A tea garden small green leafhopper prediction method and system based on a population dynamic model
By using population dynamics models and multi-source monitoring assimilation technology, a digital twin of tea gardens was constructed to achieve self-evolutionary control efficacy calibration and multi-objective strategy optimization. This solved the problems of insufficient accuracy and extensive control strategies in the prediction of tea green leafhopper infestation, and improved the accuracy of prediction and the adaptability of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 江西省经济作物研究所
- Filing Date
- 2026-03-10
- Publication Date
- 2026-06-09
AI Technical Summary
Existing technologies for predicting the infestation of the green leafhopper in tea gardens suffer from problems such as insufficient prediction accuracy, inadequate characterization of control efficacy and resistance, extensive control strategies, and difficulty in sharing knowledge among multiple tea gardens. This results in prediction results that are highly subjective, poorly adaptable, and lack multi-objective evaluation and cross-garden knowledge collaboration.
A prediction method based on a population dynamics model is adopted, which is combined with multi-source intelligent monitoring data for assimilation to construct a digital twin of tea gardens. This enables self-evolving prevention efficacy calibration and multi-objective strategy optimization. Furthermore, knowledge collaboration among multiple tea gardens is achieved through federated learning, forming a self-learning, self-adaptive, and self-repairing prevention and control system.
It improves the accuracy and reliability of pest forecasting, realizes the linkage simulation of tea garden ecosystem, has the ability to self-evolving prevention efficacy and resistance identification, supports multi-objective strategy optimization and cross-garden knowledge sharing, and enhances the adaptability and scalability of the system.
Abstract
Claims
1. A method for predicting the small green leafhopper in tea gardens based on a population dynamics model, characterized in that, include: S1. Obtain information on the target tea garden, including plot information, tea tree variety and phenological information, historical insect data, overwintering insect population survey data, and daily meteorological data. Divide the tea garden into several spatial sub-regions, determine the overwintering insect population density in each sub-region, and convert it into an overwintering base coefficient. S2. Based on the biological characteristics of the green leafhopper, its life cycle is divided into multiple stages, including at least overwintering stage, egg, nymph, and adult. A development rate function and a natural mortality function with temperature as the independent variable are established for each stage. The reproduction rate, development rate, and mortality rate are corrected by combining tea tree phenological factors to construct a structured population dynamic model of the insect stages. S3. Input daily meteorological data into the population dynamics model, use the overwintering base coefficient of each sub-region as the initial condition to drive the insect state transition, and obtain the basic insect population dynamic trajectory of each sub-region without considering prevention and control and extreme weather disturbances. S4. Deploy insect monitoring terminals and meteorological sensors in representative sub-regions, collect insect population observation data and environmental data at preset time intervals, and establish a mapping relationship between monitoring terminals and spatial sub-regions to form insect population observation time series for each sub-region. S5. Abstract chemical control, biological control, and physical control operations into control events, and abstract heavy rainfall, continuous high temperature, sudden temperature drop, and long-term cloudy and rainy weather into weather events. Form an event library containing event type, occurrence time, sub-region of effect, and disturbance parameters. Trigger control events and weather events according to the time axis during model iteration, correct the state of basic insect population dynamics, and obtain the prior population state of each sub-region. S6. Compare the prior population state with the insect population observation data at the corresponding time. When the deviation exceeds the preset threshold, call the assimilation algorithm to adjust the number of each insect stage as a whole or by insect stage, obtain the posterior population state and use it as the initial condition for the next time step, and realize the closed-loop correction between model prediction and actual observation. S7. At any assimilation time, using the current posterior population state as the initial condition, and combining the meteorological forecast data for a future period with the no-control, conventional control and candidate control schemes, a forward simulation is performed to obtain the insect population prediction curve for each scheme. Based on the economic threshold of the green leafhopper, tea tree phenology and the safe interval for tea picking, the overthreshold area, overthreshold duration and insect population peak are calculated to form a control window score and risk level, and output recommended control start and end times and control intensity suggestions.
2. The method for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 1, characterized in that, In step 6, the assimilation algorithm includes at least one of weighted update, filtering, or Bayesian estimation; The assimilation weights are adaptively adjusted based on the magnitude of the deviation between the predicted and observed values. When the deviation exceeds the second threshold for multiple consecutive steps, the parameters of the development rate function or the natural mortality function are corrected online.
3. The method for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 1, characterized in that, In step 5, the prevention and control event includes at least one of the following: spraying chemical agents, releasing natural enemies, hanging yellow sticky traps, or setting up physical prevention and control facilities; For different prevention and control events, a corresponding insect stage's prevention effectiveness vector and prevention effectiveness decay function over time are pre-defined. When a prevention and control event is triggered, the number of target insect stages is reduced, and the population state is continuously corrected according to the decay function in subsequent time steps.
4. The method for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 1, characterized in that, In step 5, the weather event includes at least one of the following: heavy rainfall, sustained high temperatures, sudden temperature drops, and prolonged overcast and rainy weather; For different weather events, perturbation coefficients are set for the development rate of each insect stage, the natural mortality rate, and the migration probability. During the weather event, the perturbation coefficients are used to correct the parameters of the population dynamics model.
5. The method for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 1, characterized in that, In step 7, the prevention and control window score is calculated based on at least one or more of the following indicators: The time integral exceeding the economic threshold, peak insect population density, degree of conflict with the safe harvesting interval, and number of control operations; The prevention and control plan is then divided into different risk levels based on the scoring results.
6. The method for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 1, characterized in that, Furthermore, it also includes: Statistical analysis was conducted on the insect population prediction curves and scoring results corresponding to different time periods and different candidate schemes to form a prevention and control scheme library that matches the tea garden management preferences. In subsequent predictions, the schemes with better historical performance in the scheme library were given priority recommendation.
7. A tea garden green leafhopper prediction system based on a population dynamics model according to any one of claims 1-6, characterized in that, This includes server-side components, data storage units, communication networks, and pest monitoring terminals, meteorological data acquisition terminals, and user terminals deployed in the tea garden; among them, The server-side includes at least: The basic data management module is used to manage and store information on tea garden plots and spatial sub-regions, tea tree varieties and phenological information, insect data, overwintering insect population survey data, historical and forecast meteorological data, and control operation records. The population dynamics model module is used to construct a structured population dynamics model of insects based on the overwintering base coefficient, daily meteorological data and tea tree phenological information, and output the basic insect population dynamic trajectory of each spatial sub-region. The monitoring data access module is used to receive insect population observation data and environmental data uploaded by the insect monitoring terminal and the meteorological acquisition terminal, perform time alignment and preprocessing on the data, and organize it into insect population observation time series for each sub-region according to the "monitoring terminal - sub-region" mapping relationship; The event library and scheduling module are used to abstract prevention and control operations into prevention and control events and meteorological processes into weather events, forming an event library. During model runtime, event information is sent to the population dynamic model module in chronological order to correct the population state and obtain the prior population state. The assimilation and state reconstruction module is used to receive prior population state and insect population observation data, calculate the deviation between the predicted value and the observed value, and call the assimilation algorithm to adjust the number of each insect state when the deviation exceeds the preset threshold, obtain the posterior population state and write it back to the population dynamic model module. The prediction and evaluation module is used to perform forward simulations of multiple scenarios after each assimilation update, using the posterior population state as the initial condition, combined with future weather forecast data and candidate control schemes, to calculate the control window score and risk level and generate recommended control schemes. The human-computer interaction and visualization module is used to display the dynamics of the insect population, spatial distribution, risk level, and evaluation results of prevention and control plans to users, and to receive user settings and adjustments to parameters and plans; Each module communicates and collaborates via a system bus and / or network to perform the method described in any one of claims 1 to 6.
8. The method and system for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 7, characterized in that, The insect monitoring terminal includes at least one of an automatic counting yellow sticky trap and an image recognition-based trapping device; The monitoring data access module is also used to identify and count insect images, and compare and fuse the identification results with the automatic counting results.
9. The method and system for predicting the small green leafhopper in tea gardens based on a population dynamics model according to claim 7, characterized in that, The prediction and evaluation module is also used to perform statistical analysis on the prediction results of different time periods and different candidate schemes, form a prevention and control scheme library, and mark the schemes with better historical performance as the preferred schemes for users to call.
10. A tea garden green leafhopper prediction system based on a population dynamics model according to any one of claims 7-9, characterized in that, The server also includes a report generation module, which generates a report document containing pest prediction curves, control windows, recommended solutions, and risk descriptions based on the prediction and assessment results.