A drone spraying method

Abstract

The present invention relates to a drone spraying method, which comprises the following steps: firstly, a drone is used to collect images of farmland after transplanting or cultivation; a plant deep learning target detection algorithm model is constructed to identify and locate plants; a spray head real-time movement path planning algorithm model is constructed; the field images taken by the drone are processed using the plant deep learning target detection algorithm model, and the processed field images are further processed using the spray head real-time movement path planning algorithm model to obtain a spray head real-time movement path planning task command; the spray head real-time movement path planning task command is uploaded to the drone to perform field task spraying operations; spray heads and a spray drive mechanism for driving the spray heads to slide left and right are provided on both sides of the drone, and after receiving the spray head real-time movement path planning task command, the control system in the drone drives the spray heads on the left and right sides to move through the spray drive mechanism to adjust the positions of the spray heads in real time.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicles (UAVs), and in particular to a UAV spraying method. Background Art

[0002] The emergence of agricultural drones has significantly improved efficiency in agricultural production. Currently, these drones primarily perform tasks such as spraying pesticides and fertilizing, saving significant manpower and reducing operation time. However, current mainstream drone applications do not fully achieve the desired results, and the design of the spraying systems can lead to waste of pesticides and fertilizers. Therefore, a new, automated, efficient, and accurate drone spraying method is urgently needed.

[0003] Among them, regarding the drone spraying method, for example, the invention patent with the authorization announcement number CN115649450A discloses "a spraying drone with an adjustable spraying range", and the spraying drone includes a drone body, a medicine box and a plurality of nozzles are installed on the drone body, the medicine box and the nozzles are connected by a liquid pump, and all the nozzles on the drone body form an inner ring nozzle array and an outer ring nozzle array. A switching mechanism is also installed on the drone body, and the switching mechanism enables only one of the inner ring array and the outer ring nozzle array to perform the spraying operation. By setting two inner and outer ring nozzle arrays and switching different nozzle array operations through the switching mechanism, the spraying range of the nozzle can be effectively changed, and since there is no need to set a complex connecting rod structure, the spraying drone is lightweight as a whole and has low cost.

[0004] For example, the utility model patent with authorization announcement number CN217918400U discloses "a spraying drone with adjustable spraying range". A pair of electric telescopic rods are connected to the bottom of the drone body, and a nozzle is connected to the end of each telescopic section of the electric telescopic rod. The nozzle is connected to a water supply pipe, the water supply pipe is connected to the water outlet pipe of the water pump, and the water inlet pipe of the water pump is connected to the bottom of the medicine box. This allows the spraying drone to adjust its spraying range, has relative flexibility, and can adapt to more terrains.

[0005] However, the above-mentioned spraying drones and spraying drones with adjustable spraying range have the following shortcomings:

[0006] (1) The above-mentioned spraying drone is designed with two groups of nozzle arrays. By setting a switching mechanism to switch between different groups of nozzle arrays, a certain degree of spraying range adjustment can be achieved. Compared with traditional fixed nozzle type drones, the spraying range of the spraying drone is adjustable. However, since the relative spraying ranges of the two groups of nozzle arrays are still limited, the field adaptability is not strong and precise spraying cannot be achieved in practice.

[0007] (2) The above-mentioned spraying drone with adjustable spraying range has a larger spraying range by designing an adjustable telescopic rod. As the telescopic section of the electric telescopic rod is extended or retracted, the spraying range of the spraying drone increases or decreases accordingly, making it convenient to adjust the spraying range according to different terrains, greatly improving the flexibility and versatility of the spraying drone. However, this spraying drone with adjustable spraying range needs to adjust the position of the telescopic rod according to the terrain each time. The position of the telescopic rod is fixed during the flight, so it is a semi-fixed spraying device. Because the plant belt in the farmland is not a straight line, this semi-fixed spraying still has the disadvantages of poor spraying effect and poor spraying accuracy. Summary of the Invention

[0008] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a drone spraying method. The drone spraying method is used to solve the problem that the existing spraying head is relatively fixed and the spraying range cannot be adjusted in real time, resulting in poor spraying effect and poor accuracy of pesticides and fertilizers. The drone spraying method has the advantages of high accuracy, high efficiency and saving of pesticides and fertilizers.

[0009] The technical solution of the present invention to solve the above technical problems is:

[0010] A drone spraying method comprises the following steps:

[0011] S1. Use drones to collect images of farmland after transplanting or planting.

[0012] S2. Build a plant deep learning target detection algorithm model to identify and locate plants;

[0013] S3. Constructing a real-time movement path planning algorithm model for the spray head based on the plant position, the drone position, the spray head position, the initial velocity of the spray material leaving the spray head at the critical state, the drone flight speed, and the flight altitude;

[0014] S4. Processing the field images taken by the drone using a plant deep learning target detection algorithm model, and then processing the processed field images using a spray head real-time movement path planning algorithm model to obtain a spray head real-time movement path planning task command;

[0015] S5. Upload the real-time movement path planning task command of the spraying head to the UAV to perform field spraying operations; wherein, the left and right sides of the UAV are provided with spraying heads and a spraying drive mechanism for driving the spraying heads to slide left and right. After the control system in the UAV receives the real-time movement path planning task command of the spraying head, the spraying drive mechanism drives the spraying heads on the left and right sides to move, so as to adjust the position of the spraying heads in real time.

[0016] Preferably, in step S1, the drone should be a drone with high-precision positioning function, the camera pixels of the drone meet the standard requirements for shooting, and the quality of the captured images can be used by the plant deep learning target detection algorithm model to accurately identify the plants.

[0017] Preferably, in step S2, the construction of the plant deep learning target detection algorithm model comprises the following steps:

[0018] S21. Using drones to collect sufficient farmland images and pre-processing the farmland images;

[0019] S22. Set up the Anaconda environment and download the YOLO model;

[0020] S23. Divide the collected farmland image samples into training and test sets in a ratio of 7:3, and use labeling tools to label the plants in the farmland images taken by the drone;

[0021] S24: Train the model and test the training results. When the accuracy meets the requirements, the deep learning target detection model is obtained.

[0022] Preferably, in step S3, a vertical force analysis is performed on the sprayed object in combination with the plant position, the drone position, the spray head position, the critical initial velocity of the sprayed object leaving the spray head, the drone flight speed, and the drone flight altitude information. According to the drone's flight altitude and the critical initial velocity of the sprayed object leaving the spray head, combined with the distance-acceleration-time relationship formula, the movement time t of the sprayed object in the vertical direction is calculated, and the critical distance between the spray head and the plant in the operating direction is inferred through the movement time t. At the same time, a horizontal rectangular coordinate system is established with the drone's operating direction as the y-axis and the spray head movement direction as the x-axis. The sprayed object is subjected to horizontal force analysis and velocity decomposition, the spray head movement speed and movement direction under the critical state at a certain time are determined, and a real-time movement path planning algorithm model for the spray head is constructed.

[0023] Preferably, at the starting point of each plant line, the spray heads are located on the left and right sides of the drone body. Analysis of the critical state of the sprayed material during drone spraying shows that there are two different situations:

[0024] (1) When the spray head does not need to move, the spray material has a movement speed relative to the plant line on the surface of the farmland. At this time, the critical state of the spray material when it is separated from the spray head is analyzed. The spray material makes a parabolic motion parallel to the direction of the UAV's flight. Among them, the flight height of the UAV is H, the height of the plant is h, and the horizontal distance between the UAV and the plant is X. Then the height difference between the UAV and the plant is Y, where Y = Hh; through the formula Y = V0t + 1 / 2a0t 2 , where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; calculate the movement time t of the sprayed material in the vertical direction; then use the formula S = V1t + 1 / 2a1t 2 , where V1 is the initial velocity in the horizontal direction, a1 is the acceleration of the spray material in the horizontal direction, and S is the critical distance. If the horizontal distance X between the drone's spray head and the plant in the operating direction is less than the critical distance S, it means that the targeted spraying of the plant has been completed. Next, the position of the spray head can be adjusted to carry out targeted spraying of the next plant.

[0025] (2) The spray material moves in a horizontal parabolic direction that is not the UAV's forward direction, that is, the spray head needs to be adjusted and moved. In this case, the plant growth position deviates from the plant belt line, and the spray head position needs to be adjusted; the critical state of separation between the spray material and the spray head is analyzed, among which,

[0026] The flight height of the drone is H, the height of the plant is h, and the displacement between the plant and the spray head in the horizontal coordinate system is X. The displacement X is decomposed into the displacement X1 along the y-axis and the displacement X2 along the x-axis. The height difference between the drone and the plant is Y, where Y = Hh. The formula Y = V0t + 1 / 2a0t 2 Calculate the vertical movement time t of the sprayed material; where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; Calculate the vertical movement time t of the sprayed material;

[0027] Then according to the formula S1=V1t+1 / 2a1t 2 Calculate the displacement of the sprayed material in the y-axis direction within time t, where V1 is the initial velocity in the y-axis direction and a1 is the acceleration of the sprayed material in the y-axis direction;

[0028] Then according to the formula S2=V2t+1 / 2a2t 2 Calculate the x-axis displacement of the spray material within time t, where V2 is the x-axis speed of the spray head. This speed V2 and speed V1 form a composite speed V3 pointing to the plant to be sprayed in the horizontal coordinate system, and a1 is the acceleration of the spray material in the x-axis direction.

[0029] Therefore, the critical distance

[0030] As the drone moves forward and the speed and position of the spraying head are adjusted, when X=S, it means that the targeted spraying of the plant has been completed; then the position of the spraying head can be adjusted to carry out targeted spraying of the next plant.

[0031] By adjusting the speed and position of the spray head in real time under two different situations, the spray head slides left and right according to the position of the plant, so that the spray head can accurately target each plant.

[0032] Preferably, in step S4, combined with the farmland images taken by the drone, the plant deep learning target detection algorithm model, and the spray head real-time movement path planning algorithm model, the plant deep learning target detection algorithm model is used to perform plant target detection on the farmland images taken by the drone, identify and detect the plants in the farmland images, and according to the plant positions, combined with the spray head real-time movement path planning algorithm model, obtain the spray head real-time movement path planning task command set, the spray head real-time movement path planning task command set is a set of commands for each plant, and is arranged and combined according to the position order of the plants during operation, the spray head real-time movement path planning task command set includes instructions for the drone flight altitude, speed, and direction during the operation, and controls the spray head to slide left and right in real time at the required different movement speeds according to the command plan formulated according to the spray head real-time movement path planning algorithm model during the execution of the task, so that the spray head can accurately target and spray each plant in the farmland.

[0033] Preferably, the spraying drive mechanism includes fixed rods arranged on both sides of the fuselage of the UAV and a linear drive mechanism for driving the spraying head to move on the fixed rods, wherein the spraying head is slidably connected to the fixed rods.

[0034] Preferably, the linear drive mechanism adopts a linear drive motor or an electric push rod.

[0035] Compared with the prior art, the present invention has the following significant effects:

[0036] The drone spraying method of the present invention offers advantages such as high spraying accuracy, flexibility, specificity, targeting, and adaptability. It incorporates image processing technology to identify and locate plants, and establishes a mathematical relationship model based on parameters such as flight speed. This provides real-time path planning commands for different spray heads for different fields. Compared to traditional drone spraying methods with fixed spray heads, this method not only allows for real-time movement of the spray heads but also provides more targeted spraying task planning for different fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 Schematic diagram of the flow of the drone spraying method of the present invention.

[0038] Figure 2 Schematic diagram of the mathematical analysis model for different motion states of the sprayed object; in the figure: 1 is the flight altitude, 2 is the motion state trajectory of the sprayed object when the spray head is stationary, 3 is the plant that needs to adjust the position of the spray head for spraying operations, 4 is the velocity component in the forward direction of the UAV, 5 is the velocity component in the moving direction of the spray head, 6 is the motion state trajectory of the sprayed object when the spray head is moving, 7 is the sprayed object, 8 is the distance between the sprayed object and the plant in the direction of the plant belt line, and 9 is the plant that can be sprayed without adjusting the position of the spray head.

[0039] Figure 3 This is a diagram showing the execution effect of the task command for real-time movement path planning of the spraying head; in the figure: 10 is the plant, and 11 is the drone whose spraying head can move in real time.

[0040] Figure 4 It is a simplified structural diagram of the UAV; in the figure: 12 is the UAV blade, 13 is the UAV arm, 14 is the UAV main body, 15 is the slidable spray head, and 16 is the fixing rod. DETAILED DESCRIPTION

[0041] The present invention will be described in further detail below with reference to the embodiments and drawings, but the embodiments of the present invention are not limited thereto.

[0042] See also Figure 1-Figure 4 The drone spraying method of the present invention comprises the following steps:

[0043] S1. Use drones to collect images of farmland after transplanting or planting.

[0044] S2. Build a plant deep learning target detection algorithm model to identify and locate plants;

[0045] S3. Constructing a real-time movement path planning algorithm model for the spray head based on the plant position, the drone position, the spray head position, the initial velocity of the spray material leaving the spray head at the critical state, the drone flight speed, and the flight altitude;

[0046] S4. Processing the field images taken by the drone using a plant deep learning target detection algorithm model, and then processing the processed field images using a spray head real-time movement path planning algorithm model to obtain a spray head real-time movement path planning task command;

[0047] S5. Upload the real-time movement path planning task command of the spraying head to the UAV to perform field spraying operations; wherein, the left and right sides of the UAV are provided with spraying heads and a spraying drive mechanism for driving the spraying heads to slide left and right. After the control system in the UAV receives the real-time movement path planning task command of the spraying head, the spraying drive mechanism drives the spraying heads on the left and right sides to move, so as to adjust the position of the spraying heads in real time.

[0048] See also Figure 1-Figure 4 In step S1, the drone should be a drone with high-precision positioning function. The camera pixels of the drone meet the standard requirements for shooting, and the quality of the captured images can be used by the plant deep learning target detection algorithm model to accurately identify plants.

[0049] See also Figure 1-Figure 4 In step S2, the construction of the plant deep learning target detection algorithm model includes the following steps:

[0050] S21. Using drones to collect sufficient farmland images and pre-processing the farmland images;

[0051] S22. Set up the Anaconda environment and download the YOLO model;

[0052] S23. Divide the collected farmland image samples into training and test sets in a ratio of 7:3, and use labeling tools to label the plants in the farmland images taken by the drone;

[0053] S24: Train the model and test the training results. When the accuracy meets the requirements, the deep learning target detection model is obtained.

[0054] See also Figure 1-Figure 4 In step S3, the vertical force analysis of the sprayed material is performed based on the plant position, the UAV position, the spray head position, the critical initial velocity of the sprayed material when it leaves the spray head, the UAV flight speed, and the UAV flight altitude information. According to the UAV flight altitude and the critical initial velocity of the sprayed material when it leaves the spray head, combined with the distance-acceleration-time relationship formula, the vertical movement time t of the sprayed material is calculated, and the critical distance between the spray head and the plant in the operation direction is inferred from the movement time t. At the same time, a horizontal rectangular coordinate system is established with the UAV operation direction as the y-axis and the spray head movement direction as the x-axis. The horizontal force analysis and velocity decomposition of the sprayed material are performed to determine the spray head movement speed and movement direction under the critical state at a certain time, and to construct a real-time movement path planning algorithm model for the spray head.

[0055] See also Figure 1-Figure 4At the starting point of each plant line, the spray heads are located on the left and right sides of the drone's fuselage. Analysis of the critical state of the sprayed material during drone spraying shows that there are two different situations:

[0056] (1) When the spray head does not need to move, the spray material has a movement speed relative to the plant line on the surface of the farmland. At this time, the critical state of the spray material when it is separated from the spray head is analyzed. The spray material makes a parabolic motion parallel to the direction of the UAV's flight. Among them, the flight height of the UAV is H, the height of the plant is h, and the horizontal distance between the UAV and the plant is X. Then the height difference between the UAV and the plant is Y, where Y = Hh; through the formula Y = V0t + 1 / 2a0t 2 , where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; calculate the movement time t of the sprayed material in the vertical direction; then use the formula S = V1t + 1 / 2a1t 2 , where V1 is the initial velocity in the horizontal direction, a1 is the acceleration of the spray material in the horizontal direction, and S is the critical distance. If the horizontal distance X between the drone's spray head and the plant in the operating direction is less than the critical distance S, it means that the targeted spraying of the plant has been completed. Next, the position of the spray head can be adjusted to carry out targeted spraying of the next plant.

[0057] (2) The spray material moves in a horizontal parabolic direction that is not in the UAV's forward direction, that is, the spray head needs to be adjusted and moved. In this case, the plant growth position deviates from the plant belt line, and the spray head position needs to be adjusted; the critical state of separation between the spray material and the spray head is analyzed, among which,

[0058] The flight height of the drone is H, the height of the plant is h, and the displacement between the plant and the spray head in the horizontal coordinate system is X. The displacement X is decomposed into the displacement X1 along the y-axis and the displacement X2 along the x-axis. The height difference between the drone and the plant is Y, where Y = Hh. The formula Y = V0t + 1 / 2a0t 2 Calculate the vertical movement time t of the sprayed material; where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; Calculate the vertical movement time t of the sprayed material;

[0059] Then according to the formula S1=V1t+1 / 2a1t 2 Calculate the displacement of the sprayed material in the y-axis direction within time t, where V1 is the initial velocity in the y-axis direction and a1 is the acceleration of the sprayed material in the y-axis direction;

[0060] Then according to the formula S2=V2t+1 / 2a2t 2Calculate the x-axis displacement of the spray material within time t, where V2 is the x-axis speed of the spray head. This speed V2 and speed V1 form a composite speed V3 pointing to the plant to be sprayed in the horizontal coordinate system, and a1 is the acceleration of the spray material in the x-axis direction.

[0061] Therefore, the critical distance

[0062] As the drone moves forward and the speed and position of the spraying head are adjusted, when X=S, it means that the targeted spraying of the plant has been completed; then the position of the spraying head can be adjusted to carry out targeted spraying of the next plant.

[0063] By adjusting the speed and position of the spray head in real time under two different situations, the spray head slides left and right according to the position of the plant, so that the spray head can accurately target each plant.

[0064] See also Figure 1-Figure 4 In step S4, combined with the farmland image taken by the drone, the plant deep learning target detection algorithm model, and the spray head real-time movement path planning algorithm model, the plant deep learning target detection algorithm model is used to perform plant target detection on the farmland image taken by the drone, identify and detect the plants in the farmland image, and obtain the spray head real-time movement path planning task command set based on the plant position and the spray head real-time movement path planning algorithm model. The spray head real-time movement path planning task command set is a set of commands for each plant, and is arranged and combined in the order of the plant positions during the operation. The spray head real-time movement path planning task command set includes instructions for the drone's flight altitude, speed, and direction during the operation, and controls the spray head to slide left and right in real time at the required different movement speeds according to the command plan formulated according to the spray head real-time movement path planning algorithm model during the drone execution task, so that the spray head can accurately target and spray each plant in the farmland.

[0065] See also Figure 1-Figure 4 The spraying drive mechanism includes fixed rods arranged on both sides of the drone's fuselage and a linear drive mechanism for driving the spraying head to move on the fixed rods, wherein the spraying head is slidably connected to the fixed rods; the linear drive mechanism adopts a linear drive motor or an electric push rod.

[0066] Finally, the drone spraying method of the present invention combines image processing and uses a plant target detection algorithm model to accurately identify and locate plants in the farmland. According to the position of each plant and combined with parameters such as the drone flight speed, a real-time movement path planning algorithm model for the spray head is established, and a real-time movement path planning task command for the spray head is accurately planned for each piece of farmland. In addition, the structure of the drone spraying system is redesigned. Unlike the traditional drone spraying system with fixed spray heads, the drone described in this method has a horizontal fixed rod on each side, and a spray head on each fixed rod. The spray head can move horizontally on the fixed rod. According to the task command, as the drone moves, the spray head also adjusts its position left and right at the same time to achieve targeted precise spraying. Compared with traditional fixed spraying, the drone spraying method of the present invention implements different spray head movement planning commands for different farmlands, which is more targeted to the farmland. At the same time, the spray head can move left and right in real time, and the spraying flexibility is stronger and the accuracy is higher. It is a targeted spraying method.

[0067] The above is a preferred embodiment of the present invention, but the embodiment of the present invention is not limited to the above content. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the scope of protection of the present invention.

Claims

1. A drone spraying method, characterized in that: The following steps are involved: S1. Use drones to collect images of farmland after transplanting or planting. S2. Build a plant deep learning target detection algorithm model to identify and locate plants; S3. Constructing a real-time movement path planning algorithm model for the spray head based on the plant position, the drone position, the spray head position, the initial velocity of the spray material leaving the spray head at the critical state, the drone flight speed, and the flight altitude; S4. Processing the farmland images taken by the drone using a plant deep learning target detection algorithm model, and then processing the processed farmland images using a spray head real-time movement path planning algorithm model to obtain a real-time movement path planning task command for the spray head; S5. Upload the real-time movement path planning task command of the spraying head to the UAV to perform field spraying operations; wherein, the left and right sides of the UAV are provided with spraying heads and a spraying drive mechanism for driving the spraying heads to slide left and right. After the control system in the UAV receives the real-time movement path planning task command of the spraying head, the spraying drive mechanism drives the spraying heads on the left and right sides to move, so as to adjust the position of the spraying heads in real time.

2. The drone spraying method according to claim 1, characterized in that: In step S1, the drone should be a drone with high-precision positioning function, the camera pixels of the drone meet the standard requirements for shooting, and the quality of the captured images can be used by the plant deep learning target detection algorithm model to accurately identify the plants.

3. The drone spraying method according to claim 2, characterized in that: In step S2, the construction of the plant deep learning target detection algorithm model includes the following steps: S21. Using drones to collect sufficient farmland images and pre-processing the farmland images; S22. Set up the Anaconda environment and download the YOLO model; S23. Divide the collected farmland image samples into training and test sets in a ratio of 7:3, and use labeling tools to label the plants in the farmland images taken by the drone; S24. Train the model and test the training results. When the accuracy meets the requirements, the plant deep learning target detection algorithm model is obtained.

4. The drone spraying method according to claim 3, characterized in that: In step S3, the vertical force analysis of the sprayed material is performed in combination with the plant position, the drone position, the spray head position, the critical initial velocity of the sprayed material leaving the spray head, the drone flight speed, and the drone flight altitude information. According to the drone's flight altitude and the critical initial velocity of the sprayed material leaving the spray head, combined with the distance-acceleration-time relationship formula, the movement time t of the sprayed material in the vertical direction is calculated, and the critical distance between the spray head and the plant in the operation direction is inferred through the movement time t. At the same time, a horizontal rectangular coordinate system is established with the drone's operating direction as the y-axis and the spray head movement direction as the x-axis. The horizontal force analysis and velocity decomposition of the sprayed material are performed to determine the spray head movement speed and movement direction under the critical state at a certain time, and to construct a real-time movement path planning algorithm model for the spray head.

5. The drone spraying method according to claim 4, characterized in that: At the starting point of each plant line, the spray heads are located on the left and right sides of the drone's fuselage. Analysis of the critical state of the spray material during drone spraying reveals the following two different situations: (1) When the spray head does not need to move, the spray material has a moving speed relative to the plant line on the surface of the farmland. At this time, the critical state of the spray material when it is separated from the spray head is analyzed. The spray material makes a parabolic motion parallel to the direction of the UAV's flight. Among them, the flight height of the UAV is H, the height of the plant is h, and the horizontal distance between the UAV and the plant is X. The height difference between the UAV and the plant is Y, where Y=Hh; through the formula Y= V0t +1 / 2a0t 2 , where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; calculate the movement time t of the sprayed material in the vertical direction; then use the formula S = V1t + 1 / 2a1t 2 , where V1 is the initial velocity in the horizontal direction, a1 is the acceleration of the spray material in the horizontal direction, and S is the critical distance. If the horizontal distance X between the drone's spray head and the plant in the operating direction is less than the critical distance S, it means that the targeted spraying of the plant has been completed. Next, the position of the spray head can be adjusted to carry out targeted spraying of the next plant. (2) The spray material moves in a horizontal parabolic direction that is not the UAV's forward direction, that is, the spray head needs to be adjusted and moved. In this case, the plant growth position deviates from the plant belt line, and the spray head position needs to be adjusted; the critical state of the separation of the spray material and the spray head is analyzed, among which, The flight height of the drone is H, the height of the plant is h, and the displacement between the plant and the spray head in the horizontal coordinate system is X. The displacement X is decomposed into the displacement X1 along the y-axis and the displacement X2 along the x-axis. The height difference between the drone and the plant is Y, where Y=Hh. The formula Y= V0t +1 / 2a0t 2 Calculate the vertical movement time t of the sprayed material; where V0 is the initial velocity of the sprayed material in the vertical direction, and a0 is the acceleration of the sprayed material in the vertical direction; Calculate the vertical movement time t of the sprayed material; Then according to the formula S1= V1t +1 / 2a1t 2 Calculate the displacement of the sprayed material in the y-axis direction within time t, where V1 is the initial velocity in the y-axis direction and a1 is the acceleration of the sprayed material in the y-axis direction; Then according to the formula S2= V2t +1 / 2a2t 2 Calculate the x-axis displacement of the spray material within time t, where V2 is the x-axis speed of the spray head. This speed V2 and speed V1 form a composite speed V3 pointing to the plant to be sprayed in the horizontal coordinate system, and a1 is the acceleration of the spray material in the x-axis direction. Therefore, the critical distance ; As the drone moves forward and the speed and position of the spraying head are adjusted, when X=S, it means that the targeted spraying of the plant has been completed; then the position of the spraying head can be adjusted to carry out the targeted spraying of the next plant; By adjusting the speed and position of the spray head in real time under two different situations, the spray head slides left and right according to the position of the plant, so that the spray head can accurately target each plant.

6. The drone spraying method according to claim 5, characterized in that: In step S4, combined with the farmland images taken by the drone, the plant deep learning target detection algorithm model, and the spray head real-time movement path planning algorithm model, the plant deep learning target detection algorithm model is used to perform plant target detection on the farmland images taken by the drone, identify and detect the plants in the farmland images, and according to the plant positions, combined with the spray head real-time movement path planning algorithm model, obtain the spray head real-time movement path planning task command set, the spray head real-time movement path planning task command set is a set of commands for each plant, and is arranged and combined according to the position order of the plants during operation, the spray head real-time movement path planning task command set includes instructions for the drone's flight altitude, speed, and direction during the operation, and controls the spray head to slide left and right in real time at the required different movement speeds according to the command plan formulated according to the spray head real-time movement path planning algorithm model during the execution of the task, so that the spray head can accurately target and spray each plant in the farmland.

7. The drone spraying method according to claim 6, characterized in that: The spraying drive mechanism includes fixed rods arranged on both sides of the fuselage of the UAV and a linear drive mechanism for driving the spraying head to move on the fixed rods, wherein the spraying head is slidably connected to the fixed rods.

8. The drone spraying method according to claim 7, characterized in that: The linear drive mechanism adopts a linear drive motor or an electric push rod.

Images