Agricultural pest ai early warning and unmanned aerial vehicle cluster precision control system and method

Abstract

The application discloses an agricultural pest AI early warning and unmanned aerial vehicle cluster precise prevention and treatment system and method, relates to the technical field of agricultural AI early warning, and collects crop pictures and weather data, identifies the types, distribution and diffusion of pests, predicts pest abnormal conditions and the development direction of the pests according to weather data of a future set time period, collects flight data of unmanned aerial vehicles in the development direction of the pests, predicts abnormal pesticide spraying conditions of the unmanned aerial vehicles according to the weather data and the flight data, analyzes pest prevention and treatment abnormal conditions according to the pest abnormal conditions and the abnormal pesticide spraying conditions, and schedules the number of unmanned aerial vehicles according to crop growth stages and the pest prevention and treatment abnormal conditions. The application combines pest abnormal prediction and abnormal unmanned aerial vehicle pesticide spraying in the development direction of the pests to determine the optimal number of unmanned aerial vehicles to be deployed according to the crop growth stages, is favorable for optimizing resource allocation, and realizes precise prevention and treatment of farmland.

Description

Technical Field

[0001] This invention relates to the field of agricultural AI early warning technology, specifically to an agricultural pest AI early warning and drone swarm precision control system and method. Background Technology

[0002] AI-powered early warning systems for agricultural pests collect multi-source data from the fields using IoT sensors, remote sensing technology, and drone patrols. These data are then fused and analyzed using artificial intelligence algorithms to enable advanced and quantitative predictions of pest occurrence periods, quantities, ranges, and spread trends. Under the unified scheduling of a task-level control system, multiple drones work as a collaborative whole, autonomously planning paths, coordinating operations, and collecting data based on the spatial prescription map provided by the AI ​​early warning system. This allows for precise application of pesticides to high-incidence areas.

[0003] Existing technologies mostly analyze the allocation of drones based on factors such as farmland area, crop type, and drone spraying efficiency. However, they lack consideration for the spraying efficiency of drones in the direction of pest spread. When drone spraying anomalies happen to fall on the pest spread path, it will cause the pests to spread rampantly, thereby reducing the efficiency of agricultural pest control.

[0004] To address the aforementioned issues, this invention proposes an AI-based early warning system and method for agricultural pest control using drone swarms. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention proposes an AI-based early warning and drone swarm precision control system and method for agricultural pests. This invention determines pest anomalies by combining pest anomaly prediction with abnormal pesticide spraying by drones in the direction of pest development, thereby determining the optimal number of drones to deploy based on the crop growth stage. This is beneficial for optimizing resource allocation and achieving precision control in farmland.

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

[0007] The method of AI-based early warning and precision control of agricultural pests using drone swarms includes the following specific steps:

[0008] The system continuously collects crop images and weather data within a set time period to identify the types, distribution, and spread of pests.

[0009] Predict abnormal pest situations and the development direction of pests based on weather data for a future time period.

[0010] Collect flight data of drones along the direction of pest development, and predict abnormal pesticide spraying situations of drones based on weather data and flight data for a future time period.

[0011] Analyze abnormal pest control situations based on abnormal pest and pesticide spraying conditions, and schedule the number of drones according to crop growth stages and abnormal pest control conditions.

[0012] Preferably, the step of continuously collecting crop images and weather data within a set time period to identify the types, distribution, and spread of pests includes the following specific steps:

[0013] Crop images and weather data are continuously collected within a set time period, including temperature, humidity, wind direction, and wind speed.

[0014] Train a farmland pest identification model by inputting crop images, generate a pest distribution heat map, and automatically identify the pest species, population density, distribution location, diffusion direction, and diffusion speed.

[0015] Preferably, the step of predicting pest anomalies and the development direction of pests based on weather data for a future set time period includes the following specific steps:

[0016] Obtain weather data for a future time period;

[0017] Construct a spatiotemporal diffusion model for pests, inputting the current pest distribution heatmap, diffusion rate, and weather data for a future set time period, and outputting a predicted pest distribution map for the future set time period;

[0018] Based on the pest prediction distribution map, abnormal pest data and the development direction of pests are extracted. The abnormal pest data includes the number of abnormal pest areas.

[0019] The pest anomaly value is obtained by dividing the number of abnormal pest areas by the total number of pest prediction areas.

[0020] Preferably, the process of collecting flight data of the drone in the direction of pest development and predicting abnormal pesticide spraying by the drone based on weather data and flight data for a future time period includes the following specific steps:

[0021] Collect flight data of the drone in the direction of pest development, including motor load, pumping pressure and spraying amount per unit time;

[0022] Obtain the standard motor load, standard pumping pressure, and standard spray volume per unit time of the drone at the time of manufacture. Obtain abnormal values ​​of motor load by dividing the motor load by the standard motor load, abnormal values ​​of pumping pressure by dividing the pumping pressure by the standard pumping pressure, and abnormal values ​​of spray volume by dividing the spray volume per unit time by the standard spray volume per unit time. Obtain abnormal values ​​of drone flight by weighted summation of abnormal values ​​of motor load, pumping pressure, and spray volume.

[0023] Collect wind speed and the maximum wind resistance speed of the drone for a future set time period, and obtain the wind speed anomaly value by dividing the wind speed for the future set time period by the maximum wind resistance speed of the drone.

[0024] The abnormal values ​​of pesticide spraying by the drone are obtained by multiplying the abnormal values ​​of drone flight and wind speed.

[0025] Preferably, the step of analyzing abnormal pest control situations based on abnormal pest conditions and abnormal pesticide spraying situations, and scheduling the number of drones according to crop growth stages and abnormal pest control situations, includes the following specific steps:

[0026] The abnormal values ​​of pest control are obtained by weighted summation of the abnormal values ​​of pests and pesticide spraying.

[0027] Preset abnormal thresholds for pest control based on crop growth stages;

[0028] The optimal number of pest control drones is obtained by dividing the abnormal value of pest control by the abnormal threshold of pest control and then multiplying it by the standard number of drones configured.

[0029] An AI-powered early warning and drone swarm precision control system for agricultural pests is used to implement AI-powered early warning and drone swarm precision control methods for agricultural pests, including:

[0030] The pest identification module is used to continuously collect crop images and weather data within a set time period to identify the types, distribution, and spread of pests.

[0031] The pest prediction module is used to predict abnormal pest situations and the development direction of pests based on weather data for a future time period.

[0032] The pesticide spraying anomaly analysis module is used to collect flight data of drones in the direction of pest development and predict pesticide spraying anomalies of drones based on weather data and flight data for a future time period.

[0033] The anomaly analysis module is used to analyze anomalies in pest control based on abnormal pest conditions and abnormal pesticide spraying conditions.

[0034] The drone scheduling module is used to schedule the number of drones based on crop growth stages and abnormal pest control situations.

[0035] An electronic device includes a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes the above-mentioned method for AI early warning and precise control of agricultural pests by calling the computer program stored in the memory.

[0036] A computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the aforementioned method for AI-based early warning and precision control of agricultural pests using drone swarms.

[0037] Compared with existing technologies, the beneficial effects of this invention are as follows: It continuously collects crop images and weather data within a set time period to identify the types, distribution, and spread of pests; predicts pest anomalies and their development direction based on weather data for a future set time period; collects flight data of drones along the pest development direction; predicts abnormal pesticide spraying by drones based on weather and flight data for a future set time period; analyzes abnormal pest control based on pest and pesticide spraying anomalies; and schedules the number of drones according to crop growth stages and pest control anomalies. This invention determines pest anomalies by combining pest anomaly prediction and abnormal pesticide spraying by drones along the pest development direction, thereby determining the optimal number of drones to deploy based on crop growth stages. This is beneficial for optimizing resource allocation and achieving precise pest control in farmland. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a schematic diagram of the process of the agricultural pest AI early warning and drone swarm precision control method of the present invention;

[0040] Figure 2 This is a schematic diagram of the pest anomaly prediction process of the present invention;

[0041] Figure 3 This is a schematic diagram of the abnormal prediction process for pesticide spraying by drones according to the present invention;

[0042] Figure 4 This is a schematic diagram of the structure of the agricultural pest AI early warning and drone swarm precision control system of the present invention;

[0043] Figure 5 This is a schematic diagram of the electronic device structure of the present invention. Detailed Implementation

[0044] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention are within the scope of protection of the present invention.

[0045] Example 1

[0046] like Figure 1 As shown, one embodiment of the present invention provides a method for AI-based early warning and precision control of agricultural pests using drone swarms, which includes the following specific steps:

[0047] The system continuously collects crop images and weather data within a set time period to identify the types, distribution, and spread of pests.

[0048] In this embodiment, continuously collecting crop images and weather data within a set time period to identify the types, distribution, and spread of pests includes the following specific steps:

[0049] Crop images and weather data are continuously collected within a set time period, including temperature, humidity, wind direction, and wind speed.

[0050] In this embodiment, a drone equipped with a high-definition camera is used to fly along a preset route and continuously take pictures of the farmland in the global and local areas within a set time period. Meteorological data collection and integration equipment is deployed in the farmland to collect indicators such as temperature, humidity, wind speed, wind direction, light intensity and leaf surface humidity in real time, and synchronize the image collection and meteorological data collection timestamps.

[0051] Train a farmland pest identification model by inputting crop images, generate a pest distribution heat map, and automatically identify the pest species, population density, distribution location, diffusion direction, and diffusion speed.

[0052] In this embodiment, to improve the model's recognition accuracy, the collected images undergo batch processing such as size standardization, image enhancement, and noise reduction. Agricultural experts are invited to use annotation tools to label the pests in the images. Each image corresponds to an XML file that records the location and type of the pests. The labeled images are divided into training, validation, and test sets in an 8:1:1 ratio. Transfer learning is performed using the weights of a farmland pest recognition model pre-trained on a large dataset. The training set is input into the farmland pest recognition model for training. The farmland pest recognition model learns the mapping relationship from image pixels to pest types and locations. The training process is monitored through the validation set to prevent overfitting, and the model performance is evaluated using the test set.

[0053] The trained farmland pest identification model was deployed to a cloud server to automatically identify pests in batches from newly collected images. The identification results of each image were bound to its GPS information. GIS software was used to assign different weights to each GPS point based on its pest population. A continuous and smooth pest distribution heat map was generated using a kernel density estimation algorithm. Blue areas represent low pest density, green-yellow areas represent medium pest density, and orange-red areas represent high pest density. Pest data was correlated with meteorological data through timestamps and location information. Random forest analysis was used to analyze the quantitative relationship between different meteorological factors and the degree of pest occurrence. For example, the pest distribution heat map showed that rice planthoppers were densely clustered in the northeast corner of a farmland, with a significantly higher pest population density than other areas. Combined with the fact that this area is close to a water ditch, the humidity is consistently above 85%, and the temperature is stable at 25-28℃, it is the most suitable environment for rice planthopper reproduction.

[0054] Obtain two consecutive pest distribution heatmaps, measure the farthest distance from the pest center point to the edge of the pest spread area in the two pest distribution heatmaps, and the spread rate is obtained by dividing the difference between the two farthest distances by the time interval.

[0055] Obtain all newly added pest areas, calculate their azimuth angles relative to the original center, and draw an azimuth histogram. The peak direction is the main diffusion direction.

[0056] Predict abnormal pest situations and the development direction of pests based on weather data for a future time period.

[0057] like Figure 2 As shown, in this embodiment, predicting pest anomalies and the development direction of pests based on weather data for a future set time period includes the following specific steps:

[0058] Obtain weather data for a future time period;

[0059] Construct a spatiotemporal diffusion model for pests, inputting the current pest distribution heatmap, diffusion rate, and weather data for a future set time period, and outputting a predicted pest distribution map for the future set time period;

[0060] In this embodiment, farmland is divided into a grid consisting of N x N square cells in GIS. Each cell represents a fixed-size area of ​​the field. The current-time pest distribution heat map is overlaid on the grid. An initial pest density value M0(i, j) is assigned to each cell (i, j). The initial pest density value can be the average density of the heat map within the cell. Weather data for a future time period is processed into a format aligned with the grid. For isometric data such as wind direction, spatial interpolation is used to distribute the data to each cell to obtain the wind direction of each cell. The above steps convert continuous input data into a discrete format that the model can process.

[0061] Using the logistic growth model to simulate the reproduction of a population within a cell, the logistic growth model can be represented as: ,in, Let be the insect population density value of cell (i, j). For diffusion rate, The maximum insect population density supported by the crop within a cell, i.e., the carrying capacity, is determined by long-term monitoring of insect population density in a control field without pesticide interference, recording the maximum value under stable conditions. By setting the interval between the future time period and the current time period, this model characterizes the nonlinear saturation characteristics of population growth.

[0062] The calculation of external migration volume simulates the migration of pests from neighboring cells. The formula for calculating external migration volume can be expressed as: ,in, Let be the insect population density value of the source cell (m, n). Let be the diffusion probability function, used to quantify the influence of wind direction and speed on pest migration. It indicates that the diffusion probability is positively correlated with wind speed, negatively correlated with wind direction and the angle between the source cell (m, n) and the direction from cell (i, j), and positively correlated with diffusion speed. This reflects that pests preferentially diffuse downwind and the greater the wind speed, the farther they diffuse. Let be the direction angle from source cell (m, n) to cell (i, j). For wind speed and direction, For wind speed, This is an empirical coefficient for wind direction. Ensure leeward diffusion. This is an empirical coefficient for wind speed. To enhance the effect of wind speed, drone aerial photography migration trajectories were collected to record the migration probability of pests under different wind speeds and wind direction angles. Nonlinear regression or machine learning was used to fit functions. With the wind direction angle fixed, the power-law relationship between migration probability and wind speed was analyzed to estimate empirical wind speed coefficients. With the wind speed fixed, the cosine relationship between migration probability and wind direction angle was analyzed to estimate empirical wind direction coefficients. This is a diffusion kernel based on wind direction and velocity, used to describe the orientation angle of pests from source cell (m, n) to cell (i, j) as a function of distance d and distance d. The attenuation mode, The attenuation coefficient is determined by monitoring the migration trajectories of pests, statistically analyzing the percentage of migrating individuals at different distances, fitting a negative exponential function to the frequency distribution of migration distances, and then obtaining the attenuation coefficient through maximum likelihood estimation. Let be the distance between cell (i, j) and source cell (m, n). It is an exponential function with the real number e as its base;

[0063] The insect population density of cell (i,j) in a future time period is the sum of internal growth and external migration, i.e. ;

[0064] The processed grid data M0(i, j) is loaded into the model as the initial state. A time step is set, and the model is iterated until the predetermined prediction duration is reached. After the model finishes running, the insect population density grid for the future set time period is output. Each predicted grid is converted back to a standard spatial data format. In the GIS platform, the predicted insect population density values ​​are re-rendered as a heatmap, with the color scheme matching the input heatmap. Figure 1 This allows for the visualization of several consecutive days' forecast heatmaps, providing an intuitive view of the predicted distribution of pests, high-risk areas, and the spread paths, speed, and direction of pests.

[0065] Based on the pest prediction distribution map, abnormal pest data and the development direction of pests are extracted. The abnormal pest data includes the number of abnormal pest areas.

[0066] In this embodiment, the insect population density values ​​of all grids within the insect pest prediction distribution map area are obtained, the deviation value of the insect population density value of each grid is calculated from the insect population density value of the corresponding grid in the current insect distribution heat map, and the grids that exceed the preset deviation threshold are marked as insect pest abnormal areas, and the number of insect pest abnormal areas is counted.

[0067] Calculate the centroid of the current pest distribution heatmap and the predicted pest distribution map, which is the weighted average position of the insect population density of all grids. Connect the centroids of the two time points to obtain the vector that represents the direction of pest development. Divide the length of this vector by the interval between the two time points to obtain the average diffusion rate.

[0068] The pest anomaly value is obtained by dividing the number of abnormal pest areas by the total number of pest prediction areas.

[0069] In this embodiment, by analyzing abnormal pest areas, the outbreak point of pests can be identified, and the area requiring emergency intervention can be quickly located.

[0070] Collect flight data of drones along the direction of pest development, and predict abnormal pesticide spraying situations of drones based on weather data and flight data for a future time period.

[0071] like Figure 3 As shown, in this embodiment, collecting flight data of the drone along the direction of pest development and predicting abnormal pesticide spraying by the drone based on weather data and flight data for a future set time period includes the following specific steps:

[0072] Collect flight data of the drone in the direction of pest development, including motor load, pumping pressure and spraying amount per unit time;

[0073] In this embodiment, the direction of pest development represents the future spread path. If the drone's spraying happens to fall on this path, it will cause the pest to spread rampantly.

[0074] Obtain the standard motor load, standard pumping pressure, and standard spray volume per unit time of the drone at the time of manufacture. Obtain abnormal values ​​of motor load by dividing the motor load by the standard motor load, abnormal values ​​of pumping pressure by dividing the pumping pressure by the standard pumping pressure, and abnormal values ​​of spray volume by dividing the spray volume per unit time by the standard spray volume per unit time. Obtain abnormal values ​​of drone flight by weighted summation of abnormal values ​​of motor load, pumping pressure, and spray volume.

[0075] In this embodiment, the motor load is obtained by reading real-time current / power data through the motor controller, reflecting the load status of the drone's power system. The pumping pressure is obtained by monitoring the pipeline pressure in real time through a pressure sensor, reflecting the stability of the drone's pesticide delivery. If the pressure fluctuation is too large, it may be due to nozzle blockage, pipeline leakage, changes in pesticide viscosity, etc. The amount of pesticide sprayed per unit time is obtained by measuring the pesticide flow rate through a flow meter, reflecting the uniformity of pesticide spraying by the drone. Abnormal spraying may be due to insufficient pressure, nozzle wear, changes in pesticide concentration, etc.

[0076] Collect wind speed and the maximum wind resistance speed of the drone for a future set time period, and obtain the wind speed anomaly value by dividing the wind speed for the future set time period by the maximum wind resistance speed of the drone.

[0077] The abnormal values ​​of pesticide spraying by the drone are obtained by multiplying the abnormal values ​​of drone flight and wind speed.

[0078] In practical implementation, drones often cause pesticide waste due to drifting, missed spraying, overspraying, and pesticide leakage. Timely warnings of problems such as motor overload and abnormal pump pressure can improve overall operation efficiency.

[0079] Analyze abnormal pest control situations based on abnormal pest and pesticide spraying situations, and schedule the number of drones according to crop growth stages and abnormal pest control situations.

[0080] In this embodiment, analyzing abnormal pest control situations based on abnormal pest conditions and pesticide spraying situations, and scheduling the number of drones according to crop growth stages and abnormal pest control situations includes the following specific steps:

[0081] The abnormal values ​​of pest control are obtained by weighted summation of the abnormal values ​​of pests and pesticide spraying.

[0082] Preset abnormal thresholds for pest control based on crop growth stages;

[0083] In this embodiment, during specific implementation, the seedlings of crops in the seedling stage are delicate with a small photosynthetic area, so it is necessary to prevent a small number of pests from damaging the growth point. The stems and leaves of crops in the vegetative growth stage are vigorous and have strong resistance, so it is necessary to control the insect population density to avoid competition for nutrients. The flowers / fruits of crops in the reproductive growth stage are sensitive and affect yield and quality, so it is necessary to avoid pesticide damage that causes flower and fruit drop. The grains / fruits of crops in the mature stage are developing and have strong resistance, so it is necessary to control pests but consider pesticide residues.

[0084] The optimal number of pest control drones is obtained by dividing the abnormal value of pest control by the abnormal threshold of pest control and then multiplying it by the standard number of drones configured.

[0085] In this embodiment, the standard number of drones is the initial preset number of drones. The weights and thresholds are set as follows: historical crop images, historical weather data, and historical drone flight data are collected from several farmlands. The above data is substituted into each step of this embodiment to obtain the abnormal value of pest control and the optimal number of drones. The historical actual number of drones and the optimal number of drones are input into the pre-trained fitting software for fitting, and the values ​​of the weights and thresholds with the highest judgment accuracy are output.

[0086] In this embodiment, insufficient drones may lead to operation delays, missing the best opportunity, greatly reducing the prevention and control effect, and even causing crop yield reduction. On the other hand, too many drones may result in idle and wasted resources. Calculating the optimal number of drones in advance based on prevention and control needs allows for reasonable resource allocation, enabling rapid response and precise strikes.

[0087] Example 2

[0088] like Figure 4 As shown, the agricultural pest AI early warning and drone swarm precision control system is implemented based on the above-mentioned agricultural pest AI early warning and drone swarm precision control method, including:

[0089] The pest identification module is used to continuously collect crop images and weather data within a set time period to identify the types, distribution, and spread of pests.

[0090] The pest prediction module is used to predict abnormal pest situations and the development direction of pests based on weather data for a future time period.

[0091] The pesticide spraying anomaly analysis module is used to collect flight data of drones in the direction of pest development and predict pesticide spraying anomalies of drones based on weather data and flight data for a future time period.

[0092] The anomaly analysis module is used to analyze anomalies in pest control based on abnormal pest conditions and abnormal pesticide spraying conditions.

[0093] The drone scheduling module is used to schedule the number of drones based on crop growth stages and abnormal pest control situations.

[0094] Example 3

[0095] like Figure 5 As shown, this embodiment provides an electronic device, including a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes the above-mentioned method for AI early warning and precise control of agricultural pests by calling the computer program stored in the memory.

[0096] The electronic device can vary considerably depending on its configuration or performance. It may include one or more Central Processing Units (CPUs) and one or more memories, wherein the memory stores at least one computer program, which is loaded and executed by the processor to implement the agricultural pest AI early warning and drone swarm precision control method provided in the above-described embodiment. The electronic device may also include other components for implementing its functions. For example, it may have wired or wireless network interfaces and input / output interfaces for data input and output, which will not be elaborated upon in this embodiment.

[0097] Example 4

[0098] This embodiment proposes a computer-readable storage medium storing instructions that, when a computer program is run on a computer device, cause the computer device to execute the aforementioned method for AI-based early warning and precise control of agricultural pests by drone swarms.

[0099] For example, computer-readable storage media can be read-only memory (ROM), random access memory (RAM), compact disc read-only memory (CD-ROM), magnetic tape, floppy disk, and optical data storage devices.

[0100] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0101] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, they generate in whole or in part the flow or function according to the embodiments of the present invention. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired network and / or wireless network. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more sets of available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. The semiconductor media can be solid-state drives.

[0102] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed in this invention can be implemented in electronic hardware or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this invention.

[0103] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of units is only one, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical or other forms.

[0104] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units. That is, they may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0105] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0106] The preferred embodiments of the present invention disclosed above are only for the purpose of illustrating the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to specific implementation methods. Obviously, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for precise control of agricultural pests using AI-based early warning and drone swarms, characterized in that... The specific steps include the following: The system continuously collects crop images and weather data within a set time period to identify the types, distribution, and spread of pests. Predict abnormal pest situations and the development direction of pests based on weather data for a future time period. The specific content is as follows: Obtain weather data for a future time period; Construct a spatiotemporal diffusion model for pests, inputting the current pest distribution heatmap, diffusion rate, and weather data for a future set time period, and outputting a predicted pest distribution map for the future set time period; The specific steps are as follows: Farmland is divided into a grid consisting of N x N square cells in GIS, with each cell representing a fixed-size area. The current-time pest distribution heatmap is overlaid on the grid, and an initial pest density value M0(i, j) is assigned to each cell (i, j). The initial pest density value is the average density of the heatmap within that cell. Weather data for a future time period is processed into a format aligned with the grid. For wind direction areal data, spatial interpolation is used to distribute the data to each cell to obtain the wind direction of each cell. The above steps convert continuous input data into a discrete format that the model can process. The logistic growth model is used to simulate the reproduction of the population within a cell. The logistic growth model is represented as follows: ,in, Let be the initial insect population density value for cell (i, j). For diffusion rate, The maximum insect population density supported by the crop within a cell, i.e., the carrying capacity, is determined by long-term monitoring of insect population density in a control field without pesticide interference, recording the maximum value under stable conditions. By setting the interval between future and current time periods, this model characterizes the nonlinear saturation characteristics of population growth. The internal growth insect population density of cell (i,j) for a future time period; The external migration rate is calculated to simulate the migration of pests from neighboring cells. The formula for calculating the external migration rate is as follows: ,in, Let be the insect population density value of the source cell (m, n). Let be the diffusion probability function, used to quantify the influence of wind direction and speed on pest migration. It indicates that the diffusion probability is positively correlated with wind speed, negatively correlated with wind direction and the angle between the source cell (m, n) and the direction from cell (i, j), and positively correlated with diffusion speed. This reflects that pests preferentially diffuse downwind and the greater the wind speed, the farther they diffuse. Let be the direction angle from source cell (m, n) to cell (i, j). For wind speed and direction, For wind speed, This is an empirical coefficient for wind direction. Ensure leeward diffusion. This is an empirical coefficient for wind speed. To enhance the effect of wind speed, drone aerial photography migration trajectories were collected to record the migration probability of pests under different wind speeds and wind direction angles. Nonlinear regression or machine learning was used to fit functions. With the wind direction angle fixed, the power-law relationship between migration probability and wind speed was analyzed to estimate empirical wind speed coefficients. With the wind speed fixed, the cosine relationship between migration probability and wind direction angle was analyzed to estimate empirical wind direction coefficients. This is a diffusion kernel based on wind direction and velocity, used to describe the orientation angle of pests from source cell (m, n) to cell (i, j) as a function of distance d and distance d. The attenuation mode, The attenuation coefficient is determined by monitoring the migration trajectories of pests, statistically analyzing the percentage of migrating individuals at different distances, fitting a negative exponential function to the frequency distribution of migration distances, and then obtaining the attenuation coefficient through maximum likelihood estimation. Let be the distance between cell (i,j) and source cell (m,n). Let be an exponential function with base e. The external migration insect population density of cell (i,j) for a future time period; The insect population density of cell (i,j) in a future time period is the sum of internal growth and external migration, i.e. ; The processed grid data M0(i, j) is loaded into the model as the initial state. The time step is set, and the model is iterated until the predetermined prediction duration is reached. After the model finishes running, the insect population density grid for the future set time period is output. Each prediction result grid is converted back to the standard spatial data format. In the GIS platform, the predicted insect population density value is re-rendered as a heat map. The color scheme should be consistent with the input heat map. The predicted heat map for several consecutive days is visualized to intuitively show the predicted distribution of the pest, high-risk areas, and the spread path, speed, and direction of the pest. Based on the pest prediction distribution map, abnormal pest data and the development direction of pests are extracted. The abnormal pest data includes the number of abnormal pest areas. The pest anomaly value is obtained by dividing the number of abnormal pest areas by the total number of pest prediction areas. Collect flight data of drones along the direction of pest development, and predict abnormal pesticide spraying situations of drones based on weather data and flight data for a future time period. The specific steps include the following: Collect flight data of the drone in the direction of pest development, including motor load, pumping pressure and spraying amount per unit time; Obtain the standard motor load, standard pumping pressure, and standard spray volume per unit time of the drone at the time of manufacture. Obtain abnormal values ​​of motor load by dividing the motor load by the standard motor load, abnormal values ​​of pumping pressure by dividing the pumping pressure by the standard pumping pressure, and abnormal values ​​of spray volume by dividing the spray volume per unit time by the standard spray volume per unit time. Obtain abnormal values ​​of drone flight by weighted summation of abnormal values ​​of motor load, pumping pressure, and spray volume. Collect wind speed and the maximum wind resistance speed of the drone for a future set time period, and obtain the wind speed anomaly value by dividing the wind speed for the future set time period by the maximum wind resistance speed of the drone. The abnormal values ​​of pesticide spraying by the drone are obtained by multiplying the abnormal values ​​of drone flight and wind speed. Analyze abnormal pest control situations based on abnormal pest and pesticide spraying conditions, and schedule the number of drones according to crop growth stages and abnormal pest control conditions.

2. The method for AI-based early warning and precision control of agricultural pests using drone swarms as described in claim 1, characterized in that, The process of continuously collecting crop images and weather data within a set time period to identify the types, distribution, and spread of pests includes the following specific steps: Crop images and weather data are continuously collected within a set time period, including temperature, humidity, wind direction, and wind speed. Train a farmland pest identification model by inputting crop images, generate a pest distribution heat map, and automatically identify the pest species, population density, distribution location, diffusion direction, and diffusion speed.

3. The method for AI-based early warning and precision control of agricultural pests using drone swarms as described in claim 2, characterized in that, The process of analyzing abnormal pest control situations based on abnormal pest and pesticide spraying conditions, and scheduling the number of drones according to crop growth stages and abnormal pest control conditions, includes the following specific steps: The abnormal values ​​of pest control are obtained by weighted summation of the abnormal values ​​of pests and pesticide spraying. Preset abnormal thresholds for pest control based on crop growth stages; The optimal number of pest control drones is obtained by dividing the abnormal value of pest control by the abnormal threshold of pest control and then multiplying it by the standard number of drones configured.

4. An agricultural pest AI early warning and drone swarm precision control system, used to implement the agricultural pest AI early warning and drone swarm precision control method as described in any one of claims 1-3, characterized in that, include: The pest identification module is used to continuously collect crop images and weather data within a set time period to identify the types, distribution, and spread of pests. The pest prediction module is used to predict abnormal pest situations and the development direction of pests based on weather data for a future time period. The pesticide spraying anomaly analysis module is used to collect flight data of drones in the direction of pest development and predict pesticide spraying anomalies of drones based on weather data and flight data for a future time period. The anomaly analysis module is used to analyze anomalies in pest control based on abnormal pest conditions and abnormal pesticide spraying conditions. The drone scheduling module is used to schedule the number of drones based on crop growth stages and abnormal pest control situations.

5. An electronic device, characterized in that, include: A memory and a processor, wherein the memory stores a computer program that can be called by the processor, and the processor executes the agricultural pest AI early warning and drone swarm precision control method according to any one of claims 1-3 by calling the computer program stored in the memory.

6. A computer-readable storage medium, characterized in that: The device stores instructions that, when executed on a computer, cause the computer to perform the agricultural pest AI early warning and drone swarm precision control method as described in any one of claims 1-3.

Images