Unmanned aerial vehicle pesticide application operation method based on simulation model

Abstract

The invention discloses an unmanned aerial vehicle pesticide application operation method based on a simulation model. The method comprises the following steps: constructing an unmanned aerial vehicle spraying simulation analysis model; according to the model, simulating the deposition amount, drift amount and wetting degree of the pesticide liquid on the plant leaves under various different environment wind speeds, unmanned aerial vehicle flight attitudes and spraying inclination angle conditions, and obtaining a data set; constructing a mapping relation model, and training the mapping relation model through the obtained data set; environment wind speed data of the farmland during operation are collected in real time through a ground wind speed sensor; importing farmland wind speed data acquired in real time during operation and expected values of the deposition amount, the wetting degree and the drift amount into the trained mapping relation model to obtain a flight attitude angle and a spray inclination angle of the unmanned aerial vehicle; and applying the obtained flight attitude angle and spray inclination angle of the unmanned aerial vehicle to the unmanned aerial vehicle for real-time operation. Operation parameters can be adjusted in real time, so that the better deposition amount and wetting degree are obtained, and the pesticide application operation effect of the unmanned aerial vehicle is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) operations, and in particular to a UAV pesticide application method based on a simulation model. Background Art

[0002] Plant protection drone spraying technology has become an important tool for field weed control due to its advantages such as high efficiency, strong terrain adaptability, and high pesticide utilization rate. However, existing technologies still have significant drawbacks in practical applications, mainly manifested in insufficient droplet deposition and poor adaptability to dynamic environments. These lead to low effective utilization and severe droplet drift, which directly affects weed control effectiveness and environmental safety.

[0003] The reasons for the above defects are as follows:

[0004] 1) Limitations of Simulation Models: Current simulation models are mostly based on idealized static wind field assumptions and fail to truly reflect the impact of complex airflow in the field on droplet motion. Factors such as wind speed, direction, and turbulence in the field environment can significantly alter droplet motion and deposition distribution. Existing models fail to fully account for these dynamic factors, resulting in significant deviations between simulation results and actual operational scenarios.

[0005] 2) Fixed simulation parameters: Existing technologies often use fixed simulation parameters, failing to tailor droplet penetration and deposition to the specific characteristics of rice plants. Factors such as rice plant density, height, and leaf morphology directly affect droplet penetration and deposition, but existing technologies fail to dynamically adjust for these characteristics, effectively preventing droplet deposition efficiency.

[0006] 3) Lack of multi-physics coupling mechanisms: Existing technologies lack systematic research on the multi-physics coupling mechanisms of wind field, droplet, and plant. In actual operations, the movement and deposition of droplets are influenced by multiple factors, including the wind field, plant structure, and the droplet's own characteristics. Existing technologies fail to effectively couple these factors, resulting in simulation results that fail to accurately reflect actual operation scenarios, which in turn affects spraying effectiveness and environmental safety.

[0007] In summary, the existing plant protection drone spraying technology has obvious deficiencies in simulation models, parameter settings, and multi-physical field coupling mechanisms, and urgently needs to be improved to increase droplet deposition and enhance dynamic environmental adaptability, thereby improving weed control effects and environmental safety. Summary of the Invention

[0008] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a drone spraying operation method based on a simulation model.

[0009] To achieve the above objectives, the technical solutions provided by the present invention are:

[0010] A method for applying pesticides by a drone based on a simulation model, comprising:

[0011] Construct a UAV spray simulation analysis model that takes into account the environmental wind field, rice plants, UAV flight posture, and spray inclination angle;

[0012] Based on the constructed drone spray simulation analysis model, the deposition, drift, and wetness of the liquid on the plant leaves under various environmental wind speeds, drone flight postures, and spray inclination conditions were simulated to obtain a data set;

[0013] Construct a nonlinear mapping relationship model between ambient wind speed, UAV flight attitude, and spray inclination and deposition amount, drift amount, and wetness, and train the mapping relationship model using the obtained data set;

[0014] Use ground wind speed sensors to collect real-time environmental wind speed data during farmland operations;

[0015] The real-time wind speed data collected during operation and the expected values ​​of deposition, wetness, and drift are imported into the trained mapping relationship model to derive the UAV flight attitude angle and spray inclination angle.

[0016] The obtained UAV flight attitude angle and spray inclination angle are applied to the UAV for real-time operation.

[0017] Furthermore, a UAV spray simulation analysis model was constructed that took into account the environmental wind field, rice plants, UAV flight posture, and spray inclination angle, including:

[0018] Constructing the UAV 3D model and external flow field includes creating the UAV fuselage, rotors, and nozzles, setting the computational domain size, and the rotor's rotational domain size. After modeling is complete, the model is imported into Design Modeler for pre-processing and simplification, and finally, tetrahedral meshing is performed on the entire model using the ICEM mesh generation tool.

[0019] Construct a rice plant model. Set the rice material according to its specific properties and divide it into the main plant and leaves. Adjust parameters including plant height, number of leaves, leaf inclination, and leaf area. Then fix and assemble it in 3D modeling software. Then import it into Abaqus to set material properties, set the elastic model, Poisson's ratio, and density. After the definition is completed, mesh it using the ICEM mesh generation tool. The entire plant is meshed with hexahedral elements. The main body and leaves are connected with multiple nodes. At the same time, the mesh at the connection is optimized to avoid distorted elements.

[0020] Laying out the rice model, including fixed boundary conditions, which include the row and plant spacing, so that each rice plant is independent of each other;

[0021] Set up the physical field, including defining the ambient wind speed, ambient boundary conditions, ambient wind field settings, UAV flight parameter settings, and overall gravity settings;

[0022] Set the spray droplet parameters and perform discrete phase settings, where the set spray droplet parameters include spray particle size, spray flow rate, spray inclination angle, and droplet initial velocity; perform discrete phase settings including using the Euler-Lagrange coupling model, solving the fluid phase using the transient NS equations, tracking the droplets using the DMP model in the discrete phase, and setting the momentum exchange between the droplets and the air.

[0023] Furthermore, the parameters in the UAV spray simulation analysis model are sorted out. The parameters include environmental wind speed parameters, UAV parameters, and spray parameters. The UAV parameters include flight altitude, flight speed, and flight attitude; the spray parameters include spray particle size, spray flow rate, and spray inclination; Latin hypercube sampling is performed on these seven parameters.

[0024] Furthermore, based on the constructed drone spray simulation analysis model, the deposition amount, drift amount, and wetness of the liquid on the plant leaves were simulated under various environmental wind speeds, drone flight postures, and spray inclination angles. The flight altitude, flight speed, spray particle size, and spray flow rate were all fixed values, while the environmental wind speed, drone flight posture, and spray inclination angle were all variables.

[0025] The simulation analysis process using the UAV spray simulation analysis model includes:

[0026] Solve the set simulation, export the discrete phase result file, build a 3D sedimentation cloud map based on the collision data and sedimentation density, and map it to a 3D spatial grid;

[0027] Deposition density = droplet mass / grid volume, then integrate the deposition density of all grids to obtain the deposition amount;

[0028] Droplets outside the overall plant area are defined as drift droplets. The droplet trajectories are derived for cluster analysis and coloring visualization, and the total mass of all droplets outside the defined range is extracted to obtain the drift amount.

[0029] The percentage of moisture content in the leaf surface grid is taken and calculated in combination with the droplet collision data. When the deposition mass of a single grid is greater than the set value, it is judged to be wet. The wetness degree = the number of wet grids / the total number of grids, thus obtaining the wetness degree of the leaf;

[0030] The parameter data of the UAV spray simulation analysis model and the simulation result data including the deposition amount, drift amount and wetness degree of the liquid medicine on the plant leaves are preprocessed to obtain a data set.

[0031] Furthermore, the mapping relationship model is trained using the obtained data set, including:

[0032] The obtained dataset was split into training set and test set in a ratio of 8:2 to verify the performance and avoid overfitting;

[0033] The mapping relationship model design uses a multi-layer perceptron, which includes three input layers corresponding to the ambient wind speed, the UAV's flight attitude angle, and the spray inclination angle, four hidden layers, and three task output layers corresponding to the deposition amount, drift amount, and wetness. The nonlinear interaction characteristics of the input parameters are extracted through the hidden layers, and the output layers are independent outputs for synchronous prediction.

[0034] The mapping relationship model is trained using the segmented training set, where a dynamic weighted multi-task loss function is used, Ltotal = αL1+βL2+γL3, where the weights α, β, and γ are adaptively adjusted according to the error. Finally, the model performance is verified using the test set.

[0035] Furthermore, ground wind speed sensors are used to collect real-time wind speed data from farmlands during operations, including:

[0036] Reasonably deploy ground wind speed sensors in the fields to obtain field wind speed in real time, perform data fusion and screening, and eliminate instantaneous disturbances;

[0037] The obtained wind speed is detected and transmitted in real time via wireless network to the environmental wind speed data.

[0038] Furthermore, the process of obtaining the UAV flight attitude angle and spray inclination angle includes:

[0039] The acquired ambient wind speed data is combined with the previously set expected values ​​of deposition, drift, and wetness in the trained mapping relationship model. Through the optimal parameter combination, the required UAV flight attitude angle and spray inclination angle are calculated.

[0040] The inversely calculated flight attitude angle and spray inclination angle are imported into the UAV control system.

[0041] Furthermore, the obtained UAV flight attitude angle and spray inclination angle are applied to the UAV, including:

[0042] The inversely calculated flight attitude angle and spray inclination angle data are loaded into the UAV's executable instructions, the nozzle rotation angle is adjusted through the motor drive, and the attitude angle and forward speed of the UAV are adjusted through the UAV control system; this process is repeated continuously until the spraying is completed.

[0043] Compared with the existing technology, the principles and advantages of this technical solution are as follows:

[0044] This technical solution first constructs a drone spray simulation analysis model that takes into account the environmental wind field, rice plants, drone flight attitude, and spray inclination angle; then, based on the constructed drone spray simulation analysis model, the deposition, drift, and wetness of the liquid medicine on the plant leaves under various environmental wind speeds, drone flight attitudes, and spray inclination angles are simulated to obtain a data set; then a mapping relationship model is constructed for the nonlinear mapping of environmental wind speed, drone flight attitude, and spray inclination angle to deposition, drift, and wetness, and the mapping relationship model is trained using the obtained data set; then, the environmental wind speed data of the farmland during operation is collected in real time through a ground wind speed sensor; then, the farmland wind speed data collected in real time during operation and the expected values ​​of deposition, wetness, and drift are imported into the trained mapping relationship model to obtain the drone flight attitude angle and spray inclination angle; finally, the obtained drone flight attitude angle and spray inclination angle are applied to the drone for real-time operation.

[0045] The drone spray simulation analysis model constructed by this technical solution can truly reflect the impact of complex airflow in the field on droplet movement. It combines the regulatory effect of rice plant characteristics on droplet penetration and deposition, supplements the wind field-droplet-plant multi-physical field coupling mechanism, and can adjust operation parameters in real time to obtain better deposition amount and wetness, greatly improving the effect of drone spraying operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the services required for use in the embodiments or the prior art descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0047] Figure 1 This is a principle flow chart of a UAV pesticide application method based on a simulation model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The present invention will be further described below in conjunction with specific embodiments:

[0049] like Figure 1 As shown, the drone spraying method based on the simulation model described in this embodiment includes the following steps:

[0050] S1. Construct a UAV spray simulation analysis model that takes into account the environmental wind field, rice plants, UAV flight posture, and spray inclination angle;

[0051] The specific process of this step is as follows:

[0052] Constructing a 3D UAV model and external flow field involves creating the UAV fuselage, rotors, and nozzles. The computational domain size is set to 20m × 20m × 10m, with the rotor's rotation domain size set to a radius of 3m and a height of 0.5m. After modeling is completed using SolidWorks, the model is imported into Design Modeler for model pre-processing and simplification. Finally, tetrahedral meshing is performed on the entire model using the ICEM mesh generation tool.

[0053] Construct a rice plant model. Set the rice material according to its specific properties and divide it into the main plant and leaves. Adjust parameters including plant height, number of leaves, leaf inclination, and leaf area. Then fix and assemble it in 3D modeling software. Then import it into Abaqus to set material properties, set the elastic model, Poisson's ratio, and density. After the definition is completed, mesh it using the ICEM mesh generation tool. The entire plant is meshed with hexahedral elements. The main body and leaves are connected with multiple nodes. At the same time, the mesh at the connection is optimized to avoid distorted elements.

[0054] The rice model was laid out, including fixed boundary conditions, which included the row and plant spacing of the plants, so that each rice plant was independent of each other, with an array distance of 30 cm and an overall range of 3m × 5m;

[0055] Set up the physical field, including defining the ambient wind speed of 0.5m / s to 5.0m / s, environmental boundary conditions (the bottom is a wall boundary, the top is a zero pressure inlet, and the side is a zero pressure outlet), ambient wind field settings (using the SST k-ω model inside the boundary layer, the k-ε model in the external flow area, and setting the inlet wind speed), UAV flight parameter settings (including flight altitude of 1.0m to 3.0m; flight speed of 3 to 8m / s, flight attitude angle of -15° to 15°, rotor speed of 6000r / min, and adjacent rotors are forward and reverse), and overall gravity settings (setting the gravity acceleration to 9.81m / s 2 vector direction);

[0056] The spray droplet parameters and discrete phase settings are set, wherein the set spray droplet parameters include spray particle size 100μm~300μm, spray flow rate 1.0~5.0L / min, spray inclination 0°~30°, and droplet initial velocity 0m / s-12m / s; the discrete phase setting includes using the Euler-Lagrange coupling model, the fluid phase is specifically solved by the transient NS equation, the discrete phase uses the DMP model to track the droplets, and the momentum exchange between the droplets and the air is set.

[0057] S2. Based on the constructed drone spray simulation analysis model, simulate the deposition, drift, and wetness of the liquid medicine on the plant leaves under various environmental wind speeds, drone flight postures, and spray inclination conditions to obtain a data set;

[0058] The specific process of this step includes:

[0059] The parameters in the UAV spray simulation analysis model were organized. These parameters include ambient wind speed, UAV parameters, and spray parameters. UAV parameters include flight altitude, flight speed, and flight attitude; spray parameters include spray particle size, spray flow rate, and spray inclination. Latin hypercube sampling was performed on these seven parameters (flight altitude, flight speed, spray particle size, and spray flow rate are all fixed values, while ambient wind speed, UAV flight attitude, and spray inclination are all variables). First, a reasonable range of values ​​for each parameter was determined, for example, ambient wind speed 0m / s to 5m / s. The sample size was then determined. For example, if N = 100, each parameter was divided into 100 equal-width intervals, and one value was randomly selected from each interval. The 100 sampled values ​​for each parameter were randomly permuted and finally combined. If N = 100, 100 data sets could be collected. This effectively covers multiple factors and reduces correlation between data.

[0060] The simulation analysis process using the UAV spray simulation analysis model includes:

[0061] Solve the set simulation, export the discrete phase result file, build a 3D sedimentation cloud map based on the collision data and sedimentation density, and map it to a 3D spatial grid;

[0062] Deposition density = droplet mass / grid volume, then integrate the deposition density of all grids to obtain the deposition amount;

[0063] Droplets outside the overall plant area are defined as drift droplets. The droplet trajectories are derived for cluster analysis and coloring visualization, and the total mass of all droplets outside the defined range is extracted to obtain the drift amount.

[0064] The percentage of moisture content in the leaf surface grid is taken and calculated in combination with the droplet collision data. When the deposition mass of a single grid is greater than the set value, it is judged to be wet. The wetness degree = the number of wet grids / the total number of grids, thus obtaining the wetness degree of the leaf;

[0065] The parameter data of the UAV spray simulation analysis model and the simulation result data including the deposition amount, drift amount and wetness degree of the liquid medicine on the plant leaves are preprocessed to obtain a data set.

[0066] S3, constructing a nonlinear mapping relationship model of environmental wind speed, UAV flight posture and spray inclination with deposition amount, drift amount and wetness, and training the mapping relationship model with the obtained data set;

[0067] In this step, the mapping relationship model is trained using the obtained dataset, including:

[0068] The obtained dataset was split into training set and test set in a ratio of 8:2 to verify the performance and avoid overfitting;

[0069] The mapping relationship model design uses a multi-layer perceptron, which includes three input layers corresponding to the ambient wind speed, the UAV's flight attitude angle, and the spray inclination angle, four hidden layers, and three task output layers corresponding to the deposition amount, drift amount, and wetness. The nonlinear interaction characteristics of the input parameters are extracted through the hidden layers, and the output layers are independent outputs for synchronous prediction.

[0070] The mapping relationship model is trained using the segmented training set, where a dynamic weighted multi-task loss function is used, Ltotal = αL1+βL2+γL3, where the weights α, β, and γ are adaptively adjusted according to the error. Finally, the model performance is verified using the test set.

[0071] S4. Use ground wind speed sensors to collect real-time wind speed data from farmland during operation, including:

[0072] Reasonably deploy ground wind speed sensors in the fields to obtain field wind speed in real time, perform data fusion and screening, and eliminate instantaneous disturbances;

[0073] The obtained wind speed is detected and transmitted in real time via wireless network to the environmental wind speed data.

[0074] S5. Importing the real-time collected farmland wind speed data and the expected values ​​of deposition, wetness, and drift into the trained mapping relationship model to derive the UAV flight attitude angle and spray inclination angle;

[0075] In this step, the process of obtaining the UAV flight attitude angle and spray inclination angle includes:

[0076] The acquired ambient wind speed data is combined with the previously set expected values ​​of deposition, drift, and wetness in the trained mapping relationship model. Through the optimal parameter combination, the required UAV flight attitude angle and spray inclination angle are calculated.

[0077] The inversely calculated flight attitude angle and spray inclination angle are imported into the UAV control system.

[0078] S6, loading the inversely calculated flight attitude angle and spray inclination angle data into the executable instructions of the UAV, adjusting the nozzle rotation angle through the motor drive, and adjusting the attitude angle and forward speed of the UAV through the UAV control system;

[0079] S7. Continuously repeat steps S4-S6 until the application of the pesticide is completed.

[0080] The embodiments described above are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, any changes made based on the shape and principle of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for spraying pesticides using a drone based on a simulation model, characterized in that: include: Construct a UAV spray simulation analysis model that takes into account the environmental wind field, rice plants, UAV flight posture, and spray inclination angle; Based on the constructed drone spray simulation analysis model, the deposition, drift, and wetness of the liquid on the plant leaves under various environmental wind speeds, drone flight postures, and spray inclination conditions were simulated to obtain a data set; Construct a nonlinear mapping relationship model between ambient wind speed, UAV flight attitude, and spray inclination and deposition amount, drift amount, and wetness, and train the mapping relationship model using the obtained data set; Use ground wind speed sensors to collect real-time environmental wind speed data during farmland operations; The real-time wind speed data collected during operation and the expected values ​​of deposition, wetness, and drift are imported into the trained mapping relationship model to derive the UAV flight attitude angle and spray inclination angle. The obtained UAV flight attitude angle and spray inclination angle are applied to the UAV for real-time operation.

2. The method for spraying pesticides using a drone based on a simulation model according to claim 1, wherein: Construct a UAV spray simulation analysis model that takes into account the environmental wind field, rice plants, UAV flight posture, and spray inclination angle, including: Constructing the UAV 3D model and external flow field includes creating the UAV fuselage, rotors, and nozzles, setting the computational domain size, and the rotor's rotational domain size. After modeling is complete, the model is imported into Design Modeler for pre-processing and simplification, and finally, tetrahedral meshing is performed on the entire model using the ICEM mesh generation tool. Construct a rice plant model. Set the rice material according to its specific properties and divide it into the main body of the plant and leaves. Adjust parameters including plant height, number of leaves, leaf inclination, and leaf area. Then fix and assemble it in 3D modeling software. Then import it into Abaqus to set material properties, set the elastic model, Poisson's ratio, and density. After the definition is completed, mesh it using the ICEM mesh generation tool. The entire plant is meshed with hexahedral elements. The main body and leaves are connected with multiple nodes. At the same time, the mesh at the connection is optimized to avoid distorted elements. Laying out the rice model, including fixed boundary conditions, which include the row and plant spacing, so that each rice plant is independent of each other; Set up the physical field, including defining the ambient wind speed, ambient boundary conditions, ambient wind field settings, drone flight parameter settings, and overall gravity settings; Set the spray droplet parameters and perform discrete phase settings, where the set spray droplet parameters include spray particle size, spray flow rate, spray inclination angle, and droplet initial velocity; perform discrete phase settings including using the Euler-Lagrange coupling model, solving the fluid phase using the transient NS equations, tracking the droplets using the DMP model in the discrete phase, and setting the momentum exchange between the droplets and the air.

3. The method for spraying pesticides using a drone based on a simulation model according to claim 2, wherein: The parameters in the UAV spray simulation analysis model are sorted out, including environmental wind speed parameters, UAV parameters, and spray parameters. UAV parameters include flight altitude, flight speed, and flight attitude; spray parameters include spray particle size, spray flow rate, and spray inclination; Latin hypercube sampling is performed on these seven parameters.

4. The method for spraying pesticides using a drone based on a simulation model according to claim 3, wherein: Based on the constructed drone spray simulation analysis model, the deposition amount, drift amount, and wetness of the liquid on the plant leaves were simulated under various environmental wind speeds, drone flight postures, and spray inclination angles. The flight altitude, flight speed, spray particle size, and spray flow rate were all fixed values, while the environmental wind speed, drone flight posture, and spray inclination angle were all variables. The simulation analysis process using the UAV spray simulation analysis model includes: Solve the set simulation, export the discrete phase result file, build a 3D sedimentation cloud map based on the collision data and sedimentation density, and map it to a 3D spatial grid; Deposition density = droplet mass / grid volume, then integrate the deposition density of all grids to obtain the deposition amount; Droplets outside the overall plant area are defined as drift droplets. The droplet trajectories are derived for cluster analysis and coloring visualization, and the total mass of all droplets outside the defined range is extracted to obtain the drift amount. The percentage of moisture content in the leaf surface grid is taken and calculated in combination with the droplet collision data. When the deposition mass of a single grid is greater than the set value, it is judged to be wet. The wetness degree = the number of wet grids / the total number of grids, thus obtaining the wetness degree of the leaf; The parameter data of the UAV spray simulation analysis model and the simulation result data including the deposition amount, drift amount and wetness degree of the liquid medicine on the plant leaves are preprocessed to obtain a data set.

5. The method for spraying pesticides using a drone based on a simulation model according to claim 1, wherein: The mapping relationship model is trained using the obtained data set, including: The obtained dataset was split into training set and test set in a ratio of 8:2 to verify the performance and avoid overfitting; The mapping relationship model design uses a multi-layer perceptron, which includes three input layers corresponding to the ambient wind speed, the UAV's flight attitude angle, and the spray inclination angle, four hidden layers, and three task output layers corresponding to the deposition amount, drift amount, and wetness. The nonlinear interaction characteristics of the input parameters are extracted through the hidden layers, and the output layers are independent outputs for synchronous prediction. The mapping relationship model is trained through the segmented training set, in which the dynamic weighted multi-task loss function is used during training. 总 =αL1+βL2+γL3, where the weights α, β, and γ are adaptively adjusted according to the error, and finally the model performance is verified by the test set.

6. The method for spraying pesticides using a drone based on a simulation model according to claim 1, wherein: Ground wind speed sensors are used to collect real-time wind speed data during farmland operations, including: Reasonably deploy ground wind speed sensors in the fields to obtain field wind speed in real time, perform data fusion and screening, and eliminate instantaneous disturbances; The obtained wind speed is detected and transmitted in real time via wireless network to the environmental wind speed data.

7. The method for spraying pesticides using a drone based on a simulation model according to claim 1, wherein: The process of obtaining the UAV flight attitude angle and spray inclination angle includes: The acquired ambient wind speed data is combined with the previously set expected values ​​of deposition, drift, and wetness in the trained mapping relationship model. Through the optimal parameter combination, the required UAV flight attitude angle and spray inclination angle are calculated. The inversely calculated flight attitude angle and spray inclination angle are imported into the UAV control system.

8. The method for spraying pesticides using a drone based on a simulation model according to claim 1, wherein: The obtained UAV flight attitude angle and spray inclination angle are applied to the UAV, including: The inversely calculated flight attitude angle and spray inclination angle data are loaded into the UAV's executable instructions, the nozzle rotation angle is adjusted through the motor drive, and the attitude angle and forward speed of the UAV are adjusted through the UAV control system; this process is repeated continuously until the spraying is completed.

Images