Automatic seeding method and system based on optimal path of unmanned aerial vehicle

Abstract

The present invention relates to the field of three-dimensional position control technology and proposes an automatic sowing method and system based on an optimal drone path. The method comprises: planning a drone sowing path based on the sowing area, recording the drone's flight speed, altitude, and wind speed at each moment, and testing to obtain an air resistance correction coefficient; obtaining the seed drop point position at the current moment based on the pre-sowing position; obtaining the ideal seed drop point position at the current moment in a windless environment, and then obtaining the seed drop point deviation at the current moment; obtaining predicted wind speeds at various future moments; obtaining several planned drone paths at the current moment; comparing the drone sowing paths to quantify the offset evaluation values ​​of each planned path; obtaining a stable sowing evaluation value for each planned path; and obtaining the optimal planned path for drone sowing. The present invention aims to address the problem of uneven sowing and seed offset caused by wind speed during drone sowing.

Description

Technical Field

[0001] The present invention relates to the field of three-dimensional position control technology, and in particular to an automatic seeding method and system based on an optimal path of an unmanned aerial vehicle. Background Art

[0002] With the modernization of global agricultural production, drone seeding, as an efficient and flexible seeding method, has been widely used in large-scale farmland planting, wasteland greening, and ecological restoration. Traditional drone seeding technology typically relies on preset routes and fixed operating parameters to achieve rapid coverage of the operating area, greatly improving seeding efficiency and reducing labor costs. The widespread use of this technology has not only reduced agricultural labor intensity but also promoted the development of precision agriculture, making farmland management more intelligent and automated, bringing revolutionary changes to future agricultural production methods.

[0003] Existing drone seeding operations typically use a fixed flight speed and constant seeding rate, failing to dynamically adjust to real-time environmental changes within the operation area. This presents certain limitations. Specifically, traditional systems execute seeding tasks solely based on preset routes and lack the ability to perceive and process uncertainties such as wind speed, wind direction, and birds in real time. This results in an inability to promptly adjust flight paths and seeding strategies when encountering sudden wind currents, airflow disturbances, or dynamic obstacles (such as birds). This can lead to seeding offsets, duplications, or omissions, impacting the uniformity and accuracy of seeding results. Furthermore, traditional systems separate path generation from seeding strategies, failing to achieve joint optimization based on real-time environmental information and failing to balance operational efficiency and seeding uniformity in dynamic environments. Therefore, it is urgent to propose a drone automated seeding method and system that can combine real-time wind speed sensing and path optimization technology to dynamically adjust flight and seeding strategies, thereby significantly improving operational quality and seeding uniformity. Summary of the Invention

[0004] The present invention provides an automatic sowing method and system based on the optimal path of a drone to solve the problem of uneven sowing and seed deviation caused by wind speed during the existing drone sowing process. The technical solutions adopted are as follows:

[0005] The present invention proposes an automatic seeding method based on an optimal path of a drone, which includes the following steps:

[0006] Plan the drone's sowing path based on the sowing area, record the drone's flight speed, flight altitude, and wind speed at each moment, and test to obtain the air resistance correction coefficient;

[0007] Based on the current drone flight speed and wind speed, the air resistance correction coefficient is combined with the drone's flight altitude to obtain the current seed drop point position before seed sowing. The influence of the drone's flight speed and flight altitude on seed sowing in a windless state is analyzed to obtain the current ideal seed drop point position in a windless state, and then the seed drop point deviation at the current moment is obtained. The wind speed is calculated using the LSTM model to obtain the predicted wind speed at each future moment.

[0008] Through DWA path planning, several planned paths for the drone at the current moment and the planned speed of the drone at several moments in each planned path are obtained. Combined with the predicted wind speed at each moment in the future, the deviation of the seed drop point at each moment in each planned path is obtained, and then the offset evaluation value of each planned path is quantified by comparing the drone's sowing path. The deviation between the planned speed and flight speed of the drone in each planned path is compared to obtain the stable sowing evaluation value of each planned path.

[0009] Based on the offset evaluation value and stable spreading evaluation value of each planned path, the optimal planned path for UAV spreading is obtained.

[0010] Optionally, the seed landing position at the current moment can be obtained by:

[0011]

[0012] in, Indicates the current time The seed placement location, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time wind speed, Indicates the current time The flight speed of the drone, Indicates the flight altitude of the drone. represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

[0013] Optionally, the method of obtaining the ideal windless seed landing position at the current moment includes:

[0014] Under ideal conditions, the flight speed of the drone is , combined with the current direction of the drone, the current moment The ideal flight speed of the drone is ; The ideal seed landing position at the current moment without wind The calculation method is:

[0015]

[0016] in, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time The ideal flight speed of the drone, Indicates the flight altitude of the drone. represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

[0017] Optionally, the seed drop point deviation at the current moment can be obtained by:

[0018] The vector difference obtained by subtracting the ideal seed landing point position without wind from the seed landing point position at the current moment is used as the seed landing point offset at the current moment.

[0019] Optionally, the predicted wind speed at each future moment is obtained by:

[0020] Obtain a large amount of historical wind speeds, train the LSTM model, and obtain a trained wind speed prediction model; input the wind speeds at the current moment and several previous moments into the trained wind speed prediction model, and output the predicted wind speeds at each future moment.

[0021] Optionally, the seed drop point deviation at each moment in each planned path can be obtained by:

[0022] For the planning paths, based on the predicted wind speed at each moment in the future and the The planned speed of the UAV at each moment in the future on the planned path is combined with the flight position of the UAV at each moment in the future to obtain the first The deviation of the seed placement point at each moment in the future under the planned path.

[0023] Optionally, the offset evaluation value of each planned path is obtained by:

[0024] No. The future under the planning path The seed placement deviation at this moment is , in The future of the planning path The flight position of the drone at the moment is ; Get the The future of the planning path The vector sum of the deviations between the UAV flight position and the seed drop point at each moment , as the first The future of the planning path The comprehensive sowing deviation at the moment is obtained, the coordinate point corresponding to the comprehensive sowing deviation in the horizontal plane coordinate system is obtained, and the minimum distance between the coordinate point and the UAV sowing path is obtained as the first The future of the planning path The offset of the sowing plan at a certain moment;

[0025] Based on the The sum of the sowing plan offsets at all future moments in the planned path is obtained The offset evaluation value of each planned path is negatively correlated with the sum of the sowing plan offsets.

[0026] Optionally, the stable seeding evaluation value of each planned path is obtained by:

[0027] For the Planning path, the future The planned speed of the UAV at the moment is expressed as , combined with the flight direction at that moment, the ideal flight speed at that moment is obtained ;No. Stable spreading evaluation value of the planned paths The calculation method is:

[0028]

[0029] in, Indicates the number of moments in the planned path; Indicates the The future of the planning path The planned speed of the drone at that moment, Indicates the The future of the planning path The ideal flight speed at a given moment, Indicates the The future of the planning path The module length of the ideal flight speed at a certain moment; Represents an exponential function with a natural constant as its base.

[0030] Optionally, the method of obtaining the optimal planned path for UAV sowing based on the offset evaluation value and the stable sowing evaluation value of each planned path includes the following specific methods:

[0031] The information entropy of the offset evaluation values ​​of all planned paths is calculated as the offset information entropy; the information entropy of the stable sowing evaluation values ​​of all planned paths is calculated as the stable sowing information entropy; the two information entropies are weighted normalized to obtain the offset weight and stable sowing weight respectively;

[0032] For the The offset evaluation value and the stable spreading evaluation value are weighted summed according to the corresponding weights, and the result is taken as the first The planning path corresponding to the maximum value of the priority of all planning paths is taken as the optimal planning path at the current moment.

[0033] The present invention also proposes an automatic seeding system based on the optimal path of a drone, which includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0034] The beneficial effects of the present invention are as follows: the present invention estimates the actual seed placement position by comprehensively analyzing the influence of the drone's flight speed, flight altitude and real-time wind speed on the seed sowing process at the current moment, and combining the air resistance correction coefficient obtained by testing. At the same time, according to the ideal state in which the drone's flight speed remains consistent and there is no wind during the sowing process, the seed placement deviation at the current moment is quantified, and the wind speed change is predicted to obtain the predicted wind speed in the future. In the subsequent speed space adjustment process, the seed placement deviation at each subsequent moment is analyzed in combination with the predicted wind speed, and the deviation impact on seed sowing under the speed space adjustment is comprehensively evaluated; a number of planned paths are generated for the current moment through the DWA path planning algorithm, and the planned speed and predicted wind speed of the drone in each planned path are used to calculate the seed placement deviation. Wind speed, comprehensively analyze the changes in the drone's position in the planned path and the deviation of the landing point during the seed sowing process, and thus construct a deviation evaluation value to reflect the extent to which the seed sowing process under the planned path can return to the corresponding seed sowing position of the original planned path; at the same time, consider whether the speed component of the drone's planned speed in the planned path along the original planned path direction remains consistent to ensure the stability and uniformity of the drone sowing process; by analyzing the impact of the drone's flight speed and wind speed on the deviation of the seed sowing landing point, combined with DWA path planning and wind speed prediction, adjustments are made in the speed space to ensure that the drone can still maintain a smaller deviation than the original planned path under the influence of wind speed changes, while ensuring the stability and uniformity of sowing, significantly improving the operation quality of drone automatic sowing. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0036] Figure 1 A schematic flow chart of an automatic seeding method based on an optimal drone path provided by one embodiment of the present invention;

[0037] Figure 2 Schematic diagram of the drone's sowing path. DETAILED DESCRIPTION

[0038] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0039] See also Figure 1 , which shows a flow chart of an automatic seeding method based on an optimal path of a drone provided by an embodiment of the present invention, the method comprising the following steps:

[0040] Step S001: Plan the drone sowing path based on the sowing area, record the drone's flight speed, flight altitude and wind speed at each moment, and test to obtain the air resistance correction coefficient.

[0041] The purpose of this example is to consider the impact of wind speed on seed drop during the automatic seeding process of a drone. The LSTM model is used to predict wind speed, and combined with DWA path planning, the drone's flight path and seeding strategy are dynamically adjusted to ensure that the seed drop point is as close as possible to the drone's sowing path under the influence of wind speed, thereby ensuring stable and efficient seeding operations. First, it is necessary to obtain the drone's flight speed and wind speed at each moment, and at the same time record the drone's flight altitude during seeding operations.

[0042] Specifically, before the operation begins, the operator first imports the farmland boundary file through aerial images to determine the sowing area. According to the sowing width of the aircraft and the operation requirements (such as route spacing and overlap rate), the sowing area is divided into parallel operation routes. Usually, an "S-shaped" or "U-shaped" path mode is adopted to cover the entire sowing area to ensure uniform sowing, and the UAV sowing path is obtained, such as Figure 2As shown in the figure, the drone sowing path starts from point A and flies along the white solid line to point B, and then forms an S-shaped drone sowing path along the white solid line. It should be noted that to simplify planning, obstacles (such as trees, canals, and buildings) are usually pre-eliminated in the boundary file or manually avoided, and the operation path does not consider dynamic factors (such as wind speed and birds). Finally, the system quickly generates a global path based on a regular geometric layout, providing a standardized flight route for sowing operations.

[0043] Furthermore, a sampling time interval is set. In this embodiment, the sampling time interval is 0.1 seconds for collection, and the flight speed of the drone at each moment during the flight is recorded. At the same time, a wind speed sensor is installed to record the wind speed in real time, where the flight speed and wind speed are both vectors with directions, and the sampling time intervals are the same; in this embodiment, the flight altitude of the drone during the sowing operation is set to 3 meters.

[0044] It should be further explained that in actual sowing, the falling motion of seeds is affected not only by gravity and wind, but also by factors such as air resistance, seed shape and density. Due to these factors, the horizontal speed attenuation (or acceleration) of seeds deviates from the theoretical calculation results, so a correction coefficient needs to be introduced in the simplified model for predicting the landing point.

[0045] Specifically, a field test is conducted in the sowing area. By comparing the theoretically predicted landing point of the seeds (based on the current position, flight speed, wind speed and falling time) with the actual detected landing point coordinates, the deviation between the two is calculated, and then a correction coefficient is obtained by using data fitting (such as the least squares method) as the air resistance correction coefficient. This is a test before the drone starts working. The specific process is existing technology and will not be repeated in this embodiment.

[0046] It should be noted that in the actual sowing process, due to the influence of real-time uncertain factors such as wind speed, the sowing uniformity under the traditional sowing method is insufficient, and some seeds will deviate from the preset planting area; it is necessary to make real-time local adjustments to the optimal path of the sowing drone based on the real-time monitoring of wind speed by the wind speed sensor to improve the sowing efficiency and uniformity of sowing.

[0047] Step S002: Based on the current UAV flight speed and wind speed, the air resistance correction coefficient and the UAV flight altitude are used to obtain the seed landing position at the current moment. The influence of the UAV flight speed and flight altitude on seed sowing in a windless state is analyzed to obtain the ideal seed landing position at the current moment, and then obtain the seed landing point deviation at the current moment. The wind speed is used to obtain the predicted wind speed at each future moment through the LSTM model.

[0048] It should be noted that since the operating height of agricultural sowing is relatively low, usually 1.5-3 meters, the sowing situation is not sensitive to vertical airflow. The wind speed obtained by the wind speed sensor in this embodiment is two-dimensional data, that is, wind speed data on the horizontal plane; and the movement of seeds from the drone sowing point to the ground is theoretically a complex aerodynamic problem. In actual sowing control, in order to quickly predict the seed landing point while ensuring sowing uniformity, this embodiment considers using simple linear superposition to quickly obtain an approximate landing point for use in path planning and sowing adjustment.

[0049] Preferably, in one embodiment of the present invention, based on the current UAV flight speed and wind speed, the air resistance correction coefficient and the UAV flight altitude are used to obtain the seed landing position at the current moment for the seed sowing position, including the specific method of:

[0050] In the process of generating the UAV sowing path for the sowing area, a two-dimensional plane coordinate system has been constructed for the sowing area, and the current time is recorded. The projection position of the drone on the horizontal plane (The projection position is a coordinate, which is directly represented by a vector pointing from the origin to the coordinate, which is convenient for vector addition calculation with the velocity vector). The seed placement location The calculation method is:

[0051]

[0052] in, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time wind speed, Indicates the current time The flight speed of the drone, Indicates the flight altitude of the drone (the flight altitude of the drone is the flight altitude set during the sowing process. In this embodiment, the drone maintains a constant altitude flight). represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

[0053] It should be noted that the seeds begin to fall from the current projection of the drone as the starting point, and the initial velocity is the vector sum of the drone's flight speed and the wind speed at the current moment. The air resistance correction coefficient and the estimated falling time are used to estimate the seed landing point.

[0054] It should be further explained that in the actual sowing process, regardless of the direction of the drone's movement, the drone's flight speed remains consistent, that is, the drone's flight speed along the direction of the drone's movement always remains constant. , it is necessary to quantify the ideal landing point of seeds after sowing in a windless state, so as to provide a basis for analyzing the impact of speed changes on seed deviation in the subsequent UAV speed space and path planning process.

[0055] Preferably, in one embodiment of the present invention, the influence of the flying speed and altitude of the drone on seed sowing in the windless state is analyzed to obtain the ideal seed landing position in the windless state at the current moment, and then obtain the current moment The specific method of seed placement deviation is as follows:

[0056] Under ideal conditions, the flight speed of the drone is (scalar without direction), combined with the current direction of the drone, the current moment The ideal flight speed of the drone is , then the ideal landing position of the seed without wind at the current moment is The calculation method is:

[0057]

[0058] in, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time The ideal flight speed of the drone, Indicates the flight altitude of the drone. represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

[0059] Furthermore, the vector difference obtained by subtracting the ideal seed landing position without wind from the seed landing position at the current moment is used as the seed landing offset at the current moment, that is, ,in For the current moment The seed drop offset, For the current moment The seed placement location, For the current moment The ideal landing location for seeds without wind.

[0060] What needs to be explained is that in the windless state, combined with the constant rational flight speed of the UAV, the ideal seed landing point position in the sowing process is estimated according to the calculation method of the seed landing point position. The vector difference between the seed landing point position and the ideal landing point position in the windless state reflects the offset of the seed landing point under the influence of the change in the UAV flight speed and the wind speed, providing a basis for the subsequent analysis of the seed landing point changes in the speed space adjustment.

[0061] Preferably, in one embodiment of the present invention, the wind speed is obtained by using an LSTM model to predict the wind speed at each moment in the future, including the following specific methods:

[0062] It should be noted that during the seed sowing process, the seed landing position under the influence of wind speed and the ideal landing position in the windless state are obtained respectively, and the seed landing point deviation is calculated based on this. In the subsequent real-time adjustment of the drone flight path, it is necessary to minimize the impact of wind speed on sowing, so as to adjust the drone position and speed in real time to adapt to wind speed changes. It is necessary to predict future wind speed changes based on historical wind speeds. This embodiment considers using an LSTM model to predict wind speed.

[0063] Specifically, a large amount of historical wind speed is obtained, and the sampling time interval of the wind speed is 0.1 second. In this embodiment, the LSTM model is trained with every 35 seconds as a data set, and the historical wind speed of the first 30 seconds in the data set is input into the model as a training data set. The output result uses the historical wind speed of the last 5 seconds, and the model is trained using the mean square error loss function to obtain a trained wind speed prediction model. The specific LSTM model training process is an existing method and will not be repeated in this embodiment. The wind speed at the current moment and the previous 30 seconds is input into the trained wind speed prediction model, and the predicted wind speed at each moment in the next 5 seconds is output.

[0064] At this point, the actual seed placement position is estimated by considering the comprehensive influence of the drone's flight speed, flight altitude, and real-time wind speed on the seed sowing process at the current moment, combined with the air resistance correction coefficient that has been tested. At the same time, based on the ideal state of the drone's flight speed remaining consistent and there being no wind during the sowing process, the seed placement deviation at the current moment is quantified. At the same time, the wind speed change is predicted to obtain the predicted wind speed in the future. In the subsequent speed space adjustment process, the predicted wind speed is combined to analyze the seed placement deviation at each subsequent moment, and the deviation impact on seed sowing under speed space adjustment is comprehensively evaluated.

[0065] Step S003: Use DWA path planning to obtain several planned paths for the drone at the current moment, as well as the planned speed of the drone at several moments in each planned path. Combined with the predicted wind speed at each moment in the future, the deviation of the seed placement point at each moment in each planned path is obtained, and then the offset evaluation value of each planned path is quantified by comparing the drone's sowing path. The deviation between the planned speed and flight speed of the drone in each planned path is compared to obtain the stable sowing evaluation value of each planned path.

[0066] It should be noted that when performing local path planning, the traditional DWA algorithm generates a series of possible speed pairs based on the current speed and acceleration limits, predicts the trajectory of each speed pair over a period of time, and scores the predicted trajectory based on various cost functions (such as obstacle avoidance, target arrival, and speed deviation). The speed pair with the lowest cost is then selected as the current control instruction. However, when combined with the traditional DWA local path planning algorithm, after obtaining the real-time deviation size, the original DWA local path planning method can be adjusted based on the deviation amount, thereby improving the seeding performance under the subsequent optimal path.

[0067] Preferably, in one embodiment of the present invention, a plurality of planned paths of the drone at the current moment and the planned speed of the drone at a plurality of moments in each planned path are obtained through DWA path planning, including the following specific methods:

[0068] The DWA path planning algorithm obtains several planned paths based on the current drone flight speed. This embodiment plans paths for the next 50 moments, or the drone's flight position for the next 5 seconds. During the DWA path planning process, each path has a corresponding linear velocity and angular velocity. The drone's flight direction is the tangent direction of the corresponding position on the planned path. The speed of the drone at each moment in each planned path is obtained as the planned speed of the drone at each moment. It should be noted that in this embodiment, since the drone maintains a constant and relatively low altitude, only the speed change in the horizontal plane is considered.

[0069] It should be further explained that, considering the influence of wind speed during the sowing process of the drone, the actual optimal path is not the pre-planned path. If sowing is carried out along the original optimal path, the sown seeds will be unevenly distributed. That is, it is necessary to build an offset evaluation function and obtain an offset evaluation value based on the calculation method of the seed drop point deviation amount, combined with the predicted wind speed and the sowing position changes caused by the drone's flight process, and compared with the original planned path (drone sowing path) based on the planned path of the traditional DWA algorithm.

[0070] Preferably, in one embodiment of the present invention, the predicted wind speed at each moment in the future is combined to obtain the seed placement deviation at each moment in each planned path, and then the offset evaluation value of each planned path is quantified by comparing the drone sowing path. The specific method includes:

[0071] For the Planned paths, according to the current time The calculation method of the seed placement deviation is given by the predicted wind speed at each moment in the future and the planned speed of the drone. The deviation of the seed placement point at each moment in the future under the planned path (the ideal seed placement point without wind is not affected by the speed space adjustment and DWA path planning), where The future under the planning path The seed placement deviation at the moment is , in The future of the planning path The flight position of the drone at the moment (the starting point of seed sowing) is ; Get the The future of the planning path The vector sum of the deviations between the UAV flight position and the seed drop point at each moment , as the first The future of the planning path The comprehensive sowing deviation at the moment is obtained, the coordinate point corresponding to the comprehensive sowing deviation in the horizontal plane coordinate system is obtained, and the minimum distance between the coordinate point and the UAV sowing path is obtained as the first The future of the planning path The sowing plan offset at the moment; The offset evaluation value of the planned path The calculation method is:

[0072]

[0073] in, Indicates the number of moments in planning the path. This embodiment has set ; Indicates the The future of the planning path The offset of the sowing plan at a certain moment; Represents an exponential function with a natural constant as the base. In this embodiment, Model to present inverse proportional relationship and normalization processing, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0074] It should be noted that the seed drop point deviation is the deviation caused by the influence of the drone's planned speed and predicted wind speed during the seed falling process, and the drone's flight position is the planned position change at the corresponding moment of the planned path. The vector sum integrates the change in the drone's own position and the deviation of the affected seeds to obtain the comprehensive sowing deviation. Based on the comprehensive sowing deviation, the position is compared with the original planned path (drone sowing path) to construct an offset evaluation function. The smaller the sowing planning offset reflected by the minimum distance in the cumulative result at each moment, the smaller the deviation of the planned path sowing compared with the original planned path, and the larger the offset evaluation value, that is, the higher the offset evaluation, the smaller the offset of the corresponding planned path, that is, in the process of sowing along the planned path, under the influence of the comprehensive planned speed and wind speed, the seed drop point is more likely to return to the original planned path.

[0075] It should be further explained that, based on the offset evaluation function, in order to ensure the uniformity of seed sowing in the flight direction of the UAV, it is necessary to compare the planned speed at each moment on the planned path with the ideal flight speed of the UAV, that is, in the direction of the original planned path, the speed component of the UAV should be kept consistent with the initial speed as much as possible to ensure a stable and uniform sowing process. In this way, a stable sowing evaluation function is constructed to obtain the stable sowing evaluation value of each path.

[0076] Preferably, in one embodiment of the present invention, the deviation between the planned speed and the flight speed of the drone in each planned path is compared to obtain the stable seeding evaluation value of each planned path, including the specific method of:

[0077] For the Planning path, the future The planned speed of the UAV at the moment is expressed as , combined with the flight direction at that moment, the ideal flight speed at that moment is obtained , then Stable spreading evaluation value of the planned paths The calculation method is:

[0078]

[0079] in, Indicates the number of moments in the planned path; Indicates the The future of the planning path The planned speed of the drone at that moment, Indicates the The future of the planning path The ideal flight speed at a given moment, Indicates the The future of the planning path The module length of the ideal flight speed at a certain moment; Represents an exponential function with a natural constant as the base. In this embodiment, Model to present inverse proportional relationship and normalization processing, As the input of the model, the implementer can set the inverse proportional function and normalization function according to the actual situation.

[0080] What needs to be explained is that It represents the projected speed scalar of the UAV's planned speed in the direction of the original planned path at the corresponding moment of the planned path. The smaller the difference between the planned speed scalar and the ideal flight speed scalar, the greater the possibility of ensuring stable seeding by flying at a uniform speed along the original planned direction, and thus the greater the stable seeding evaluation value.

[0081] At this point, the DWA path planning algorithm generates several planned paths for the current moment. Based on the planned speed and predicted wind speed of the drone in each planned path, the changes in the drone's position in the planned path and the deviation of the landing point during the seed sowing process are comprehensively analyzed. In this way, a deviation evaluation value is constructed to reflect the extent to which the seed sowing process under the planned path can return to the corresponding seed sowing position of the original planned path. At the same time, the speed component of the drone's planned speed in the planned path along the direction of the original planned path is considered to ensure the stability and uniformity of the drone sowing process.

[0082] Step S004: Based on the offset evaluation value and stable sowing evaluation value of each planned path, the optimal planned path for drone sowing is obtained.

[0083] It should be noted that, based on the offset evaluation value and the stable spreading evaluation value, a comprehensive analysis is needed to obtain the optimal planning path, and the entropy value of the two evaluation values ​​reflects the change of the planning path on the two evaluation values. The larger the entropy value, the greater the difference in the evaluation values ​​corresponding to different planning paths. As the optimal planning path screening process, this evaluation value should be referred to more, that is, a small change in the planning path will cause a large change in the evaluation value, so as to assign weights to the two evaluation values.

[0084] Specifically, the information entropy of the offset evaluation values ​​of all planned paths is calculated as the offset information entropy; similarly, the information entropy of the stable sowing evaluation values ​​of all planned paths is calculated as the stable sowing information entropy; the two information entropies are weighted normalized to obtain the offset weight and the stable sowing weight respectively; for the first The offset evaluation value and the stable spreading evaluation value are weighted summed according to the corresponding weights, and the result is taken as the first The preferred degree of each planned path; the planned path corresponding to the maximum preferred degree of all planned paths is taken as the optimal planned path at the current moment.

[0085] Furthermore, after the current moment, the drone moves along the optimal planned path and sows seeds. The optimal planned path in this embodiment includes a total of 50 moments, or 5 seconds. After completing the optimal planned path, the predicted wind speed is re-obtained according to the above method, and several planned paths are obtained through the DWA algorithm. The optimal planned path is selected, and so on, to complete the automatic sowing of the drone along the optimal planned path in the sowing area.

[0086] By analyzing the impact of drone flight speed and wind speed on seed placement deviation, combined with DWA path planning and wind speed prediction, adjustments were made in speed space to ensure that the drone can maintain a smaller deviation from the original planned path under the influence of wind speed changes, while ensuring the stability and uniformity of seeding, significantly improving the quality of drone automated seeding operations.

[0087] Another embodiment of the present invention provides an automatic seeding system based on an optimal path of a drone, the system comprising a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, steps S001 to S004 of the above method are implemented.

[0088] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The automatic seeding method based on the optimal path of the drone is characterized by: The method comprises the following steps: Plan the drone's sowing path based on the sowing area, record the drone's flight speed, flight altitude, and wind speed at each moment, and test to obtain the air resistance correction coefficient; Based on the current drone flight speed and wind speed, the air resistance correction coefficient is combined with the drone's flight altitude to obtain the current seed drop point position before seed sowing. The influence of the drone's flight speed and flight altitude on seed sowing in a windless state is analyzed to obtain the current ideal seed drop point position in a windless state, and then the seed drop point deviation at the current moment is obtained. The wind speed is calculated using the LSTM model to obtain the predicted wind speed at each future moment. Through DWA path planning, several planned paths for the drone at the current moment and the planned speed of the drone at several moments in each planned path are obtained. Combined with the predicted wind speed at each moment in the future, the deviation of the seed placement point at each moment in each planned path is obtained, and then the offset evaluation value of each planned path is quantified by comparing the drone's sowing path. The deviation of the drone's planned speed in each planned path from the drone's ideal flight speed is compared to obtain the stable sowing evaluation value of each planned path. Based on the offset evaluation value and stable spreading evaluation value of each planned path, the optimal planned path for UAV spreading is obtained; The specific method for obtaining the seed placement deviation at each moment in each planned path is as follows: For the planning paths, based on the predicted wind speed at each moment in the future and the The planned speed of the UAV at each moment in the future on the planned path is combined with the flight position of the UAV at each moment in the future to obtain the first The deviation of the seed placement point at each moment in the future under the planned path; The specific method for obtaining the stable seeding evaluation value of each planned path is as follows: For the Planning path, the future The planned speed of the UAV at the moment is expressed as , combined with the flight direction at that moment, the ideal flight speed at that moment is obtained ;No. Stable spreading evaluation value of the planned paths The calculation method is: in, Indicates the number of moments in the planned path; Indicates the The future of the planning path The planned speed of the drone at that moment, Indicates the The future of the planning path The ideal flight speed at a given moment, Indicates the The future of the planning path The module length of the ideal flight speed at a certain moment; Represents an exponential function with a natural constant as its base.

2. The automatic sowing method based on the optimal path of a drone according to claim 1, characterized in that: The specific method for obtaining the seed landing position at the current moment is as follows: in, Indicates the current time The seed placement location, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time wind speed, Indicates the current time The flight speed of the drone, Indicates the flight altitude of the drone. represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

3. The automatic sowing method based on the optimal path of a drone according to claim 2, characterized in that: The specific method for obtaining the ideal windless seed landing position at the current moment is as follows: Under ideal conditions, the flight speed of the drone is , combined with the current direction of the drone, the current moment The ideal flight speed of the drone is ; The ideal seed drop location at the current moment without wind The calculation method is: in, Indicates the current time The projection position of the drone on the horizontal plane, represents the air resistance correction factor, Indicates the current time The ideal flight speed of the drone, Indicates the flight altitude of the drone. represents the acceleration due to gravity, Represents the estimated falling time of seeds during the sowing process.

4. The automatic seeding method based on the optimal path of a drone according to claim 3 is characterized in that: The specific method for obtaining the seed landing point deviation at the current moment is as follows: The vector difference obtained by subtracting the ideal seed landing point position without wind from the seed landing point position at the current moment is used as the seed landing point deviation at the current moment.

5. The automatic seeding method based on the optimal path of a drone according to claim 1, characterized in that: The specific method for obtaining the predicted wind speed at each future moment is as follows: Obtain a large amount of historical wind speeds, train the LSTM model, and obtain a trained wind speed prediction model; input the wind speeds at the current moment and several previous moments into the trained wind speed prediction model, and output the predicted wind speeds at each future moment.

6. The automatic seeding method based on the optimal path of a drone according to claim 1, characterized in that: The specific method for obtaining the offset evaluation value of each planned path is as follows: No. The future under the planning path The seed placement deviation at this moment is , in The future of the planning path The flight position of the drone at the moment is ; Get the The future of the planning path The vector sum of the deviations between the UAV flight position and the seed landing point at each moment is recorded as , and take the vector sum as the first The future of the planning path The comprehensive sowing deviation at the moment is obtained, the coordinate point corresponding to the comprehensive sowing deviation in the horizontal plane coordinate system is obtained, and the minimum distance between the coordinate point and the sowing path of the UAV is obtained as the first The future of the planning path The offset of the sowing plan at a certain moment; Based on the The sum of the sowing plan offsets at all future moments in the planned path is obtained The offset evaluation value of each planned path is negatively correlated with the sum of the sowing plan offsets.

7. The automatic seeding method based on the optimal path of a drone according to claim 1, characterized in that: The specific method for obtaining the optimal planned path for UAV sowing based on the offset evaluation value and stable sowing evaluation value of each planned path is as follows: The information entropy of the offset evaluation values ​​of all planned paths is calculated as the offset information entropy; the information entropy of the stable sowing evaluation values ​​of all planned paths is calculated as the stable sowing information entropy; the two information entropies are weighted normalized to obtain the offset weight and stable sowing weight respectively; For the The offset evaluation value and the stable spreading evaluation value are weighted summed according to the corresponding weights, and the result is taken as the first The planning path corresponding to the maximum value of the priority of all planning paths is taken as the optimal planning path at the current moment.

8. An automatic seeding system based on an optimal path of a drone, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the automatic sowing method based on the optimal path of the drone are implemented as described in any one of claims 1 to 7.

Images