest risk index; Furthermore, in the above technical solution, the meteorological sensitivity coefficient is specifically: , Among them, t0 is the start time of data acquisition, t is the current time, k is the temperature response attenuation coefficient, T air (t) is the air temperature at time t, T opt is the optimum temperature for crop growth, M(t) is the air humidity at time t, M base The optimum humidity for crop growth, I light (t) is the light intensity at time t, I thresh is the light intensity threshold, P rain，max is the single maximum precipitation, e is a natural constant, and ln is the logarithm with e as the base.

[0027] It should be noted that the formula design of the meteorological sensitivity coefficient is based on the dynamic coupling mechanism and time accumulation effect of environmental factors on pests and diseases, and its logical structure is progressive: first, the numerator of the formula captures the continuous effect of environmental factors during the crop growth cycle through integral operations, emphasizing the difference between short-term fluctuations and long-term trends; second, triple nonlinear coupling is designed in the integral term, and the temperature response adopts an exponential decay function. Express the sensitivity of pests and diseases to temperature deviation. For example, the activity is highest in the optimal temperature range and decays rapidly when deviated. The humidity effect is realized through logarithmic scaling. For example, high humidity promotes the germination of pathogen spores, and low humidity inhibits the activities of pests. The penalty term of the light intensity threshold I thresh is associated with the weakening of crop resistance in low-light environments; finally, the denominator uses the maximum single precipitation P rain，max +1 to characterize the precipitation inhibition effect, forming a dynamic balance where the numerator promotes and the denominator inhibits.

[0028] Furthermore, in the above technical solution, the crop resistance coefficient is specifically: , where α, β, γ, and δ are weight factors, α + β + γ + δ = 1 and α, β, γ, and δ ∈ [0, 1], Chl is the relative chlorophyll content, Ψ stem is the stem water potential, Ψ wilt is the critical wilting water potential, F min is the minimum fluorescence in dark adaptation, A lesion is the proportion of the typical lesion area, F v is the variable fluorescence, and the calculation formula is F v =F m -F o , where F m is the maximum fluorescence, and F o is the initial fluorescence.

[0029] It should be noted that in the formula design of the crop resistance coefficient, the numerator and denominator respectively represent the dynamic balance of the positive driving factors and negative inhibitory factors of resistance. The numerator part is composed of the relative chlorophyll content and the difference between the stem water potential and the wilting critical water potential: the chlorophyll content reflects the functional integrity of the photosynthetic organs through the weight α, and the water potential difference is quantified by β to represent the buffering ability of the crop to water stress such as drought. The two together reflect the potential of the crop to maintain growth through photosynthesis and water homeostasis; the denominator includes the photosynthetic efficiency variation term and the disease damage term δ×A lesion : where F v =F m -F o defines the ratio of variable fluorescence to maximum fluorescence, which is regulated by γ to represent the response sensitivity of photosystem II to biotic or abiotic stresses; the proportion of the typical lesion area A lesionDirectly correlate the degree of pathological damage through the weight δ, and the two jointly reflect the weakening effect of the antagonism between functional impairment and energy dissipation under adversity. The setting of the four weight factors α, β, γ, and δ: α represents the contribution degree of the relative chlorophyll content Chl to the target index R. If chlorophyll is the main influencing factor, such as the assessment of photosynthesis in healthy plants, α can be set between 0.3 and 0.5. If the plant is under stress, such as drought or disease, the chlorophyll content may decrease, and α needs to be reduced; β reflects the influence of the plant's water status, that is, the difference between the stem water potential and the wilting critical value. In the study of drought or water stress, β can be set between 0.2 and 0.4. If the research focus is on water relations, β needs to be increased. If the water condition is stable, β can be reduced; γ measures the photosynthetic efficiency , in the research related to photosynthesis, γ can be set between 0.2 and 0.3. If the parameters in are sensitive to the target index, such as the early detection of stress, γ can be increased; δ represents the influence of diseases on the target index. In the study of healthy plants, δ can be set to a lower value, which can be set between 0 and 0.1. In the study of diseases, δ needs to be increased, which can be set between 0.3 and 0.5.

[0030] Furthermore, in the above technical solution, the probability coefficient of pest and disease outbreaks is specifically: , where λ, η, and ρ are adjustment factors, D pest is the pest population density, S spores is the concentration of pathogen spores, A lesion is the proportion of the area of typical disease spots, Chl is the relative chlorophyll content, Chl crit is the chlorophyll critical value, T air is the air temperature, T pest is the optimal temperature for pest activity, and e is the natural constant.

[0031] It should be noted that in the formula design of the probability coefficient of pest and disease outbreaks, the core driving factors are composed of the pest population density D pest and the concentration of pathogen spores S spores , the proportion of the area of typical disease spots A lesion . Among them, the pest population density and the concentration of pathogen spores form the basic driving term through direct multiplication, reflecting the direct impact of their synergistic effect on the outbreak of pests and diseases. As the pest is the transmission vector of the pathogen, the larger the product value of the pest population density and the concentration of pathogen spores, the easier it is to form the basic conditions for the prevalence of pests and diseases. At the same time, the proportion of the area of typical disease spots passes through the exponential function Perform non - linear amplification to reflect the spatial expansion effect of the disease: The higher the proportion of the lesion area, it not only indicates that the host tissue has been severely infected, but is more likely to accelerate the epidemic outbreak process through the cumulative effect of secondary transmission or tissue damage. This exponential growth characteristic highlights the significant strengthening effect of the disease diffusion stage on the outbreak probability; The plant resistance regulation term consists of the relative chlorophyll content Chl and the critical value Chl crit and the denominator term adopts the Sigmoid function form, implying the following logic: When Chl > Chl crit , that is, when the plant is healthy, the denominator approaches 1, and the whole term approaches 1, suppressing the outbreak probability; When Chl < Chl crit , that is, when the plant is weak, the denominator approaches 0, and the whole term approaches 0, significantly reducing the outbreak probability; The temperature adaptability correction term consists of the air temperature T air and the optimal temperature T of the pest pest . The correction term also adopts the Sigmoid function, with the following effects: When T air ≈T pest , the correction term approaches 1, and the temperature is suitable for pest activities; When T air <<T pest or T air >>T pest , the correction term approaches 0, and the unsuitable temperature inhibits the pest; The roles of the adjustment factors λ, η, and ρ: λ controls the sensitivity of the influence of the lesion area. The larger the value, the stronger the amplification effect of the change in the lesion area on the probability; η adjusts the steepness of the transition near the plant resistance threshold. The larger the value, the more drastic the probability change when the relative chlorophyll content approaches the critical value; ρ controls the response width of the temperature adaptability. The larger the value, the faster the probability decreases when the temperature deviates from the optimal value; The setting of the adjustment factors λ, η, and ρ: If the probability increases by n times for every doubling of the lesion area of a certain disease, then λ = ln(n), where λ ∈ [0.1, 5]; If the relative chlorophyll content is m times lower than the critical value, then η = 10×m, where η ∈ [0.5, 10]; If the pest is active within the range of the optimal temperature ±T °C, then ρ = , where ρ ∈ [0.1, 1].

[0032] Furthermore, in the above technical solution, the pest and disease risk quantification model is specifically: , where , ω1, ω2, and ω3 are dynamic weight coefficients, C is the meteorological sensitivity coefficient, R is the crop resistance coefficient, P is the pest and disease outbreak probability coefficient, τ is the historical data correction factor, Q hist is the pest and disease risk index in the same historical period, and e is the natural constant.

[0033] It should be noted that if the historical data is stable for a long time, such as the pest and disease patterns are the same in the same period of a certain area for many years, and the changes in meteorology and crop varieties are small, τ can be set between 2 and 5, with the aim of making the historical pattern have a more significant impact on the current prediction; if the historical data fluctuates greatly, such as abnormal climate and frequent variety updates in recent years, and the historical pattern fails, τ can be set between 0.1 and 1, with the aim of reducing the interference of historical data; It should be noted that the setting of the initial values of the dynamic weight coefficients ω1, ω2, and ω3: In the case of a meteorological-driven pest and disease scenario, the initial values can be set as ω1 = 0.4, ω2 = 0.3, ω3 = 0.3; in the case of a pest-source-dominated scenario, the initial values can be set as ω1 = 0.3, ω2 = 0.3, ω3 = 0.4; in the case of a scenario in the promotion area of disease-resistant varieties, the initial values can be set as ω1 = 0.3, ω2 = 0.4, ω3 = 0.3; in the case of a scenario in an area with frequent extreme climates, the initial values can be set as ω1 = 0.5, ω2 = 0.25, ω3 = 0.25.

[0034] S4. The early warning decision-making terminal triggers multi-level early warning rules according to the pest and disease risk index, and generates a customized early warning plan including suggestions for emergency measures; Furthermore, in the above technical solution, the multi-level early warning rules include: If the pest and disease risk index Q ≥ Q1, trigger a first-level early warning and take chemical emergency measures; If Q2 ≤ Q < Q1 for the pest and disease risk index, trigger a second-level early warning and implement biological control; If Q3 ≤ Q < Q2 for the pest and disease risk index, trigger a third-level early warning and start enhanced inspections; If 0 ≤ Q < Q3 for the pest and disease risk index, trigger a fourth-level early warning and recommend routine monitoring.

[0035] It should be noted that Q1, Q2, and Q3 are risk index critical values, Q1 = 0.8, Q2 = 0.6, Q3 = 0.3; It should be noted that the chemical emergency measures are based on the pest and disease types to match a recommended list of highly effective and low-toxic chemical agents. For example, chlorantraniliprole for Lepidoptera pests and pyraclostrobin for fungal diseases, generate a drone spraying path plan using the three-dimensional coordinates of the farmland to ensure uniform coverage of the agent, and combine weather forecast data to provide the best prevention and control window period such as no rain in the next 24 hours to maximize the efficacy; The biological control specifically includes releasing natural enemy organisms such as Trichogramma and Aphidiidae for pest species, applying microbial agents such as Bacillus subtilis and Trichoderma, using plant-derived extracts such as azadirachtin and matrine, and combining the adjustment of the farmland vegetation structure, such as intercropping insect-repellent plants, to enhance the natural pest control ability; For the enhanced inspection, the frequency of manual field inspections needs to be increased from the regular once a week to once every two to three days. Key areas such as the undersides of leaves and the bases of stems of crops should be observed and data recorded. At the same time, intelligent devices such as insect situation forecasting lights and multi-spectral cameras are used to collect data in real time, automatically identify indicators such as the number of adult pests and the trend of disease spot expansion, issue immediate warnings when abnormalities are found, and compare the inspection data with historical models to evaluate the development trend of pests and diseases.

[0036] Furthermore, in the above technical solution, the customized early warning plan includes: Matching a list of recommended pesticides for different types of pests and diseases; Generating a drone spraying path plan based on the three-dimensional coordinates of the farmland; Providing the best prevention and control window period in combination with weather forecast data.

[0037] It should be noted that matching a list of recommended pesticides for different types of pests and diseases specifically means automatically retrieving and matching corresponding highly effective and low-toxic chemical pesticides or biological pesticides from the agricultural knowledge graph database built into the early warning decision-making terminal according to the types of pests and diseases output by the intelligent analysis terminal, such as the types of pests and the categories of pathogens. For example, for Lepidoptera pests such as cotton bollworms and corn borers, selective insecticides such as chlorantraniliprole and emamectin benzoate are recommended; for fungal diseases such as powdery mildew and downy mildew, fungicides such as pyraclostrobin and tebuconazole are recommended; for bacterial diseases, agents such as thiodiazole copper and zhongshengmycin are recommended. This process achieves precise matching through the association between pest and disease characteristic data, such as pest population density and typical disease spot morphology, and the labels in the pesticide library. Generating a drone spraying path plan based on the three-dimensional coordinates of the farmland specifically means using the three-dimensional geographical coordinate data of the farmland plot, combining parameters such as the row spacing and canopy height of crop planting, dividing the farmland into several grid units through the geographic information system platform of the early warning decision-making terminal, setting different application rates according to the pest and disease risk index of each unit, and then using the Dijkstra algorithm or genetic algorithm to plan the shortest flight route to ensure that the drone evenly covers all high-risk areas and dynamically corrects the path in combination with the image data transmitted back by the drone in real time to improve the control efficiency and reduce pesticide waste. Providing the best prevention and control window period in combination with weather forecast data specifically means that the early warning decision-making terminal connects to the real-time data interface of the meteorological department to obtain weather forecast information such as air temperature, relative humidity, precipitation probability, and wind speed for the next 3 - 7 days, and generates and pushes a prevention and control time window suggestion including specific dates and time periods to users through analysis in combination with rules such as avoiding rainfall periods in the next 24 hours, selecting a spraying temperature of 15 - 28 °C suitable for most chemical pesticides, and a wind speed not exceeding level 4.

[0038] S5. The feedback verification terminal receives the feedback data after the user implements the prevention and control measures and dynamically updates the model parameters.

[0039] Further, in the above technical solution, the specific implementation manner of the feedback verification terminal is as follows Figure 3 shown as: A1. Obtain the disease lesion regression rate and pest population survival rate after pesticide application through the intelligent monitoring terminal and the data processing terminal; A2. The feedback verification terminal reversely corrects the model parameters according to the disease lesion regression rate and the pest population survival rate, and iteratively optimizes the meteorological sensitivity coefficient, the crop resistance coefficient, and the pest and disease outbreak probability coefficient; A3. According to the correction result, re - allocate the weights of the meteorological sensitivity coefficient, the crop resistance coefficient, and the pest and disease outbreak probability coefficient.

[0040] It should be noted that the disease lesion regression rate , where A lesion，before is the proportion of the disease lesion area before pesticide application, and A lesion，after is the proportion of the disease lesion area after pesticide application; the pest population survival rate , where N surv is the initial number of the pest population counted before pesticide application, and N init is the number of survivors counted after pesticide application; Further, in the above technical solution, the rules of iterative optimization include: Correction of the meteorological sensitivity coefficient: C new =C old ×(1 - η C ×S surv ), where η C is the adjustment factor, S surv is the pest survival rate, C old is the meteorological sensitivity coefficient before iterative optimization, and C new is the meteorological sensitivity coefficient after iterative optimization; Correction of the crop resistance coefficient: R new =R old ×(1 + η R ×(1 - S surv )) where η R is the adjustment factor, S surv is the pest survival rate, R old is the crop resistance coefficient before iterative optimization, and R new is the crop resistance coefficient after iterative optimization; Correction of the pest and disease outbreak probability coefficient: P new =P old ×(1 + η P ×(1 - R rec )) Among them, η P is the adjustment factor, R rec is the rate of lesion regression, P old is the probability coefficient of pest and disease outbreak before iterative optimization, P new is the probability coefficient of pest and disease outbreak after iterative optimization; The implementation method of the A3 step is as follows: Compare the corrected model value with the model prediction value and calculate the error: , where Q predicted is the model prediction value, Q actual is the corrected model value; Combine the error and the corrected coefficient value to update the weight coefficient: , , , where E total =(C new +R new +P new )×(1 + E).

[0041] It should be noted that the setting of the adjustment factor η C : If the meteorological conditions have a significant impact on the prevention and control effect, such as in rainy and high-temperature areas, then increase η C , which can be set between 0.3 and 0.5; if the meteorological conditions are stable or have a small impact, then reduce η C , which can be set between 0.05 and 0.15; The setting of the η R : If the crop resistance is crucial for the prevention and control effect, such as in the promotion area of disease-resistant varieties, then increase η R , which can be set between 0.3 and 0.5; if the general growth of the crop is weak, then reduce η R , which can be set between 0.1 and 0.2; The setting of the η P : If the disease spreads rapidly, such as a high concentration of pathogen spores, then increase η P , which can be set between 0.4 and 0.6; if the pest is the main problem or the disease is stable, then reduce η P , which can be set between 0.1 and 0.3.

[0042] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not used to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An artificial intelligence-based early warning method for crop pests and diseases, characterized in that, It includes an intelligent monitoring terminal, a data processing terminal, an intelligent analysis terminal, an early warning decision-making terminal, and a feedback verification terminal. Specifically, it includes the following steps: S1. The intelligent monitoring terminal regularly collects multi-dimensional environmental data through the Internet of Things sensor array deployed in the farmland during the crop growth cycle to form a multi-source fusion data set; S2. The data processing terminal performs outlier removal and normalization processing on the multi-source fusion data set to generate a standardized training data set; S3. The intelligent analysis terminal performs multi-modal feature fusion analysis on the standardized training data set to obtain a meteorological sensitivity coefficient, a crop resistance coefficient, and a pest and disease outbreak probability coefficient, and inputs the three coefficients into the pest and disease risk quantification model to output the pest and disease risk index; S4. The early warning decision-making terminal triggers multi-level early warning rules according to the pest and disease risk index, and generates a customized early warning plan including suggestions for emergency measures; S5. The feedback verification terminal receives the feedback data after the user implements the prevention and control measures, and dynamically updates the model parameters.

2. The method for warning of crop diseases and insect pests based on artificial intelligence according to claim 1, wherein, The multi-source fusion data set includes a meteorological environment data set, a crop physiological data set, and a pest and disease characteristic data set; the meteorological environment data set includes the three-dimensional coordinates of the farmland, air temperature, relative humidity, light intensity, and precipitation; the crop physiological data set includes the relative chlorophyll content, initial fluorescence, maximum fluorescence, and stem water potential; the pest and disease characteristic data set includes the pest population density, the concentration of pathogen spores, and the proportion of the area of typical disease spots.

3. The method for warning of crop diseases and insect pests based on artificial intelligence according to claim 1, characterized in that, The meteorological sensitivity coefficient is specifically: ， Among them, t0 is the starting time of data acquisition, t is the current time, k is the temperature response attenuation coefficient, T air (t) is the air temperature at time t, T opt is the optimal temperature for crop growth, M(t) is the air humidity at time t, M base is the optimal humidity for crop growth, I light (t) is the light intensity at time t, I thresh is the light intensity threshold, P rain，max is the maximum single precipitation, e is the natural constant, and ln is the logarithm with base e.

4. A method for warning of crop pests and diseases based on artificial intelligence according to claim 1, characterized in that The crop resistance coefficient is specifically: ， Among them, α, β, γ, and δ are weighting factors, α + β + γ + δ = 1 and α, β, γ, and δ ∈ [0, 1], Chl is the relative chlorophyll content, Ψ stem is the stem water potential, Ψ wilt is the critical value of wilting water potential, F min is the minimum fluorescence under dark adaptation, A lesion is the proportion of the area of typical lesions, F v is the variable fluorescence, and the calculation formula is F v = F m - F o , where F m is the maximum fluorescence, and F o is the initial fluorescence.

5. The method for warning of crop pests and diseases based on artificial intelligence according to claim 1, characterized in that, The pest and disease outbreak probability coefficient is specifically: ， Among them, λ, η, and ρ are adjustment factors, D pest is the density of the pest population, S spores is the concentration of pathogen spores, A lesion is the proportion of the area of typical lesions, Chl is the relative chlorophyll content, Chl crit is the chlorophyll critical value, T air is the air temperature, T pest is the optimal temperature for pest activity, and e is the natural constant.

6. The method for warning of crop pests and diseases based on artificial intelligence according to claim 1, characterized in that, The pest and disease risk quantification model is specifically: ， Among them, , ω1, ω2 and ω3 are dynamic weight coefficients, C is the meteorological sensitivity coefficient, R is the crop resistance coefficient, P is the pest and disease outbreak probability coefficient, τ is the historical data correction factor, Q hist is the pest and disease risk index in the same historical period, and e is the natural constant.

7. The method for warning of crop diseases and insect pests based on artificial intelligence according to claim 1, characterized in that, The multi-level early warning rules include: If the pest and disease risk index Q≥Q1, trigger a first-level early warning and take chemical emergency measures; If Q2≤Q<Q1, trigger a second-level early warning and implement biological control; If Q3≤Q<Q2, trigger a third-level early warning and start enhanced inspections; If 0≤Q<Q3, trigger a fourth-level early warning and recommend routine monitoring.

8. A method for warning of crop pests and diseases based on artificial intelligence according to claim 1, characterized in that, The customized early warning plan includes: Match a list of recommended pesticides for different pest and disease types; Generate a drone spraying path plan based on the three-dimensional coordinates of the farmland; Provide the best prevention and control window period in combination with weather forecast data.

9. The method for warning of crop diseases and insect pests based on artificial intelligence according to claim 1, wherein The specific implementation method of the feedback verification terminal is: A1. Obtain the disease spot regression rate and pest population survival rate after spraying through the intelligent monitoring terminal and the data processing terminal; A2. The feedback verification terminal reversely corrects the model parameters according to the disease spot regression rate and the pest population survival rate, and iteratively optimizes the meteorological sensitivity coefficient, the crop resistance coefficient, and the pest and disease outbreak probability coefficient; A3. According to the correction results, reallocate the weights of the meteorological sensitivity coefficient, the crop resistance coefficient, and the pest and disease outbreak probability coefficient.

10. A method for warning of crop pests and diseases based on artificial intelligence according to claim 9, characterized in that, The rules for iterative optimization include: Correction of the meteorological sensitivity coefficient: C new =C old ×(1 - η C ×S surv ) Among them, η C is the adjustment factor, S surv is the survival rate of pests, C old is the meteorological sensitivity coefficient before iterative optimization, C new is the meteorological sensitivity coefficient after iterative optimization; Correction of the crop resistance coefficient: R new =R old ×(1 + η R ×(1 - S surv )) Among them, η R is the adjustment factor, S surv is the pest survival rate, R old is the crop resistance coefficient before iterative optimization, and R new is the crop resistance coefficient after iterative optimization; Correction of the pest and disease outbreak probability coefficient: P new = P old × (1 + η P × (1 - R rec )) Among them, η P is the adjustment factor, R rec is the lesion regression rate, P old is the probability coefficient of pest and disease outbreak before iterative optimization, and P new is the probability coefficient of pest and disease outbreak after iterative optimization; The implementation method of step A3 is: Compare the corrected model value with the model prediction value and calculate the error: ， Among them, Q predicted is the model prediction value, and Q actual is the corrected model value; Combine the error and the corrected coefficient value to update the weight coefficient: ， ， ， Among them, E total =(C new +R new +P new )×(1 + E).

Images