Agricultural unmanned aerial vehicle inspection path correction method and system

Abstract

The invention discloses an agricultural unmanned aerial vehicle inspection path correction method and system, and belongs to the technical field of agricultural unmanned aerial vehicles. The method comprises the following steps: acquiring real-time sequence image data of a to-be-inspected farmland area; spatial distribution information and a three-dimensional digital elevation model of target crops in the to-be-inspected farmland area are obtained; the routing inspection task complexity of each path planning unit is obtained; obtaining predicted power consumption of a plurality of inspection paths in the future; and constructing an optimal inspection path model, and generating an optimal inspection correction path. According to the invention, by acquiring the real-time sequence image data and accurately extracting the crop distribution information and the topographic information, the multi-target optimization model is constructed to comprehensively consider the routing inspection coverage rate and the energy consumption, so that the technical effects of accurately correcting the routing inspection path, improving the routing inspection efficiency and optimizing the energy consumption management are achieved; the problems that in the prior art, unmanned aerial vehicle routing inspection path planning is not accurate, and the coverage rate and energy consumption cannot be balanced are solved.

Description

Technical Field

[0001] This invention relates to the field of agricultural drone technology, and in particular to a method and system for correcting inspection paths of agricultural drones. Background Technology

[0002] With the advancement of agricultural modernization, agricultural drones are widely used in farmland monitoring, crop management, and other operations. Traditional agricultural inspection methods mainly rely on manual on-site inspections, which suffer from problems such as low efficiency, limited coverage, and terrain limitations. Drone inspections can effectively overcome these drawbacks, quickly obtain information on large areas of farmland, and improve the level of precision in agricultural management.

[0003] Existing agricultural drone inspection systems typically operate along pre-set fixed routes. These routes are generally generated manually or through simple algorithms, making them ill-suited to the complex and ever-changing farmland environment. In actual inspections, the failure to fully consider factors such as variations in crop spatial distribution, terrain undulations, and the drone's own flight status often results in unsatisfactory inspection coverage, with some areas potentially being inspected repeatedly or missed. Furthermore, inaccurate estimations of drone energy consumption can lead to problems such as insufficient battery power forcing drones to return to base and interrupting inspection missions, or carrying excessive battery power resulting in resource waste. These systems fail to optimize energy consumption while ensuring inspection quality, thus failing to meet the urgent need for efficient and precise agricultural inspection operations. Summary of the Invention

[0004] This application provides an agricultural drone inspection path correction method and system, which solves the problems in the prior art where drone inspection path planning cannot adapt to complex farmland environments and where it is difficult to balance inspection coverage and energy consumption optimization. It achieves accurate path correction based on a multi-objective optimization model, thereby improving the technical effects of inspection coverage and optimizing energy consumption management.

[0005] This application provides a method for correcting the inspection path of an agricultural drone, including: acquiring real-time sequence image data of the farmland area to be inspected using an agricultural drone;

[0006] Based on real-time sequence image data, spatial distribution information and three-dimensional digital elevation model of target crops in the farmland to be inspected are obtained.

[0007] Based on the spatial distribution information of the target crop and the three-dimensional digital elevation model, multiple UAV path planning units are divided to obtain the inspection task complexity of each path planning unit.

[0008] Based on the complexity of the inspection task and the current flight status parameters of the agricultural drone, the predicted power consumption of multiple future inspection paths is obtained.

[0009] By combining the real-time remaining power and predicted power consumption of agricultural drones, an optimal inspection path model is constructed, and an optimal inspection correction path is generated.

[0010] Further steps to obtain spatial distribution information and a three-dimensional digital elevation model of the target crops within the farmland to be inspected include:

[0011] A shared feature extraction network is created, and the real-time sequence image data is input into the shared feature extraction network to obtain a shared depth feature map;

[0012] Based on a shared feature extraction network, a crop segmentation decoder and an elevation prediction decoder are constructed.

[0013] The shared deep feature map is input into the crop segmentation decoder and the elevation prediction decoder;

[0014] The pixel-level category mask of the target crop is obtained through the crop segmentation decoder, which serves as the spatial distribution information of the target crop.

[0015] The pixel-level relative elevation values ​​are obtained by the elevation prediction decoder, and the set of pixel-level relative elevation values ​​is used as the three-dimensional digital elevation model of the target crop.

[0016] Furthermore, based on the spatial distribution information of the target crop and the three-dimensional digital elevation model, the steps for dividing the UAV path planning into multiple units include:

[0017] The spatial distribution information of the target crop is converted into a first grid map, where each first grid cell records the crop cover status.

[0018] The three-dimensional digital elevation model is converted into a second raster map, where each second raster cell records the elevation value;

[0019] Spatial alignment is performed on the first and second raster maps. The first and second raster cells that overlap with the same geographical location are denoted as the third raster cell.

[0020] The farmland area is divided into multiple rectangular path planning units according to the preset size. Each path planning unit consists of multiple third grid units arranged in a continuous manner.

[0021] Furthermore, the steps to obtain the inspection task complexity for each path planning unit include:

[0022] The inspection task complexity of each path planning unit is calculated using the inspection task complexity calculation formula.

[0023] The formula for calculating the complexity of the inspection task is as follows:

[0024] ;

[0025] In the formula, For the first The complexity of the inspection task for each path planning unit In the first The normalized standard inspection power required for the target crop identified within each path planning unit. For normalization Local terrain gradient of each path planning unit, , The preset weighting factors are obtained through dynamic adjustment. It is a natural constant.

[0026] Further, the dynamic adjustment steps include:

[0027] By using a variable weighting function , Make dynamic adjustments;

[0028] The variable weighting function is:

[0029] ;

[0030] In the formula, This refers to the current real-time remaining battery power of the standardized agricultural drone. This refers to the total battery capacity of the standardized agricultural drone when fully charged. It is a positive number.

[0031] Furthermore, the steps to obtain the predicted power consumption for multiple future inspection paths include:

[0032] The predicted power consumption for multiple future inspection paths is calculated using the predicted power consumption calculation formula.

[0033] The formula for calculating the predicted power consumption is as follows:

[0034] ;

[0035] In the formula, Predicting power consumption for inspection routes. This represents the start time of the inspection route. This is the end time of the inspection route. For the total mass of agricultural drones, It is the acceleration due to gravity. The air density at the location of the agricultural drone. For the horizontal flight speed of agricultural drones, This is the reference cross-sectional area for agricultural drones. for The climb angle of agricultural drones at any given moment. In order to be in The operational power required to perform the corresponding inspection task at any time. The basic operating power for agricultural drones, This is the drag coefficient.

[0036] Furthermore, the steps for constructing the optimal inspection path model include:

[0037] Define the inspection coverage objective function as follows:

[0038] ;

[0039] In the formula, For path The set of path planning units covered by the inspection. For the first The area of ​​each path planning unit. For the first The importance weight of each path planning unit For path Overall inspection coverage score;

[0040] Define the energy consumption objective function for agricultural drones as follows:

[0041] ;

[0042] In the formula, For path The Middle Predicting power consumption based on the path of the joke. For path Total predicted energy consumption For path The total number of sub-path segments;

[0043] Establish a power constraint condition for the drone, wherein the drone power constraint condition is as follows:

[0044] ;

[0045] In the formula, Real-time remaining battery power for agricultural drones. The amount of electricity required for the safe return of agricultural drones;

[0046] The Pareto optimal solution set satisfying the energy constraint is obtained by using a multi-objective evolutionary algorithm, i.e., the solution is:

[0047] ;

[0048] In the formula, The set of Pareto optimal paths that satisfy the energy constraints. To maximize inspection coverage under constraints, To minimize the power consumption target under constraints;

[0049] The optimal inspection path is selected from the Pareto optimal solution set as the final output. The optimal inspection path is:

[0050] ;

[0051] In the formula, is the optimal inspection path, and is the path length. Energy efficiency indicators.

[0052] This application provides an agricultural drone inspection path correction system to implement an agricultural drone inspection path correction method, including: a data acquisition module, a model building module, a task complexity calculation module, a predicted power consumption acquisition module, and an inspection correction path generation module.

[0053] The data acquisition module is used to acquire real-time sequence image data of the farmland to be inspected through agricultural drones.

[0054] The model building module is used to obtain the spatial distribution information and three-dimensional digital elevation model of the target crops in the farmland area to be inspected based on real-time sequence image data.

[0055] The task complexity calculation module is used to divide the target crop into multiple UAV path planning units based on the spatial distribution information and three-dimensional digital elevation model, and obtain the inspection task complexity of each path planning unit.

[0056] The predicted power consumption acquisition module is used to obtain the predicted power consumption of multiple future inspection paths based on the complexity of the inspection task and the current flight status parameters of the agricultural drone.

[0057] The inspection correction path generation module is used to combine the real-time remaining power of the agricultural drone with the predicted power consumption to construct an optimal inspection path model and generate an optimal inspection correction path.

[0058] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0059] 1. By creating a shared feature extraction network and constructing a crop segmentation decoder and an elevation prediction decoder, the spatial distribution information and three-dimensional digital elevation model of the target crop can be accurately and efficiently obtained from real-time sequence image data. This enables precise perception of the farmland environment and effectively solves the problem that existing technologies cannot accurately obtain farmland crop information and terrain information, resulting in inaccurate inspection path planning.

[0060] 2. By dividing the UAV path planning into units and calculating the complexity of the inspection task, and combining it with a multi-objective optimization model, the inspection coverage and energy consumption factors are comprehensively considered, thereby optimizing the inspection path planning. This avoids the problems of unsatisfactory inspection coverage, energy waste or insufficient energy consumption caused by relying solely on fixed routes or simple planning methods in existing technologies, thus improving inspection efficiency and extending the single operation time of UAVs.

[0061] 3. By dynamically adjusting the weighting factors, the complexity calculation of the inspection task can be dynamically adjusted according to the real-time remaining power of the UAV. This allows for the rational allocation of inspection resources when power is limited, prioritizing important tasks. Consequently, it enables flexible adjustment of the inspection path to adapt to different power states, effectively solving the problem in existing technologies that cannot dynamically adjust the inspection strategy according to the real-time status of the UAV, and improving the adaptability of UAV inspection. Attached Figure Description

[0062] Figure 1 A flowchart of an agricultural drone inspection path correction method provided in this application embodiment;

[0063] Figure 2 This is a schematic diagram of the structure of an agricultural drone inspection path correction system provided in an embodiment of this application. Detailed Implementation

[0064] This application provides an agricultural drone inspection path correction method and system, which solves the problems in the prior art where drone inspection path planning cannot adapt to complex farmland environments and where it is difficult to balance inspection coverage and energy consumption optimization. By accurately extracting crop distribution and terrain information from real-time sequence image data, dividing the path planning units and calculating the complexity of the inspection task, a multi-objective optimization model is constructed to comprehensively balance coverage and energy consumption, achieving accurate path correction and efficient inspection operations, improving inspection coverage and optimizing energy consumption management.

[0065] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0066] like Figure 1 The diagram shown is a flowchart of an agricultural drone inspection path correction method provided in an embodiment of this application. The method is applied to an agricultural drone inspection path correction system and includes the following steps: acquiring real-time sequence image data of the farmland to be inspected using an agricultural drone.

[0067] Based on real-time sequence image data, spatial distribution information and three-dimensional digital elevation model of target crops in the farmland to be inspected are obtained.

[0068] Based on the spatial distribution information of the target crop and the three-dimensional digital elevation model, multiple UAV path planning units are divided to obtain the inspection task complexity of each path planning unit.

[0069] Based on the complexity of the inspection task and the current flight status parameters of the agricultural drone, a preset nonlinear power consumption model is invoked to obtain the predicted power consumption for multiple future inspection paths.

[0070] By combining the real-time remaining power and predicted power consumption of agricultural drones, an optimal inspection path model is constructed, an optimal inspection correction path is generated, and the agricultural drone is controlled to fly along the optimal inspection correction path; the optimal path aims to maximize inspection coverage and minimize total energy consumption.

[0071] Further steps to obtain spatial distribution information and a three-dimensional digital elevation model of the target crops within the farmland to be inspected include:

[0072] A shared feature extraction network is created, and the real-time sequence image data is input into the shared feature extraction network. The shared feature extraction network performs feature extraction processing on the real-time sequence image data to obtain a shared depth feature map.

[0073] Based on a shared feature extraction network, a crop segmentation decoder and an elevation prediction decoder are constructed.

[0074] The crop segmentation decoder is based on a convolutional neural network architecture. It performs pixel-by-pixel classification on the shared deep feature map through convolutional layers, pooling layers, and upsampling layers, and outputs a pixel-level class mask for the target crop.

[0075] The elevation prediction decoder, based on a CNN architecture, utilizes convolutional layers, pooling layers, and fully connected layers to perform regression calculations on shared deep feature maps, outputting pixel-level relative elevation values. These two components are used to obtain the spatial distribution information of the target crop and a 3D digital elevation model, respectively, providing a basis for subsequent UAV path planning and correction.

[0076] The shared deep feature map is input into the crop segmentation decoder and the elevation prediction decoder;

[0077] The shared deep feature map is processed by the crop segmentation decoder, and pixel-by-pixel classification calculation is performed to obtain the pixel-level category mask of the target crop as the spatial distribution information of the target crop.

[0078] The shared depth feature map is processed by the elevation prediction decoder, and elevation regression calculation is performed to obtain pixel-level relative elevation values. The set of pixel-level relative elevation values ​​is used as the three-dimensional digital elevation model of the target crop.

[0079] Furthermore, based on the spatial distribution information of the target crop and the three-dimensional digital elevation model, the steps for dividing the UAV path planning into multiple units include:

[0080] The spatial distribution information of the target crop is converted into a first grid map, where each first grid cell records the crop cover status.

[0081] The three-dimensional digital elevation model is converted into a second raster map, where each second raster cell records the elevation value;

[0082] Spatial alignment is performed on the first and second raster maps. The first and second raster cells that overlap with the same geographical location are denoted as the third raster cell.

[0083] The farmland area is divided into multiple rectangular path planning units according to the preset size. Each path planning unit consists of multiple third grid units arranged in a continuous manner.

[0084] Furthermore, the steps to obtain the inspection task complexity for each path planning unit include:

[0085] The inspection task complexity of each path planning unit is calculated using the inspection task complexity calculation formula.

[0086] The formula for calculating the complexity of the inspection task is as follows:

[0087] ;

[0088] In the formula, For the first The complexity of the inspection task for each path planning unit In the first The normalized standard inspection power required to identify the target crop within each path planning unit is determined by the crop type, the required inspection speed, and the sensor operating mode. For normalization The local terrain gradient, i.e., the steepness of the slope, of each path planning unit is derived from the three-dimensional digital elevation model. , The preset weighting factors are obtained through dynamic adjustment. It is a natural constant.

[0089] After the calculation is completed, the complexity of the inspection task is used as one of the input parameters of the nonlinear power consumption model.

[0090] Further, the dynamic adjustment steps include:

[0091] By using a variable weighting function , Make dynamic adjustments;

[0092] The variable weighting function is:

[0093] ;

[0094] In the formula, This refers to the current real-time remaining battery power of the standardized agricultural drone. This refers to the total battery capacity of the standardized agricultural drone when fully charged. It is a positive number.

[0095] Furthermore, the steps to obtain the predicted power consumption for multiple future inspection paths include:

[0096] The predicted power consumption for multiple future inspection paths is calculated using the predicted power consumption calculation formula.

[0097] The formula for calculating the predicted power consumption is as follows:

[0098] ;

[0099] In the formula, Predicting power consumption for inspection routes. This represents the start time of the inspection route. This is the end time of the inspection route. For the total mass of agricultural drones, It is the acceleration due to gravity. The air density at the location of the agricultural drone. For the horizontal flight speed of agricultural drones, This is the reference cross-sectional area for agricultural drones. for The climb angle of agricultural drones at any given moment. In order to be in The operational power required to perform the corresponding inspection task at any time. The basic operating power for agricultural drones, This is the drag coefficient.

[0100] Furthermore, the steps for constructing the optimal inspection path model include:

[0101] Define the inspection coverage objective function as follows:

[0102] ;

[0103] In the formula, For path The set of path planning units covered by the inspection. For the first The area of ​​each path planning unit. For the first The importance weight of each path planning unit For path Overall inspection coverage score;

[0104] Define the energy consumption objective function for agricultural drones as follows:

[0105] ;

[0106] In the formula, For path The Middle Predicting power consumption based on the path of the joke. For path Total predicted energy consumption For path The total number of sub-path segments;

[0107] Establish a power constraint condition for the drone, wherein the drone power constraint condition is as follows:

[0108] ;

[0109] In the formula, Real-time remaining battery power for agricultural drones. The amount of electricity required for the safe return of agricultural drones;

[0110] The Pareto optimal solution set satisfying the energy constraint is obtained by using a multi-objective evolutionary algorithm, i.e., the solution is:

[0111] ;

[0112] In the formula, The set of Pareto optimal paths that satisfy the energy constraints. To maximize inspection coverage under constraints, To minimize the power consumption target under constraints;

[0113] The optimal inspection path is selected from the Pareto optimal solution set as the final output. The optimal inspection path is:

[0114] ;

[0115] In the formula, is the optimal inspection path, and is the path length. Energy efficiency index (unit: square meters / joule, representing the area covered by a unit of energy consumption).

[0116] like Figure 2The diagram shown is a structural schematic of an agricultural drone inspection path correction system provided in an embodiment of this application, used to implement the agricultural drone inspection path correction method. The agricultural drone inspection path correction system provided in this embodiment includes: a data acquisition module, a model building module, a task complexity calculation module, a predicted power consumption acquisition module, and an inspection correction path generation module.

[0117] The data acquisition module is used to acquire real-time sequence image data of the farmland to be inspected through agricultural drones.

[0118] The model building module is used to obtain the spatial distribution information and three-dimensional digital elevation model of the target crops in the farmland area to be inspected based on real-time sequence image data.

[0119] The task complexity calculation module is used to divide the target crop into multiple UAV path planning units based on the spatial distribution information and three-dimensional digital elevation model, and obtain the inspection task complexity of each path planning unit.

[0120] The predicted power consumption acquisition module is used to obtain the predicted power consumption of multiple future inspection paths based on the complexity of the inspection task and the current flight status parameters of the agricultural drone.

[0121] The inspection correction path generation module is used to combine the real-time remaining power of the agricultural drone with the predicted power consumption to construct an optimal inspection path model and generate an optimal inspection correction path.

[0122] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0123] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0124] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0125] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0126] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0127] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for correcting inspection paths using agricultural drones, characterized in that, The following steps are involved: Real-time sequential image data of the farmland to be inspected is obtained through agricultural drones. Based on real-time sequence image data, spatial distribution information and three-dimensional digital elevation model of target crops in the farmland to be inspected are obtained. Based on the spatial distribution information of the target crop and the three-dimensional digital elevation model, multiple UAV path planning units are divided to obtain the inspection task complexity of each path planning unit. Based on the complexity of the inspection task and the current flight status parameters of the agricultural drone, the predicted power consumption of multiple future inspection paths is obtained. By combining the real-time remaining power and predicted power consumption of agricultural drones, an optimal inspection path model is constructed, and an optimal inspection correction path is generated.

2. The method for correcting inspection paths of agricultural drones as described in claim 1, characterized in that, The steps to obtain spatial distribution information and a three-dimensional digital elevation model of target crops within the farmland to be inspected include: A shared feature extraction network is created, and the real-time sequence image data is input into the shared feature extraction network to obtain a shared depth feature map; Based on a shared feature extraction network, a crop segmentation decoder and an elevation prediction decoder are constructed. The shared deep feature map is input into the crop segmentation decoder and the elevation prediction decoder; The pixel-level category mask of the target crop is obtained through the crop segmentation decoder, which serves as the spatial distribution information of the target crop. The pixel-level relative elevation values ​​are obtained by the elevation prediction decoder, and the set of pixel-level relative elevation values ​​is used as the three-dimensional digital elevation model of the target crop.

3. The method for correcting inspection paths of agricultural drones as described in claim 1, characterized in that, Based on the spatial distribution information of the target crop and a three-dimensional digital elevation model, the steps for dividing the UAV path planning unit include: The spatial distribution information of the target crop is converted into a first grid map, where each first grid cell records the crop cover status. The three-dimensional digital elevation model is converted into a second raster map, where each second raster cell records the elevation value; Spatial alignment is performed on the first and second raster maps. The first and second raster cells that overlap with the same geographical location are denoted as the third raster cell. The farmland area is divided into multiple rectangular path planning units according to the preset size. Each path planning unit consists of multiple third grid units arranged in a continuous manner.

4. The method for correcting inspection paths of agricultural drones as described in claim 1, characterized in that, The steps to obtain the inspection task complexity for each path planning unit include: The inspection task complexity of each path planning unit is calculated using the inspection task complexity calculation formula. The formula for calculating the complexity of the inspection task is as follows: ； In the formula, For the first The complexity of the inspection task for each path planning unit In the first The normalized standard inspection power required for the target crop identified within each path planning unit. For normalization Local terrain gradient of each path planning unit, , The preset weighting factors are obtained through dynamic adjustment. It is a natural constant.

5. The method for correcting inspection paths of agricultural drones as described in claim 4, characterized in that, The steps for dynamic adjustment include: By using a variable weighting function , Make dynamic adjustments; The variable weighting function is: ； In the formula, This refers to the current real-time remaining battery power of the standardized agricultural drone. This refers to the total battery capacity of the standardized agricultural drone when fully charged. It is a positive number.

6. The method for correcting inspection paths of agricultural drones as described in claim 1, characterized in that, The steps to obtain the predicted power consumption for multiple future inspection paths include: The predicted power consumption for multiple future inspection paths is calculated using the predicted power consumption calculation formula. The formula for calculating the predicted power consumption is as follows: ； In the formula, Predicting power consumption for inspection routes. This represents the start time of the inspection route. This is the end time of the inspection route. For the total mass of agricultural drones, It is the acceleration due to gravity. The air density at the location of the agricultural drone. For the horizontal flight speed of agricultural drones, This is the reference cross-sectional area for agricultural drones. for The climb angle of agricultural drones at any given moment. For The operational power required to perform the corresponding inspection task at any time. The basic operating power for agricultural drones, This is the drag coefficient.

7. The method for correcting the inspection path of an agricultural drone as described in claim 1, characterized in that, The steps to construct the optimal inspection path model include: Define the inspection coverage objective function as follows: ； In the formula, For path The set of path planning units covered by the inspection. For the first The area of ​​each path planning unit. For the first The importance weight of each path planning unit For path Overall inspection coverage score; Define the energy consumption objective function for agricultural drones as follows: ； In the formula, For path The Middle Predicting power consumption based on the path of the joke. For path Total predicted energy consumption For path The total number of sub-path segments; Establish a power constraint condition for the drone, wherein the drone power constraint condition is as follows: ； In the formula, Real-time remaining battery power for agricultural drones. The amount of electricity required for the safe return of agricultural drones; The Pareto optimal solution set satisfying the energy constraint is obtained by using a multi-objective evolutionary algorithm, i.e., the solution is: ； In the formula, The set of Pareto optimal paths that satisfy the energy constraints. To maximize inspection coverage under constraints, To minimize the power consumption target under constraints; The optimal inspection path is selected from the Pareto optimal solution set as the final output. The optimal inspection path is: ； In the formula, is the optimal inspection path, and is the path length. Energy efficiency indicators.

8. An agricultural drone inspection path correction system, used to implement the agricultural drone inspection path correction method according to any one of claims 1-7, characterized in that, include: Data acquisition module, model building module, task complexity calculation module, predicted power consumption acquisition module, and inspection correction path generation module; The data acquisition module is used to acquire real-time sequence image data of the farmland to be inspected through agricultural drones. The model building module is used to obtain the spatial distribution information and three-dimensional digital elevation model of the target crops in the farmland area to be inspected based on real-time sequence image data. The task complexity calculation module is used to divide the target crop into multiple UAV path planning units based on the spatial distribution information and three-dimensional digital elevation model, and obtain the inspection task complexity of each path planning unit. The predicted power consumption acquisition module is used to obtain the predicted power consumption of multiple future inspection paths based on the complexity of the inspection task and the current flight status parameters of the agricultural drone. The inspection correction path generation module is used to combine the real-time remaining power of the agricultural drone with the predicted power consumption to construct an optimal inspection path model and generate an optimal inspection correction path.

Images