An unmanned aerial vehicle pesticide precision spraying method and system for intelligent identification of farmland areas

Abstract

The application provides a kind of unmanned aerial vehicle pesticide precision spraying method and system for intelligent identification of farmland area, and relates to the technical field of agricultural unmanned aerial vehicle pesticide application, the method comprises obtaining the image data of target farmland and the unmanned aerial vehicle flight parameters associated with the image data;According to the image data, the weed distribution area in the target farmland is identified, and a variable spraying prescription map corresponding to the geographical location of the farmland is generated;Wherein, the variable spraying prescription map includes the pesticide application parameters of different weed distribution areas;According to the variable spraying prescription map and the unmanned aerial vehicle flight parameters, the flight path of the unmanned aerial vehicle is planned;Based on the flight path, the unmanned aerial vehicle is controlled to execute differentiated pesticide spraying on different weed distribution areas according to the pesticide application parameters in the variable spraying prescription map, and the spraying result is obtained.According to the spatial difference of weed distribution area, precision variable spraying is carried out, which avoids the overuse or deficiency of pesticide caused by uniform spraying method, improves the utilization efficiency of pesticide and the precision of spraying operation.

Description

Technical Field

[0001] This invention relates to the field of agricultural drone spraying technology, and in particular to a method and system for precise spraying of pesticides by drones with intelligent identification of farmland areas. Background Technology

[0002] In recent years, with the rapid development of modern agriculture, the application of automation technology in agricultural production has received widespread attention. In crop production, such as rice cultivation, pesticide spraying is a crucial field management step, and its operation directly impacts agricultural production efficiency and ecological environmental safety. Traditional pesticide spraying methods rely primarily on manual operation and uniform application. While this can meet basic pest control needs to some extent, its lack of adaptability to varying weed cover levels in different areas often leads to over-spraying. Over-spraying not only wastes pesticides but can also cause pesticide drift and soil runoff, polluting the surrounding environment and affecting normal crop growth, even leading to yield reduction.

[0003] With the rapid development of drone technology, drone-based pesticide spraying in rice paddies is gradually becoming an important direction for precision agriculture. Drone pesticide spraying not only enables automated operation but also significantly improves operational efficiency. However, most existing drone pesticide spraying operations still rely on uniform spraying, making it difficult to perform precise control based on spatial differences in weed coverage in rice paddies.

[0004] Therefore, improving the accuracy of pesticide spraying has become a pressing technical challenge in rice production. Summary of the Invention

[0005] This invention provides a method and system for precise pesticide spraying by drones based on intelligent identification of farmland areas, which solves the technical problem in the prior art of differentiated pesticide spraying based on the weed coverage in farmland.

[0006] On the one hand, this invention provides a method for precise pesticide spraying by drone using intelligent identification of farmland areas, comprising: Acquire image data of the target farmland and the drone flight parameters associated with the image data; Based on image data, weed distribution areas in the target farmland are identified, and a variable spraying prescription map corresponding to the farmland's geographical location is generated; the variable spraying prescription map includes application parameters for different weed distribution areas; Based on the variable spraying prescription map and the drone flight parameters, plan the drone's flight path; Based on the flight path, the drone is controlled to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map, and the spraying results are obtained.

[0007] Optionally, acquire image data of the target farmland and associated UAV flight parameters, including: The drone is controlled to fly along a preset trajectory, acquire initial images of the target farmland, and record the drone flight parameters corresponding to the initial images; the drone flight parameters include the drone's position coordinates and flight attitude data. The initial image is preprocessed and georegistered to generate orthophoto data of the land parcel with geographic coordinates, which serves as the image data.

[0008] Optionally, identifying the weed distribution area in the target farmland based on image data includes: Image data is input into a pre-trained deep learning semantic segmentation model to obtain pixel-level classification results; in the classification results, each pixel is classified as weeds, crops, or field ridges. For pixels classified as weeds in the classification results, perform connected component area analysis and select connected components with an area greater than a preset area threshold as weed distribution areas.

[0009] Optionally, generate a variable spraying prescription map corresponding to the geographical location of the farmland, including: Based on the preset grid resolution, the target farmland is divided into multiple grid units; Based on the classification results, the weed coverage in each grid unit is calculated. Based on the weed coverage of each grid cell, determine the target unit area effective ingredient application rate for that grid cell; Based on the preset minimum coverage spray volume constraint and maximum allowable pesticide concentration constraint, the effective ingredient application amount per unit area of ​​the target grid cell is calculated to obtain the target pesticide concentration and target spray volume of the grid cell. Based on the target pesticide concentration and target spray volume, a variable spraying prescription map corresponding to the geographical location of the farmland is generated.

[0010] Optionally, based on the weed cover of each grid cell, the target unit area application rate of the active ingredient for that grid cell is determined, including: Obtain a pre-set standard reference effective ingredient application rate; wherein, the standard reference effective ingredient application rate corresponds to a pre-set reference weed coverage range; Based on the preset correspondence between weed coverage intervals and mapping coefficients, the mapping coefficient corresponding to the coverage interval to which the weed coverage of the current grid cell belongs is determined; wherein, the mapping coefficient is a weight value used to convert the standard reference effective ingredient application rate into the target effective ingredient application rate per unit area. Multiply the mapping coefficient by the standard reference effective component application amount to obtain the target effective component application amount per unit area of ​​the current grid cell.

[0011] Optionally, based on preset minimum coverage spray volume constraints and maximum allowable pesticide concentration constraints, the effective ingredient application rate per unit area is calculated to obtain the target pesticide concentration and target spray volume for the grid cell, including: When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray rate is less than or equal to the maximum allowable pesticide concentration, the minimum coverage spray rate is taken as the target spray rate for that grid cell, and the ratio is taken as the target pesticide concentration for that grid cell. When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray amount is greater than the maximum allowable pesticide concentration, the maximum allowable pesticide concentration is taken as the target pesticide concentration for that grid cell, and the ratio of the effective ingredient application rate per unit area to the maximum allowable pesticide concentration is taken as the target spray amount for that grid cell.

[0012] Optionally, based on the variable spraying prescription map and the drone flight parameters, the drone's flight path is planned, including: Unify the variable spraying prescription map and the UAV flight parameters to the same spatial coordinate system; Under the same spatial coordinate system, the workable area, the restricted area and the restricted spraying buffer area are determined according to the variable spraying prescription map, which serve as the feasible domain constraints for path planning; With the goal of minimizing operational costs, a sequence of waypoints is generated as the initial flight path of the UAV based on feasible domain constraints and pre-acquired task constraints. Based on the variable spraying prescription map, variable spraying control commands corresponding to the spraying parameters are generated for each waypoint in the initial flight path to obtain the flight path.

[0013] Optionally, based on the flight path, the drone is controlled to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map, including: The drone's current location information is acquired in real time, and the current operating area is determined based on the current location information; Obtain the application parameters corresponding to the current work area from the variable spraying prescription map; the application parameters include the target pesticide concentration and the target spray volume; Control the spray concentration and spray flow rate based on the target pesticide concentration and target spray volume.

[0014] Optionally, controlling the spray flow rate also includes: Obtain the wind speed and direction in the area where the target farmland is located; Determine the crosswind component based on wind speed, wind direction, and the drone's flight direction; Determine the rotation compensation angle of the injection direction based on the crosswind component; Control the injection direction to rotate according to the rotation compensation angle, so that the injection direction is compensated relative to the oncoming wind direction; When the wind speed exceeds a preset threshold, a drift suppression strategy is implemented; wherein, the drift suppression strategy includes at least one of the following: reducing flight altitude, reducing flight speed, reducing spray flow rate, and suspending spraying.

[0015] Optionally, the rotation compensation angle of the injection direction is determined based on the crosswind component, including: Based on the crosswind component, the preset droplet settling velocity, and the drone's spraying height, the horizontal drift distance of the droplets during the flight time from the spray point to the crop canopy is determined; where the flight time is the ratio of the spraying height to the droplet settling velocity. Obtain the equivalent horizontal initial velocity of the jet; Based on the horizontal drift distance and the equivalent horizontal initial velocity, the rotational compensation angle used to counteract the crosswind component is calculated; wherein, the direction of the rotational compensation angle is to deflect the spray direction towards the upwind side of the incoming wind direction, so as to minimize the offset of the actual landing point of the liquid relative to the target working area.

[0016] On the other hand, the present invention also provides a drone-based precision pesticide spraying system for intelligent identification of farmland areas, comprising: An intelligent proportional pesticide spraying device, mounted on the body of a drone, includes a water chamber, a pesticide chamber, and a mixing chamber. The water chamber stores pure water, the pesticide chamber stores pesticides, and the mixing chamber forms a pesticide solution of the target concentration. The water chamber and the mixing chamber are connected by a first pipeline equipped with a first solenoid valve. The pesticide chamber and the mixing chamber are connected by a second pipeline equipped with a second solenoid valve. The spraying execution module is located below the drone body, including a rotatable nozzle and connected to the output end of the mixing chamber; The controller, mounted on the drone body, is communicatively connected to the first solenoid valve, the second solenoid valve, and the spraying execution module, and is used to execute the drone-based precise pesticide spraying method for intelligent identification of farmland areas as described above.

[0017] This invention provides a method and system for precise pesticide spraying using unmanned aerial vehicles (UAVs) based on intelligent identification of farmland areas. The method identifies weed distribution areas in a target farmland based on image data and generates a variable spraying prescription map corresponding to the farmland's geographical location. The variable spraying prescription map includes pesticide application parameters for different weed distribution areas. Based on the variable spraying prescription map and UAV flight parameters, the method plans the UAV's flight path. Based on the flight path, the UAV is controlled to perform differentiated pesticide spraying on different weed distribution areas according to the pesticide application parameters in the variable spraying prescription map, obtaining the spraying results. This method achieves precise variable spraying based on the spatial differences in weed distribution areas, avoiding over- or under-application of pesticides caused by uniform spraying methods, thus improving pesticide utilization efficiency and spraying accuracy.

[0018] Furthermore, by combining a deep learning semantic segmentation model with connected component area analysis, this invention achieves pixel-level identification and noise removal of weed distribution areas in farmland, avoiding the problems of difficulty in accurately distinguishing weeds from crops and susceptibility to isolated noise interference, thus improving the accuracy of weed distribution area identification.

[0019] Furthermore, this invention divides grid units based on weed coverage and introduces a mapping coefficient to dynamically convert the standard reference effective ingredient application rate into the target effective ingredient application rate per unit area. This avoids the problem that uniform spraying methods cannot adjust the dosage according to the spatial differences in weed coverage, and achieves precise matching between the application rate and the weed distribution density, thus avoiding pesticide waste or insufficient control.

[0020] Furthermore, under the constraints of minimum coverage spray volume and maximum allowable pesticide concentration, the present invention performs a collaborative calculation of the effective ingredient application amount per unit area to obtain the target pesticide concentration and target spray volume for each grid cell. This avoids the risk of phytotoxicity caused by excessive concentration in high coverage areas and the waste of resources caused by excessive spray volume in low coverage areas, thus ensuring the safety and economy of pesticide application.

[0021] Furthermore, this invention unifies the variable spraying prescription map and UAV flight parameters into the same spatial coordinate system, and generates a flight path that minimizes the operation cost based on the constraint-aware maximum-minimum ant colony algorithm. This avoids the problem that path planning does not fully consider the spatial differences in drug dosage and the resource constraints of UAVs, thereby improving operation efficiency and safety.

[0022] Furthermore, this invention obtains wind speed and direction in real time, calculates the crosswind component, and determines the rotation compensation angle of the spray direction. It actively counteracts crosswind drift by utilizing the horizontal velocity component generated by the nozzle rotation, thus avoiding the problem of reduced spraying accuracy caused by pesticide droplet drift under wind interference. This significantly improves the effective deposition rate of pesticides and reduces environmental pollution to non-target areas.

[0023] Furthermore, by setting up a water chamber, a pesticide chamber, and a mixing chamber, this invention achieves online dynamic adjustment of the target pesticide concentration, avoiding the problem that drone spraying systems cannot adjust the pesticide concentration and spray volume in real time according to the prescription map. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating the method for precise pesticide spraying by drones based on intelligent identification of farmland areas provided in this embodiment of the invention. Figure 2 This is a schematic diagram of the structure of the drone-based precision pesticide spraying system for intelligent identification of farmland areas provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the central control system layout according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the internal structure of the intelligent proportional pesticide water tank according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the external structure of the intelligent proportional pesticide water tank according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the anti-entanglement rotating spraying device according to an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0027] Figure 1 This is a flowchart illustrating the method for precise pesticide spraying by drones based on intelligent identification of farmland areas, provided in an embodiment of the present invention.

[0028] See Figure 1 The method for precise pesticide spraying by drones based on intelligent identification of farmland areas may include the following steps: Step 110: Obtain image data of the target farmland and UAV flight parameters associated with the image data.

[0029] Specifically, acquiring image data of the target farmland and associated UAV flight parameters may include: Step 1: Control the drone to fly along the preset trajectory, collect initial images of the target farmland, and record the drone flight parameters corresponding to the initial images; the drone flight parameters may include the drone's position coordinates and flight attitude data. Specifically, the preset trajectory can include flight altitude, flight speed, and forward and lateral overlap rates. The forward overlap rate refers to the proportion of overlap between two adjacent images along the flight direction, while the lateral overlap rate refers to the proportion of lateral overlap between images of adjacent flight paths. Initial images of the target farmland can be acquired using an RGB camera mounted below the UAV. The UAV's position coordinates can include longitude, latitude, and altitude. Flight attitude data can include heading angle, pitch angle, and roll angle. The initial images can include shooting time, exposure parameters, and focal length parameters.

[0030] Step 2: Preprocess and georegister the initial image to generate orthophoto data of the land parcel with geographic coordinates, which will serve as the image data. Specifically, the initial images can be quality-assessed to remove anomalous images. Anomalous images can include: blurred images due to inaccurate focus or motion blur, overexposed images with lost details due to overexposure, and underexposed images where shadow details are unrecognizable due to underexposure. Preprocessing can include distortion correction, brightness and color consistency correction, noise suppression, and contrast normalization. Georegistration refers to assigning geographic coordinates to the preprocessed images. Specifically, this can include: projecting each image into a unified geographic coordinate system based on the location and attitude information recorded by the UAV, and generating an orthophoto with geographic coordinates covering the entire target area through image stitching technology. Orthophotos are geometrically corrected images that eliminate perspective distortion caused by factors such as camera tilt and terrain undulations, giving the image a uniform scale and accurately reflecting the planar position and shape of features. The orthophoto can also be cropped according to the actual boundaries of the target farmland to remove irrelevant areas outside the plot, such as roads, ditches, and surrounding vegetation.

[0031] Step 120: Identify the weed distribution areas in the target farmland based on the image data, and generate a variable spraying prescription map corresponding to the geographical location of the farmland; wherein, the variable spraying prescription map may include application parameters for different weed distribution areas. It can be understood that the application parameters may include the spraying dosage and / or the concentration level of the pesticide solution.

[0032] Specifically, identifying the distribution area of ​​weeds in a target farmland based on image data can include: Step 1: Input the image data into a pre-trained deep learning semantic segmentation model to obtain pixel-level classification results; where each pixel in the classification results is classified as weeds, crops, or field ridges. Specifically, deep learning semantic segmentation models refer to deep neural network models that are trained with parameters using a large number of labeled samples. The training process may include: collecting a large number of rice paddy image samples, labeling weeds, crops (rice), and paddy ridges in the images at the pixel level, and then training the neural network using the labeled dataset to learn the mapping relationship between image features and pixel category labels. Commonly used semantic segmentation network architectures include U-Net, DeepLabv3+, and SegFormer.

[0033] Step 2: Perform connected component area analysis on the pixels classified as weeds in the classification results, and select connected components with an area greater than a preset area threshold as weed distribution areas.

[0034] Specifically, connected component area analysis (CBI) involves performing connectivity detection on pixels classified as "weeds" in pixel-level classification results. It combines adjacent (four-connected or eight-connected) "weed" pixels into independent connected regions and calculates the number of pixels or actual area contained in each region. Since the classification results output by deep learning models may contain isolated noise pixels (such as a single pixel misclassified as a weed), CBI can identify these noise points and remove them through area filtering, avoiding misclassification caused by scattered noise points.

[0035] Generating variable spraying prescription maps corresponding to the geographical location of farmland can include: Step 1: Divide the target farmland into multiple grid units based on the preset grid resolution; Specifically, the preset grid resolution can be set according to factors such as the drone's spray pattern and the spatial distribution of weeds.

[0036] Step 2: Based on the classification results, calculate the weed coverage in each grid unit; Specifically, weed coverage refers to the ratio of the number of weed pixels within a grid to the total number of pixels.

[0037] Step 3: Determine the target unit area effective ingredient application rate for each grid cell based on the weed coverage. Step 4: Based on the preset minimum coverage spray volume constraint and maximum allowable pesticide concentration constraint, calculate the effective ingredient application amount per unit area of ​​the target grid cell to obtain the target pesticide concentration and target spray volume. Step 5: Based on the target pesticide concentration and target spray volume, generate a variable spraying prescription map corresponding to the geographical location of the farmland.

[0038] Step 130: Based on the variable spraying prescription map and the drone flight parameters, plan the drone's flight path.

[0039] Specifically, planning the drone's flight path based on the variable spraying prescription map and drone flight parameters can include: Step 1: Unify the variable spraying prescription map and the UAV flight parameters to the same spatial coordinate system; As previously mentioned, the variable application prescription map can be a digital map corresponding to the geographical location of the field, recording the application parameters for each grid cell. Unifying it to the same spatial coordinate system ensures accurate spatial correspondence between the two. This same spatial coordinate system can be a projected coordinate system, preferably a UTM projected coordinate system.

[0040] Step 2: Under the same spatial coordinate system, determine the workable area, the restricted area, and the restricted spraying buffer area based on the variable spraying prescription map, which serve as the feasible domain constraints for path planning; The no-spray buffer zone is a safe transition area set up between the workable area and the restricted area. Within this area, drones can refrain from spraying or reduce the spraying dosage. The purpose of the no-spray buffer zone is to prevent pesticides from drifting into the restricted area, thereby avoiding contamination of non-target areas.

[0041] Step 3: With the goal of minimizing operational costs, generate a sequence of waypoints as the initial flight path for the UAV based on feasible domain constraints and pre-acquired task constraints; It is understandable that task constraints can generally include one or more of the following: battery power threshold, pesticide remaining amount threshold, maximum range, maximum number of waypoints, minimum waypoint spacing, maximum and minimum flight speed, minimum turning radius, and minimum safe distance. Minimizing operational costs can include one or more of the following: total flight distance, total operation time, energy consumption cost, turning cost, dosage switching cost, and near-obstacle / boundary risk cost. A waypoint sequence refers to a series of spatial points including longitude, latitude, and altitude.

[0042] Specifically, to meet the resource and safety constraints of UAV operations, this embodiment uses the constraint-aware max-min ant colony algorithm (MMAS) to optimize and generate flight paths.

[0043] (1) Modeling and Constraints: Discretize the workable area into a set of candidate waypoints. The path is represented as a sequence of waypoints. Any flight segment The length is The course changes to The speed of the flight segment is The task constraints can be written as feasibility conditions, and can include at least the following formula (1): (1); in, Minimum segment length; Minimum flight speed; Maximum flight speed; Minimum turning radius; Minimum safe distance; For obstacles; For the segment To the obstacle The distance. The number of waypoints is less than the maximum number of waypoints; the total distance is less than or equal to the maximum distance. Resource constraints: Assume energy consumption is estimated as... The equivalent flow rate of the spraying system at the calculated dosage is Then, the following formula (2) is shown: (2); in , For flight segment time; This represents the upper limit of usable energy corresponding to the battery capacity threshold. This represents the upper limit of the amount of pesticide that can be used, corresponding to the pesticide remaining threshold.

[0044] (2) Objective function: Construct sub-objectives It may include at least one or more of the following: total flight distance, total operation time, energy consumption cost, turning cost, dose switching cost, and near obstacle / boundary risk cost, and normalize them as shown in the following formula (3): (3); The overall cost function is defined as follows (4): (4); Where m is the index number of different targets. The original value of the target; The minimum value of the objective; The maximum value of the target; It should be a very small positive number to prevent the denominator from being zero; This is the normalized value of the m-th objective. The comprehensive cost function; The weights after normalization for the m-th objective; Penalty is the penalty coefficient. To constrain the penalty term, if a path violates any constraint, a larger penalty is imposed to suppress it during optimization.

[0045] (3) Transfer rules: Let the pheromone be The heuristic factor is For any ant at a waypoint The optional set is defined as: ;in, Let j be the set of waypoints that can be selected as the next waypoint j at waypoint i; Candidate waypoints; For the set of candidate waypoints; The transition probability is shown in formula (5) below: (5); in, For the transition probability; and These are their respective weighting coefficients; for Any waypoint in the; The heuristic factor is preferably associated with short distances, few turns, few dose switching, and distance from obstacles / boundaries, as shown in the following formula (6): (6); in Risk items related to proximity to obstacles / boundaries; The dose level for waypoint j; The dose setting for waypoint i; , , These are the weighting coefficients for each type of cost.

[0046] (4) Pheromones Update (MMAS): After each iteration, pheromone evaporation and enhancement are performed, as shown in the following formula (7): (7); Only for the contemporary best feasible path The flight segment is enhanced as shown in the following formula (8): (8); Restrict pheromones to To avoid premature convergence. For the updated pheromones; The pheromone efficaciousness coefficient; For pheromone increment; This is the optimal feasible path in the present era; The pheromone enhancement coefficient; The comprehensive cost of the optimal feasible path; This represents the minimum pheromone level. This represents the maximum pheromone value. When the maximum number of iterations is reached or the overall cost converges, the feasible waypoint sequence with the minimum overall cost is output as the UAV flight path, and variable spraying control commands corresponding to the waypoint sequence are generated.

[0047] Step 4: Based on the variable spraying prescription map, generate variable spraying control commands corresponding to the spraying parameters for each waypoint in the initial flight path to obtain the flight path.

[0048] In this step, the flight path satisfies both spatial constraints and resource security constraints, with the optimization goal of minimizing operational costs, thereby improving the safety, economy, and efficiency of UAV operations.

[0049] Furthermore, based on the variable application prescription map, variable application control commands corresponding to the application parameters are generated for each waypoint in the initial flight path, which may include: The variable application prescription map and the initial flight path are in the same spatial coordinate system. The variable application prescription map is rasterized in the same spatial coordinate system to obtain prescription raster cells, and a prescription vector element layer is established. The position of the element in the prescription vector element layer corresponds to the center point of the prescription raster cell, and a drug application parameter field is set for each element. The initial flight path is sampled or interpolated at equal intervals in the same spatial coordinate system to obtain several path sampling points; For each path sampling point, spatial matching is performed in the prescription vector feature layer to determine the prescription feature corresponding to the path sampling point, and the medication parameter field of the prescription feature is assigned to the corresponding path sampling point to obtain several fusion points. Generate waypoint sequences and corresponding variable spraying control commands based on fusion points.

[0050] Step 140: Flight path. Control the drone to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map, and obtain the spraying results.

[0051] In this step, after obtaining the spraying results, a spraying report can be generated based on the results. The spraying report may include the sprayed area, spraying dosage, flight distance, etc.

[0052] In this embodiment, weed distribution areas in the target farmland are identified based on image data, and a variable spraying prescription map corresponding to the farmland's geographical location is generated. The variable spraying prescription map includes application parameters for different weed distribution areas. Based on the variable spraying prescription map and UAV flight parameters, the UAV's flight path is planned. Based on the flight path, the UAV is controlled to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map, obtaining the spraying results. This achieves precise variable spraying based on the spatial differences in weed distribution areas, avoiding over- or under-application of pesticides caused by uniform spraying methods, and improving pesticide utilization efficiency and spraying accuracy.

[0053] In one embodiment of this specification, determining the target unit area application rate of the active ingredient corresponding to each grid cell based on the weed coverage of that grid cell may include: Step 1: Obtain the pre-set standard reference effective ingredient application rate; wherein, the standard reference effective ingredient application rate corresponds to the pre-set reference weed coverage range; Specifically, the standard reference active ingredient application rate refers to the standard value of pesticide active ingredient application per unit area for a reference weed coverage range (e.g., a medium coverage range with weed coverage of 10%-20%). This standard reference value is a benchmark application rate determined based on pesticide use specifications, weed control test data, and agricultural production practice experience.

[0054] Step 2: Based on the preset correspondence between weed coverage intervals and mapping coefficients, determine the mapping coefficient corresponding to the coverage interval to which the weed coverage of the current grid cell belongs; wherein, the mapping coefficient is a weight value used to convert the standard reference effective ingredient application rate into the target effective ingredient application rate per unit area. Specifically, for example, very low coverage (less than 1%) corresponds to 0 or 0.3; low coverage (greater than or equal to 1%, less than 5%) corresponds to 0.5; relatively low coverage (greater than or equal to 5%, less than 10%) corresponds to 0.75; medium coverage (greater than or equal to 10%, less than 20%) corresponds to 1.0; relatively high coverage (greater than or equal to 20%, less than 35%) corresponds to 1.2; and high coverage (greater than or equal to 35%) corresponds to 1.35. When it is high coverage, the maximum allowable amount of active ingredient can be compared with the target amount of active ingredient per unit area determined according to the mapping coefficient, and then the minimum value of the two can be taken.

[0055] Step 3: Multiply the mapping coefficient by the standard reference effective component application rate to obtain the target effective component application rate per unit area of ​​the current grid cell. This is shown in formula (9) below: (9); in, These are the mapping coefficients; This is the standard reference dosage of the active ingredient. This refers to the index number of the grid cell.

[0056] Specifically, the target unit area effective ingredient application rate refers to the amount of pesticide effective ingredient that needs to be applied to the current grid cell to achieve effective weed control.

[0057] In this embodiment, a quantitative mapping between weed coverage and the target effective ingredient application rate per unit area is realized, so that the target effective ingredient application rate per unit area can be adjusted according to the change of weed coverage. This avoids pesticide waste or insufficient control caused by using a uniform application rate due to ignoring the differences in weed coverage, and improves the scientificity and accuracy of determining the application rate.

[0058] In one embodiment of this specification, the application rate of the effective ingredient per unit area is calculated based on preset minimum coverage spray volume constraints and maximum allowable pesticide concentration constraints to obtain the target pesticide concentration and target spray volume for the grid cell. This may include: Step 1: When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray rate is less than or equal to the maximum allowable pesticide concentration, the minimum coverage spray rate is taken as the target spray rate for that grid cell, and the ratio is taken as the target pesticide concentration for that grid cell. Step 2: When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray amount is greater than the maximum allowable pesticide concentration, the maximum allowable pesticide concentration is taken as the target pesticide concentration for that grid cell, and the ratio of the effective ingredient application rate per unit area to the maximum allowable pesticide concentration is taken as the target spray amount for that grid cell.

[0059] The target application rate of active ingredient per unit area can be expressed by the following formula (10): (10); in, The effective ingredient dosage per unit area; This refers to the concentration of the drug solution in that area. This refers to the amount of liquid sprayed per unit area in this region; This refers to the index number of the grid cell.

[0060] Specifically, define the minimum coverage spray volume. Maximum permissible concentration of drug solution For each grid: The target concentration is shown in the following formula (11): (11); The target spray volume is shown in the following formula (12): (12).

[0061] First scenario: Low coverage: if This indicates that the minimum spray volume is maintained. In this case, simply increasing the concentration is sufficient. .

[0062] The second scenario: very high coverage. if ; This means that if you still insist on spraying That would require adjusting the concentration too high, which is unsafe. At this point, .

[0063] The dosage results are written into the prescription data structure to form a variable spraying prescription map corresponding to the geographical location of the field. The prescription map can be a raster prescription map and / or a vector prescription map, and a dosage field is set for each grid cell or feature to identify the dosage level.

[0064] In this embodiment, the synergistic calculation of the effective ingredient application rate per unit area is achieved under the dual constraints of minimum coverage spray volume and maximum allowable pesticide concentration. This ensures that in areas with low or medium weed coverage, the application requirements are met by increasing the pesticide concentration, avoiding unnecessary increases in spray volume. Conversely, in areas with high weed coverage, the risk of soil and crop damage due to excessively high pesticide concentrations is avoided by capping the concentration and increasing the spray volume. Table 2 below shows the relationship between weed coverage and pesticide application.

[0065] Table 2 Relationship between weed coverage and pesticide application In one embodiment of this specification, controlling a drone to perform differentiated pesticide spraying on different weed distribution areas based on the flight path and according to the application parameters in the variable spraying prescription map may include: Step 1: Obtain the current location information of the drone in real time, and determine the current working area based on the current location information; Step 2: Obtain the application parameters corresponding to the current work area from the variable spraying prescription map; the application parameters may include the target pesticide concentration and the target spray volume. Step 3: Control the spray concentration and spray flow rate according to the target pesticide concentration and target spray volume.

[0066] In this embodiment, the variable spraying prescription map records the application parameters for different areas, which are generally the application parameters corresponding to each grid cell. This embodiment enables the UAV to identify its current work area in real time during flight and accurately match it with the application parameters on the prescription map, thereby improving the response speed and execution accuracy of pesticide spraying.

[0067] In one embodiment of this specification, controlling the spray flow rate may further include: Step 1: Obtain the wind speed and direction in the area where the target farmland is located; Step 2: Determine the crosswind component based on wind speed, wind direction, and the drone's flight direction; The crosswind component refers to the component of the wind speed vector perpendicular to the direction of the UAV's flight.

[0068] Step 3: Determine the rotation compensation angle of the spray direction based on the crosswind component; Step 4: Control the spray direction to rotate according to the rotation compensation angle, so that the spray direction is compensated relative to the oncoming wind direction; Specifically, the spray direction compensates for the wind direction. This involves rotating the pesticide nozzle to direct the spray towards the wind direction (headwind), thus giving the pesticide an initial velocity against the wind and counteracting the drift caused by crosswinds. For example, when a crosswind blows from the left, rotating the nozzle to the left directs the spray towards the left (headwind), causing the pesticide to spray to the left and gain an initial velocity to the left, canceling out the rightward crosswind. Conversely, when a crosswind blows from the right, rotating the nozzle to the right directs the spray towards the right (headwind), causing the pesticide to spray to the right and gain an initial velocity to the right, canceling out the leftward crosswind. This compensation reduces the lateral drift distance of the droplets in the air.

[0069] Step 5: When the wind speed exceeds the preset threshold, execute the drift suppression strategy; wherein, the drift suppression strategy may include at least one of the following: reducing flight altitude, reducing flight speed, reducing spray flow rate, and suspending spraying.

[0070] In this embodiment, the nozzle direction is adaptively adjusted based on real-time wind field information. The horizontal velocity component generated by the nozzle rotation actively counteracts crosswind drift, reducing the lateral drift distance of pesticide droplets in the air, improving the effective deposition rate of pesticides, and reducing environmental pollution to surrounding non-target areas.

[0071] In one embodiment of this specification, determining the rotation compensation angle of the injection direction based on the crosswind component may include: Step 1: Based on the crosswind component, the preset droplet settling velocity, and the drone's spraying height, determine the horizontal drift distance of the droplets during the flight time from the spray point to the crop canopy; where the flight time is the ratio of the spraying height to the droplet settling velocity. Specifically, horizontal drift distance refers to the lateral displacement of the droplets during their flight time under the influence of the crosswind component after they are ejected.

[0072] Step 2: Obtain the equivalent horizontal initial velocity of the jet; The equivalent horizontal initial velocity is the initial velocity component in the horizontal direction when the droplet is ejected from the nozzle.

[0073] Step 3: Calculate the rotational compensation angle to counteract the crosswind component based on the horizontal drift distance and the equivalent horizontal initial velocity; wherein, the direction of the rotational compensation angle is to make the spray direction deflect to the upwind side of the incoming wind direction, so as to minimize the offset of the actual landing point of the liquid relative to the target working area.

[0074] In this embodiment, a rotation compensation angle is calculated to counteract the crosswind component based on the horizontal drift distance and the equivalent horizontal initial velocity, so as to minimize the offset of the actual landing point of the droplets relative to the target working area, thereby minimizing environmental pollution to the surrounding non-target areas while ensuring the effective deposition rate of pesticides.

[0075] Prescription pathway fusion and variable spray sequence generation: Unify the variable application prescription map and flight path into the same projected coordinate system (preferably UTM), rasterize the prescription map and generate prescription point elements at the center of the raster; sample the flight path at preset intervals to obtain path sampling points, perform nearest neighbor search or spatial matching based on search radius for each path sampling point to obtain fusion points, and assign the prescription dose field to the corresponding path point to form a variable application sequence.

[0076] The variable spraying sequence is compensated and corrected by combining flight parameters such as flight speed, flight altitude and flight direction. For example, the dose switching point is advanced / delayed along the flight direction based on the response delay of the spraying system.

[0077] Spray control: When the drone flies over a specific work area, the Jetson Nano 101 reads the target concentration for that area in the variable spray prescription map. ; Let the current volume of the mixing chamber 301 be... The current concentration is The concentration of pesticide solution in the high-concentration pesticide chamber 307 is: The first pilot-operated diaphragm solenoid valve 314 controls the pure water flow rate as follows: The second pilot-operated diaphragm solenoid valve 315 controls the flow rate of high-concentration pesticides. The pure water valve is open for a duration of time. The high-concentration pesticide valve opening time is The volume of pure water added is shown in formula (13): (13); The volume of high-concentration pesticide added is shown in the following formula (14): (14); Since the concentration of pure water is approximately zero, the concentration in the adjusted mixing chamber... Satisfy the following formula (15): (15); Jetson Nano 101 based on target concentration Current concentration Current volume and preset traffic and Determine the opening time of the first pilot-operated diaphragm solenoid valve 314 and the second pilot-operated diaphragm solenoid valve 315. and This allows for the adjustment of the target concentration.

[0078] When it is necessary to increase the spraying intensity, the opening time of the second pilot-operated diaphragm solenoid valve 315 is increased, so that more high-concentration pesticide solution is input from the high-concentration pesticide chamber 307 to the mixing chamber 301; when it is necessary to reduce the spraying intensity, the opening time of the first pilot-operated diaphragm solenoid valve 314 is increased, so that more pure water is input from the water chamber 306 to the mixing chamber 301.

[0079] After the concentration is adjusted, the stirring motor 308 drives the stirring rod 303 and the stirring rod impeller 304 to rotate, quickly mixing the liquid in the mixing chamber 301. After mixing, the small water pump 320 starts, delivering the target concentration of the drug solution to the slip ring stator end 326 and the slip ring rotor end 327, and spraying it out through the mist nozzle 318.

[0080] The control commands are processed with time delay compensation and smoothing to compensate for the response delay of the spraying system and limit frequent dose switching; the control spraying actuator completes the spraying according to the control commands and records the spraying log data (spraying time, location coordinates, dose, spraying flow rate, etc.).

[0081] Wind field acquisition and filtering: During the spraying process, wind speed and direction information are acquired in real time, and a wind speed vector is established under the geographic coordinate system (east-north), as shown in the following formula (16): (16); in, Sampling time; No. Wind speed vector at each sampling time; This represents the eastward wind speed vector; This represents the northward wind speed vector; To reduce control jitter caused by instantaneous gusts, the wind speed vector is low-pass filtered as shown in the following formula (17): (17); in, For the first Filtered wind speed vector at each sampling time; For the first Filtered wind speed vector at each sampling time; Filter coefficients; calculate wind speed scalar. . For wind speed scalar; Let be the vector magnitude.

[0082] Relative wind calculation and crosswind component extraction: Let the ground velocity vector of the UAV be The relative wind vector is then shown in formula (18): (18); in, The speed of the drone relative to the air; This is the ground velocity vector of the UAV; This represents the ground velocity component of the UAV in the east-west direction; This represents the ground velocity component of the UAV in the north-south direction; Let the drone's heading angle be The heading unit vector and the normal unit vector are shown in formulas (19) and (20) respectively: (19); (20); The components of the relative wind in the heading direction and the crosswind direction are shown in the following formula (21): (twenty one); in The symbol is used to indicate that the crosswind is blowing from the left or right of the heading. This is a unit vector pointing in the direction of the drone's flight, with a length of 1; It is a normal vector and a unit vector; The relative wind component in the heading direction represents the projection of the relative wind onto the UAV's flight path; a positive value indicates a tailwind, and a negative value indicates a headwind. The crosswind component represents the projection of the relative wind onto the direction perpendicular to the heading. A positive value indicates that the wind is blowing from the left, and a negative value indicates that the wind is blowing from the right.

[0083] Droplet flight time and drift distance estimation: Let the spraying height (relative to the crop canopy height) be h, and the equivalent settling velocity of the droplets be... The flight time of the droplets from the nozzle to the canopy. The estimate is shown in the following formula (22): (twenty two); Corresponding lateral drift distance The estimate is shown in the following formula (23): (twenty three); The preferred spraying height h is... Equivalent settlement velocity The droplet size can be preset or determined by referring to a table, preferably... (For larger particles, use a higher settling velocity; for smaller particles, use a lower settling velocity).

[0084] Calculation and execution of nozzle rotation compensation angle: Let the equivalent initial velocity of the nozzle ejected onto the horizontal plane (formed by nozzle pressure, downwash flow field, etc.) be... (unit The nozzle rotation angle is (Rotating about the vertical axis), the upwind horizontal velocity component generated by the nozzle rotation can be used to counteract crosswind drift. Preferably, if the compensation objective is to make the droplet lateral drift approach zero, then formula (24) is satisfied: (twenty four); Therefore, the nozzle rotation angle command is obtained (with upwind side compensation), as shown in formula (25): (25); This is the nozzle rotation angle command. The maximum allowable rotation angle for the nozzle mechanism; when When using saturation limiting To avoid high-frequency vibration, the nozzle actuation angle... First-order tracking is used, as shown in formula (26): (26); in, The angular velocity of the nozzle rotation; To track gain, Map the angle error to Equivalent initial velocity This can be obtained through ground calibration or system parameter settings, preferably... .

[0085] In some implementations, formula (27) may incorporate a compensation coefficient. To prevent overcompensation: (27).

[0086] Wind speed threshold linkage: when When this occurs, a drift suppression strategy is triggered, which includes at least one or more of the following: reducing flight altitude, reducing flight speed, reducing spray flow rate, and / or suspending spraying; when If necessary, spraying should be suspended or operations interrupted and the crew returned to base until the wind speed returns to within the threshold range before resuming operations. This is the wind speed stopping threshold. This is the wind speed warning threshold.

[0087] In some other embodiments of the present invention, the method further includes: Obstacle detection sensors installed on the drone are used to detect obstacles near the drone's flight path in real time; the obstacle detection sensors include one or more of lidar, visual sensors, and ultrasonic sensors; Based on the detection results, obtain the spatial location and size information of the obstacle, and calculate the relative distance and / or relative height difference between the obstacle and the drone; When the relative distance and / or relative height difference does not meet the preset safety threshold, the obstacle avoidance condition is determined to be triggered. When obstacle avoidance conditions are triggered, the drone is controlled to perform obstacle avoidance actions. Obstacle avoidance actions include adjusting flight altitude, changing flight speed and / or changing flight heading to avoid obstacles, and locally replanning the flight path or generating temporary detour waypoints. During obstacle avoidance, the variable spraying control instructions are synchronously corrected according to the variable spraying prescription map. Synchronous corrections include pausing spraying, reducing the spraying dose, or resuming the corresponding dose spraying after returning to the original path.

[0088] Based on the same concept, this invention also provides a drone-based precision pesticide spraying system with intelligent identification of farmland areas. The drone-based precision pesticide spraying system with intelligent identification of farmland areas provided by this invention will be described below. The drone-based precision pesticide spraying system with intelligent identification of farmland areas described below can be referred to in correspondence with the drone-based precision pesticide spraying method with intelligent identification of farmland areas described above.

[0089] A drone-based precision pesticide spraying system with intelligent identification of farmland areas includes: Intelligent proportional pesticide spraying module and spraying execution module installed on the drone body; An intelligent proportional pesticide spraying device, mounted on the body of a drone, includes a water chamber, a high-concentration pesticide chamber, and a mixing chamber. The water chamber stores pure water, the high-concentration pesticide chamber stores high-concentration pesticide, and the mixing chamber forms the pesticide solution of the target concentration. The water chamber and the mixing chamber are connected by a first pipeline equipped with a first solenoid valve. The high-concentration pesticide chamber and the mixing chamber are connected by a second pipeline equipped with a second solenoid valve. The spraying execution module is located below the drone body, including a nozzle, and is connected to the output end of the mixing chamber; the nozzle can rotate at different angles; The controller, mounted on the drone body, is communicatively connected to the first solenoid valve, the second solenoid valve, and the spraying execution module, and is used to execute any of the above-mentioned drone-based precision pesticide spraying methods based on regional intelligent recognition.

[0090] like Figures 2 to 6As shown, the system in this embodiment consists of a controller 1, a drone 2, an intelligent proportional pesticide spraying device 3, and an RGB camera 4. The drone 2 serves as the flight platform, and its structure is not limited in this embodiment. The controller 1 is mounted on the drone 2 and is used to perform image recognition, positioning calculation, path planning, concentration calculation, and spraying control. The RGB camera 4 is mounted below the drone 2 and is used to collect image data of the work area. The intelligent proportional pesticide spraying device 3 is mounted below the drone 2 and is used to dynamically mix and spray pesticides according to the control commands from the controller 1.

[0091] like Figure 4 As shown, controller 1 includes a Jetson Nano 101, a UM982 module 102, a battery 103, and an anemometer 104. The Jetson Nano 101 is used to run deep learning recognition models, variable spraying prescription map generation algorithms, path planning algorithms, and proportional mixing control algorithms; the UM982 module 102 is used to output the real-time position coordinates of the UAV; the anemometer 104 is used to measure the real-time wind speed and direction in the operating area; and the battery 103 is used to power controller 1.

[0092] like Figure 4 and Figure 5 As shown, the intelligent proportional pesticide spraying device 3 includes a mixing chamber 301, a stirring rod bearing 302, a stirring rod 303, a stirring rod impeller 304, a coupling 305, a water chamber 306, a high-concentration pesticide chamber 307, a stirring motor 308, a water pipe connector for the mixing chamber 309, a first water pipe connector 310, a second water pipe connector 311, a sealing gasket 312, a threaded connector 313, a first pilot-operated diaphragm solenoid valve 314, a second pilot-operated diaphragm solenoid valve 315, a second pipeline 316, a first pipeline 317, a small water pump 320, a mounting lug 321, and a water tank bottom plate 322.

[0093] The system comprises three chambers: a water chamber 306 for storing pure water, a high-concentration pesticide chamber 307 for storing high-concentration pesticide solution, and a mixing chamber 301 for forming a pesticide solution of the target concentration. A first pilot-operated diaphragm solenoid valve 314 is positioned between the water chamber 306 and the mixing chamber 301 to control the entry of pure water into the mixing chamber 301; a second pilot-operated diaphragm solenoid valve 315 is positioned between the high-concentration pesticide chamber 307 and the mixing chamber 301 to control the entry of high-concentration pesticide into the mixing chamber 301. A threaded connector 313 and a sealing gasket 312 are used to achieve a sealed connection between the solenoid valves and the chambers, preventing liquid leakage.

[0094] A stirring motor 308 drives a stirring rod 303 to rotate via a coupling 305. A stirring rod impeller 304 is mounted at the lower end of the stirring rod 303. The stirring rod 303 is supported in the mixing chamber 301 by a stirring rod bearing 302, and is used to quickly mix the drug solution after concentration adjustment. One or more small water pumps 320 are connected downstream of the mixing chamber 301, preferably four small water pumps 320. The small water pumps 320 transport the mixed drug solution to the spraying execution module through the water pipe connector 309 of the mixing chamber.

[0095] like Figure 6 As shown, the anti-entanglement rotary spraying device includes a mist nozzle 318, a slip ring stator end water pipe connector 319, a slip ring rotor end drive motor 323, a rotor drive input shaft 324, a slip ring connector 325, a slip ring stator end 326, and a slip ring rotor end 327.

[0096] The liquid medicine output from the small water pump 320 is transported through the mixing chamber water pipe connector 309 to the slip ring stator end water pipe connector 319, and then enters the slip ring stator end 326. The liquid medicine is transported through the slip ring connector 325 to the slip ring rotor end 327 and finally to the mist nozzle 318. The slip ring rotor end drive motor 323 drives the rotor drive input shaft 324 to rotate, which in turn drives the slip ring rotor end 327 and the mist nozzle 318 to rotate and spray. Because the liquid is transmitted through the slip ring stator end 326 and the slip ring rotor end 327, even if the mist nozzle 318 rotates continuously, it will not cause the external infusion pipeline to become entangled.

[0097] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units.

[0098] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for precise pesticide spraying using unmanned aerial vehicles (UAVs) with intelligent identification of farmland areas, characterized in that, include: Acquire image data of the target farmland and UAV flight parameters associated with the image data; The image data is used to identify the weed distribution areas in the target farmland and generate a variable spraying prescription map corresponding to the geographical location of the farmland; wherein, the variable spraying prescription map includes the application parameters for different weed distribution areas; Based on the variable spraying prescription map and the UAV flight parameters, plan the flight path of the UAV; Based on the flight path, the drone is controlled to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map, and the spraying results are obtained.

2. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 1, characterized in that, The step of identifying the weed distribution area in the target farmland based on the image data includes: The image data is input into a pre-trained deep learning semantic segmentation model to obtain pixel-level classification results; wherein, each pixel in the classification results is classified as weeds, crops, or field ridges. For pixels classified as weeds in the classification results, connected component area analysis is performed, and connected components with an area greater than a preset area threshold are selected as weed distribution areas.

3. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 2, characterized in that, The generation of variable spraying prescription maps corresponding to the geographical location of farmland includes: Based on the preset grid resolution, the target farmland is divided into multiple grid units; Based on the classification results, the weed coverage in each grid cell is calculated. Based on the weed coverage of each grid cell, determine the target unit area effective ingredient application rate for that grid cell; Based on the preset minimum coverage spray volume constraint and maximum allowable pesticide concentration constraint, the effective ingredient application amount per unit area of ​​the target grid cell is calculated to obtain the target pesticide concentration and target spray volume of the grid cell. Based on the target pesticide concentration and the target spray volume, a variable spraying prescription map corresponding to the geographical location of the farmland is generated.

4. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) with intelligent identification of farmland areas according to claim 3, characterized in that, The step of determining the target unit area effective ingredient application rate for each grid cell based on the weed coverage includes: Obtain a pre-set standard reference effective ingredient application rate; wherein the standard reference effective ingredient application rate corresponds to a pre-set reference weed coverage range; Based on the preset correspondence between weed coverage intervals and mapping coefficients, the mapping coefficient corresponding to the coverage interval to which the weed coverage of the current grid cell belongs is determined; wherein, the mapping coefficient is a weight value used to convert the standard reference effective ingredient application rate into the target effective ingredient application rate per unit area. Multiplying the mapping coefficient by the standard reference effective ingredient application amount yields the target effective ingredient application amount per unit area for the current grid cell.

5. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 3, characterized in that, The step of calculating the effective ingredient application rate per unit area based on preset minimum coverage spray volume constraints and maximum allowable pesticide concentration constraints to obtain the target pesticide concentration and target spray volume for the grid cell includes: When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray rate is less than or equal to the maximum allowable pesticide concentration, the minimum coverage spray rate is taken as the target spray rate for that grid cell, and the ratio is taken as the target pesticide concentration for that grid cell. When the ratio of the effective ingredient application rate per unit area to the minimum coverage spray amount is greater than the maximum allowable pesticide concentration, the maximum allowable pesticide concentration is taken as the target pesticide concentration for that grid cell, and the ratio of the effective ingredient application rate per unit area to the maximum allowable pesticide concentration is taken as the target spray amount for that grid cell.

6. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 1, characterized in that, The step of planning the flight path of the UAV based on the variable spraying prescription map and the UAV flight parameters includes: Unify the variable spraying prescription map and the UAV flight parameters to the same spatial coordinate system; Under the same spatial coordinate system, the workable area, the restricted area and the restricted spraying buffer area are determined according to the variable spraying prescription map, which serve as the feasible domain constraints for path planning; With the goal of minimizing operational costs, a sequence of waypoints is generated as the initial flight path of the UAV based on the feasible domain constraints and pre-acquired task constraints. Based on the variable spraying prescription map, variable spraying control commands corresponding to the application parameters are generated for each waypoint in the initial flight path to obtain the flight path.

7. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 1, characterized in that, Based on the flight path, controlling the drone to perform differentiated pesticide spraying on different weed distribution areas according to the application parameters in the variable spraying prescription map includes: The current location information of the drone is acquired in real time, and the current working area is determined based on the current location information; Obtain the application parameters corresponding to the current work area in the variable spraying prescription map; wherein, the application parameters include the target pesticide concentration and the target spray volume; The spray concentration and spray flow rate are controlled based on the target drug concentration and the target spray volume.

8. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) with intelligent identification of farmland areas according to claim 7, characterized in that, The control of the spray flow rate also includes: Obtain the wind speed and direction in the area where the target farmland is located; The crosswind component is determined based on the wind speed, wind direction, and the flight direction of the UAV; Based on the crosswind component, determine the rotation compensation angle of the injection direction; The injection direction is controlled to rotate according to the rotation compensation angle, so that the injection direction is compensated relative to the oncoming wind direction; When the wind speed exceeds a preset threshold, a drift suppression strategy is executed; wherein, the drift suppression strategy includes at least one of reducing flight altitude, reducing flight speed, reducing spray flow rate, and suspending spraying.

9. The method for precise pesticide spraying by unmanned aerial vehicle (UAV) based on intelligent identification of farmland areas according to claim 8, characterized in that, The step of determining the rotational compensation angle of the injection direction based on the crosswind component includes: Based on the crosswind component, the preset droplet settling velocity, and the spraying height of the UAV, the horizontal drift distance of the droplets during the flight time from the spray point to the crop canopy is determined; wherein, the flight time is the ratio of the spraying height to the droplet settling velocity. Obtain the equivalent horizontal initial velocity of the jet; Based on the horizontal drift distance and the equivalent horizontal initial velocity, a rotational compensation angle is calculated to counteract the crosswind component; wherein the direction of the rotational compensation angle is such that the spray direction is biased towards the upwind side of the incoming wind direction, so as to minimize the offset of the actual landing point of the liquid relative to the target working area.

10. A drone-based precision pesticide spraying system for intelligent identification of farmland areas, characterized in that, include: An intelligent proportional pesticide spraying device, mounted on the body of a drone, includes a water chamber, a pesticide chamber, and a mixing chamber. The water chamber stores pure water, the pesticide chamber stores pesticides, and the mixing chamber forms a pesticide solution of the target concentration. The water chamber and the mixing chamber are connected by a first pipeline equipped with a first solenoid valve. The pesticide chamber and the mixing chamber are connected by a second pipeline equipped with a second solenoid valve. The spraying execution module is located below the drone body, including a nozzle, and is connected to the output end of the mixing chamber; The controller, mounted on the drone body, is communicatively connected to the first solenoid valve, the second solenoid valve, and the spraying execution module, and is used to execute the drone-based precise pesticide spraying method for intelligent identification of farmland areas as described in any one of claims 1 to 9.

Images