A precision spraying system for forest pests and diseases based on drones

Abstract

The present invention provides a precision spraying system for forest pests and diseases based on drones, which relates to the field of pest and disease control technology. The system comprises: a real-time monitoring module for real-time monitoring of forest areas; a division and identification module for performing detection and classification, generating data classification results, dividing and identifying forest areas, and determining forest area grid parameters; a diffusion prediction module for predicting the spread of pests and diseases and obtaining pest and disease diffusion coefficients; a synchronous update module for constructing a predicted spread trend of pests and diseases; and a precision spraying module for formulating spraying tasks, executing and evaluating the effects, performing dynamic feedback optimization, generating spraying optimization tasks, and performing precision spraying of pests and diseases. The present invention solves the technical problem that the prior art often has difficulty in accurately distinguishing the subtle differences between pests and diseases and healthy plants, especially when the pests and diseases are more dispersed or hidden, resulting in missed spraying or uneven spraying, which in turn leads to insufficient spraying effects.

Description

Technical Field

[0001] The present invention relates to the technical field of pest control, and in particular to a precision spraying system for forest pests based on an unmanned aerial vehicle (UAV). Background Art

[0002] Using drones to monitor crops in real time, identify pests and diseases, and conduct precise spraying can greatly improve the efficiency and effectiveness of pest control. However, in terms of forest pest control, although drone technology has also been gradually introduced, due to the particularity of forest areas, such as relatively complex terrain, sparse vegetation, and different types of trees, existing technologies still have certain limitations.

[0003] On the one hand, in forest areas, due to the large variety and density of trees and vegetation, the monitoring and identification of pests and diseases are more complicated than those of agricultural crops. Existing monitoring technologies often find it difficult to accurately distinguish the subtle differences between pests and diseases and healthy plants, especially when the distribution of pests and diseases is relatively scattered or hidden, which may lead to spraying omissions or uneven spraying. On the other hand, traditional spraying mainly relies on regular and fixed-dose spraying of pesticides, which is difficult to adjust according to the spread dynamics of pests and diseases. In complex forest environments, the spread pattern of pests and diseases is not fixed and is affected by multiple factors (such as climate change, forest type, terrain, etc.), making the spread path difficult to accurately predict. Summary of the Invention

[0004] This application provides a drone-based precision spraying system for forest pests and diseases, aiming to solve the technical problem that existing technologies often have difficulty in accurately distinguishing the subtle differences between pests and diseases and healthy plants, especially when the pests and diseases are distributed more dispersedly or hidden, resulting in missed spraying or uneven spraying, which in turn leads to insufficient spraying effect.

[0005] The present application discloses a precision spraying system for forest pests and diseases based on a drone, the system comprising: a real-time monitoring module for performing real-time monitoring of a forest area through an integrated sensor device group of the drone, and obtaining a forest real-time monitoring sensor data set; a division and identification module for performing detection and classification based on the forest real-time monitoring sensor data set, generating a data classification result, dividing and identifying the forest area according to the data classification result, and determining the forest area grid parameters; a diffusion prediction module for traversing the forest area grid parameters to perform pest and disease diffusion prediction and obtain a pest and disease diffusion coefficient; a synchronous update module for synchronously updating the forest area grid parameters according to the pest and disease diffusion coefficient and constructing a predicted pest and disease spread trend; a precision spraying module for formulating a spraying task based on the predicted pest and disease spread trend in combination with the forest area, performing an effect evaluation on the spraying task by executing the spraying task through a drone, performing dynamic feedback optimization on the spraying task according to the evaluation result, generating a spraying optimization task, and performing precision spraying of pests and diseases on the forest area according to the spraying optimization task.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0007] By using the integrated sensor equipment group carried by the drone to monitor the forest area in real time, the forest monitoring data set can be obtained quickly and accurately, providing timely and accurate data support, and providing a basis for subsequent spraying decisions; using the real-time monitoring sensor data for detection and classification, and dividing the forest area based on the classification results, the forest area is divided into multiple grids, and the pest and disease density, health status and other characteristics of each grid are identified. This fine area division enables the spraying operation to more accurately locate high-risk areas and improve the efficiency and accuracy of spraying; by traversing the forest area grid parameters, the spread of pests and diseases is predicted, and the pest and disease diffusion coefficient is calculated, which provides spatial and temporal warnings for subsequent spraying tasks, so that the spraying tasks can be covered first Cover areas where the disease may spread faster or at higher risks; based on the synchronous update of the pest and disease diffusion coefficient, a predicted pest and disease spread trend is constructed. This trend can reflect the spread of pests and diseases in different areas in real time, so that the spraying task can flexibly respond to changes in pests and diseases in forest areas and improve prevention and control efficiency; through a comprehensive analysis of the predicted spread trend of pests and diseases and forest areas, precise spraying tasks are formulated, and spraying tasks are carried out by drones. The spraying effect is dynamically evaluated, and the spraying tasks are optimized and adjusted based on the evaluation results to ensure that the spraying dose matches the actual pest and disease density, thereby improving the spraying efficiency and pest and disease control effect. This dynamic feedback optimization ensures that the spraying task can be flexibly adjusted according to the real-time monitoring and evaluation results, so that the spraying operation is continuously improved to adapt to changes in pests and diseases.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A schematic structural diagram of a drone-based precision spraying system for forest pests and diseases provided in an embodiment of the present application.

[0010] Figure 2 This is a schematic diagram of the structure of a division and identification module in a drone-based precision spraying system for forest pests and diseases provided in an embodiment of the present application.

[0011] Explanation of the accompanying drawings: real-time monitoring module 10, division and recognition module 20, diffusion prediction module 30, synchronous update module 40, precision spraying module 50, semantic segmentation unit 21, ratio calculation unit 22, density classification unit 23, feature extraction unit 24, feature addition unit 25. DETAILED DESCRIPTION

[0012] The embodiments of the present application provide a drone-based precision spraying system for forest pests and diseases, which solves the technical problem that the existing technology often has difficulty in accurately distinguishing the subtle differences between pests and diseases and healthy plants, especially when the pests and diseases are distributed more dispersedly or hidden, resulting in missed spraying or uneven spraying, which in turn leads to insufficient spraying effect.

[0013] After introducing the basic principles of this application, various non-limiting embodiments of this application will be specifically described below in conjunction with the accompanying drawings. It should be understood that the specific embodiments described here are only used to explain this application and are not used to limit this application.

[0014] like Figure 1 As shown, the embodiment of the present application provides a precision spraying system for forest pests and diseases based on a drone, the system comprising:

[0015] The real-time monitoring module 10 is used to perform real-time monitoring of the forest area through the integrated sensor equipment group of the UAV to obtain a forest real-time monitoring sensor data set.

[0016] The drone's integrated sensing suite is equipped with a variety of sensors, including but not limited to visual sensors, infrared sensors, lidar sensors, meteorological sensors, and positioning systems. These sensors can cover different types of monitoring needs, such as pest and disease detection, forest health monitoring, and environmental change tracking. The drone regularly flies and monitors forest areas in real time. The data collected by all sensors is centrally processed to form a multi-dimensional forest real-time monitoring sensor dataset, encompassing outputs from different sensor types. This dataset is used for subsequent analysis and decision-making.

[0017] The division and identification module 20 is used to perform detection and classification based on the forest real-time monitoring sensor data set, generate data classification results, divide and identify the forest area according to the data classification results, and determine the forest area grid parameters.

[0018] Real-time forest monitoring sensor data sets are analyzed, and machine learning or deep learning algorithms are used to detect and classify the data. This process can employ different technical approaches depending on the specific task. Specifically, for image data, deep learning algorithms are used for semantic segmentation to identify different objects in the image and label areas with pests and diseases, healthy vegetation, and disease-free areas. Features are extracted from sensor data, such as those of pests and diseases and color changes in trees, and classified using data mining techniques to generate different forest area states. These processes ultimately produce a data classification result that divides forest areas into different categories, such as healthy vegetation, pests and diseases, and areas requiring increased monitoring.

[0019] Based on the classification results, spatial data, such as GPS location information, is used to spatially partition the forest area, allowing the status of each area to be visualized on a map. This spatial partitioning divides the forest area into multiple small grid cells, each representing a specific area, such as a small forest patch. By analyzing the data classification results, the status of each grid cell can be identified and appropriate treatment measures assigned. For example, if the density of pests and diseases in a particular grid cell is too high, drone spraying operations can be prioritized in subsequent steps.

[0020] For each grid cell, its corresponding grid parameters are calculated and determined. The grid parameters include the area of ​​the grid, the surrounding environment, the density of pests and diseases, the forest vegetation type, etc. These grid parameters are the basic data for subsequent pest and disease prediction, spraying tasks, monitoring and evaluation, etc.

[0021] The diffusion prediction module 30 is used to traverse the grid parameters of the forest area to predict the spread of pests and diseases and obtain the pest and disease diffusion coefficient.

[0022] By analyzing the characteristic data of each grid one by one and predicting the spread of pests and diseases, we can predict how pests and diseases will spread from known infected areas to the surrounding areas. For example, if some grid cells are already infected by pests and diseases, by analyzing the characteristics of the cells, we can predict how the pests and diseases will spread to adjacent cells. The prediction process takes into account multiple factors that affect the spread, such as meteorological conditions, geographical characteristics, vegetation type, type of pests and diseases, etc. These factors determine the speed and range of the spread of pests and diseases in different areas.

[0023] Through diffusion prediction, the pest and disease diffusion coefficient of each grid unit is calculated. This coefficient reflects the degree of pest and disease diffusion around the grid unit. The larger the value, the stronger the potential for pest and disease spread in the area.

[0024] The synchronous updating module 40 is used to synchronously update the grid parameters of the forest area according to the pest and disease diffusion coefficient, and to construct a predicted pest and disease spreading trend.

[0025] The obtained pest and disease diffusion coefficient is used to synchronously update each grid cell. This means that the original forest area grid parameters will be adjusted according to the diffusion potential of pests and diseases. For example, the diffusion coefficient of some areas is larger, indicating that the possibility of pests and diseases spreading is higher. Therefore, the grid parameters of these areas will be updated to a higher risk level. The updated grid parameters will include the diffusion coefficient to support subsequent spraying tasks, monitoring of key areas and other decisions.

[0026] Based on the updated results of the pest and disease diffusion coefficient, a predicted pest and disease spread trend is constructed. This trend can be displayed in the form of a spatial distribution map, showing the spread trajectory of pests and diseases in the future time period. Furthermore, the predicted trend can be displayed through color coding (such as red for high-risk areas and green for low-risk areas) or numerical values, which helps to determine the dynamic situation of pest and disease spread and provide a basis for subsequent precision spraying, monitoring and other control measures.

[0027] The precision spraying module 50 is used to formulate spraying tasks based on the predicted spread trend of the pests and diseases in combination with the forest area, evaluate the effect of the spraying tasks by executing the spraying tasks through drones, perform dynamic feedback optimization on the spraying tasks based on the evaluation results, generate spraying optimization tasks, and perform precision spraying of pests and diseases in the forest area according to the spraying optimization tasks.

[0028] In forest areas, based on the predicted spread trend of pests and diseases, we identify which areas have the most serious pests and diseases and which areas have a higher risk of transmission. Based on this information, we preliminarily formulate spraying tasks, including planning the spraying area and spraying amount, as well as planning the flight path, flight altitude, spraying altitude, etc. in combination with the forest area, to ensure that the pests and diseases in each area are reasonably controlled.

[0029] The drones fly according to the established spraying missions and spray pests. After the spraying is completed, the effect is evaluated. Specifically, the pest density is evaluated. Through subsequent monitoring data analysis, the changes in pest density after spraying are evaluated to determine whether the spraying has achieved the expected effect, for example, whether the pest density has significantly decreased after spraying; the coverage effect is evaluated, the spraying coverage area is monitored, and it is evaluated whether the spraying has fully covered the target area. If some areas are not effectively covered, subsequent spraying may be required.

[0030] Based on the spraying effectiveness evaluation results, feedback optimization is performed on the spraying mission, including adjusting the spray dosage, such as increasing or decreasing the spray volume; adjusting the spraying area and path to focus on covering missed areas; and adjusting the flight altitude and spraying height to ensure the uniformity and effectiveness of the pesticide distribution. Based on the optimized spraying mission, the drone once again performs precision spraying. Precision spraying ensures that each area receives sufficient pesticide by optimizing factors such as the spray dosage, spraying path, and flight altitude, while avoiding resource waste and negative environmental impacts. This not only improves spraying efficiency, but also ensures the effectiveness and environmental friendliness of pest control.

[0031] Furthermore, if Figure 2 As shown, the partition identification module 20 includes:

[0032] The semantic segmentation unit 21 is used to obtain a forest pest and disease monitoring dataset and a forest healthy plant monitoring dataset by performing semantic segmentation on the forest real-time monitoring sensor dataset; the ratio calculation unit 22 is used to perform ratio calculation based on the forest pest and disease monitoring dataset and the forest healthy plant monitoring dataset to obtain a pest and disease density value; the density classification unit 23 is used to perform density classification on the forest area according to the pest and disease density value and determine multiple density levels; the feature extraction unit 24 is used to perform feature extraction on the forest pest and disease monitoring dataset based on the multiple density levels and determine multiple pest and disease area features; the feature adding unit 25 is used to add the multiple pest and disease area features to the data classification result.

[0033] Semantic segmentation is a computer vision technology that aims to classify each pixel in an image. Through semantic segmentation, different areas in the image can be identified as different categories, for example, separating the pest and disease areas from the healthy plant areas in the image. In this step, the semantic segmentation algorithm is used to process the forest real-time monitoring sensor dataset, such as using a convolutional neural network to perform pixel-level classification of the image. By training the model, it is possible to identify the pest and disease areas and the healthy plant areas in the image. Through semantic segmentation, the forest real-time monitoring sensor dataset is divided into a forest pest and disease monitoring dataset and a forest healthy plant monitoring dataset. Among them, the forest pest and disease monitoring dataset contains areas infected by pests and diseases, which may appear as abnormalities in color, shape or other features in the image or sensor data; the forest healthy plant monitoring dataset contains healthy vegetation areas, which are free of pests and diseases or damage.

[0034] Ratio calculations based on pixel counts, areas, etc. are performed on the forest pest and disease monitoring dataset and the forest health plant monitoring dataset. For example, the pest and disease density value is equal to the area or number of pixels of the pest and disease area divided by the total area or total number of pixels of the forest area. Through this ratio calculation, the density value of pests and diseases in the forest area is obtained. This density value measures the distribution and severity of pests and diseases in the forest area. The higher the density value, the more serious the infection level of pests and diseases in the area, and stronger prevention and control measures may be needed.

[0035] Based on the density values ​​of pests and diseases, forest areas are divided into different density levels according to preset thresholds. For example, areas with density values ​​below a certain threshold can be classified as low-density levels. In low-density areas, pests and diseases are relatively light and the infected area is small; areas with density values ​​within a certain range can be classified as medium-density levels. In medium-density areas, pests and diseases have a certain spread and may require routine monitoring or prevention and control; areas with higher density values ​​are classified as high-density levels. In high-density areas, pests and diseases are serious and immediate measures need to be taken, such as spraying pesticides or other control measures.

[0036] Further analysis of the forest pest monitoring dataset, based on regions at different density levels, extracted regional characteristics of pests and diseases for each density level. These characteristics include: spatial distribution characteristics of pests and diseases, including the morphology, boundaries, and distribution patterns of pest and disease areas, which help identify distribution trends of pest and disease areas; temporal characteristics of pests and diseases, such as the changing trends of pests and diseases at different time points, reflecting the speed and direction of pest spread; environmental characteristics, such as temperature, humidity, and soil type, which may be closely related to the occurrence and spread of pests and diseases; and pest and disease species characteristics. These extracted characteristics support a better understanding of the causes and transmission pathways of pests and diseases, providing a basis for pest and disease prevention and control.

[0037] The extracted multiple pest and disease regional characteristics are added to the data classification results to provide data support and decision-making basis for subsequent pest and disease control measures.

[0038] Furthermore, the partition identification module 20 includes:

[0039] An identification unit is used to traverse the forest area to construct a forest geographic coordinate system, synchronize the characteristics of the multiple pest and disease areas to the forest geographic coordinate system for identification, and set multiple area boundaries; a fuzzy segmentation unit is used to fuzzy segment the forest area according to the multiple area boundaries to obtain multiple pest and disease areas; a screening unit is used to screen the forest area based on the multiple pest and disease areas to determine multiple healthy plant areas; a spatial aggregation unit is used to spatially aggregate the multiple pest and disease areas with the multiple healthy plant areas to construct the forest area grid parameters.

[0040] The forest area is spatially traversed to construct a forest geographic coordinate system. A geographic coordinate system refers to a reference system used to represent locations on the earth, usually defined by longitude and latitude. In this step, the constructed forest geographic coordinate system is used to map the forest area data into a unified spatial framework.

[0041] Synchronizing the extracted multiple pest and disease regional features into the forest geographic coordinate system ensures that each pest and disease regional feature has a clear geographic location, facilitating spatial analysis. These pest and disease regional features are then identified and mapped to the geographic information of the forest area. Based on the synchronized pest and disease regional features, multiple regional boundaries are set within the forest area. These boundaries are used to distinguish different pest and disease areas and serve as the basis for subsequent regional segmentation and management. Boundary settings can be based on factors such as pest and disease density, type, and diffusion trends to help identify areas that require special attention or spraying.

[0042] Fuzzy segmentation is a spatial data processing technique used to divide an area into multiple sub-areas without requiring strict boundaries. Unlike traditional precise segmentation, fuzzy segmentation allows for a certain degree of overlap or fuzzy areas between regions. Fuzzy segmentation uses set regional boundaries to fuzzily segment forest areas, i.e., fuzzy regional divisions are performed based on the characteristics of pest and disease areas. Each pest and disease area may partially overlap with adjacent areas, or some areas may have lower pest and disease severity. After fuzzy segmentation, the forest area is divided into multiple pest and disease areas, corresponding to different pest and disease densities or different types of pests and diseases. The divided pest and disease areas provide detailed spatial information for subsequent management and operations.

[0043] Based on the identified pest and disease areas, screening operations are carried out, which means removing these pest and disease areas from the entire forest area. The goal is to find healthy plant areas that are not affected by pests and diseases. Healthy plant areas refer to those areas that have no pests and diseases or have very mild pests and diseases. The vegetation in these areas is relatively healthy and does not require special intervention, or requires lighter monitoring and management.

[0044] Spatial aggregation is the process of combining information from different areas based on their spatial locations. Multiple pest and disease areas are spatially aggregated with multiple healthy plant areas to form a comprehensive forest area grid. Corresponding parameters are assigned to each grid cell, including the area of ​​the grid, the surrounding environment, the density of pests and diseases, the forest vegetation type, etc. These grid parameters serve as the basic data for subsequent pest and disease prediction, spraying tasks, monitoring and evaluation, and other processes.

[0045] Furthermore, the diffusion prediction module 30 includes:

[0046] A geographically weighted regression unit is used to perform geographically weighted regression on the multiple pest and disease areas to generate pest and disease distribution data; an impact analysis unit is used to perform impact analysis on the multiple healthy plant areas based on the pest and disease distribution data to obtain multiple impact factors; a diffusion analysis unit is used to perform diffusion analysis based on the multiple impact factors in combination with the pest and disease distribution data to determine multiple diffusion patterns; a diffusion path determination unit is used to traverse the multiple pest and disease areas based on the multiple diffusion patterns to determine multiple diffusion paths; a diffusion calculation unit is used to match the multiple diffusion paths with the multiple diffusion patterns, perform diffusion calculation on the forest area grid parameters, and obtain the pest and disease diffusion coefficient.

[0047] Geographically weighted regression is a spatial statistical method used to analyze spatial heterogeneity in geographic data. It analyzes spatial location differences and studies the relationship between independent variables and dependent variables, and can calculate different regression parameters for each geographic location. In this step, geographically weighted regression is performed on multiple pest and disease areas, that is, the distribution of pests and diseases is modeled and analyzed through changes in spatial location. By considering the influencing factors of different locations, the distribution patterns of pests and diseases, as well as the spatial relationship between pests and diseases and environmental factors, are identified. Through geographically weighted regression, pest and disease distribution data are generated to reflect the spatial distribution of pests and diseases in forest areas. These data include the intensity, diffusion path and potential impact area of ​​pests and diseases in different regions.

[0048] An impact analysis of healthy plant areas is conducted based on pest and disease distribution data. This involves assessing whether the spread of pests and diseases within forest areas will affect healthy plants, and the extent to which these healthy plants are affected by the spread of pests and diseases. During the analysis, multiple influencing factors are identified. Influencing factors refer to factors that affect the spread of pests and diseases or plant health, including but not limited to climatic factors, geographical factors, vegetation factors, and the types and transmission routes of pests and diseases. These influencing factors help quantify the risk of healthy plant areas being affected by pests and diseases.

[0049] Combining influencing factors and pest distribution data, a diffusion analysis of pests and diseases is conducted to determine how pests and diseases spread within forest areas, especially the extent of their impact in healthy plant areas. The goal of diffusion analysis is to identify the paths, speeds, and modes of pest spread in different areas. Through diffusion analysis, multiple diffusion patterns are determined, which reflect the different paths and characteristics of pest spread. For example, in the radial diffusion pattern, pests and diseases spread outward from a central area, similar to the circular diffusion pattern; in the linear diffusion pattern, pests and diseases spread along a certain path or in a certain direction, such as along forest trails or along the banks of rivers; in the group diffusion pattern, pests and diseases spread in local clusters, which are usually closely related to vegetation types, terrain, etc.

[0050] Based on different diffusion patterns, we traverse multiple pest and disease areas and determine specific diffusion paths for each area. Each pest and disease area will have one or more possible diffusion paths, which help describe how the pests and diseases spread.

[0051] The resulting diffusion paths are matched with diffusion patterns. This matching process is based on the pest's propagation characteristics and the type of diffusion pattern. Different types of diffusion patterns may have different impacts on the corresponding diffusion paths. The goal of the matching is to ensure that each diffusion path corresponds to the most appropriate diffusion pattern. Based on the matched diffusion paths and diffusion patterns, a diffusion calculation is performed on the grid parameters of the forest area. That is, by calculating the impact range, diffusion speed, and degree of pest diffusion, a pest diffusion coefficient is assigned to each grid cell. The pest diffusion coefficient represents the intensity of pest diffusion in a particular area. A larger coefficient indicates a faster pest spread and a wider impact range; a smaller coefficient indicates a slower pest spread and a smaller impact range.

[0052] Furthermore, the precision spraying module 50 includes:

[0053] A flight simulation unit is used to perform flight simulation on a UAV based on a forest area and determine multiple flight simulation parameters, wherein the multiple flight simulation parameters include a flight simulation altitude and a flight simulation speed; a matching unit is used to match the multiple pest and disease areas according to the multiple density levels and generate pest and disease matching results; a spraying calculation unit is used to perform spraying calculation according to the flight simulation altitude and the pest and disease matching results and determine spraying concentration information; a descending processing unit is used to perform descending processing on the pest and disease matching results and generate a sequence of areas to be sprayed; an integration unit is used to perform integration according to the sequence of areas to be sprayed combined with the spraying concentration information by executing the flight simulation speed by a UAV and determine spraying dosage data; a first data adding unit is used to add the spraying dosage data to the spraying task.

[0054] The flight of UAV is simulated in the forest area so that the flight simulation parameters, including flight simulation altitude and flight simulation speed, can be determined according to the specific conditions of the forest area and the mission requirements. The flight simulation altitude refers to the flight altitude of the UAV when performing the spraying mission. The choice of flight altitude will be optimized according to the specific characteristics of the forest (such as the height of trees, the density of vegetation, etc.). A flight altitude that is too low may result in the failure of the pesticide to cover a large enough area, and a flight altitude that is too high may result in uneven distribution of the pesticide. The flight simulation speed refers to the flight speed of the UAV. The flight speed determines the time it takes for the UAV to complete the spraying mission and the accuracy of the pesticide distribution. Pest and disease areas with different density levels may require different flight speeds to ensure the efficiency and effectiveness of the spraying mission.

[0055] According to different density levels, the pest and disease areas in the forest area are matched with the corresponding density levels. For high-density areas, higher spraying doses need to be matched because the spread of pests and diseases in these areas is more serious. For medium-density and low-density areas, the spraying dose can be reduced accordingly. The generated pest and disease matching results include the density level corresponding to each pest and disease area, thereby determining the spraying needs for different areas. Through the pest and disease matching results, specific area division and spraying dose information can be provided for the spraying task.

[0056] Based on the flight simulation altitude and the results of the pest and disease matching, spraying calculations are performed to determine the appropriate spraying concentration. The spraying concentration directly affects the coverage effect of the pesticide and the prevention and control effect of pests and diseases. Specifically, areas with pests and diseases of different densities require different concentrations of spraying doses. For example, high-density areas require higher spraying concentrations to ensure effective control of pests and diseases in these areas; medium and low-density areas require lower spraying concentrations, which can avoid excessive use of pesticides while achieving the effect of controlling pests and diseases; the flight simulation altitude determines the distribution range and uniformity of the pesticide. A lower flight altitude may cause the pesticide to be concentrated in a smaller area, while a higher flight altitude may cause the pesticide to be distributed too widely and the concentration to be dilute. By combining the flight simulation altitude and the results of the pest and disease matching, the spraying concentration is determined to ensure the efficiency and accuracy of the spraying task in areas with different densities.

[0057] Processing the pest and disease matching results in descending order means placing areas with higher pest and disease density in front. The pest and disease matching results after descending order are converted into a sequence of areas to be sprayed. Each area in the sequence will be processed in turn according to its pest and disease density and spraying requirements. This sequence provides a clear execution order for subsequent spraying tasks.

[0058] The spraying task is performed using the flight simulation speed of the drone and the sequence of areas to be sprayed. The spraying efficiency of each area is calculated based on the priority of the areas to be sprayed and the flight simulation speed. Areas with higher flight speeds can complete spraying faster, while areas with lower flight speeds may require more time to complete the spraying task. The actual spraying dose of each area is calculated based on the density level of different areas and the flight simulation speed. The determination of the spraying dose not only needs to consider the concentration of the pesticide, but also needs to be comprehensively calculated based on the size of the area, the density of pests and diseases, and the flight speed. Combining the above factors, the spraying dose data for each area is generated, that is, the specific amount of pesticide required for each area to be sprayed.

[0059] The calculated spraying dosage data is added to the drone's spraying mission, and the drone performs operations according to the spraying dosage data provided in the mission to ensure that pests and diseases are effectively controlled.

[0060] Furthermore, the precision spraying module 50 includes:

[0061] An identification unit is used to identify the pests and diseases according to the predicted spread trend in combination with the multiple diffusion paths, and generate multiple identification data to be sprayed; a flight analysis unit is used to traverse the forest geographic coordinate system and perform flight analysis according to the multiple identification data to be sprayed in combination with the multiple flight simulation parameters to determine the flight attitude parameters of the UAV; a dynamic analysis unit is used to retrieve the historical environmental record data set of the forest area, perform dynamic analysis based on the historical environmental record data set, and construct a forest environment change curve map; a path planning unit is used to perform path planning based on the forest environment change curve map in combination with the UAV flight attitude parameters to generate spraying path data; a second data adding unit is used to add the spraying path data to the spraying task.

[0062] Combined with the predicted spread trends and diffusion paths of pests and diseases, areas to be sprayed are identified within the forest area. These areas are where pests and diseases spread faster, or are high-risk areas predicted to spread in the future. Spraying identification data is generated in these areas, indicating that these areas need to be sprayed. These identification data can provide a basis for subsequent spraying tasks.

[0063] Traversing the forest geographic coordinate system, the coordinate system provides the precise coordinates of each area, ensuring that the drone's flight path and target area can be accurately calibrated when planning flight missions. The identification data to be sprayed is combined with the forest geographic coordinate system to accurately calibrate the areas that need to be sprayed. Each area to be sprayed will have a clear geographical location and pest and disease density information, providing data support for subsequent flight analysis.

[0064] Flight simulation parameters include flight altitude and flight speed. Flight analysis is performed in combination with the identification data to be sprayed and the flight simulation parameters. The purpose is to determine the flight attitude parameters of the drone when performing the spraying task to ensure that the spraying operation can accurately cover the area to be sprayed. The flight attitude parameters include the flight angle, pitch angle, roll angle, etc. of the drone, which determine the heading and altitude of the drone. According to the location of the area to be sprayed, the flight simulation parameters and the needs of the target area, the flight attitude parameters are calculated to ensure that the drone can perform the spraying task with a suitable attitude, maintain a stable flight trajectory, and avoid uneven distribution of the pesticide.

[0065] Retrieve historical environmental record datasets for forest areas. These data contain environmental data such as climate, temperature, humidity, wind speed, precipitation, and sunshine in the forest area over the past period of time. Based on the historical environmental datasets, use dynamic analysis methods to evaluate environmental change trends over the past period of time. For example, certain weather conditions (such as excessive temperature or humidity) may promote the spread of pests and diseases. Based on the results of dynamic analysis, draw forest environmental change curves to show the environmental change trends of the forest area over a period of time. These curves usually show how factors such as temperature, humidity, and precipitation change over time.

[0066] The forest environment change curve provides basic data on changes in environmental conditions, and the UAV flight attitude parameters provide flight trajectory requirements. The spraying path is planned based on the environmental change trend and flight attitude parameters. Among them, the environmental change curve suggests that pests and diseases in certain areas may spread rapidly in the future, so these areas need to be sprayed first, and the flight attitude parameters ensure that the UAV can cover these areas with an appropriate flight attitude. The spraying path is generated based on the above analysis results. These paths guide the UAV on how to cover the pest and disease areas when performing spraying tasks and how to adjust the flight trajectory under different environmental conditions to ensure that the pesticide can be evenly sprayed to the target area. For example, high-density pest and disease areas require lower flight altitudes and slower flight speeds, while low-density areas can adopt higher flight altitudes and faster speeds.

[0067] Add the spraying path data to the drone's spraying mission, and the drone will start spraying according to this mission. Through a reasonably planned path, it can ensure that pests and diseases in each area are effectively controlled while avoiding waste of pesticides.

[0068] Furthermore, the precision spraying module 50 includes:

[0069] A density assessment unit is used to perform pest density assessment on the spraying dosage data by executing the spraying task through a drone, and generate a pest reduction ratio value; a coverage assessment unit is used to perform pest coverage assessment on the spraying path data by executing the spraying task through a drone, and generate a pest coverage ratio value; an effect assessment unit is used to perform effect assessment based on the pest reduction ratio value combined with the pest coverage ratio value, and generate a spraying effect score; a dynamic feedback optimization unit is used to add the spraying effect score to the evaluation result to perform dynamic feedback optimization on the spraying task, and generate the spraying optimization task.

[0070] The drone's sensors are used to obtain pest and disease distribution data before and after spraying. The data before spraying reflects the original distribution of pests and diseases, while the data after spraying reflects the changes in pests and diseases. By comparing the distribution of pests and diseases before and after spraying, the rate of change of pests and diseases is calculated. Based on the above calculation results, a pest and disease reduction ratio value is generated. This ratio value represents the effectiveness of the spraying task. The higher the value, the better the spraying effect and the greater the proportion of pest and disease reduction.

[0071] Drone-mounted sensors monitor the sprayed area and assess pest and disease coverage. This evaluates the effectiveness of the spraying. Evaluation metrics include coverage, uniformity, and utilization. Specifically, the sprayer checks whether the intended pest and disease area is covered. The coverage should align with the intended sprayed area to ensure no areas are missed. The sprayer also assesses the uniformity of the sprayer during the spraying process, ensuring that each pest and disease area receives the spray evenly. Uneven spray concentrations or coverage can lead to poor pest and disease control in some areas. Furthermore, the sprayer assesses whether the sprayer is effectively used for pest control, rather than wasted or drifting to unrelated areas. A high utilization rate indicates that the sprayed dose is being used effectively, reducing waste and environmental burden. Based on these assessments, a pest and disease coverage ratio is calculated, representing the effectiveness of the spraying. Higher values ​​indicate wider coverage, better uniformity, and higher utilization.

[0072] Combining the Pest Reduction and Coverage values ​​provides a comprehensive reflection of spraying effectiveness. A higher Pest Reduction value indicates effective spraying in reducing pests, while a higher Coverage value indicates wide spray coverage, even distribution of the pesticide, and effective control. Combining these two values ​​creates a Spraying Effectiveness Score, which can be a weighted average or composite score, reflecting the overall effectiveness of the spraying task.

[0073] Dynamic feedback optimization of spraying tasks is carried out based on the spraying effect score. For example, the spraying amount or concentration of the pesticide is adjusted according to the effect score, especially in areas with high density of certain pests and diseases. Through this feedback mechanism, the execution effect of the spraying task can be continuously improved, and the spraying efficiency and accuracy can be enhanced.

[0074] Furthermore, the dynamic feedback optimization unit includes:

[0075] A scoring judgment channel is used to set an expected spraying threshold and judge whether the spraying effect score meets the expected spraying threshold; a tracking and monitoring channel is used to continuously track and monitor the forest area based on the spraying task if the spraying effect score meets the expected spraying threshold, generate dynamic tracking and monitoring data, and continuously feed back the dynamic tracking and monitoring data to the spraying task; an abnormality analysis channel is used to generate a feedback instruction if the spraying effect score does not meet the expected spraying threshold, perform abnormality analysis on the spraying task through the feedback instruction, and determine the spraying dose abnormality data and / or the spraying path abnormality data; a dynamic adjustment channel is used to use the spraying dose abnormality data and / or the spraying path abnormality data as an index, feed it back to the spraying task for dynamic adjustment, and generate the spraying optimization task.

[0076] The expected spraying threshold is the target value for the spraying effectiveness score, set based on actual needs, mission objectives, and predetermined standards. This threshold reflects the minimum requirement for the spraying task. Only when this threshold is reached can the spraying task be considered successful and achieve the expected pest control effect. If the spraying effectiveness score reaches or exceeds the expected spraying threshold, it indicates that the spraying task is effective and the pest control effect has met the expectations. Conversely, if the spraying effectiveness score is lower than the expected spraying threshold, it means that the spraying task is ineffective, and there may be problems with the spraying dosage, coverage, uniformity, etc., which require adjustment and optimization.

[0077] If the spraying effectiveness score meets the expected spraying threshold, the spraying mission has achieved its intended results. Continued follow-up monitoring will then be conducted to ensure that pests and diseases do not recur or spread over a period of time. Dynamic tracking monitoring data is collected through sensors that continuously monitor the forest area after the spraying mission is completed, collecting real-time data on pest and disease distribution, environmental changes, and more. This dynamic tracking monitoring data is continuously fed back into the spraying mission to ensure that the spraying mission continues to meet the actual needs of the forest area and effectively control pests and diseases over the long term. This dynamic feedback not only improves the immediate effectiveness of the spraying mission but also optimizes resource allocation and prevention strategies for subsequent spraying.

[0078] If the spraying effectiveness score falls below the expected spraying threshold, the spraying task failed to meet expectations. This may be due to issues with the spray volume, coverage, or uniformity of the pesticide. Feedback instructions are automatically generated for unsatisfactory spraying results. This is used to identify anomalies in the spraying task and guide subsequent adjustments and optimizations. The feedback instructions will include a detailed description of the problem, such as insufficient spraying dosage or inaccurate spraying path, to help analyze the cause and propose solutions.

[0079] The actual data of the spraying task is reviewed and analyzed according to the feedback instructions to find out the problems. The abnormal analysis content includes abnormal spraying dosage data and abnormal spraying path data. Among them, abnormal spraying dosage data, for example, insufficient or excessive amount of medicine sprayed in certain areas, failing to meet the predetermined spraying volume standard; abnormal spraying path data, for example, the spraying path does not cover all target areas, or there is a deviation in the path planning, resulting in missed spraying or uneven spraying in certain areas.

[0080] Abnormal spray dosage data and / or abnormal spray path data are used as indexes and fed back to the spraying task. This means that this abnormal data will serve as a key basis for triggering adjustments to the spraying task. For example, if the spray dosage in certain areas is too low, the spray dosage in these areas will be increased; if the path deviates, the spray path will be replanned to ensure complete coverage of all areas requiring spraying. The purpose of dynamic adjustment is to ensure that the spraying task can be flexibly adjusted based on real-time monitoring and feedback to adapt to various environmental changes and spraying effects. Through dynamic adjustment, an optimized spraying task is generated to guide subsequent spraying operations, making the spraying task more accurate and efficient while ensuring maximum pest control effectiveness.

[0081] In summary, the drone-based precision spraying system for forest pests and diseases provided by the embodiments of the present application has the following technical effects:

[0082] By using the integrated sensor equipment group carried by the drone to monitor the forest area in real time, the forest monitoring data set can be obtained quickly and accurately, providing timely and accurate data support, and providing a basis for subsequent spraying decisions; using the real-time monitoring sensor data for detection and classification, and dividing the forest area based on the classification results, the forest area is divided into multiple grids, and the pest and disease density, health status and other characteristics of each grid are identified. This fine area division enables the spraying operation to more accurately locate high-risk areas and improve the efficiency and accuracy of spraying; by traversing the forest area grid parameters, the spread of pests and diseases is predicted, and the pest and disease diffusion coefficient is calculated, which provides spatial and temporal warnings for subsequent spraying tasks, so that the spraying tasks can be covered first Cover areas where the disease may spread faster or at higher risks; based on the synchronous update of the pest and disease diffusion coefficient, a predicted pest and disease spread trend is constructed. This trend can reflect the spread of pests and diseases in different areas in real time, so that the spraying task can flexibly respond to changes in pests and diseases in forest areas and improve prevention and control efficiency; through a comprehensive analysis of the predicted spread trend of pests and diseases and forest areas, precise spraying tasks are formulated, and spraying tasks are carried out by drones. The spraying effect is dynamically evaluated, and the spraying tasks are optimized and adjusted based on the evaluation results to ensure that the spraying dose matches the actual pest and disease density, thereby improving the spraying efficiency and pest and disease control effect. This dynamic feedback optimization ensures that the spraying task can be flexibly adjusted according to the real-time monitoring and evaluation results, so that the spraying operation is continuously improved to adapt to changes in pests and diseases.

[0083] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application is not limited to the embodiments shown herein, but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A precision spraying system for forest pests and diseases based on drones, characterized by: The system comprises: A real-time monitoring module is used to monitor the forest area in real time through the integrated sensor equipment group of the UAV and obtain the forest real-time monitoring sensor data set; a division and identification module, configured to perform detection and classification based on the forest real-time monitoring sensor data set, generate data classification results, divide and identify the forest area according to the data classification results, and determine forest area grid parameters; A diffusion prediction module is used to traverse the grid parameters of the forest area to predict the spread of pests and diseases and obtain a pest and disease diffusion coefficient; A synchronous updating module, configured to synchronously update the grid parameters of the forest area according to the pest and disease diffusion coefficient, and construct a predicted pest and disease spread trend; A precision spraying module is used to formulate spraying tasks based on the predicted spread trend of the pests and diseases and in combination with the forest area, evaluate the effectiveness of the spraying tasks by executing the tasks through drones, perform dynamic feedback optimization on the spraying tasks based on the evaluation results, generate spraying optimization tasks, and perform precision spraying of pests and diseases in the forest area according to the spraying optimization tasks; The partition identification module includes: a semantic segmentation unit, configured to perform semantic segmentation on the forest real-time monitoring sensor dataset, identify pest and disease areas and healthy plant areas in the image using a training model, and divide the forest real-time monitoring sensor dataset into a forest pest and disease monitoring dataset and a forest healthy plant monitoring dataset through semantic segmentation, thereby obtaining a forest pest and disease monitoring dataset and a forest healthy plant monitoring dataset; a ratio calculation unit, configured to calculate a ratio based on the number of pixels and the area of ​​the forest pest and disease monitoring dataset and the forest healthy plant monitoring dataset to obtain a pest and disease density value; A density classification unit, used to classify the forest area according to the pest and disease density value and determine multiple density levels; a feature extraction unit, configured to extract features from the forest pest and disease monitoring dataset based on the multiple density levels, and determine multiple pest and disease area features; A feature adding unit, configured to add the plurality of pest and disease area features to the data classification result; an identification unit, configured to traverse the forest area to construct a forest geographic coordinate system, synchronize the plurality of pest and disease area features to the forest geographic coordinate system for identification, and set a plurality of area boundaries; a fuzzy segmentation unit, configured to perform fuzzy segmentation on the forest area according to the plurality of area boundaries to obtain a plurality of pest and disease areas; a screening unit for screening the forest area based on the plurality of pest and disease areas to determine a plurality of healthy plant areas; The spatial aggregation unit is used to spatially aggregate the multiple pest and disease areas with the multiple healthy plant areas to construct the forest area grid parameters.

2. The UAV-based precision spraying system for forest pests and diseases according to claim 1, characterized in that: The diffusion prediction module includes: A geographically weighted regression unit, configured to perform geographically weighted regression on the plurality of pest and disease areas to generate pest and disease distribution data; an impact analysis unit, configured to perform an impact analysis on the plurality of healthy plant areas based on the pest and disease distribution data to obtain a plurality of impact factors; a diffusion analysis unit, configured to perform diffusion analysis based on the plurality of influencing factors in combination with the pest distribution data to determine a plurality of diffusion patterns; a diffusion path determining unit, configured to determine a plurality of diffusion paths by traversing the plurality of pest and disease areas based on the plurality of diffusion patterns; A diffusion calculation unit is used to match the multiple diffusion paths with the multiple diffusion patterns, perform diffusion calculation on the forest area grid parameters, and obtain the pest diffusion coefficient.

3. The UAV-based precision spraying system for forest pests and diseases as claimed in claim 2, characterized in that: The precision spraying module comprises: a flight simulation unit, configured to perform flight simulation on the UAV based on the forest area and determine a plurality of flight simulation parameters, wherein the plurality of flight simulation parameters include a flight simulation altitude and a flight simulation speed; a matching unit, configured to match the plurality of pest and disease areas according to the plurality of density levels and generate a pest and disease matching result; a spraying calculation unit, configured to perform spraying calculations according to a matching result between the simulated flight altitude and the pests and diseases, and determine spraying concentration information; A descending order processing unit, configured to perform descending order processing on the pest and disease matching results to generate a sequence of areas to be sprayed; An integration unit is configured to integrate the flight simulation speed executed by the drone according to the sequence of the areas to be sprayed and the spraying concentration information to determine spraying dosage data; A first data adding unit is used to add the spraying dosage data to the spraying task.

4. The UAV-based precision spraying system for forest pests and diseases as claimed in claim 3, characterized in that: The precision spraying module includes: An identification unit, configured to identify the pests and diseases according to the predicted spread trend in combination with the multiple diffusion paths, and generate multiple identification data to be sprayed; A flight analysis unit is configured to traverse the forest geographic coordinate system, perform flight analysis according to the plurality of identification data to be sprayed and the plurality of flight simulation parameters, and determine the flight attitude parameters of the UAV; A dynamic analysis unit is used to retrieve a historical environmental record data set of a forest area, perform dynamic analysis based on the historical environmental record data set, and construct a forest environment change curve chart; a path planning unit, configured to perform path planning based on the forest environment change curve diagram and the UAV flight attitude parameters to generate spraying path data; The second data adding unit is configured to add the spraying path data to the spraying task.

5. The UAV-based precision spraying system for forest pests and diseases as claimed in claim 4, characterized in that: The precision spraying module comprises: A density assessment unit, configured to assess the pest density based on the spraying dosage data by using a drone to perform the spraying task, and generate a pest reduction ratio value; a coverage assessment unit, configured to assess the pest coverage of the spraying path data by executing the spraying task through a drone, and generate a pest coverage ratio value; An effect evaluation unit, configured to perform effect evaluation based on the pest reduction ratio value and the pest coverage ratio value to generate a spraying effect score; A dynamic feedback optimization unit is used to add the spraying effect score to the evaluation result to perform dynamic feedback optimization on the spraying task to generate the spraying optimization task.

6. The UAV-based precision spraying system for forest pests and diseases according to claim 5, characterized in that: The dynamic feedback optimization unit includes: A scoring judgment channel is used to set an expected spraying threshold and judge whether the spraying effect score meets the expected spraying threshold; a tracking and monitoring channel for continuously tracking and monitoring the forest area based on the spraying task if the spraying effect score meets the expected spraying threshold, generating dynamic tracking and monitoring data, and continuously feeding the dynamic tracking and monitoring data back to the spraying task; an abnormality analysis channel, for generating a feedback instruction if the spraying effect score does not meet the expected spraying threshold, and performing abnormality analysis on the spraying task through the feedback instruction to determine abnormal spraying dosage data and / or abnormal spraying path data; The dynamic adjustment channel is used to use the abnormal spraying dosage data and / or the abnormal spraying path data as indexes, feed them back to the spraying task for dynamic adjustment, and generate the spraying optimization task.

Images