A sowing control system for agricultural plant protection drones

Abstract

The present invention discloses a sowing control system for an agricultural plant protection drone, which relates to the field of agricultural intelligent equipment technology. The control system includes a sowing flow field dynamic modeling and prediction module, a posture control module, a speed matching module, a path compensation module, a visual monitoring module, and a control strategy module. The sowing flow field dynamic modeling and prediction module constructs a flow field dynamic mapping model based on flight altitude, downward pressure airflow velocity, and sowing particle size and density. It combines real-time flight data to predict particle landing point offsets and identify aerodynamic coupling risk areas. The present invention achieves adaptive control of the sowing process through flow field modeling and prediction, posture adjustment, trajectory compensation, speed control, and visual feedback, effectively preventing aerial heaping and nonlinear collapse, avoiding the risks of seedling burning, germination failure, and agricultural residue caused by over-dense sowing, improving operation uniformity and agronomic effects, and possessing practical value and promotion prospects.

Description

Technical Field

[0001] The present invention relates to the technical field of agricultural intelligent equipment, and in particular to a sowing control system of an agricultural plant protection drone. Background Art

[0002] The sowing control system for agricultural plant protection drones refers to an intelligent sowing operation system integrated into agricultural unmanned aerial vehicles. It is mainly used to achieve precise, efficient, and controllable delivery of pesticides, seeds, fertilizers, and other materials in farmland. The system typically includes a flight control module, a sowing actuator, a sensor monitoring device, an environmental perception unit, and an intelligent control algorithm. It can dynamically adjust the flight path, sowing speed, and delivery volume based on crop type, terrain environment, meteorological data, and operational requirements, achieving precise control of the sowing range, uniformity, and operational efficiency. The system is widely used in scenarios such as large-scale agricultural planting, pest control, and soil improvement. It has the advantages of low operating costs, a high degree of automation, and strong environmental adaptability. It is an important component of modern precision agriculture and green plant protection.

[0003] The existing technology has the following shortcomings: When agricultural plant protection drones are performing granular fertilizer or seed spreading operations, when the flight altitude is within a specific critical range and the downward pressure of the drone forms a nonlinear interference relationship with the particle size and density parameters, it is easy to cause the spread particles to exhibit abnormal dynamic behavior in the air. Specifically, due to the unstable coupling between particle size, airflow intensity, and gravitational acceleration, the spread particles are unable to penetrate the airflow belt of the aircraft in time during the falling process, resulting in short-term suspension, rotation, or aggregation below the operation path, thereby forming a local "air pile" phenomenon. When these retained particles are affected by sudden disturbances or the collapse of the airflow structure, they will fall to a specific area on the ground in the form of nonlinear collapse, resulting in a serious exceedance of the local spread density in the target area. This abnormal spreading behavior will directly lead to a series of serious agricultural consequences: on the one hand, excessive local fertilizer concentration may cause drastic changes in ion concentration in the soil, leading to crop root burns, the so-called "seedling burn" phenomenon; on the other hand, dense accumulation of seeds may cause post-sowing suffocation, germination failure, abnormal nutrient competition and other problems, seriously affecting the normal growth and development of crops; at the same time, if it is a high-activity or high-concentration pesticide spreading task, it may cause soil toxicity accumulation, environmental pollution or agricultural product residue risks, significantly reducing operational safety and agronomic effects.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to a person of ordinary skill in the art. Summary of the Invention

[0005] The purpose of the present invention is to provide a sowing control system for agricultural plant protection drones, which realizes adaptive regulation of the sowing process through flow field modeling prediction and attitude adjustment, trajectory compensation, speed control and visual feedback, effectively prevents aerial pile-up and nonlinear collapse, avoids the risks of seedling burning, germination failure and pesticide residue caused by over-dense sowing, improves operation uniformity and agronomic effects, has practical value and promotion prospects, and solves the problems in the above-mentioned background technology.

[0006] To achieve the above objectives, the present invention provides the following technical solutions: a sowing control system for an agricultural plant protection UAV, comprising a sowing flow field dynamic modeling and prediction module, an attitude control module, a speed matching module, a path compensation module, a visual monitoring module, and a control strategy module:

[0007] The seeding flow field dynamic modeling and prediction module builds a dynamic flow field mapping model based on flight altitude, downward pressure airflow velocity, seeding particle size and density. It combines real-time flight data to predict particle landing point deviation and identify aerodynamic coupling risk areas.

[0008] The attitude control module adjusts the flight attitude angle and the orientation angle of the spreading device based on the prediction results, so that the particle release direction avoids the vortex area of ​​the downward pressure airflow and reduces the aerodynamic retention effect;

[0009] The speed matching module, based on the matching relationship between particle size and wind speed, controls the initial velocity of particle release and reduces the probability of resonance by adjusting the driving frequency of the sowing motor and the outlet cross-sectional structure;

[0010] The path compensation module performs trajectory compensation in high-risk areas, adjusts flight altitude or offsets the path to form a disturbance suppression zone to prevent particle aggregation and collapse;

[0011] The visual monitoring module uses a visual device to collect images of actual particle distribution, generate a thermal map of the offset rate, and modify the spreading parameters to optimize distribution uniformity.

[0012] The control strategy module calculates the control factor based on the error between the actual landing point and the predicted trajectory, and jointly adjusts the release speed, direction and flight altitude to achieve precise control and dynamic optimization of multi-parameter spreading.

[0013] Preferably, the specific steps of constructing the dynamic mapping model of the spreading flow field are as follows:

[0014] Collect flight altitude, flight speed, attitude angle, rotor speed, particle size, particle density, particle initial velocity, gravitational acceleration and disturbance coefficient parameters;

[0015] Establish a three-dimensional model of the spreading flow field based on the nonlinear fluid dynamics equations and the particle motion trajectory prediction formula;

[0016] Use the real-time flight data collected by the flight control system to dynamically drive the model and calibrate its parameters;

[0017] Based on the probability density distribution of particle landing points output by the model, high-risk areas of aerodynamic coupling during the seeding process are identified.

[0018] Preferably, the specific steps of adjusting the flight attitude angle and the spreading device orientation angle are as follows:

[0019] Based on real-time flight parameters and aerodynamic flow field models, high interference areas in the rotor downflow area are identified and a three-dimensional spatial interference map is constructed;

[0020] Adjust the pitch, roll, and yaw angles of the agricultural plant protection drone so that the axis of the sowing trajectory deviates from the main axis direction of the high interference area;

[0021] The spray direction angle of the spreading device is adjusted in conjunction to make the particle release direction away from the core area of ​​air flow disturbance;

[0022] By comparing the spreading trajectory data with the image recognition results, feedback is provided to optimize the flight attitude angle and the spray direction angle to achieve spreading trajectory stability control.

[0023] Preferably, the specific steps of controlling the initial velocity of particle release are as follows:

[0024] Construct a particle size and wind speed matching model to determine the ideal initial release velocity range for particles of different sizes at the target wind speed;

[0025] Collect information on particle size and current wind speed, and adjust the drive frequency of the spreading motor to ensure that the particles obtain a matching initial release velocity;

[0026] Adjust the opening size and direction angle of the particle outlet cross-section structure to achieve dynamic control of the release path;

[0027] Based on the comparison between the initial release velocity and the downward pressure airflow velocity, the resonance risk is assessed and fine-tuning compensation control is performed.

[0028] Preferably, the specific steps of performing the flight trajectory compensation operation in the high coupling risk area are as follows:

[0029] Based on the seeding flow field model and particle landing point prediction results, high coupling risk areas are identified and mapped to the flight path;

[0030] Before flying to a high coupling risk area, control the agricultural plant protection drone to temporarily increase its flight altitude to form a disturbance dilution layer;

[0031] Control the agricultural plant protection drone to perform lateral path deviation operations so that the sowing trajectory avoids the core disturbance area;

[0032] Collect image data of the sowing results and perform distribution analysis, and optimize trajectory compensation parameters based on particle landing point offset feedback.

[0033] Preferably, the specific steps for monitoring the falling trajectory of the seeded particles are as follows:

[0034] The visual acquisition device installed on the agricultural plant protection drone is used to collect particle falling images in real time;

[0035] Perform image recognition processing on the collected images, extract the coordinates of the particle landing points and generate a particle distribution heat map;

[0036] Compare the thermal map with the theoretical sowing trajectory, calculate the sowing particle landing point deviation rate and establish a deviation model;

[0037] Based on the results of the offset model, the spreading release speed, release direction angle and flight attitude angle are corrected to optimize the spreading distribution uniformity.

[0038] Preferably, the specific steps of constructing a sowing offset control factor based on the error value between the actual landing position of the sowing particles and the predicted trajectory, and adjusting the release speed, direction angle and flight altitude accordingly are as follows:

[0039] The particle landing point coordinate data is collected through the visual monitoring module, and the landing point error of each particle is obtained by combining it with the predicted trajectory model. The landing point error is defined as follows:

[0040] , where and is the actual landing coordinate of the particle in the working area, and are the predicted landing coordinates of the particle, is the single particle landing error;

[0041] In the time period The average spreading error value of all sampled particles is calculated by the following expression:

[0042] , where is the number of sampled particles, It is The landing point error of each particle is is the average spreading error value;

[0043] In order to achieve standardized control of the error, the error tolerance threshold is introduced and the normalized offset control factor is constructed. The calculation formula is as follows:

[0044] , where is the maximum acceptable error threshold, is the offset regulating factor;

[0045] Obtaining the offset control factor After that, the adaptive adjustment process of the spreading control parameters is started, covering the three core variables of particle release speed, release direction angle and flight height. The control expression is as follows:

[0046] , where Is the current sowing cycle The initial release velocity of the particles, is the particle release rate adjustment sensitivity coefficient, is the release rate of particles in the next cycle after adjustment, is the particle release direction angle of the spreading device in the current spreading cycle, is the release direction angle adjustment sensitivity coefficient, is the angle adjustment response smoothing coefficient, is the release direction angle of the particles in the next cycle after adjustment, is the flight altitude of the current spreading cycle, is the flight altitude adjustment amplitude coefficient, It is the flying height of the particle in the next cycle after adjustment.

[0047] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0048] This invention constructs a mapping model of the spreading flow field and aerodynamic interference to identify risk areas in advance. It also effectively avoids particle aggregation and retention through attitude adjustment and trajectory compensation. Furthermore, a speed control mechanism that matches particle size to wind speed and a visual feedback correction function dynamically adjusts release parameters, making the spreading process adaptive and self-learning. Ultimately, this system effectively avoids the problems of excessive localized spreading density caused by "aerial heaping" and "nonlinear collapse," reduces agricultural risks such as seedling burn, germination failure, and pesticide residue contamination caused by uneven spreading, and improves agronomic results, crop quality, and operational safety. It possesses significant practical value and industrial application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction to the drawings required for use in the embodiments will be given below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0050] Figure 1 The figure is a module schematic diagram of a sowing control system for an agricultural plant protection UAV of the present invention.

[0051] Figure 2This is a flowchart of the working principle of a sowing control system for an agricultural plant protection drone of the present invention.

[0052] Figure 3 This is a logic diagram of a sowing control system for an agricultural plant protection UAV according to the present invention. DETAILED DESCRIPTION

[0053] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of this disclosure will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0054] The present invention provides Figure 1 The sowing control system for an agricultural plant protection UAV shown in the figure includes a sowing flow field dynamic modeling and prediction module, an attitude control module, a speed matching module, a path compensation module, a visual monitoring module, and a control strategy module:

[0055] The seeding flow field dynamic modeling and prediction module constructs a seeding flow field dynamic mapping model that integrates flight altitude parameters, rotor downforce velocity parameters, and seeding particle size and density parameters. This model is driven by real-time flight data to predict the trend of seeding particle landing point deviation and identify aerodynamic coupling risk areas during the seeding process.

[0056] To address the complex coupling between flight altitude, rotor downforce, and particle physics during the spreading of granular materials (such as fertilizers, seeds, or pesticides) by agricultural plant protection drones, which can lead to problems such as offset particle landing points and uneven sedimentation, a dynamic modeling and prediction method for the spreading flow field is proposed as the core prediction submodule of the system. This method integrates multiple real-time flight data with the physical parameters of the spreading particles and establishes a continuously evolving dynamic mapping model to enable a priori identification of high-risk spreading areas. The specific implementation is as follows:

[0057] First, a basic physical modeling framework for the spreading flow field was established. By obtaining parameters such as the agricultural plant protection UAV's current flight altitude, flight speed, attitude angle, and rotor speed, a velocity field distribution function was constructed to reflect the characteristics of the UAV's downward pressure airflow. Furthermore, the physical parameters of the particles, such as particle size, density, shape coefficient, and discharge velocity, were used as input variables. Gravitational acceleration, airflow disturbance parameters, and the influence of the airframe structure were introduced to construct a preliminary three-dimensional mathematical model of the spreading aerodynamic flow field. This model utilizes a nonlinear fluid dynamics system coupled with a particle trajectory prediction formula to theoretically simulate the particle's falling path and the distribution trend of the landing area in the airflow.

[0058] Secondly, to address the idealization and static characteristics of basic modeling, real-time flight data is introduced as a dynamic driving factor to dynamically calibrate and iteratively optimize the model. Specifically, this involves collecting external disturbances such as the flight altitude change rate, the real-time rotor thrust curve, and wind speed and direction sensor data in real time through the flight control system. Furthermore, the dynamic drag coefficient, airflow disturbance intensity distribution, and particle drop path offset function in the prediction model are weighted and corrected using a sliding time window, combined with the initial velocity of the particle seeding and the angle between the flight direction during operation. This dynamic driving mechanism enables adaptive convergence of the model, improving its prediction accuracy and versatility in different flight scenarios.

[0059] Third, a multivariable landing point deviation prediction algorithm was constructed based on the dynamically calibrated spreading flow field model. This algorithm comprehensively considers the deviation trends of particle trajectory under the synergistic effects of multiple factors, including airflow disturbances, particle size changes, flight attitude deviations, and sudden changes in operating altitude. Using data fitting and Monte Carlo probabilistic simulation, it estimates the expected landing point probability density distribution of the spread particles at the current moment. Combining farmland plot maps with target spreading area information, it calculates the probability threshold for particles to deviate from the target area. This threshold is used to delineate high-risk spreading areas for aerodynamic coupling, providing a reference for adjusting the spread path and parameters.

[0060] Finally, based on the above-mentioned landing point prediction results, spatial identification and marking of high-risk areas are performed. The identified landing point deviation abnormal areas are projected onto the current operation path, and a two-dimensional risk heat map is constructed, with color gradients used to represent the seeding uniformity risk level of each area. The system inputs this heat map into the seeding control core module for subsequent seeding attitude adjustment, flight path compensation, and parameter correction modules to carry out strategic linkage. At the same time, the module also has self-learning capabilities. By comparing the differences between historical seeding effect data and predicted deviation results, it can reversely adjust and retrain the flow field model, thereby achieving adaptive model evolution and enhanced prediction accuracy in long-term seeding operations.

[0061] In summary, this implementation not only integrates aerodynamic parameters and particle characteristics in theoretical modeling, but also incorporates real-time flight data-driven and adaptive iterative optimization mechanisms in engineering implementation. It also features high-risk spreading area identification and model evolution capabilities, significantly improving the accuracy and uniformity of granular material spreading operations by agricultural plant protection drones in complex environments. This dynamic spreading flow field modeling and prediction technology provides stable data support for the control strategies of subsequent modules and is one of the fundamental core elements of this invention for achieving precise spreading control.

[0062] The attitude control module performs seeding attitude control operations based on the prediction results. By adjusting the flight attitude angle of the agricultural plant protection UAV and the orientation angle of the seeding device, the seeding particle release direction avoids the high interference area of ​​the rotor downward pressure airflow area, thereby reducing the aerodynamic retention coupling between the seeding particles and the airflow;

[0063] When agricultural plant protection drones are flying at low altitude, the downward pressure airflow area generated by the rotor has high disturbance intensity and instability. The seeding particles are prone to stagnation, rotation and nonlinear motion in this area, which in turn leads to problems such as landing point deviation and local accumulation. To improve the accuracy and stability of seeding particle delivery, a seeding attitude control mechanism based on aerodynamic interference prediction results is proposed. By adjusting the drone's flight attitude angle and the seeding device's orientation angle in a coordinated manner, the particle release direction deviates from the high interference area, thereby weakening the impact of airflow interference. The specific implementation method is as follows:

[0064] First, obtain the aerodynamic interference distribution information under the current flight state. Based on the dynamic mapping model of the seeding flow field constructed in the previous step, combined with real-time flight parameters and environmental disturbance data, calculate the disturbance intensity distribution field in the UAV rotor airflow area. By identifying the spatial areas below the rotor where high-pressure vortices, backflows, or overlapping disturbances exist, determine the high-risk interference areas within this area where seeding particles may be trapped or collapsed. This spatial interference area is represented by three-dimensional coordinates and projected into the coordinate system of the UAV's current attitude and motion direction to form a spatial interference mapping template for attitude control.

[0065] Secondly, based on the spatial distribution of the interference zone, active adjustment operations of the flight attitude angle are performed. The pitch, roll, and yaw angles of the drone are controlled by the flight control system to achieve slight adjustments of the fuselage around the three axes, so that the axis of the seeding trajectory deviates from the main axis direction of the interference zone. Under the premise of ensuring flight stability, the acceptable attitude adjustment range is dynamically calculated, and the prediction model is used to continuously monitor the airflow distribution state after regulation to ensure that the seeding particles can cross the downward pressure airflow boundary and form a smooth falling path. At the same time, in order to avoid flight control instability caused by sudden changes in attitude, a slowly changing control logic is also set to make the adjustment action continuous and gradual, thereby improving the safety and response accuracy of flight control.

[0066] Third, the orientation angle of the spreading device is adjusted synchronously. The spreading device is composed of a nozzle assembly or a spreading outlet structure installed on a rotatable bracket, and has a certain degree of angle adjustment capability. According to the current posture adjustment result and the position of the interference zone, the spray orientation angle of the spreading device is set to the position with the largest angle with the main direction of the disturbance, so that the particle release path forms a projection direction that crosses the boundary in the shortest distance. This adjustment process is controlled by the spreading control unit and the posture perception unit in a coordinated manner, which can achieve real-time angle matching, and combined with the speed adjustment function of the spreading motor, further refine the dynamic coordinated adjustment capability of the particle release speed and direction, so as to ensure that the spreading particles have the optimal power path in the initial stage of release and avoid the high-disturbance core area.

[0067] Finally, the system monitors the effect of attitude control and automatically adjusts it through feedback. Image recognition and data comparison are performed on the particle trajectories before and after the control. The error between the actual particle landing range and the predicted landing point is extracted. Combined with the airflow disturbance heat map, the current attitude angle and injection angle parameters are fine-tuned and optimized. Over multiple rounds of spreading, the system continuously learns the correlation between attitude adjustment effects and spreading uniformity, constructing a parameter-effect correlation model that can be used to rapidly generate optimal attitude control instructions for subsequent spreading tasks, thereby achieving both improved spreading accuracy and control stability.

[0068] In summary, this embodiment establishes a highly intelligent sowing attitude adjustment mechanism by implementing four steps: spatial interference mapping, dynamic flight attitude control, intelligent adjustment of the sowing device's orientation, and feedback optimization. This effectively prevents sowing particles from entering areas of high interference from the rotor's downward airflow, ensuring the stability of the sowing trajectory and accurate control of the landing point. This significantly improves the reliability and agronomic effectiveness of agricultural plant protection drones in complex aerodynamic environments. This technical solution boasts a rational structure, high control accuracy, and strong intelligent adjustment capabilities, making it a key component of the present invention.

[0069] The speed matching module adjusts the release speed of the seeding particles based on the matching relationship between particle size and wind speed. This includes adjusting the drive frequency of the seeding motor and the cross-sectional structure of the seeding particle outlet to fine-tune the initial release speed of seeding particles of different particle sizes and reduce the probability of resonance between the seeding particles and the downward pressure airflow of the rotor.

[0070] To effectively control the release dynamics of particles of varying sizes during UAV seeding operations and mitigate the potential for velocity resonance, accumulation, or nonlinear offset within the rotor's downdraft, a method for adaptively adjusting particle release velocity based on the particle size-wind speed matching relationship was proposed. By dynamically adjusting the seeding motor's drive frequency and the seeding outlet's structural parameters, this method ensures that particles of varying sizes and densities achieve the optimal initial velocity at the moment of release, improving their penetration and achieving highly stable seeding control.

[0071] The specific implementation is as follows:

[0072] First, a model is constructed to match the relationship between particle size and wind speed. Through preliminary experiments or flight sampling, the inertial response time, hovering time, and actual landing point data of particles of different particle sizes after release under different flight wind speed conditions are obtained, and a particle size-wind speed-falling behavior response matrix is ​​constructed. Based on this matrix, a set of mathematical functions is established to describe the ideal initial velocity range required for particles under unit wind speed and specific particle size conditions. This model serves as the control core and is loaded into the sowing parameter control system before sowing to determine the release velocity target for particles of different particle sizes in real time.

[0073] Secondly, based on the real-time identification of particle size and current wind speed information, the spreader motor drive frequency is dynamically adjusted. The spreader motor typically controls a rotary or vibrating conveying structure, and its drive frequency determines the conveying speed and initial kinetic energy of the particles per unit time. The target drive frequency value is calculated using the material property information and flight environment data collected by the spreader control unit, and the motor input voltage or pulse signal duty cycle is adjusted without changing the stability of the entire machine to achieve fine control of the drive frequency. This adjustment method has a microsecond response speed and can achieve rapid switching between different areas during the spread process, ensuring that the system dynamically adapts to the spread needs of particles of different particle sizes.

[0074] Third, implement the linkage adjustment of the cross-sectional structure of the sowing particle outlet. The sowing outlet is the most critical spatial channel in the particle release path. Its shape, size and particle release direction directly affect the initial velocity distribution and the emission angle characteristics. In the present invention, the outlet structure adopts a variable cross-sectional mechanism design, such as an electrically controlled openable and closable slit, a flexible deformable thin-walled pipe, or a discharging component that can rotate to adjust the nozzle angle. According to the current particle size and aerodynamic conditions, the system automatically calculates the required outlet area and flow pattern, and adjusts the geometric dimensions and direction parameters of the discharge port through a micro motor or a shape memory alloy drive mechanism to achieve dynamic optimization matching of the particle release cross-section, thereby further controlling the release speed characteristics.

[0075] Finally, the initial velocity and resonance risk assessment feedback process is executed. By comparing the deviation between the current initial release velocity and the particle target velocity range through the model, combined with the current downward pressure airflow velocity field information of the rotor, it is calculated in real time whether the particles are likely to form resonance or a speed matching zone with the downward pressure airflow after release. If the deviation exceeds the set threshold, the system will automatically adjust the motor frequency or outlet structure to the next fitting gear and perform fine-tuning control. At the same time, the landing point detection module is deployed during the sowing process for verification. By feedback on the actual landing point position offset and trajectory change trend, the matching model parameters are further iteratively optimized to achieve continuous evolution and self-learning updates of the particle release control strategy.

[0076] In summary, this implementation utilizes four steps: particle size-wind speed matching modeling, motor frequency regulation, outlet structure deformation adjustment, and resonance risk feedback. This method effectively improves the drop penetration performance of particles of varying sizes, reduces the probability of resonance or stagnation within the rotor airflow, and ensures accurate, stable, and agronomically effective seeding operations. It represents a key execution control step within the technical system of this invention.

[0077] The path compensation module performs flight trajectory compensation operations in identified high-coupling risk areas, adjusting the flight trajectory of the agricultural plant protection drone, causing the drone to temporarily increase its flight altitude or laterally deviate its flight path in the area, thereby creating an airflow disturbance suppression zone and preventing the seeded particles from forming clusters in the air.

[0078] A feedforward flight trajectory compensation control method is proposed to address the high-risk areas of particle suspension, aggregation, and even collapse caused by the interference and superposition effects between the flight path and the rotor's downward airflow during the seeding operation of agricultural plant protection drones. Based on seeding flow field modeling and risk identification results, this method proactively deflects the path or adjusts the altitude in high-coupling risk areas. By constructing a disturbance-suppressing flight attitude, this method breaks the retention cycle of seeded particles within a specific space and suppresses the aggregation and accumulation phenomenon at the source.

[0079] This implementation includes the following four steps:

[0080] First, a spatial identification and path mapping model for coupling risk areas was constructed. Based on the dynamic modeling of the spreading flow field and the prediction of particle landing points, the three-dimensional coordinate information of areas with high coupling disturbance intensity was extracted and mapped onto the flight mission path. By analyzing the overlap between the flight path and high-risk areas, the spatial overlap between each section of the operation path and the risk area was calculated, and a coupling risk level table was generated. Based on the terrain data and crop distribution characteristics of the target operation plot, flight areas that require priority avoidance or dynamic adjustment were delineated, providing a basis for the dynamic planning of subsequent flight trajectories.

[0081] Secondly, a temporary flight altitude adjustment strategy is implemented. Before entering the identified high-coupling risk area, the minimum effective ascent altitude is calculated based on the system's built-in safe flight altitude limit and airflow disturbance influencing factors, and the flight control system automatically controls the flight platform to climb to the set altitude range in a stable attitude. By increasing the flight altitude, the range of action of the rotor's downward airflow on the surface can be changed, thereby forming a disturbance dilution layer in the originally high-interference area, weakening the retention conditions after particle release, and improving the permeability of the fall. At the same time, the system monitors the stability of the flight attitude during the ascent to prevent rapid climbs from causing flight trajectory deviations or load imbalance problems.

[0082] Third, combined with the flight altitude adjustment, the path lateral offset control operation is performed within the range allowed by the space. By planning the offset trajectory of the drone's flight attitude on the lateral plane, the flight path is controlled to shift a specific distance to the non-interference core area, so that the sowing device and the interference core area are staggered, thereby achieving the purpose of weakening the eddy current intervention. The path offset strategy is dynamically generated based on the plot boundaries, operation coverage and terrain obstacle avoidance requirements to ensure that the continuity of the sowing operation is not disrupted. During the path offset process, the system jointly adjusts the release angle and speed of the sowing device to ensure that the particle delivery always covers the target sowing area, avoiding leakage due to path changes.

[0083] Finally, the trajectory compensation control effectiveness evaluation and feedback optimization process are executed. The system uses the post-operation image monitoring module to perform particle distribution imaging and identification in the spreading trajectory area. It compares the uniformity of landing points and the distribution changes in abnormal areas before and after compensation to determine the effectiveness of the compensation strategy. If a tendency for clustering and piling is still observed in the compensated area, feature mapping analysis is performed between the flight trajectory data and the spreading results, and the path adjustment model parameter library is updated. After multiple operation cycles, the system establishes a data model correlating trajectory compensation with spreading effectiveness, enabling adaptive optimization of the path compensation algorithm and intelligent evolution of the spreading control logic.

[0084] In summary, this implementation forms a complete and dynamically adaptable flight trajectory compensation control method by building a coupled risk identification model, implementing a temporary altitude increase strategy, implementing lateral path deviation control, and performing a feedback optimization process. This method effectively addresses the issues of particle retention and collapse in areas of strong rotor airflow disturbance, improving the stability and uniformity of particle distribution during seeding operations. It is a key step in achieving refined seeding control for UAVs and exhibits high reliability, robustness, and engineering feasibility.

[0085] The visual monitoring module monitors the falling trajectory of the seeding particles. It uses the visual acquisition device installed on the agricultural plant protection drone to collect the actual scattered images of the seeding particles in real time. It then generates a heat map of the seeding particle landing point deviation rate through an image processing algorithm, which is used to reversely correct the current seeding parameters to maintain the uniformity of the seeding distribution.

[0086] To address potential abnormalities in the actual seeding process of agricultural plant protection drones, such as landing point deviation, uneven seeding, and particle aggregation, a seeding particle trajectory monitoring method based on a visual acquisition device and image processing algorithms is proposed. By constructing an image acquisition system, a heat map generation mechanism, and a parameter feedback and correction mechanism during the seeding process, the actual landing point distribution of seeded particles can be captured in real time. This deviation is fed back to the seeding control system in the form of a visual heat map, allowing the seeding parameters to be adjusted in the opposite direction, achieving continuous optimization of seeding uniformity and dynamic stability.

[0087] The method includes the following four implementation steps:

[0088] First, the construction and deployment of the seeding image acquisition system was completed. A high-frame-rate, wide-angle, industrial-grade visual acquisition device, coupled with a low-latency image processing hardware platform, is installed beneath the agricultural plant protection drone. This device is used to capture the particle's trajectory and dispersion in real time after release. By selecting appropriate viewing angles and exposure timings, this visual system ensures stable capture of the particle's trajectory in the air and its distribution after landing, even under conditions of high flight speeds and variable lighting conditions. Furthermore, by configuring a particle color contrast enhancement filter and an ambient background separation algorithm, the effects of ground texture interference and flight vibration on image recognition are effectively suppressed, providing accurate and clear basic data for subsequent image calculations.

[0089] Secondly, real-time processing and feature extraction of the seeding images are performed. Data collected by the image acquisition system is continuously transmitted to the onboard processing unit, where it is processed using an embedded image recognition algorithm. First, the particle trajectories in the image sequence are identified and located, and the final landing coordinates and distribution density of each particle are calculated. The landing data from multiple image frames are then fused, removing abnormal disturbance points to form a statistically significant particle distribution map. Using a spatial grid, particle density is calculated per unit area to construct a two-dimensional heat map matrix. This matrix uses color gradient coding to represent the seeding density of different areas, generating a "sowing particle landing point deviation rate heat map" that reflects the difference in distribution between the actual seeding effect and the theoretical trajectory.

[0090] Third, based on the thermal map data, the system performs offset analysis and fitting modeling of the spreading parameters. The system compares the current thermal map with the theoretical spreading trajectory model and extracts the offset rate values ​​for each area. By establishing a spatial mapping function, the actual delivery error model under the current spreading parameters is calculated, including key indicators such as particle initial velocity deviation, release angle drift, and attitude angle disturbance effects. The system further uses a fitting algorithm to analyze the formation mechanism of abnormal landing point density areas, distinguishing between spreading anomalies caused by aerodynamic disturbances, changes in particle properties, or attitude deviations, providing an accurate basis for the selection of subsequent control strategies.

[0091] Finally, feedback corrections to the spreading parameters are performed. Based on the above analysis results, the system automatically generates a set of correction parameters, including a release speed correction value, a release direction angle correction value, and a flight attitude angle fine-tuning value, and inputs them into the spreading control module to adjust the parameters for the next round of spreading operations in real time. If the system detects that the offset trend has continuous or regional characteristics, the path planning module can also synchronously update the flight trajectory and attitude plan to match the new spreading control logic. At the same time, the visual feedback mechanism has self-learning capabilities and can construct historical spreading heat maps and corresponding correction parameters as data pair samples for subsequent intelligent spreading control strategy model training, thereby improving the automatic adaptability under multivariate interference.

[0092] To summarize, this embodiment forms a closed-loop sowing monitoring and control optimization system through four steps: particle trajectory image acquisition, sowing heat map construction, landing point offset modeling and parameter feedback correction. It can realize accurate identification of particle landing point behavior and real-time dynamic adjustment of sowing parameters during the actual sowing process, effectively improve sowing uniformity and operation accuracy, and provide key support for agricultural plant protection drones to achieve high-quality automatic sowing operations in complex farmland environments.

[0093] The control strategy module constructs a seeding offset control factor based on the error between the actual landing point of the seeding particles and the predicted trajectory. Based on this control factor, it dynamically adjusts the parameter relationship between the seeding particle release speed, release direction angle, and flight altitude, thereby establishing a precise seeding control logic in a multivariable parameter space and achieving continuous dynamic optimization of seeding uniformity.

[0094] This step aims to achieve closed-loop precision control and dynamic adaptive adjustment of the spreading system, and is the most critical "decision-making execution core" in the entire agricultural plant protection drone spreading control method. During the spreading operation, the actual particle landing point is often affected by multiple factors such as flight attitude disturbances, airflow changes, and uneven particle size distribution, resulting in discrepancies between the actual particle landing point and the predicted trajectory. This step first calculates the error between the actual landing point and the predicted trajectory to quantify the degree of spreading deviation and further constructs a spreading deviation control factor. This factor, as a feedback control parameter, directly regulates the coupling relationship between the core spreading control parameters: particle release rate, release direction angle, and flight altitude. This mechanism enables the system to dynamically adjust spreading behavior based on the feedback results of each spreading operation, gradually approaching the optimal spreading state.

[0095] Through this process, the system established a multivariable parameter space control logic based on the principle of "error-factor-parameter adjustment." This allows for real-time adjustments in complex aerodynamic environments while also adapting to variations in seeding conditions depending on crop type, field type, and operational requirements. Ultimately, this control logic enables the system to continuously optimize seeding uniformity, avoid localized misses and piles, and improve agronomic results and operational safety. It also possesses high scalability and self-learning capabilities, making it a key component in achieving intelligent seeding in precision agriculture.

[0096] Based on the error between the actual landing point of the seeded particles and the predicted trajectory, the seeding offset control factor is constructed and the release speed, direction angle, and flight altitude are adjusted accordingly. The specific steps are as follows:

[0097] The particle landing point coordinate data is collected through the visual monitoring module, and the landing point error of each particle is obtained by combining it with the predicted trajectory model. The landing point error is defined as follows:

[0098] , where and It is the actual landing point coordinates of the particles in the working area (obtained by the visual system), indicating the horizontal and vertical coordinates of the particles in the two-dimensional ground coordinate system when they land in actual operation. It is used to evaluate the actual particle spreading results and serves as the basis for accuracy assessment. and It is the predicted landing point coordinate of the particle, based on the current spreading parameters (speed, angle, height) and the theoretical landing point predicted by the flow field model, used to compare the actual landing point and analyze the accuracy of the control parameter setting. The single particle landing point error represents the two-dimensional distance between the actual landing point and the predicted landing point. The unit is usually meters (m). It is the accuracy indicator of the minimum particle size and the basis for error accumulation calculation.

[0099] In the time period The average spreading error value of all sampled particles is calculated by the following expression:

[0100] , where is the number of sampled particles, which indicates the total number of particles involved in the error analysis within a flight seeding cycle, and is used to ensure the stability of statistical data. The larger it is, the more reliable the error mean is. It is The landing point error of each particle is used to support the calculation of the average error, which is the key measure of the system's seeding stability. is the average spreading error value, which represents the average value of the landing point errors of all sampled particles in the current operating area;

[0101] In order to achieve standardized control of the error, the error tolerance threshold is introduced and the normalized offset control factor is constructed. The calculation formula is as follows:

[0102] , where The maximum acceptable error threshold is the upper limit of the allowable sowing error preset by the system. It is usually set according to the crop tolerance and fertilizer density requirements. The unit is meter. It is used to define the "tolerance boundary" for determining whether the control parameters need to be adjusted. is the offset control factor, which is the normalized value of the current seeding offset relative to the maximum allowable error. It is a dimensionless coefficient and the control signal source for the control system to adjust the release speed, angle and height. , then the sowing parameters need to be adjusted immediately;

[0103] The offset control factor It is the core driving variable of the subsequent parameter adjustment process and is used to quantify the degree of deviation between the current spreading state and the ideal spreading state.

[0104] Obtaining the offset control factor After that, the adaptive adjustment process of the spreading control parameters is started, covering the three core variables of particle release speed, release direction angle and flight height. The control expression is as follows:

[0105] , where Is the current sowing cycle The initial release velocity of the particles determines whether the particles can penetrate the downward pressure airflow area of ​​the rotor to prevent suspension or retention. The larger the value, the stronger the particle inertia and the enhanced ability to resist airflow disturbances. is the particle release rate adjustment sensitivity coefficient, which controls The impact on the speed adjustment range, the larger the value, the stronger the response; through historical training dynamic setting, is the release rate of particles in the next cycle after adjustment, It is the particle release direction angle of the spreading device in the current spreading cycle. It adjusts the particle throwing direction to avoid disturbed airflow areas or high-risk paths. The positive and negative offsets can adjust the left / right throwing angle. It is the sensitivity coefficient of the release direction angle adjustment, which controls the maximum amplitude of the angle change. The value range is set according to the mechanical structure limit of the operating equipment. It is the angle adjustment response smoothing coefficient, which prevents the angle adjustment response from being too fast and causing system oscillation. It is usually set between 0.5 and 2 and is obtained through experimental fitting. is the release direction angle of the particles in the next cycle after adjustment, is the flight altitude of the current spreading cycle, It is the flight height adjustment coefficient, which is used to control the response intensity of height change to sowing error, combined with safe flight restrictions and crop type settings. It is the flying height of the particle in the next cycle after adjustment.

[0106] In the spreading control model, and Two nonlinear functions are used to adjust the control response of the seeding direction angle and flight altitude, respectively. These functions are introduced not for mathematical complexity but to give the control system a "slow-changing" and "responsive" behavior that better aligns with physical behavior, avoiding abrupt jumps or over-adjustments in the control variables when the error changes.

[0107] First, It is a periodic smooth function. When the offset control factor When it is small, its value increases linearly, but as As the error increases, its growth rate slows down and approaches 1, effectively suppressing the sharp fluctuations in angle adjustment caused by excessive errors. It is used for angle adjustment in the spreading direction to "suppress" angle adjustment within a safe range under large errors while maintaining good adjustment sensitivity in the small and medium error ranges. Ultimately, it implements a nonlinear and controllable directional correction strategy, improving the system's stability under complex airflow disturbances.

[0108] second, is a logarithmic function that grows slowly. When it is smaller, the response value is smaller. As the error increases, the rate of increase gradually decreases, making this characteristic well-suited for use in flight altitude adjustment control. Altitude changes significantly impact the flight attitude stability, energy consumption, and operational safety of the seeding system, and therefore cannot be increased linearly and rapidly. By introducing a logarithmic function, the system barely adjusts altitude when the error is small, and slowly increases altitude as the error gradually increases. This achieves a safe, gradual, and stable flight altitude compensation strategy, enhancing the system's long-term adaptability to seeding errors while avoiding system instability or reduced crop coverage caused by drastic altitude fluctuations.

[0109] Through the above-mentioned parameter linkage adjustment method, the system can optimize the core control variables in real time according to the actual sowing error situation, realize dynamic correction of particle trajectory while ensuring the stability of the sowing landing point, and thus achieve the goal of continuous sowing uniformity optimization within the multivariable coupling control space.

[0110] The technical solution for the agricultural plant protection drone seeding control system enables dynamic modeling, real-time monitoring, intelligent regulation, and multi-parameter closed-loop control of the entire seeding process, from release to landing, significantly improving the accuracy and uniformity of seeding operations. By mapping the seeding flow field and aerodynamic interference, the system proactively identifies risk areas and effectively mitigates particle aggregation and retention through attitude adjustment and trajectory compensation. Furthermore, a speed control mechanism that matches particle size to wind speed and a visual feedback correction function dynamically adjusts release parameters, making the seeding process adaptive and self-learning. Ultimately, this system effectively avoids localized over-density caused by "aerial heaping" and "nonlinear collapse," reduces agricultural risks such as seedling burn, germination failure, and pesticide residue contamination caused by uneven seeding, and improves agronomic results, crop quality, and operational safety, demonstrating significant practical value and industrial application prospects.

[0111] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0112] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

[0113] It should be noted that, in this document, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprises", "comprising" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "comprising a ..." does not exclude the presence of other identical elements in the process, method, article or device that includes the element.

[0114] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0115] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0116] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0117] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0118] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0119] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0120] The above description is merely illustrative of certain exemplary embodiments of the present invention. It goes without saying that those skilled in the art will be able to modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the scope of protection of the claims.

Claims

1. An agricultural plant protection drone sowing control system, characterized in that: It includes the spreading flow field dynamic modeling and prediction module, attitude control module, speed matching module, path compensation module, visual monitoring module and control strategy module: The seeding flow field dynamic modeling and prediction module builds a dynamic flow field mapping model based on flight altitude, downward pressure airflow velocity, seeding particle size and density. It combines real-time flight data to predict particle landing point deviation and identify aerodynamic coupling risk areas. The attitude control module adjusts the flight attitude angle and the orientation angle of the spreading device based on the prediction results, so that the particle release direction avoids the downward pressure airflow vortex area; The speed matching module controls the initial velocity of particle release by adjusting the driving frequency of the spreading motor and the outlet cross-sectional structure based on the matching relationship between particle size and wind speed; The path compensation module performs trajectory compensation in high-risk areas, adjusts flight altitude or offsets the path to form a disturbance suppression zone to prevent particle aggregation and collapse; The visual monitoring module uses a visual device to collect images of actual particle distribution, generate a thermal map of the offset rate, and modify the spreading parameters to optimize distribution uniformity. The control strategy module calculates the control factor based on the error between the actual landing point and the predicted trajectory, and jointly adjusts the release speed, direction and flight altitude to achieve precise control and dynamic optimization of multi-parameter seeding; Based on the error between the actual landing point of the seeded particles and the predicted trajectory, the seeding offset control factor is constructed and the release speed, direction angle, and flight altitude are adjusted accordingly. The specific steps are as follows: The particle landing point coordinate data is collected through the visual monitoring module, and the landing point error of each particle is obtained by combining it with the predicted trajectory model. The landing point error is defined as follows: , where and is the actual landing coordinate of the particle in the working area, and are the predicted landing coordinates of the particle, is the single particle landing error; In the time period The average spreading error value of all sampled particles is calculated by the following expression: , where is the number of sampled particles, It is The landing point error of each particle is is the average spreading error value; In order to achieve standardized control of the error, the error tolerance threshold is introduced and the normalized offset control factor is constructed. The calculation formula is as follows: , where is the maximum acceptable error threshold, is the offset regulating factor; Obtaining the offset control factor After that, the adaptive adjustment process of the spreading control parameters is started, covering the three core variables of particle release speed, release direction angle and flight height. The control expression is as follows: , where Is the current sowing cycle The initial release velocity of the particles, is the particle release rate adjustment sensitivity coefficient, is the release rate of particles in the next cycle after adjustment, is the particle release direction angle of the spreading device in the current spreading cycle, is the release direction angle adjustment sensitivity coefficient, is the angle adjustment response smoothing coefficient, is the release direction angle of the particles in the next cycle after adjustment, is the flight altitude of the current spreading cycle, is the flight altitude adjustment amplitude coefficient, It is the flying height of the particle in the next cycle after adjustment.

2. The agricultural plant protection drone sowing control system according to claim 1, characterized in that: The specific steps to construct the dynamic mapping model of the spreading flow field are as follows: Collect flight altitude, flight speed, attitude angle, rotor speed, particle size, particle density, particle initial velocity, gravitational acceleration and disturbance coefficient parameters; Establish a three-dimensional model of the spreading flow field based on the nonlinear fluid dynamics equations and the particle motion trajectory prediction formula; Use the real-time flight data collected by the flight control system to dynamically drive the model and calibrate its parameters; Based on the probability density distribution of particle landing points output by the model, high-risk areas of aerodynamic coupling during the seeding process are identified.

3. The agricultural plant protection drone sowing control system according to claim 1, characterized in that: The specific steps for adjusting the flight attitude angle and the spreading device orientation angle are as follows: Based on real-time flight parameters and aerodynamic flow field models, high interference areas in the rotor downflow area are identified and a three-dimensional spatial interference map is constructed; Adjust the pitch, roll, and yaw angles of the agricultural plant protection drone so that the axis of the sowing trajectory deviates from the main axis direction of the high interference area; The spray direction angle of the spreading device is adjusted in conjunction to make the particle release direction away from the core area of ​​air flow disturbance; By comparing the spreading trajectory data with the image recognition results, feedback is provided to optimize the flight attitude angle and the spray direction angle to achieve spreading trajectory stability control.

4. The agricultural plant protection drone sowing control system according to claim 1, characterized in that: The specific steps to control the initial velocity of particle release are as follows: Construct a particle size and wind speed matching model to determine the ideal initial release velocity range for particles of different sizes at the target wind speed; Collect information on particle size and current wind speed, and adjust the drive frequency of the spreading motor to ensure that the particles obtain a matching initial release velocity; Adjust the opening size and direction angle of the particle outlet cross-section structure to achieve dynamic control of the release path; Based on the comparison between the initial release velocity and the downward pressure airflow velocity, the resonance risk is assessed and fine-tuning compensation control is performed.

5. The agricultural plant protection drone sowing control system according to claim 1, characterized in that: The specific steps for performing flight trajectory compensation in high coupling risk areas are as follows: Based on the seeding flow field model and particle landing point prediction results, high coupling risk areas are identified and mapped to the flight path; Before flying to a high coupling risk area, control the agricultural plant protection drone to temporarily increase its flight altitude to form a disturbance dilution layer; Control the agricultural plant protection drone to perform lateral path deviation operations so that the sowing trajectory avoids the core disturbance area; Collect image data of the sowing results and perform distribution analysis, and optimize trajectory compensation parameters based on particle landing point offset feedback.

6. The agricultural plant protection drone sowing control system according to claim 1, characterized in that: The specific steps for monitoring the falling trajectory of seeding particles are as follows: The visual acquisition device installed on the agricultural plant protection drone is used to collect particle falling images in real time; Perform image recognition processing on the collected images, extract the coordinates of the particle landing points and generate a particle distribution heat map; Compare the thermal map with the theoretical sowing trajectory, calculate the sowing particle landing point deviation rate and establish a deviation model; Based on the results of the offset model, the spreading release speed, release direction angle and flight attitude angle are corrected to optimize the spreading distribution uniformity.

Images