Autonomous flight path planning and image acquisition method and system for unmanned aerial vehicle

Abstract

The invention provides an unmanned aerial vehicle autonomous flight path planning and image acquisition method and system, and relates to the technical field of unmanned aerial vehicle inspection. Firstly, through intelligent three-dimensional grid division and dynamic route planning, the coverage rate and efficiency of feature region monitoring are remarkably improved; secondly, due to the integration of real-time meteorological data and historical feature data, the pertinence and adaptability of acquisition are enhanced, and high-quality images can be obtained in a complex and changeable environment; through combination of real-time processing and ground deep learning analysis, the accuracy and efficiency of feature extraction are greatly improved; and on the basis of a self-adaptive supplementary sampling mechanism of the region of interest, quick response and accurate positioning of the suspected feature region are realized, and reliable data support is provided for subsequent detailed analysis and decision making. The method is not only suitable for identification tasks of various preset features, but also has good expansibility and flexibility, and can be adjusted and optimized according to different application requirements.

Description

Technical Field

[0001] The present invention relates to the field of unmanned aerial vehicle inspection technology, and in particular to a method and system for autonomous trajectory planning and image acquisition of unmanned aerial vehicles. Background Art

[0002] The current use of drones for image acquisition and path planning is an innovative inspection technology that uses unmanned aerial vehicles (UAVs) to perform inspection, inspection and monitoring tasks, replacing or assisting traditional manual operations. This technology integrates the achievements of multiple fields such as drone platforms, automatic control technology, sensor technology, image processing technology, and data analysis and processing. Using drones for image acquisition reduces manual input and increases the scope of image acquisition. The closest existing technology to this application can use drones for image acquisition, but it still has certain limitations and shortcomings when used in complex forest environments.

[0003] In the existing technology, first, there is a lack of specific optimization for the forest environment. The existing technology is mainly aimed at scenes such as substations and mountains, and does not fully consider the particularity of the forest environment, such as complex terrain, dense vegetation and other factors; second, the route planning lacks adaptability. Most methods rely on pre-set fixed routes and cannot be dynamically adjusted according to actual flight conditions, making it difficult to cope with uncertainties in the forest environment. Third, the image acquisition process lacks intelligent control. Existing methods usually use fixed parameters for shooting, and fail to automatically adjust the shooting parameters according to the characteristics of different areas, affecting the image quality. Fourth, the data processing efficiency is low. Some methods need to transmit a large amount of data to the ground station for processing, which increases the communication burden and reduces real-time performance. Fifth, safety considerations are insufficient. In complex forest environments, existing methods lack effective obstacle avoidance and emergency processing mechanisms, and there are safety hazards. Summary of the invention

[0004] The main purpose of the present invention is to provide a method and system for autonomous trajectory planning and image acquisition of unmanned aerial vehicles, so as to solve the technical problem that the current unmanned aerial vehicle technical solution can only simply control the unmanned aerial vehicle to fly along a fixed route and collect images according to a fixed mode, but cannot meet the technical problems of adaptive flight and image acquisition in complex terrain and complex environment.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is: a method for autonomous trajectory planning and image acquisition of a UAV, comprising the following steps: S1: Obtain preset forest boundary data, and generate an initial gridded flight area according to the forest boundary data; based on the initial gridded flight area, combine terrain data and vegetation density data to obtain the current terrain undulation and vegetation distribution, and generate a three-dimensional gridded flight area; S2: In the three-dimensional grid flight area, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance; the current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized according to the safety factors to generate a real-time route; S3: The drone flies according to the real-time route and uses the onboard laser radar to detect the surrounding environment in real time and perform obstacle avoidance. S4: According to the current UAV flight altitude, speed and vegetation characteristics of the collection area, the camera's focal length, shutter speed and ISO sensitivity are adaptively adjusted to the optimal state, and the camera device is controlled to collect images; S5: Input the collected images into the deep learning model deployed on the drone, analyze and process the images in real time, and identify areas that meet the preset feature conditions; S6: Based on the recognition results, a list of areas that need to be focused on is generated. The UAV automatically plans additional routes based on the area list, and further collects and analyzes images of the areas that need to be focused on, so as to achieve accurate recognition and positioning of areas that meet the preset feature conditions.

[0006] In the preferred embodiment, step S1 specifically includes: S101: resampling the terrain data and vegetation density data based on the same resolution as the initial gridded flight area, including: assigning an elevation value to each grid point in the initial gridded flight area according to the terrain data to generate a preliminary three-dimensional grid; and adaptively adjusting the density of the preliminary three-dimensional grid based on the vegetation density data to increase the grid density in an area where the vegetation density is higher than a preset threshold; S102: Calculate the safe flight altitude of each grid point; the safe flight altitude is the sum of the larger value of the terrain altitude, the vegetation altitude, and the minimum safe altitude; S103: Perform three-dimensional Laplace smoothing on the grid, and update the height of the grid point by calculating the average height of the neighboring points around each grid point and combining the smoothing factor; S104: Perform collision detection to ensure that the safe flight altitude of all grid points is higher than the surrounding terrain and vegetation, and adjust the altitude of grid points with potential collision risks; S105: Generate an optimized three-dimensional grid flight area according to the grid data after adaptive density adjustment, safe flight altitude calculation, three-dimensional Laplace smoothing processing and collision detection adjustment.

[0007] In the preferred solution, in step S2, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance, which specifically includes the following steps: S201: constructing a three-dimensional pheromone matrix, wherein the three-dimensional pheromone matrix includes three-dimensional coordinate information of grid points; initializing parameters of the improved ant colony algorithm, including weights of pheromone importance factor, heuristic factor and height difference factor; calculating an adaptive state transition probability, wherein the adaptive state transition probability takes into account pheromone concentration, heuristic information and height difference factor; Performing path search according to the adaptive state transition probability; applying a local search strategy based on a 2-opt algorithm to the searched path, and optimizing the current path by exchanging two edges in the path; S202: updating the three-dimensional pheromone matrix using a dual pheromone update mechanism, wherein the dual pheromone update mechanism combines global optimal information and local elite solution information; S203: Dynamically adjust the parameters of the improved ant colony algorithm according to the current search results, including the weights of the pheromone importance factor, the heuristic factor and the height difference factor; based on the updated parameters and the three-dimensional pheromone matrix, repeatedly execute the path search, local search strategy, pheromone update and parameter adjustment steps until the preset termination condition is met; select the path with the shortest flight distance and covering all grid points from the multiple paths that meet the termination condition as the initial route.

[0008] In the preferred solution, the adaptive state transition probability formula is: ; In the formula, is the probability that ant k moves from the current grid point i to the next grid point j, Pheromone importance factor, is the weight of the heuristic factor, is the weight of the height difference factor, is the pheromone concentration from grid point i to grid point j, is the heuristic information, representing the expected degree from point i to point j; is the height difference factor, used to balance the change in flight altitude. are the coordinates of all possible next grid points.

[0009] In the preferred solution, in step S2, current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized according to the safety factors to generate a real-time route, specifically including: S21: Real-time collection of meteorological parameters, including temperature, humidity, wind speed and wind direction, and calculation of forest fire risk level data based on forest fire meteorological index; use Kriging spatial interpolation algorithm to process discrete meteorological data and forest fire risk level data to generate continuous meteorological parameters and fire risk level distribution map; S22: construct a multi-objective optimization function including a flight safety scoring function, a mission efficiency function and an energy consumption function, and solve the function to obtain an optimization result; wherein the safety scoring function takes into account the fire risk level and wind speed factors, the mission efficiency function evaluates the route coverage and flight time, and the energy consumption function is related to the flight distance and altitude change; and use the fuzzy hierarchical analysis method to determine the weight coefficient of each objective in the multi-objective optimization function; use an improved particle swarm optimization algorithm to solve the multi-objective optimization problem, and the improved particle swarm optimization algorithm introduces an adaptive inertia weight and a chaotic perturbation mechanism; S23: Based on the optimization results, the initial route is dynamically adjusted to generate a real-time route that takes safety factors into consideration.

[0010] In the preferred embodiment, step S3 specifically includes: The drone flies according to the real-time route and uses the lidar system to detect the surrounding environment in real time; Use real-time obstacle avoidance algorithm to handle dynamic obstacles encountered during flight and obtain real-time obstacle avoidance results; The flight path is adjusted accordingly based on real-time obstacle avoidance results.

[0011] In the preferred embodiment, step S5 specifically includes the following steps: S501: inputting the collected images into a deep learning model deployed on the drone; S502: The deep learning model based on the Transformer architecture is a multi-task model, which analyzes and processes the input image; the multi-task model includes an image reconstruction module, a Transformer encoder, a multi-scale feature fusion module and a task-specific head; the deep learning model is optimized using MixUp and CutMix data enhancement techniques, Dropout and L2 regularization, cosine annealing learning rate scheduling strategy and label smoothing technology; the deep learning model is trained using a multi-task loss function, and the multi-task loss function includes classification loss, detection loss, segmentation loss and image reconstruction loss; the deep learning model is used to identify areas that meet preset feature conditions.

[0012] In the preferred embodiment, step S6 specifically includes the following steps: S601: Based on the recognition result of the deep learning model, an adaptive key area generation algorithm is used to generate a list of areas that need to be focused on. The adaptive key area generation algorithm uses a weighted scoring function to calculate the area score. The weighted scoring function comprehensively considers the recognition confidence, area, fire risk level and geographical location importance, and introduces a time decay factor to dynamically adjust the key areas of concern; S602: Encode the generated list of key areas of concern in a compact binary format, where the binary format includes header information and multiple area data, each area data including center coordinates, a bounding box, a score, and an attribute flag; and send the encoded list of key areas of concern to the drone; S603: After receiving the list of key focus areas, the UAV automatically plans a supplementary route using a real-time route planning method based on the improved A algorithm; The improved A algorithm adopts a heuristic function that comprehensively considers distance, remaining power, communication quality and obstacle avoidance; implements an online replanning mechanism and sets replanning trigger conditions according to changes in power, communication quality and distance to obstacles; the UAV further collects images of areas that need to be focused on according to the planned supplementary routes, so as to achieve accurate identification and positioning of areas that meet the characteristics of the preset areas.

[0013] An unmanned aerial vehicle autonomous track planning and image acquisition system, used to implement the unmanned aerial vehicle autonomous track planning and image acquisition method, comprising: The three-dimensional grid generation module is used to obtain the preset forest boundary data and generate the initial grid flight area according to the forest boundary data; based on the initial grid flight area, the current terrain undulation and vegetation distribution are obtained in combination with the terrain data and the vegetation density data, and the three-dimensional grid flight area is generated; The route planning module is used to plan an initial route that covers all grids and has the shortest flight distance in a three-dimensional grid flight area using an improved ant colony algorithm; obtain current safety factors, which include meteorological data and forest fire risk level data, and dynamically optimize the initial route based on the safety factors to generate a real-time route; The obstacle avoidance module uses the drone to fly according to the real-time route and uses the airborne laser radar to detect the surrounding environment in real time to avoid obstacles. The image acquisition module is used to adaptively adjust the focal length, shutter speed and ISO sensitivity of the camera to the optimal state according to the current flight altitude, speed and vegetation characteristics of the acquisition area of ​​the drone, and control the camera to acquire images; An image processing module is used to input the collected images into a deep learning model deployed on the drone, analyze and process the images in real time, and identify areas that meet preset feature conditions; The recognition and feedback module is used to generate a list of areas that need special attention based on the recognition results. The UAV automatically plans additional routes based on the area list, conducts further image acquisition and analysis on the areas that need special attention, and realizes accurate recognition and positioning of areas that meet the preset feature conditions.

[0014] The present invention provides a method and system for autonomous trajectory planning and image acquisition of unmanned aerial vehicles. First, through intelligent three-dimensional grid division and dynamic route planning, the coverage and efficiency of feature area monitoring are significantly improved; secondly, the integration of real-time meteorological data and historical feature data enhances the pertinence and adaptability of acquisition, which helps to obtain high-quality images in complex and changeable environments; then the combination of real-time processing and ground deep learning analysis greatly improves the accuracy and efficiency of feature extraction; based on the adaptive supplementary sampling mechanism of key focus areas, rapid response and precise positioning of suspected feature areas are achieved, providing reliable data support for subsequent detailed analysis and decision-making. The present invention is not only suitable for various recognition tasks of preset features, but also has good scalability and flexibility, and can be adjusted and optimized according to different application requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below in conjunction with the accompanying drawings and embodiments: Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a structural diagram of the deep learning model of the present invention; Figure 3 It is a simulation diagram of the operation results of the method of the present invention. DETAILED DESCRIPTION

[0016] Example 1 like Figure 1-3 As shown, a method for autonomous trajectory planning and image acquisition of a UAV includes the following steps: S1: Obtain preset forest boundary data, and generate an initial gridded flight area according to the forest boundary data; based on the initial gridded flight area, combine terrain data and vegetation density data, and consider terrain undulation and vegetation distribution to generate a three-dimensional gridded flight area.

[0017] S2: In the three-dimensional grid flight area, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance; the current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized based on the safety factors to generate a real-time route.

[0018] S3: The drone flies according to the real-time route and uses the onboard laser radar to detect the surrounding environment in real time and perform obstacle avoidance.

[0019] S4: According to the current UAV flight altitude, speed and vegetation characteristics of the collection area, the camera's focal length, shutter speed and ISO sensitivity are adaptively adjusted to the optimal state, and the camera device is controlled to collect images.

[0020] S5: The collected images are input into the deep learning model deployed on the drone, and the images are analyzed and processed in real time to identify areas that meet the preset feature conditions.

[0021] S6: Based on the recognition results, a list of areas that need to be focused on is generated. The UAV automatically plans additional routes based on the area list, and further collects and analyzes images of the areas that need to be focused on, so as to achieve accurate recognition and positioning of areas that meet the preset feature conditions.

[0022] In this embodiment, a three-dimensional grid flight area is first generated, and an improved ant colony algorithm is used for route planning and dynamic optimization, real-time environmental detection and obstacle avoidance, and adaptive image acquisition is performed. The acquired images are input into a deep learning model deployed on the UAV, and the images are analyzed in real time to identify areas that meet preset feature conditions, and further analyze key areas to ensure that the UAV can fully cover the designated area, adapt to complex environments, and quickly respond to environmental changes and safety factors, thereby improving safety and flexibility. Combined with deep learning technology, the recognition rate of intelligent target areas is improved, and the efficiency and accuracy of data acquisition are improved.

[0023] The beneficial effects of the technical solution of the present invention are mainly reflected in the following aspects: Through intelligent 3D grid division and dynamic route planning, monitoring coverage and efficiency are improved. Combining onboard real-time processing and ground-based deep learning analysis, feature extraction accuracy is improved. Based on the adaptive supplementary sampling mechanism of key areas, rapid response and precise positioning of suspected feature areas are achieved.

[0024] In this embodiment, the preferred data specifically adopted are as follows: In step S1, the forest boundary GPS data is first obtained to generate an initial grid of 100m×100m. Combined with the DEM data (elevation 120-450m, slope 0-35 degrees) and vegetation density data (canopy density 0.4-0.8), a three-dimensional grid flight area with a vertical interval of 50m is generated by the trilinear interpolation method.

[0025] In step S3, the drone flies at a cruising speed of 15m / s according to the real-time route, and the laser radar system (detection range 100m, point cloud density 50 points / square meter) on board scans the surrounding environment in real time. When an obstacle is detected, the obstacle avoidance algorithm generates an obstacle avoidance path within 0.1 seconds to ensure a safe distance of more than 50m from the obstacle.

[0026] After the above steps S1-S6, the recognition results are obtained: the system generates a list of key areas of concern, including 3 suspected pests and diseases points and 2 fire potential points. The drone automatically descends to a flight altitude of 30m and performs secondary image acquisition of these areas at a speed of 5m / s. The acquisition interval is shortened to 2 seconds to obtain higher-precision monitoring data.

[0027] In the preferred embodiment, step S1 specifically includes: S101: Based on the same resolution as the initial gridded flight area, resampling the terrain data and the vegetation density data includes: assigning elevation values ​​to each grid point in the initial gridded flight area according to the terrain data to generate a preliminary three-dimensional grid; and performing adaptive density adjustment on the preliminary three-dimensional grid based on the vegetation density data to increase the grid density in areas where the vegetation density is higher than a preset threshold.

[0028] S102: Calculate the safe flight altitude of each grid point; the safe flight altitude is the sum of the larger value of the terrain altitude, the vegetation altitude and the minimum safe altitude.

[0029] S103: Perform three-dimensional Laplace smoothing on the grid, and update the height of the grid point by calculating the average height of neighboring points around each grid point and combining the smoothing factor.

[0030] S104: Perform collision detection to ensure that the safe flight altitude of all grid points is higher than the surrounding terrain and vegetation, and adjust the altitude of grid points with potential collision risks.

[0031] S105: Generate an optimized three-dimensional grid flight area according to the grid data after adaptive density adjustment, safe flight altitude calculation, three-dimensional Laplace smoothing processing and collision detection adjustment.

[0032] Furthermore, the system uses a rasterization algorithm to convert the forest boundary into an initial gridded flight area (50 m × 50 m). The terrain data (DEM) and vegetation density data (NDVI) are resampled to be consistent with the initial grid resolution. A three-dimensional grid is generated by assigning elevation values ​​to each grid point based on the terrain data. In areas where the vegetation density exceeds 60%, the grid is subdivided into small grids of 25 m × 25 m for adaptive density adjustment.

[0033] Calculate the safe flight altitude of the grid point, which is the larger value of the terrain height plus the vegetation height and the minimum safe altitude (20 meters). Optimize the grid through three-dimensional Laplace smoothing, calculate the average height of the neighboring points around each grid point, and update the grid point height in combination with the smoothing factor (0.5), and repeat until the expected smoothing effect is achieved.

[0034] Perform collision detection within a 100-meter range to ensure that the safe flight altitude of the grid points is higher than the surrounding terrain and vegetation. For grid points with collision risks, adjust the altitude to be higher than the highest point around and add a 5-meter safety margin. Based on the above processing, the final optimized three-dimensional grid flight area is generated.

[0035] Steps S101-S105, by introducing high-precision terrain data and vegetation density data, the generated three-dimensional grid flight area can accurately reflect the actual terrain characteristics and vegetation distribution, providing a reliable basis for subsequent route planning. Secondly, the adaptive density adjustment mechanism ensures higher sampling accuracy in densely vegetated areas, improving the pertinence and efficiency of data collection. Thirdly, the calculation of safe flight altitude and collision detection steps significantly improve flight safety and effectively reduce the risk of drone operation. Finally, three-dimensional Laplace smoothing optimizes the smoothness of the flight path, which is conducive to improving flight stability and data collection quality.

[0036] In the preferred solution, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance, which specifically includes the following steps: S201: Construct a three-dimensional pheromone matrix: the three-dimensional pheromone matrix contains the three-dimensional coordinate information of the grid points; initialize the parameters of the improved ant colony algorithm, including the weights of the pheromone importance factor, the heuristic factor and the height difference factor; calculate the adaptive state transition probability: the adaptive state transition probability takes into account the pheromone concentration, the heuristic information and the height difference factor.

[0037] According to: adaptive state transition probability to perform path search; apply the local search strategy based on 2-opt algorithm to the searched path, and optimize the current path by exchanging two edges in the path.

[0038] S202: The three-dimensional pheromone matrix is ​​updated using a dual pheromone update mechanism: the dual pheromone update mechanism combines global optimal information and local elite solution information.

[0039] S203: Dynamically adjust the parameters of the improved ant colony algorithm according to the current search results, including the weights of the pheromone importance factor, the heuristic factor and the height difference factor; based on the updated parameters and the three-dimensional pheromone matrix, repeatedly execute the path search, local search strategy, pheromone update and parameter adjustment steps until the preset termination condition is met; select the path with the shortest flight distance and covering all grid points from the multiple paths that meet the termination condition as the initial route.

[0040] Furthermore, after the three-dimensional gridded flight area is generated, the system uses an improved ant colony algorithm to plan the initial route.

[0041] First, the system constructs a three-dimensional pheromone matrix, which contains the three-dimensional coordinate information and initial pheromone concentration of each grid point.

[0042] Next, the system initializes the parameters of the improved ant colony algorithm, which include the weights of the pheromone importance factor α (such as 1.0), the heuristic factor β (such as 2.0), and the height difference factor γ (such as 0.5).

[0043] During the path search process, the system calculates the adaptive state transition probability. This probability takes into account three factors: the pheromone concentration from the current grid point to the next possible grid point, the Euclidean distance between the two points (as heuristic information), and the height difference. The system uses the weighted sum of these three factors to calculate the transition probability, with weights α, β, and γ respectively. For example, if the pheromone concentration from the current grid point to a neighbor point is 2.0, the Euclidean distance is 50 meters, and the height difference is 10 meters, a comprehensive score can be calculated, and the higher the score, the greater the transition probability.

[0044] Based on the calculated transition probabilities, the system performs a path search. Each "ant" starts from a randomly selected starting point and selects the next grid point based on the transition probability until all grid points are covered or the maximum step limit is reached. During this process, the system records the path and total flight distance of each ant.

[0045] The system optimizes the path of each ant using a local search strategy based on the 2-opt algorithm. Two non-adjacent edges in the path are randomly selected for exchange and connection. If the new path is shorter, the change is accepted. This process is repeated until no better local improvement can be found.

[0046] The system uses a dual pheromone update mechanism to update the three-dimensional pheromone matrix, combining the global optimal and local elite solution information. The pheromone concentration is increased by 0.1 for the global optimal path and 0.05 for the local optimal solution.

[0047] When no better solution is found in consecutive iterations, the system dynamically adjusts the α and β values ​​to balance exploration and utilization. Path search, local search strategy, pheromone update, and parameter adjustment are repeated until the maximum number of iterations (1000) is reached or there is no significant improvement after 50 consecutive iterations.

[0048] Finally, from the paths that meet the termination conditions, the path with the shortest flight distance and covering all grid points is selected as the initial route. The algorithm improves the convergence speed and solution quality through three-dimensional pheromone matrix, adaptive state transition probability and dual pheromone update mechanism.

[0049] In this embodiment, step S2 is the route planning stage.

[0050] In the three-dimensional grid space, the improved ant colony algorithm sets the population size to 50, the number of iterations to 200, and the pheromone volatility coefficient to 0.3. The algorithm considers constraints such as flight altitude and route spacing to generate an initial route with a total length of about 380 kilometers. Real-time acquisition of weather station data (temperature 26°C, relative humidity 55%, east wind 4m / s, moderate fire risk level), combined with Kriging spatial interpolation to generate a continuous meteorological field, dynamically optimize the route, and generate a real-time route that avoids high-risk areas. The specific data and effects are shown in Table 1 below.

[0051] Table 1 Route data and optimization effect table

[0052] In the preferred solution, the adaptive state transition probability formula is: ; In the formula, is the probability that ant k moves from the current grid point i to the next grid point j, Pheromone importance factor, is the weight of the heuristic factor, is the weight of the height difference factor, is the pheromone concentration from grid point i to grid point j, is the heuristic information, representing the expected degree from point i to point j; is the height difference factor, used to balance the change in flight altitude. are the coordinates of all possible next grid points.

[0053] In this embodiment, the algorithm can adaptively select the optimal path based on historical information (pheromones), current state (distance) and environmental characteristics (altitude changes). By dynamically adjusting the values ​​of α, β, and γ, the algorithm can balance global exploration and local development during the search process, thereby improving the efficiency of finding the optimal solution. In the preferred solution, the current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized according to the safety factors to generate a real-time route, specifically including: S21: Collect meteorological parameters in real time, including temperature, humidity, wind speed and wind direction, and calculate forest fire risk level data based on the forest fire meteorological index; use the Kriging spatial interpolation algorithm to process discrete meteorological data and forest fire risk level data to generate continuous meteorological parameters and fire risk level distribution maps.

[0054] S22: Construct a multi-objective optimization function including flight safety scoring function, mission efficiency function and energy consumption function, and solve it to obtain the optimization result; among them, the safety scoring function takes into account the fire risk level and wind speed factors, the mission efficiency function evaluates the route coverage and flight time, and the energy consumption function is related to the flight distance and altitude change; and use the fuzzy hierarchical analysis method to determine the weight coefficient of each objective in the multi-objective optimization function; adopt the improved particle swarm optimization algorithm to solve the multi-objective optimization problem, and the improved particle swarm optimization algorithm introduces adaptive inertia weight and chaotic perturbation mechanism.

[0055] S23: Based on the optimization results, the initial route is dynamically adjusted to generate a real-time route that takes safety factors into consideration.

[0056] In the preferred embodiment, step S3 specifically includes: The drone flies according to the real-time route and uses the lidar system to detect the surrounding environment in real time. Use real-time obstacle avoidance algorithms to handle dynamic obstacles encountered during flight and obtain real-time obstacle avoidance results. The flight path is adjusted accordingly based on real-time obstacle avoidance results.

[0057] Furthermore, through the above steps S2-S3, the system collects key meteorological parameters including temperature, relative humidity, wind speed and wind direction in real time through the automatic weather station network deployed in the forest area.

[0058] Based on the collected meteorological data, the system uses the forest fire meteorological index system to calculate the forest fire risk level data.

[0059] The system uses the Kriging spatial interpolation algorithm to process these discrete data and generate continuous meteorological parameters and fire risk level distribution maps.

[0060] Next, the system constructs a multi-objective optimization function that includes flight safety, mission efficiency, and energy consumption. The safety score function considers fire risk level and wind speed factors. The mission efficiency function evaluates route coverage and flight time. The energy consumption function is related to flight distance and altitude change.

[0061] The system uses fuzzy analytic hierarchy process to determine the weight coefficient of each objective in the multi-objective optimization function.

[0062] To solve multi-objective optimization problems, the system uses an improved particle swarm optimization algorithm. The algorithm introduces adaptive inertia weights. At the same time, the algorithm also introduces a chaotic perturbation mechanism, which increases the ability of the algorithm to jump out of the local optimum by applying small random perturbations to some particles in each iteration.

[0063] Based on the optimization results, the system dynamically adjusts the initial route to generate a real-time route that takes safety factors into consideration.

[0064] The system controls the drone to fly according to the real-time route, and uses the high-precision lidar system on board to detect the surrounding environment in real time.

[0065] During flight, the system uses real-time obstacle avoidance algorithms to handle dynamic obstacles that may be encountered.

[0066] Based on the real-time obstacle avoidance results, the system makes local adjustments to the flight path to ensure the safe flight of the drone.

[0067] In steps S2-S3 of this embodiment, first, by introducing real-time meteorological data and forest fire risk level data, the safety and adaptability of UAV route planning are significantly improved, enabling the system to respond quickly to environmental changes. Secondly, the design of the multi-objective optimization function achieves a balance between safety, mission efficiency and energy consumption, ensuring the comprehensive performance of the route. Thirdly, the improved particle swarm optimization algorithm improves the solution efficiency and result quality, enabling the system to complete route optimization in a short time. Finally, the introduction of the real-time obstacle avoidance system greatly enhances the autonomous flight capability of the UAV and improves flight safety. This dynamic optimization and real-time control method is not only suitable for forest monitoring, but can also be extended to other UAV application scenarios in complex environments, such as disaster monitoring, urban inspections, etc., and has broad practical value and promotion prospects.

[0068] Steps S4-S5 are to adaptively adjust the camera to the optimal state according to the current flight altitude, speed and vegetation characteristics of the collection area of ​​the drone, control the camera to collect images, input the images into the deep learning model, analyze and process the images in real time, and identify areas that meet the preset characteristic conditions; In this embodiment, during the flight of the drone, based on the current flight parameters and vegetation characteristics, the camera parameter control algorithm adjusts the focal length to 35 mm, the shutter speed to 1 / 1000 second, and the ISO sensitivity to 400. With 4K resolution, an image is captured every 5 seconds to ensure image clarity and an overlap rate of more than 85%, as shown in Table 2 below.

[0069] Table 2 Image quality control index table

[0070] like Figure 3 As shown in , the adaptive camera parameters (focal length, shutter speed, ISO) indicate that the system can adaptively adjust the camera settings according to the fire risk and environmental conditions in different areas. Dynamic adjustment enables the automatic adjustment of the imaging device by the drone in this embodiment during the monitoring process, making the data collected under different environmental conditions clearer and more accurate.

[0071] In the preferred embodiment, step S5 specifically includes the following steps: S501: Input the collected images into the deep learning model deployed on the drone.

[0072] S502: The deep learning model based on the Transformer architecture is a multi-task model that analyzes and processes the input image; the multi-task model includes an image reconstruction module, a Transformer encoder, a multi-scale feature fusion module and a task-specific head; the deep learning model is optimized using MixUp and CutMix data enhancement techniques, Dropout and L2 regularization, cosine annealing learning rate scheduling strategy and label smoothing technology; the deep learning model is trained using a multi-task loss function, which includes classification loss, detection loss, segmentation loss and image reconstruction loss; the deep learning model is used to identify areas that meet the preset feature conditions.

[0073] In this embodiment, the deep learning model first inputs the collected image into the preprocessing module for normalization and data enhancement. MixUp mixes two images with weights of 0.7 and 0.3, and CutMix replaces the 128×128 area in the 224×224 image.

[0074] The image reconstruction module adopts an encoder-decoder structure, contains 5 convolutional blocks, uses a 30% Dropout rate and a 0.0001 L2 regularization coefficient.

[0075] The Transformer encoder uses a multi-head self-attention mechanism, with 8 attention heads and a hidden layer dimension of 512. The position encoding uses sine-cosine encoding, and the encoder contains 6 layers of standard Transformer blocks.

[0076] The multi-scale feature fusion module adopts a feature pyramid structure to fuse features of different scales obtained from the Transformer encoder. The top-down path uses 1×1 convolution for channel adjustment, and the lateral connection uses 3×3 convolution for feature fusion. Finally, a feature map containing multi-scale semantic information is generated, with scales ranging from original size to 1 / 16 size features.

[0077] The task-specific head consists of three branches: classification, detection, and segmentation. The classification branch uses global average pooling followed by a fully connected layer to output class probabilities.

[0078] The training process adopts the cosine annealing learning rate strategy, with an initial learning rate of 0.001 and a periodic adjustment every 100 epochs. The label smoothing coefficient is set to 0.1. The weight of each task loss in the multi-task loss function is adaptively adjusted based on the performance on the validation set. The initial weight ratio is classification: detection: segmentation: reconstruction = 1:1:1:0.5. The specific image preprocessing results are shown in Table 3 below.

[0079] Table 3 Image preprocessing specifications and effects

[0080] The beneficial effects are: the use of multi-task learning improves the generalization ability of the model; the Transformer structure enhances the feature extraction effect; multi-scale feature fusion improves the detection accuracy of targets of different scales; data enhancement and adaptive loss weight optimization ensure the stability of model training and achieve efficient forest monitoring.

[0081] In the preferred embodiment, step S6 specifically includes the following steps: S601: Based on the recognition results of the deep learning model, an adaptive key area generation algorithm is used to generate a list of areas that need to be focused on. The adaptive key area generation algorithm uses a weighted scoring function to calculate the area score. The weighted scoring function comprehensively considers the recognition confidence, area area, fire risk level and geographical location importance, and introduces a time decay factor to dynamically adjust the key focus areas.

[0082] S602: Encode the generated key focus area list in a compact binary format, where the binary format includes header information and multiple area data, each area data includes center coordinates, a bounding box, a score, and an attribute flag; send the encoded key focus area list to the drone.

[0083] S603: After receiving the list of key focus areas, the UAV automatically plans a supplementary route using a real-time route planning method based on the improved A algorithm.

[0084] The improved A algorithm adopts a heuristic function that comprehensively considers distance, remaining power, communication quality and obstacle avoidance; an online replanning mechanism is implemented to set replanning trigger conditions according to changes in power, communication quality and distance to obstacles; the UAV conducts further image collection of areas that need to be focused on according to the planned supplementary routes, so as to achieve accurate identification and positioning of areas that meet the characteristics of the preset areas.

[0085] like Figure 3 As shown in the figure, it includes path planning and fire risk avoidance, error tolerance control, accurate monitoring and secondary collection of high-priority areas, and adaptive adjustment of cameras. The forest area simulated by test data includes data such as terrain height, vegetation density and fire risk level. The information in the figure specifically includes: Figure 3The middle green dotted line is the ideal path, which runs from the lower left corner to the upper right corner of the map, marking the ideal route of the drone when there is no interference from fire risk. The red solid line is the actual path after optimization in this embodiment. The actual path deviation is calculated to ensure that the drone effectively avoids high-risk points of fire risk without affecting the overall smoothness of the route. The optimized path verifies the effectiveness of route optimization by avoiding high-risk fire areas, and marks high-priority target points in fire risk areas to further verify the drone method's identification and precise positioning of key areas.

[0086] The colored background includes: 1) Altitude (first color bar): Use terrain height data to simulate terrain changes. The color changes represent different altitudes, from low to high, reflecting the undulating terrain that the route may involve.

[0087] 2) Vegetation density (second color bar): Green transparent areas indicate different densities of vegetation cover, which increase the complexity of the flight path.

[0088] 3) Fire risk level (third color bar): Displays the distribution of fire risk levels, with colors ranging from green (low fire risk) to red (high fire risk), so that the drone path can avoid fire risks.

[0089] 4) Error range (orange filled area): Based on the error range of 5 meters above and below the ideal path, it represents the tolerance error range of the route, showing the tolerance deviation space between the ideal path and the optimized path, and verifying that the drone’s navigation system can accurately avoid high-risk areas within the preset error tolerance.

[0090] Further includes: Secondary collection path (purple dashed line): The secondary collection path around the high-priority area is used for subsequent precise monitoring, which shows that drones need to pay special attention in key areas and conduct secondary collection to ensure the accuracy of fire risk monitoring data.

[0091] Camera parameters (focal length, shutter, ISO): Near the high priority area, each area shows the adaptive acquisition parameters of the camera; the differentiated setting of the parameters indicates the system's ability to adaptively adjust the acquisition under different environmental conditions, enabling the drone to collect high-quality images and data according to environmental changes.

[0092] High priority areas (red marks): Several areas with higher fire risk levels were selected on the ideal path, marked as pink dots, and avoided by the actual route. These areas were set as target areas of focus for detecting fire risk conditions.

[0093] In summary, it can be concluded that the optimized path of this embodiment effectively avoids high-priority areas and realizes the precise avoidance requirements for fire risk areas; the addition of the secondary collection path shows that this embodiment automatically detects and performs secondary collection tasks in required areas according to the fire risk level, further improving the accuracy of fire risk monitoring and greatly improving the accurate recognition rate and positioning accuracy that meet the characteristics of the preset areas.

[0094] In this embodiment, based on the recognition results of the ground station deep learning model, an adaptive key area generation algorithm is used to generate a list of areas that need to be focused on. The algorithm uses a weighted scoring function that takes into account multiple factors: the recognition confidence reflects the reliability of the model's judgment of the area, usually between 0 and 1; the area of ​​the area takes into account the scale of the problem; the importance of the geographical location takes into account the strategic value of the area, and the algorithm introduces a time decay factor to dynamically adjust the key focus areas.

[0095] After the list of key areas of interest is generated, it is transmitted to the drone: it is encoded in a compact binary format. This format includes header information and multiple area data. The header information may contain version number, timestamp, total number of areas, etc. Each area data contains the following information: center coordinates, bounding box, score, and attribute flags.

[0096] After receiving the list of key areas of interest, the drone needs to automatically plan a supplementary route for further image acquisition. Here, a real-time route planning method based on the improved A algorithm is used. The improved A algorithm uses a heuristic function that comprehensively considers multiple factors to evaluate the path cost. These factors include: distance, remaining battery, communication quality, and obstacle avoidance.

[0097] The route planning algorithm also implements an online re-planning mechanism, which sets re-planning trigger conditions based on changes in battery level, communication quality, and distance to obstacles.

[0098] According to the planned supplementary route, the drone will further collect images of areas that need special attention. During this process, the drone may adjust its flight altitude and speed to obtain clearer images.

[0099] In step S6 of this embodiment: the adaptive key area generation algorithm improves monitoring efficiency and accuracy; the compact binary coding reduces the amount of data transmission; the route planning based on the improved A algorithm ensures flight safety and monitoring effect; the online re-planning mechanism enhances the system adaptability.

[0100] Example 2 Further illustrate with reference to Example 1, Figure 2 The structure shown is a UAV autonomous track planning and image acquisition system, which is used to implement the UAV autonomous track planning and image acquisition method in Example 1, including: The three-dimensional grid generation module is used to obtain preset forest boundary data and generate an initial gridded flight area based on the forest boundary data; based on the initial gridded flight area, the current terrain undulation and vegetation distribution are obtained in combination with terrain data and vegetation density data to generate a three-dimensional gridded flight area.

[0101] The route planning module is used to plan an initial route that covers all grids and has the shortest flight distance in a three-dimensional grid flight area using an improved ant colony algorithm; obtain current safety factors, which include meteorological data and forest fire risk level data, and dynamically optimize the initial route based on safety factors to generate a real-time route.

[0102] The obstacle avoidance module uses the drone to fly according to the real-time route and adopts the airborne laser radar to detect the surrounding environment in real time to perform obstacle avoidance.

[0103] The image acquisition module is used to adaptively adjust the camera's focal length, shutter speed, and ISO sensitivity to the optimal state according to the current UAV's flight altitude, speed, and vegetation characteristics of the acquisition area, and control the camera to acquire images.

[0104] The image processing module is used to input the collected images into the deep learning model deployed on the drone, perform real-time analysis and processing on the images, and identify areas that meet the preset feature conditions.

[0105] The recognition and feedback module is used to generate a list of areas that need special attention based on the recognition results. The UAV automatically plans additional routes based on the area list, conducts further image acquisition and analysis on the areas that need special attention, and realizes accurate recognition and positioning of areas that meet the preset feature conditions.

[0106] The above embodiments are only preferred technical solutions of the present invention and should not be regarded as limiting the present invention. The protection scope of the present invention shall be the technical solutions recorded in the claims, including equivalent replacement solutions of the technical features in the technical solutions recorded in the claims. That is, equivalent replacement improvements within this scope are also within the protection scope of the present invention.

Claims

1. A method for autonomous trajectory planning and image acquisition of an unmanned aerial vehicle, characterized in that: The following steps are involved: S1: Obtain preset forest boundary data, generate an initial gridded flight area according to the forest boundary data, obtain the current terrain undulation and vegetation distribution based on the initial gridded flight area, and generate a three-dimensional gridded flight area by combining terrain data and vegetation density data; S2: In the three-dimensional grid flight area, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance; the current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized according to the safety factors to generate a real-time route; S3: The drone flies according to the real-time route and uses the onboard laser radar to detect the surrounding environment in real time and perform obstacle avoidance. S4: According to the current UAV flight altitude, speed and vegetation characteristics of the collection area, the camera's focal length, shutter speed and ISO sensitivity are adaptively adjusted to the optimal state, and the camera device is controlled to collect images; S5: Input the collected images into the deep learning model deployed on the drone, analyze and process the images in real time, and identify areas that meet the preset feature conditions; S6: Based on the recognition results, a list of areas that need to be focused on is generated. The UAV automatically plans additional routes based on the area list, and further collects and analyzes images of the areas that need to be focused on, so as to achieve accurate recognition and positioning of areas that meet the preset feature conditions.

2. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicles according to claim 1, characterized in that: Step S1 specifically includes: S101: resampling the terrain data and vegetation density data based on the same resolution as the initial gridded flight area, including: assigning an elevation value to each grid point in the initial gridded flight area according to the terrain data to generate a preliminary three-dimensional grid; and adaptively adjusting the density of the preliminary three-dimensional grid based on the vegetation density data to increase the grid density in an area where the vegetation density is higher than a preset threshold; S102: Calculate the safe flight altitude of each grid point; the safe flight altitude is the sum of the larger value of the terrain altitude, the vegetation altitude, and the minimum safe altitude; S103: Perform three-dimensional Laplace smoothing on the grid, and update the height of the grid point by calculating the average height of the neighboring points around each grid point and combining the smoothing factor; S104: Perform collision detection to ensure that the safe flight altitude of all grid points is higher than the surrounding terrain and vegetation, and adjust the altitude of grid points with potential collision risks; S105: Generate an optimized three-dimensional grid flight area according to the grid data after adaptive density adjustment, safe flight altitude calculation, three-dimensional Laplace smoothing processing and collision detection adjustment.

3. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicle according to claim 1, characterized in that: In step S2, an improved ant colony algorithm is used to plan an initial route that covers all grids and has the shortest flight distance, which specifically includes the following steps: S201: constructing a three-dimensional pheromone matrix, wherein the three-dimensional pheromone matrix includes three-dimensional coordinate information of grid points; initializing parameters of the improved ant colony algorithm, including weights of pheromone importance factor, heuristic factor and height difference factor; calculating an adaptive state transition probability, wherein the adaptive state transition probability takes into account pheromone concentration, heuristic information and height difference factor; Performing path search according to the adaptive state transition probability; applying a local search strategy based on a 2-opt algorithm to the searched path, and optimizing the current path by exchanging two edges in the path; S202: updating the three-dimensional pheromone matrix using a dual pheromone update mechanism, wherein the dual pheromone update mechanism combines global optimal information and local elite solution information; S203: Dynamically adjust the parameters of the improved ant colony algorithm according to the current search results, including the weights of the pheromone importance factor, the heuristic factor and the height difference factor; based on the updated parameters and the three-dimensional pheromone matrix, repeatedly execute the path search, local search strategy, pheromone update and parameter adjustment steps until the preset termination condition is met; select the path with the shortest flight distance and covering all grid points from the multiple paths that meet the termination condition as the initial route.

4. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicle according to claim 3, characterized in that: The formula for adaptive state transition probability is: ； In the formula, is the probability that ant k moves from the current grid point i to the next grid point j, Pheromone importance factor, is the weight of the heuristic factor, is the weight of the height difference factor, is the pheromone concentration from grid point i to grid point j, is the heuristic information, representing the expected degree from point i to point j; is the height difference factor, used to balance the change in flight altitude. are the coordinates of all possible next grid points.

5. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicles according to claim 1, characterized in that: In step S2, the current safety factors are obtained, including meteorological data and forest fire risk level data, and the initial route is dynamically optimized according to the safety factors to generate a real-time route, specifically including: S21: Real-time collection of meteorological parameters, including temperature, humidity, wind speed and wind direction, and calculation of forest fire risk level data based on forest fire meteorological index; use Kriging spatial interpolation algorithm to process discrete meteorological data and forest fire risk level data to generate continuous meteorological parameters and fire risk level distribution map; S22: construct a multi-objective optimization function including a flight safety scoring function, a mission efficiency function and an energy consumption function, and solve the function to obtain an optimization result; wherein the safety scoring function takes into account the fire risk level and wind speed factors, the mission efficiency function evaluates the route coverage and flight time, and the energy consumption function is related to the flight distance and altitude change; and use the fuzzy hierarchical analysis method to determine the weight coefficient of each objective in the multi-objective optimization function; use an improved particle swarm optimization algorithm to solve the multi-objective optimization problem, and the improved particle swarm optimization algorithm introduces an adaptive inertia weight and a chaotic perturbation mechanism; S23: Based on the optimization results, the initial route is dynamically adjusted to generate a real-time route that takes safety factors into consideration.

6. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicle according to claim 5, characterized in that: Step S3 specifically includes: The drone flies according to the real-time route and uses the lidar system to detect the surrounding environment in real time; Use real-time obstacle avoidance algorithm to handle dynamic obstacles encountered during flight and obtain real-time obstacle avoidance results; The flight path is adjusted accordingly based on real-time obstacle avoidance results.

7. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicles according to claim 1, characterized in that: Step S5 specifically includes the following steps: S501: inputting the collected images into a deep learning model deployed on the drone; S502: The deep learning model based on the Transformer architecture is a multi-task model, which analyzes and processes the input image; the multi-task model includes an image reconstruction module, a Transformer encoder, a multi-scale feature fusion module and a task-specific head; the deep learning model is optimized using MixUp and CutMix data enhancement techniques, Dropout and L2 regularization, cosine annealing learning rate scheduling strategy and label smoothing technology; the deep learning model is trained using a multi-task loss function, and the multi-task loss function includes classification loss, detection loss, segmentation loss and image reconstruction loss; the deep learning model is used to identify areas that meet preset feature conditions.

8. The method for autonomous trajectory planning and image acquisition of unmanned aerial vehicle according to claim 1, characterized in that: Step S6 specifically includes the following steps: S601: Based on the recognition result of the deep learning model, an adaptive key area generation algorithm is used to generate a list of areas that need to be focused on. The adaptive key area generation algorithm uses a weighted scoring function to calculate the area score. The weighted scoring function comprehensively considers the recognition confidence, area, fire risk level and geographical location importance, and introduces a time decay factor to dynamically adjust the key areas of concern; S602: Encode the generated list of key areas of concern in a compact binary format, where the binary format includes header information and multiple area data, each area data including center coordinates, a bounding box, a score, and an attribute flag; and send the encoded list of key areas of concern to the drone; S603: After receiving the list of key focus areas, the UAV automatically plans a supplementary route using a real-time route planning method based on the improved A algorithm; The improved A algorithm adopts a heuristic function that comprehensively considers distance, remaining power, communication quality and obstacle avoidance; implements an online replanning mechanism and sets replanning trigger conditions according to changes in power, communication quality and distance to obstacles; the UAV further collects images of areas that need to be focused on according to the planned supplementary routes, so as to achieve accurate identification and positioning of areas that meet the characteristics of the preset areas.

9. An autonomous trajectory planning and image acquisition system for unmanned aerial vehicles, characterized in that: The method for autonomous trajectory planning and image acquisition of a UAV according to any one of claims 1 to 8 comprises: The three-dimensional grid generation module is used to obtain the preset forest boundary data and generate the initial grid flight area according to the forest boundary data; based on the initial grid flight area, the current terrain undulation and vegetation distribution are obtained in combination with the terrain data and the vegetation density data, and the three-dimensional grid flight area is generated; The route planning module is used to plan an initial route that covers all grids and has the shortest flight distance in a three-dimensional grid flight area using an improved ant colony algorithm; obtain current safety factors, which include meteorological data and forest fire risk level data, and dynamically optimize the initial route based on the safety factors to generate a real-time route; The obstacle avoidance module uses the drone to fly according to the real-time route and uses the airborne laser radar to detect the surrounding environment in real time to avoid obstacles. The image acquisition module is used to adaptively adjust the focal length, shutter speed and ISO sensitivity of the camera to the optimal state according to the current flight altitude, speed and vegetation characteristics of the acquisition area of ​​the drone, and control the camera to acquire images; An image processing module is used to input the collected images into a deep learning model deployed on the drone, analyze and process the images in real time, and identify areas that meet preset feature conditions; The recognition and feedback module is used to generate a list of areas that need special attention based on the recognition results. The UAV automatically plans additional routes based on the area list, conducts further image acquisition and analysis on the areas that need special attention, and realizes accurate recognition and positioning of areas that meet the preset feature conditions.

Images