Method and system for ecological prevention and control of chinch bug

Abstract

The application discloses a method and system for ecological prevention and control of chinch bugs, and relates to the field of livestock production.The method comprises the following steps: acquiring multi-source data; constructing a feature vector based on the multi-source data, performing principal component analysis dimension reduction on the feature vector, and generating a time series feature matrix; inputting the time series feature matrix into a pre-trained time convolution network to output a predicted insect population density; calculating a risk score based on the predicted insect population density using a pest risk dynamic scoring model; dividing a risk level based on the risk score; determining corresponding prevention and control measures according to the risk level to perform ecological prevention and control of chinch bugs.The application can realize early and accurate early warning, intelligent dynamic prediction and targeted ecological prevention and control of pests, and can balance the prevention and control effect and environmental protection, thereby improving the safety and sustainability of pasture production.

Description

Technical Field

[0001] This application relates to the field of animal husbandry, and in particular to a method and system for the ecological control of mirid bugs in pasture. Background Technology

[0002] Forage grasses are an important feed source for livestock, and their yield and quality directly affect the efficiency and sustainable development of livestock production. However, forage grass production is often attacked by various pests. Among them, the forage grass mirid bug, due to its feeding and sucking behavior, causes serious damage to the leaves and roots of forage grasses, leading to stunted growth, reduced yield, and even large-scale outbreaks, resulting in significant economic losses forage grass cultivation.

[0003] Currently, the control of pasture mirid bugs mainly relies on traditional chemical pesticide spraying, which has the following prominent problems: (1) Single monitoring methods make it difficult to achieve early warning: Traditional pest monitoring relies heavily on manual field surveys, which are time-consuming and prone to errors, making it difficult to grasp the dynamics of pest occurrence and development in a timely and accurate manner. Single environmental parameters or pest density data lack multi-dimensional correlation analysis, resulting in unscientific pest risk assessment and delayed control measures.

[0004] (2) Insufficient prediction models make it difficult to achieve dynamic and accurate prediction: Existing pest predictions mostly use empirical formulas or simple statistical models, ignoring the complex spatiotemporal relationship between environmental factors and pest population dynamics. They lack the ability to assign weights to key influencing factors and intelligent learning capabilities, resulting in low prediction accuracy and real-time performance, which makes it difficult to meet the needs of modern precision agriculture.

[0005] (3) Single control method and great pressure on the ecological environment: The heavy reliance on chemical pesticides not only aggravates environmental pollution and the risk of pasture residue, but also easily leads to pesticide resistance in mirid bugs, reducing the control effect. Although biological control methods have ecological advantages, the control effect is unstable because the timing and quantity of release are difficult to control precisely.

[0006] (4) Insufficient data communication and processing capabilities make it difficult to achieve real-time intelligent decision-making: The pasture planting areas are widely distributed and the base station coverage is insufficient. The existing monitoring equipment has poor communication, high data transmission packet loss rate, and lacks edge computing capabilities, resulting in slow response to pest risk assessment and control decisions, which affects the control effect. Summary of the Invention

[0007] To address the aforementioned problems, this application provides an ecological control method and system for mirid bugs in pastureland.

[0008] To achieve the above objectives, this application provides the following solution: Firstly, this application provides an ecological control method for pasture mirid bugs, including: Acquire multi-source data; the multi-source data includes environmental data, insect infestation information, and images of insect infestation hotspots. Based on the multi-source data, feature vectors are constructed, and principal component analysis is performed on the feature vectors to reduce dimensionality, generating a time-series feature matrix; The temporal feature matrix is ​​input into a pre-trained temporal convolutional network, which outputs the predicted insect population density. Based on the predicted insect population density, a risk score is calculated using a dynamic insect risk scoring model. Risk levels are determined based on the aforementioned risk scores; Based on the risk level, corresponding prevention and control measures are determined to carry out ecological control of pasture mirid bugs.

[0009] Secondly, this application provides an ecological control system for pasture mirid bugs, which implements the above-described ecological control method for pasture mirid bugs, the system comprising: A multi-source data acquisition module is used to acquire multi-source data, including environmental data, pest information, and images of pest spread hotspots. The time-series feature matrix generation module is used to construct feature vectors based on the multi-source data, and perform principal component analysis to reduce the dimensionality of the feature vectors to generate a time-series feature matrix. The insect population density prediction membrane is used to input the temporal feature matrix into a pre-trained temporal convolutional network and output the predicted insect population density. The risk score calculation module is used to calculate the risk score based on the predicted insect population density using a dynamic insect risk scoring model. A risk level classification module is used to classify risk levels based on the risk score. The prevention and control module is used to determine the corresponding prevention and control measures based on the risk level, and to carry out ecological prevention and control of pasture mirid bugs.

[0010] According to the specific embodiments provided in this application, this application has the following technical effects: (1) Achieve multi-dimensional pest monitoring and improve early warning capabilities: By integrating three types of heterogeneous data—environment, pests, and images—multi-dimensional data support is provided for early pest identification, which helps to achieve timely understanding of pest dynamics and early warning. (2) Improve the accuracy and timeliness of insect population density prediction: Temporal convolutional networks have the ability to capture temporal dependencies. Combined with temporal feature matrices, they can effectively model the complex spatiotemporal relationship between environmental factors and insect population dynamics, thereby improving prediction accuracy and real-time performance. (3) Achieve scientific quantification and graded early warning of pest risks: Calculate risk scores through a dynamic pest risk scoring model, classify risk levels based on risk scores, transform predicted pest population density into quantifiable risk indicators, and implement graded management to provide a basis for differentiated and precise prevention and control decisions. (4) Promote ecologically-oriented precision prevention and control to reduce environmental pressure: Determine the corresponding prevention and control measures according to the risk level, carry out ecological prevention and control of pasture mirid bugs, and emphasize the implementation of "ecological prevention and control" based on the risk level; guide differentiated and appropriate prevention and control strategies through risk classification, which helps to reduce the abuse of chemical pesticides, reduce environmental pollution and residue risks, and meet the needs of sustainable development. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A flowchart illustrating an ecological control method for mirid bugs in pastureland, provided as an embodiment of this application; Figure 2 This is a schematic diagram of the trapping unit. Detailed Implementation

[0013] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0014] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0015] In one exemplary embodiment, such as Figure 1 As shown, an ecological control method for pasture mirid bugs is provided. This method is executed by computer equipment, specifically by a terminal or server alone, or by both a terminal and a server. In this embodiment, the method is described using a server as an example, and includes the following steps S1 to S6.

[0016] S1: Acquire multi-source data; the multi-source data includes environmental data, pest information, and images of pest spread hotspots.

[0017] S2: Construct feature vectors based on the multi-source data, and perform principal component analysis to reduce the dimensionality of the feature vectors to generate a time-series feature matrix.

[0018] S3: Input the temporal feature matrix into a pre-trained temporal convolutional network and output the predicted insect population density.

[0019] S4: Based on the predicted insect population density, a risk score is calculated using a dynamic insect risk scoring model.

[0020] S5: Classify risk levels based on the aforementioned risk score.

[0021] S6: Determine the corresponding prevention and control measures based on the risk level and carry out ecological control of pasture mirid bugs.

[0022] By implementing steps S1 to S6 above, multi-source data fusion is used to overcome the limitations of traditional single monitoring methods. By leveraging the intelligent learning capabilities of temporal feature extraction and temporal convolutional networks, the modeling of the complex spatiotemporal relationship between environmental factors and insect population dynamics is enhanced. This overcomes the shortcomings of inaccurate predictions and delayed responses of empirical formulas, enabling the scientific quantification and graded early warning of pest risks. This, in turn, guides differentiated and appropriate ecological control decisions, reduces excessive reliance on chemical pesticides, and improves the accuracy, timeliness, and eco-friendliness of pasture mirid bug control.

[0023] In one specific embodiment, the environmental data in step S1 includes meteorological data and soil parameters; the insect information includes the surface reflectance spectral characteristics of the pasture mirid bug and the vibration signals of nymphs feeding.

[0024] Step S1 specifically includes the following steps: S11: Meteorological data is collected by a group of meteorological sensors deployed in the pasture canopy, and soil parameters are collected by soil parameter sensors buried in the underground root system.

[0025] S12: A multispectral imager was used to identify the surface reflectance spectral characteristics of the pasture mirid bug, and an acoustic sensor was used to capture the vibration signals of the nymphs feeding.

[0026] S13: Control the cruise drone equipped with a thermal imager to cruise along a preset grid path and acquire images of hotspot areas for pest spread.

[0027] In one specific embodiment, step S2 specifically includes: The environmental data (including meteorological data and soil parameters) and insect information (including the surface reflectance spectral characteristics of pasture mirid bugs and the vibration signals of nymphs feeding) obtained in step S1 are aligned with the images of the hotspot areas of insect infestation according to a unified timestamp. For various types of data collected at the same time, the numerical data (temperature, humidity, leaf surface wetting time, soil pH, organic matter content, reflectance ratio R780 / R550, main frequency and intensity of vibration signals, and average temperature of hotspot areas, etc.) are directly used as feature components. For image data, quantitative features such as damage area, damage shape, and distribution of hotspot areas are extracted through a pre-trained convolutional neural network. All feature components are then concatenated into an original feature vector. Principal component analysis is applied to reduce dimensionality, and the sliding average within a set sliding window and the relative rate of change of adjacent sliding windows are calculated. The sliding average and the relative rate of change are arranged in chronological order to generate a temporal feature matrix, which is used as the input for the subsequent temporal convolutional network.

[0028] In a specific embodiment, step S3 specifically includes: The temporal feature matrix is ​​input into a pre-trained temporal convolutional network, which introduces an attention mechanism to assign weights to the following key factors: Effective accumulated temperature: When the effective accumulated temperature for 5 consecutive days is greater than 125℃·day, its weight is multiplied. Leaf surface wetting time: Leaf surface wetting time has an exponential relationship with egg production, and its weight = 0.28 × exp(0.05 × wetting duration). Leaf surface wetting time is the duration when the canopy humidity is >90%. Natural enemy regulatory factor: When the ratio of ladybugs to mirid bugs reaches 1:50, the probability of inhibiting population growth is >70%, and its weight is -0.12×ln(number of ladybugs / number of mirid bugs).

[0029] This application employs a 4-layer temporal convolutional network with a kernel size of 3 and an inflation factor of [1, 2, 4, 8]. Model distillation techniques are used to compress the temporal convolutional network to 15% of its original volume (e.g., from 200MB to 30MB), enabling real-time risk assessment in less than 200ms.

[0030] The temporal convolutional network is trained using historical multi-source data and corresponding historical insect population density.

[0031] In one specific embodiment, the expression for the dynamic pest risk scoring model is: RiskScore=α×Tacc+β×LWD+γ×(Nt / Nt-1)-δ×NPR Among them, RiskScore is the risk score, Tacc is the effective accumulated temperature, LWD is the leaf wet time, and N is the nitrogen content. t To predict insect population density, N t-1N is the pest population density in the previous cycle, NPR is the natural enemy regulation factor, and α, β, γ, and δ are all weight coefficients optimized by the genetic algorithm. In this application, α is 0.35, β is 0.28, γ is 0.25, and δ is 0.12.

[0032] The specific process of genetic algorithm optimization is as follows: Set the population size of the genetic algorithm to 100 individuals. Each individual consists of four weight coefficients α, β, γ, and δ. The initial population is randomly generated within the interval [0, 1] and satisfies the constraint of α + β + γ + δ = 1. Then, define the fitness function as the reciprocal of the root mean square error between the risk score RiskScore and the actual pest occurrence level during the historical time period. The actual pest occurrence level is determined based on the historical measured pest population density and the risk levels (assigned values of 0.3, 0.65, and 1.0 for low, medium, and high respectively) marked by experts. The higher the fitness value, the better the weight combination. Then, use the roulette wheel selection method to select parental individuals according to the fitness ratio, and perform crossover operations with a single-point crossover probability of 0.8. The crossover point is randomly selected. After crossover, generate offspring individuals. Then, perform Gaussian mutation on each weight coefficient of the offspring individuals with a mutation probability of 0.05, that is, add a random perturbation with a mean of 0 and a standard deviation of 0.05 to the original value, truncate the values outside the interval [0, 1], and renormalize to make the sum of the four weights equal to 1. The number of elite individuals retained in each generation of the population is 2 (the two individuals with the highest fitness are directly copied to the next generation). After 200 generations of iteration, select the individual with the highest fitness as the final optimization result. The finally obtained optimal weight coefficients are α = 0.35, β = 0.28, γ = 0.25, and δ = 0.12. This set of coefficients makes the prediction accuracy of the risk scoring model on the validation set reach more than 92%, thus completing the optimization of the weight coefficients by the genetic algorithm.

[0033] In a specific embodiment, step S5 specifically includes: When RiskScore ≤ RiskScore1, it is a low risk (green warning); RiskScore is the risk score, and RiskScore1 is the first risk score threshold; when RiskScore1 < RiskScore ≤ RiskScore2, it is a medium risk (yellow warning); RiskScore2 is the second risk score threshold; when RiskScore > RiskScore2, it is a high risk (red warning).

[0034] In this application, RiskScore1 is 0.3 and RiskScore2 is 0.65.

[0035] In a specific embodiment, step S6 specifically includes: S61: When the risk level is low, the corresponding prevention and control measures are: maintain the monitoring frequency of multi-source data, keep the trap array in standard configuration operation, and release 2 adult lacewings / m² per week. 2 To carry out preventive biological control.

[0036] S62: When the risk level is medium risk, the corresponding prevention and control measures are: increase the monitoring frequency of multi-source data, adjust the trap array to medium-efficiency mode, and spray plant endophytic bacteria preparations (concentration 5×10). 7 CFU / mL), release natural enemies differentiated according to insect age (egg stage: 3 lacewing larvae / mL). 2 Nymphal stage: 2 adult assassin bugs / m 2 The process involves inducing forage grasses to produce insect-resistant secondary metabolites through atomized spraying of plant endophytic bacteria preparations.

[0037] S63: When the risk level is high, the corresponding prevention and control measures are as follows: adjust the trap array to high-efficiency mode, increase the density of trapping points (to 4 per acre), and release natural enemies according to the age of the insects (6 lacewings per m²). 2 4 assassin bugs / m 2 Place 30 mirid bug virus bait strips per acre, spray with methyl jasmonate slow-release capsules to activate plant resistance, regulate soil microecology, and increase the application of beneficial microbial agents.

[0038] The trapping array comprises multiple detachable and combinable trapping units. For example... Figure 2 As shown, each trapping unit includes a reflector 1 and a light attractor 2; the standard configuration of the trapping device is that the reflector angle of the reflector 1 is 30° and the light intensity of the light attractor 2 is 300 lux; the medium-efficiency mode of the trapping device is that the reflector angle of the reflector 1 is 45° and the light intensity of the light attractor 2 is 400 lux; the high-efficiency mode of the trapping device is that the reflector angle of the reflector 1 is 60° and the light intensity of the light attractor 2 is 500 lux.

[0039] In the above method, the insect age is identified based on a pre-trained convolutional neural network (CNN) or support vector machine (SVM). The specific steps are as follows: (1) Data acquisition: The canopy of pasture was scanned by a multispectral imager to obtain images of leaf damage areas; vibration signals of nymphs feeding were captured by an acoustic sensor; a cruise drone equipped with a thermal imager was used to identify images of hotspots of pest spread; and temperature and humidity sensors provided time-series data of temperature and humidity. (2) Feature extraction: Extract the following from the image of the damaged area of ​​the leaf: Damage area (unit: mm) 2Damage shape (circular / striped / dotted), reflectance ratio (R780 / R550); extracted from nymph feeding vibration signals: dominant frequency (nymph: 18–20kHz, adult: 20–22kHz), signal duration and intensity; extracted from images of pest spread hotspot areas: mean temperature and distribution of hotspot areas; extracted from temperature and humidity time series data: effective accumulated temperature, leaf surface wetting time.

[0040] (3) Model inference: Input the extracted features into the insect age classification model (such as convolutional neural network CNN or support vector machine SVM) in the pre-trained model library. The model outputs the insect age stage: egg stage: mainly relies on environmental factors (temperature, humidity) to predict the hatching probability; nymph stage: relies on nymph feeding vibration signal + small area damage image + abnormal temperature area; adult stage: relies on large area damage image + high frequency vibration signal + high reflectivity ratio.

[0041] (4) Insect age output: The output is structured insect age data.

[0042] Monitoring of control effectiveness: Effectiveness evaluation begins within 24 hours of implementing control measures. The change rate of insect population density before and after control is compared, and the control efficiency is calculated as: η = (N0 - N1) / N0 × 100%. Survival rate and activity range of natural enemies are recorded. Here, η represents the control efficiency, a comprehensive indicator evaluating the control effect by the change rate of insect population density before and after the implementation of control measures. N0 represents the insect population density before the implementation of control measures (the last monitoring within 24 hours before implementation), in units of insects / m². 2 N1 represents the insect population density after the implementation of control measures (monitored within 24 to 72 hours after implementation), expressed in heads / m³. 2 This formula quantitatively characterizes the effectiveness of control measures by calculating the relative percentage decrease in insect population density; the higher the η value, the more significant the control effect.

[0043] Decision feedback loop: Analyze successful / failed control cases, extract key factors, adjust risk thresholds and response strategies, and generate control reports, including: pest occurrence trend analysis, evaluation of the effectiveness of control measures, prevention recommendations for the next stage, and cost-benefit analysis.

[0044] Equipment status monitoring: Daily self-check of sensor working status, monitoring of battery power, sending replacement reminder when it is below 20%, calibrating sensor accuracy, and automatically correcting when the deviation is >5%.

[0045] Data storage and backup: Detailed data for the past 30 days is stored locally, while historical data and model parameters are backed up in the cloud. Data compression rate reaches 70% to ensure storage efficiency.

[0046] Based on the same inventive concept, this application also provides a system for implementing the above-mentioned ecological control method for mirid bugs in pastureland. The solution provided by this system is similar to the solution described in the above method; therefore, the specific limitations of one or more embodiments of the ecological control system for mirid bugs in pastureland provided below can be found in the limitations of the ecological control method for mirid bugs in pastureland described above, and will not be repeated here.

[0047] In one exemplary embodiment, an ecological control system for pasture mirid bugs is provided, comprising the following modules.

[0048] The multi-source data acquisition module is used to acquire multi-source data, including environmental data, pest information, and images of pest spread hotspots.

[0049] The time-series feature matrix generation module is used to construct feature vectors based on the multi-source data, and perform principal component analysis to reduce the dimensionality of the feature vectors to generate a time-series feature matrix.

[0050] The insect population density prediction membrane is used to input the temporal feature matrix into a pre-trained temporal convolutional network and output the predicted insect population density.

[0051] The risk score calculation module is used to calculate the risk score based on the predicted insect population density using a dynamic insect risk scoring model.

[0052] The risk level classification module is used to classify risk levels based on the risk score.

[0053] The prevention and control module is used to determine the corresponding prevention and control measures based on the risk level, and to carry out ecological prevention and control of pasture mirid bugs.

[0054] Furthermore, the multi-source data module includes: a meteorological sensor group, a soil parameter sensor, a multispectral imager, an acoustic sensor, and a thermal imager.

[0055] Meteorological sensor array 1 is deployed in a grid pattern within the pasture canopy (20m × 20m spacing) to collect meteorological data. The array collects data every 5 minutes. It includes a temperature sensor (accuracy ±0.5℃), a humidity sensor (range 0-100%RH), and a leaf condensation sensor (capacitive, resolution 0.01mm). The meteorological data includes temperature, humidity, and leaf condensation.

[0056] Soil parameter sensors are buried in the underground root layer (15cm deep) to collect soil parameters. The soil parameter sensors include a pH sensor (range 3-9, accuracy ±0.2) and an organic matter content sensor (near-infrared spectroscopy, error <5%). Soil parameters include soil pH and soil organic matter content.

[0057] A multispectral imager is used to identify the surface reflectance spectral characteristics of mirid bugs in pasture. The imager employs dual-channel imaging in the 780nm near-infrared band and the 550nm visible light band. The reflectance ratio (R780 / R550) is used to distinguish between healthy leaves and areas damaged by mirid bug piercing and sucking. For healthy leaves, R780 / R550 ≈ 1.8-2.2; for mirid bug-damaged areas, R780 / R550 < 1.3 (detectable up to 0.5mm). 2 (Micro-damage).

[0058] The acoustic signature sensor is used to capture vibration signals from nymphs feeding. The sensor has a built-in 40kHz high-frequency microphone to capture these vibration signals (characteristic frequency band 18-22kHz).

[0059] Thermal imagers are deployed on cruise drones to acquire images of pest spread hotspots. The quadcopter cruise drone is equipped with a high-resolution thermal imager (640×512 resolution, temperature range -20-150℃), cruises along a preset route (flight altitude 5m, speed 3m / s), dynamically identifies pest spread hotspots (areas with abnormal temperature increases >2℃), and generates images of these hotspot areas.

[0060] Furthermore, the trapping array comprises multiple detachable and combinable trapping units; each trapping unit includes a reflector with a rotatable support (rotation angle 0-180°) and a light attractor. The reflector surface is coated with a high-reflectivity aluminum film (reflectivity > 95%), which, in conjunction with the light attractor, emits a light source in the phototactic band of pasture bugs (420-470nm, blue light), increasing the trapping efficiency by more than 20% compared to traditional fixed traps.

[0061] The angle of the reflector can be adjusted manually or electrically based on the real-time wind speed (obtained by a wind speed sensor) and the direction of pest activity.

[0062] Furthermore, the natural enemy coordinated release unit includes a partitioned natural enemy storage chamber and an electrically controlled release valve; the partitioned natural enemy storage chamber is equipped with a lacewing larva chamber and an assassin bug adult chamber; the electrically controlled release valve selectively opens the corresponding lacewing larva chamber or assassin bug adult chamber according to the insect age identification result, and releases natural enemies according to the age difference.

[0063] Furthermore, the microecological regulation device includes a hydraulic sprayer equipped with a dual-chamber storage tank. The first chamber of the dual-chamber storage tank stores a plant endophytic bacteria preparation (Bacillus cereus, concentration 1×10⁻⁶). 8 The second chamber of the dual-chamber storage tank stores a suspension containing methyl jasmonate sustained-release capsules (particle size 50 μm, encapsulation rate >90%). The release rate of the suspension containing methyl jasmonate sustained-release capsules is triggered by the concentration of β-glucosidase in root exudates, achieving precise induction of the insect-resistant phenotype in forage grasses.

[0064] The modules of the system provided in this application are interconnected through a LoRaWAN and NB-IoT dual-mode communication module, and self-organizing network relay transmission is used in areas without base stations (packet loss rate <0.3%).

[0065] This application achieves comprehensive and accurate monitoring of pasture mirid bug infestation through a multi-source data acquisition module; it establishes a scientific pest dynamic prediction model based on advanced machine learning algorithms and a pest risk dynamic scoring model, which greatly improves the accuracy and response speed of risk assessment; the control module integrates physical trapping, biological control, and micro-ecological regulation as three control methods, which not only effectively improves control efficiency but also takes into account ecological environmental protection and significantly reduces the amount of chemical pesticides used.

[0066] This application features a high degree of intelligence, automation, and integration, adapting to the complex and ever-changing environment in which mirid bugs occur in forage grasses. It enables early warning, precise control, and dynamic management of the pests. By effectively integrating environmental monitoring, pest identification, intelligent decision-making, and targeted control, it significantly improves the safety and sustainability of forage production and promotes the green development of the forage industry.

[0067] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0068] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0069] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0070] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0071] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for ecological control of pasture mirid bugs, characterized in that, Including: Obtain multi-source data; the multi-source data includes environmental data, pest information, and images of pest infestation hotspots; Construct a feature vector based on the multi-source data, and perform principal component analysis dimensionality reduction on the feature vector to generate a time series feature matrix; Input the time series feature matrix into a pre-trained temporal convolutional network to output the predicted pest population density; Based on the predicted pest population density, calculate the risk score using a dynamic pest risk scoring model; Divide the risk levels based on the risk score; Determine the corresponding prevention and control measures according to the risk level, and conduct ecological prevention and control of Lygus lucorum.

2. The ecological control method for mirid bugs in pastureland according to claim 1, characterized in that, The environmental data includes meteorological data and soil parameters; the pest information includes the surface reflection spectral characteristics of the Lygus lucorum body and the feeding vibration signals of nymphs; Among them, obtaining multi-source data specifically includes: Collect meteorological data through a meteorological sensor group deployed in the forage canopy, and at the same time collect soil parameters through soil parameter sensors buried in the underground root layer; Use a multispectral imager to identify the surface reflection spectral characteristics of the Lygus lucorum body, and use a voiceprint sensor to capture the feeding vibration signals of nymphs; Control a cruising unmanned aerial vehicle equipped with a thermal imager to cruise along a preset grid path to obtain images of pest infestation hotspots.

3. The ecological control method for mirid bugs in pastureland according to claim 1, characterized in that, The expression of the dynamic pest risk scoring model is: RiskScore=α×Tacc+β×LWD+γ×(N t / N t-1 )-δ×NPR Among them, RiskScore is the risk score, Tacc is the effective accumulated temperature, LWD is the leaf wet time, and N is the nitrogen content. t To predict insect population density, N t-1 The insect population density represents the previous cycle, NPR is the natural enemy regulatory factor, and α, β, γ, and δ are all weighting coefficients.

4. The ecological control method for mirid bugs in pastureland according to claim 1, characterized in that, Dividing the risk levels based on the risk score specifically includes: When RiskScore ≤ RiskScore1, it is a low risk; RiskScore is the risk score, and RiskScore1 is the first risk score threshold; When RiskScore1 < RiskScore ≤ RiskScore2, it is a medium risk; RiskScore2 is the second risk score threshold; When RiskScore > RiskScore2, it is a high risk.

5. The ecological control method for mirid bugs in pastureland according to claim 4, characterized in that, Determining the corresponding prevention and control measures according to the risk level specifically includes: When the risk level is low, the corresponding prevention and control measures are as follows: maintain the monitoring frequency of multi-source data, keep the trap array in standard configuration operation, and release 2 adult lacewings per week / m². 2 ; When the risk level is medium risk, the corresponding prevention and control measures are: increasing the monitoring frequency of multi-source data, adjusting the trapping array to operate in a medium-effective mode, spraying plant endophyte preparations, and releasing natural enemies differentially according to the insect age; When the risk level is high risk, the corresponding prevention and control measures are: adjusting the trapping array to operate in a high-effective mode, increasing the density of trapping points, releasing natural enemies differentially according to the insect age, placing a bait strip of Entomophthora grylli, and spraying methyl jasmonate sustained-release capsules.

6. The ecological control method for mirid bugs in pastureland according to claim 5, characterized in that, The trapping array includes multiple detachable and combined trapping units; each trapping unit includes a reflector and a light trap; the standard configuration of the trapping device is that the reflector angle of the reflector is 30°, and the light intensity of the light trap is 300 lux; the medium-effective mode of the trapping device is that the reflector angle of the reflector is 45°, and the light intensity of the light trap is 400 lux; the high-effective mode of the trapping device is that the reflector angle of the reflector is 60°, and the light intensity of the light trap is 500 lux.

7. An ecological control system for pasture mirid bugs, comprising the ecological control method for pasture mirid bugs as described in any one of claims 1-6, characterized in that, The system includes: A multi-source data acquisition module for acquiring multi-source data; the multi-source data includes environmental data, pest information, and images of pest infestation hotspots; The time-series feature matrix generation module is used to construct feature vectors based on the multi-source data, and perform principal component analysis to reduce the dimensionality of the feature vectors to generate a time-series feature matrix. The insect population density prediction membrane is used to input the temporal feature matrix into a pre-trained temporal convolutional network and output the predicted insect population density. The risk score calculation module is used to calculate the risk score based on the predicted insect population density using a dynamic insect risk scoring model. A risk level classification module is used to classify risk levels based on the risk score. The prevention and control module is used to determine the corresponding prevention and control measures based on the risk level, and to carry out ecological prevention and control of pasture mirid bugs.

8. The pasture mirid bug ecological control system according to claim 7, characterized in that, The environmental data includes meteorological data and soil parameters; the insect information includes the surface reflectance spectral characteristics of the pasture mirid bug and the vibration signals of nymphs feeding. The multi-source data module includes: A meteorological sensor array, deployed in the pasture canopy, is used to collect meteorological data; Soil parameter sensors are buried in the underground root zone to collect soil parameters; A multispectral imager was used to identify the surface reflectance spectral characteristics of the grass mirid bug. Acoustic sensor used to capture vibration signals from nymphs feeding; Thermal imagers, deployed on patrol drones, are used to acquire images of hotspots where pests are spreading.

9. The pasture mirid bug ecological control system according to claim 7, characterized in that, The prevention and control module includes: The trapping array comprises multiple detachable and combinable trapping units; each trapping unit includes a reflector and a light attractor. The natural enemy coordinated release unit includes a compartmentalized natural enemy storage chamber and an electrically controlled release valve; the compartmentalized natural enemy storage chamber is equipped with a lacewing larva chamber and an assassin bug adult chamber; the electrically controlled release valve selectively opens the corresponding lacewing larva chamber or assassin bug adult chamber according to the insect age identification result, and releases natural enemies according to the insect age differentiation; A microecological regulation device includes a hydraulic sprayer equipped with a dual-chamber storage tank. The first chamber of the dual-chamber storage tank stores a plant endophytic bacteria preparation, and the second chamber of the dual-chamber storage tank stores a suspension containing methyl jasmonate sustained-release capsules.

10. The pasture mirid bug ecological control system according to claim 9, characterized in that, When the risk level is low, the corresponding prevention and control measures are as follows: maintain the monitoring frequency of multi-source data, keep the trap array in standard configuration operation, and release 2 adult lacewings per week / m². 2 ; When the risk level is medium risk, the corresponding prevention and control measures are: increase the monitoring frequency of multi-source data, adjust the trap array to operate in medium-efficiency mode, activate the micro-ecological regulation device to spray plant endophytic bacteria preparations, and control the release of natural enemies by the natural enemy synergistic release unit according to the age of the insects. When the risk level is high, the corresponding prevention and control measures are as follows: adjust the trap array to operate in high-efficiency mode, increase the density of trapping points, control the release of natural enemies by the natural enemy co-release unit according to the age of the insects, control the release of mirid bug pox virus bait by the patrol drone, and activate the microecological regulation device to spray methyl jasmonate sustained-release capsules.

Images