A method and system for collecting data of an offshore wind turbine

Abstract

The present invention relates to a method and system for collecting data of offshore wind turbines, the method comprising: dividing a target offshore wind farm into a number of inspection areas, obtaining structural parameters and terrain data of wind turbines in each inspection area; constructing a UAV flight trajectory of each inspection area according to the structural parameters and terrain data of the wind turbines, and presetting a circling radius and a camera angle; controlling the UAV to circling the wind turbines in each inspection area based on the flight trajectory, circling radius and camera angle, and collecting images and point cloud data of the wind turbines during the circling flight; constructing a three-dimensional model of the wind turbine based on the images and point cloud data of the wind turbines, and constructing a three-dimensional map in combination with the terrain data, integrating the flight trajectory into the three-dimensional map, and completing the data collection of the target offshore wind farm. The present invention can realize autonomous intelligent inspection of offshore wind turbines.

Description

Technical Field

[0001] The present invention relates to the technical field of offshore wind power, and in particular to a method and system for collecting data of an offshore wind turbine generator set. Background Art

[0002] The wide distribution of offshore wind turbines and the large number of units bring many challenges to drone inspection operations. On the one hand, the endurance and range of a single drone are limited, making it difficult to complete large-scale inspection tasks in a short period of time; on the other hand, harsh offshore environmental conditions such as strong winds and typhoons will directly affect the flight posture and safety of drones. In addition, offshore wind turbines are different from land units. Their unique layout and environment require drones to have more intelligent obstacle avoidance and route planning capabilities.

[0003] When multiple drones are working together for inspection, how to achieve efficient task allocation and dynamic scheduling is also a major technical challenge. Wind turbine failures are somewhat random and sudden, and drones need to dynamically adjust inspection tasks based on real-time unit status, respective locations and other factors. However, during the scheduling process, problems such as repeated inspections and missed units by drones may occur, affecting overall operational efficiency. At the same time, the availability of drones is reduced in severe weather. How to maximize the use of limited drone resources and reasonably allocate inspection tasks is also an optimization problem that needs to be solved urgently. Therefore, the present invention proposes a method and system for collecting data from offshore wind turbines. Summary of the invention

[0004] The purpose of the present invention is to solve the above technical problems and provide a method and system for collecting data of offshore wind turbines, so as to realize autonomous intelligent inspection of wind turbines, greatly improve inspection efficiency and accuracy, and reduce the safety risks and costs of manual inspection.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A method for collecting data of an offshore wind turbine generator system, comprising:

[0007] Divide the target offshore wind farm into several inspection areas, and obtain the structural parameters and terrain data of the wind turbines in each inspection area;

[0008] Construct the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine and the terrain data, and preset the circling radius and the camera angle;

[0009] Based on the flight trajectory, circling radius and aircraft position angle, the drone is controlled to circumvent the wind turbines in each inspection area, and images and point cloud data of the wind turbines are collected during the circling flight;

[0010] A three-dimensional model of the wind turbine is constructed based on the wind turbine image and point cloud data, and a three-dimensional map is constructed in combination with the terrain data. The flight trajectory is integrated into the three-dimensional map to complete data collection of the target offshore wind farm.

[0011] Optionally, dividing the target offshore wind farm into a number of inspection areas includes:

[0012] Acquire geographic information data and a wind turbine distribution map of the target offshore wind farm, and construct a topological structure map, wherein the topological structure map uses wind turbines as nodes and distances between wind turbines as edges;

[0013] Obtain the maximum range and load capacity of the drone, and calculate the maximum single inspection range and the maximum number of wind turbines in a single inspection mission;

[0014] A graph theory algorithm is used, and the maximum single inspection range of the UAV and the maximum number of wind turbines at a single time are taken as constraints. The topological structure diagram is divided to obtain a number of connected subgraphs, that is, a number of inspection areas.

[0015] Optionally, constructing the UAV flight trajectory of each inspection area according to the wind turbine structure parameters and terrain data includes:

[0016] Constructing a three-dimensional scene model according to the structural parameters of the wind turbine generator set and terrain data;

[0017] Using an artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain a flight trajectory;

[0018] The flight trajectory is locally optimized based on collision risk detection, and the flight trajectory is globally optimized based on a particle swarm optimization algorithm to obtain a final flight trajectory.

[0019] Optionally, using an artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain a flight trajectory includes:

[0020] Acquire the positions of wind turbines and terrain obstacles based on the three-dimensional scene model;

[0021] Using the position of the wind turbine as the target point in the artificial potential field algorithm, and using the position of the terrain obstacle as the obstacle point in the artificial potential field algorithm;

[0022] The flight trajectory of the UAV is planned by dynamically adjusting the distance between the UAV and the target point and the obstacle point.

[0023] Optionally, the method for presetting the surround radius and the camera angle is:

[0024] Get the camera parameters and lidar parameters carried by the drone;

[0025] Simulate flight in the three-dimensional scene model based on the final flight trajectory, and calculate the data acquisition efficiency and quality at different circumferential radii and camera positions according to the camera parameters and lidar parameters;

[0026] Based on the data acquisition efficiency and quality, and combined with flight safety, determine the optimal circumferential radius and the optimal camera position.

[0027] Optionally, controlling the drone to fly around the wind turbines in each inspection area based on the flight trajectory, circumferential radius, and camera position includes:

[0028] Discretize the flight trajectory into a number of waypoints, where the waypoints contain position, attitude, and speed information. Among them, the position and attitude information are calculated through the circumferential radius and camera position, and the speed information is a preset speed;

[0029] After controlling the drone to take off, obtain the real-time position, real-time attitude, and real-time speed of the drone, and compare the real-time position, real-time attitude, and real-time speed with the position, attitude, and speed information in the next waypoint to obtain the real-time heading adjustment amount and the real-time speed adjustment amount;

[0030] Continuously correct the heading and speed of the drone through the real-time heading adjustment amount and the real-time speed adjustment amount until the flight trajectory ends.

[0031] Optionally, constructing a three-dimensional model of the wind turbine based on the wind turbine image and point cloud data includes:

[0032] Preprocess the wind turbine image and point cloud data respectively, and then perform fusion;

[0033] Extract the component features of the wind turbine, as well as the geometric profiles and position information of each component, based on the fused wind turbine image and point cloud data, and construct the three-dimensional model of the wind turbine.

[0034] Optionally, preprocessing the wind turbine image and point cloud data respectively, and then performing fusion includes:

[0035] Stitch multiple wind turbine images into a panoramic wind turbine image, and then perform distortion correction processing on the panoramic wind turbine image to obtain the processed panoramic image;

[0036] Perform denoising and point cloud smoothing on the point cloud data to obtain the processed point cloud data;

[0037] Register and map the processed panoramic image with the processed point cloud data to obtain a panoramic image containing point cloud data.

[0038] To further achieve the above object, the present invention also provides an offshore wind turbine data acquisition system, comprising:

[0039] The inspection area division module is used to divide the target offshore wind farm into several inspection areas and obtain the structural parameters and terrain data of the wind turbines in each inspection area;

[0040] A flight information construction module is used to construct the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine set and the terrain data, and to preset the circling radius and the aircraft position angle;

[0041] The inspection data acquisition module is used to control the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, the circling radius and the aircraft position angle, and to collect images and point cloud data of the wind turbines during the circling flight;

[0042] The inspection data processing module is used to construct a three-dimensional model of the wind turbine based on the wind turbine image and point cloud data, and to construct a three-dimensional map in combination with the terrain data, and to integrate the flight trajectory into the three-dimensional map to complete the data collection of the target offshore wind farm.

[0043] The beneficial effects of the present invention are:

[0044] The present invention first obtains the geographic information and wind turbine distribution map of the target offshore wind farm, and divides the wind farm into several inspection areas according to the maximum range and load capacity of the drone, using a graph theory algorithm. The number and spacing of the units in each area meet the operating capacity of a single drone. For each inspection area, the present invention obtains the structural parameters and surrounding terrain data of the wind turbine, plans the inspection path through the improved artificial potential field method, and generates an optimal flight trajectory that covers all wind turbines and avoids terrain obstacles. During the circling flight, the drone continuously shoots multi-angle, high-resolution unit images through the onboard camera, and the laser radar scans the wind turbine in real time to obtain dense and uniform point cloud data, transmits the acquired unit images and point cloud data to the ground station, builds a realistic three-dimensional map in the virtual scene, and integrates the drone flight trajectory, obstacle avoidance strategy and other information into the map. The present invention can realize autonomous intelligent inspection of wind turbines, greatly improve inspection efficiency and accuracy, and reduce the safety risks and costs of manual inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 The present invention is a flow chart of a method for collecting data of an offshore wind turbine generator system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0049] Although the application of drone inspection in offshore wind turbines has achieved remarkable results, it still faces some technical and environmental challenges: (1) Battery life: Offshore wind turbines are usually located in remote sea areas, and the endurance of drones is an important limiting factor. Developing drones with long flight time and improving battery technology is a future direction. (2) Climate impact: Offshore wind turbines are usually located in extreme climate conditions. How to ensure the stability of drones under weather conditions such as strong winds and heavy rains is a technical challenge. (3) Data storage and processing: The storage and real-time processing of massive data requires drones to have powerful computing and transmission capabilities. The use of drone inspections to collect data from offshore wind turbines is gradually replacing traditional manual inspection methods. Through advanced technologies such as high-definition visual imaging, thermal imaging, and lidar, drones can efficiently and accurately collect various types of data from wind turbines, help detect the health status of wind turbines, reduce failures and downtime, and improve the operational efficiency and safety of wind farms. With the continuous advancement of technology, the application prospects of drone inspections will be broader.

[0050] Therefore, this embodiment provides a method for collecting data of an offshore wind turbine generator set, comprising:

[0051] Divide the target offshore wind farm into several inspection areas, and obtain the structural parameters and terrain data of the wind turbines in each inspection area;

[0052] The flight trajectory of the drone in each inspection area is constructed according to the structural parameters of the wind turbine and the terrain data, and the circling radius and the camera angle are preset;

[0053] Based on the flight trajectory, circling radius and camera angle, the drone is controlled to circle the wind turbines in each inspection area, and images and point cloud data of the wind turbines are collected during the circling process.

[0054] Construct a 3D model of a wind turbine based on the wind turbine images and point cloud data, construct a 3D map in combination with the terrain data, fuse the flight trajectory into the 3D map, and complete the data collection of the target offshore wind farm.

[0055] Specifically, in this embodiment, the geographical information and the wind turbine distribution map of the target offshore wind farm are first obtained. According to the maximum flight range and payload capacity of the UAV, the wind farm is divided into several inspection areas by using the graph theory algorithm, and the number of wind turbines and the spacing within each area meet the operation ability of a single UAV. For each inspection area, the structural parameters of the wind turbine and the surrounding terrain data are obtained in this embodiment. An improved artificial potential field method is used to plan the inspection path, and an optimal flight trajectory covering all wind turbines and avoiding terrain obstacles is generated. During the circumferential flight, the UAV continuously takes multi-angle and high-resolution images of the wind turbine through the on-board camera, and at the same time the lidar scans the wind turbine in real time to obtain dense and uniform point cloud data. The obtained wind turbine images and point cloud data are transmitted to the ground station, and a realistic 3D map is constructed in the virtual scene, and information such as the UAV flight trajectory and obstacle avoidance strategy is fused into the map. This embodiment can realize the autonomous and intelligent inspection of wind turbines, greatly improve the inspection efficiency and accuracy, and reduce the safety risks and costs of manual inspection.

[0056] Next, in combination with Figure 1 A data collection method for an offshore wind turbine proposed in this embodiment is described in detail, which specifically includes the following steps:

[0057] Step 1: Divide the target offshore wind farm into several inspection areas, and obtain the structural parameters of the wind turbines and the terrain data in each inspection area.

[0058] Further, dividing the target offshore wind farm into several inspection areas includes:

[0059] Obtain the geographical information data and the wind turbine distribution map of the target offshore wind farm, and construct a topological structure diagram, where the topological structure diagram takes the wind turbines as nodes and the distances between the wind turbines as edges;

[0060] Obtain the maximum flight range and payload capacity of the UAV, and calculate the maximum single inspection range and the maximum single wind turbine quantity of the UAV in the inspection task;

[0061] Use the graph theory algorithm, take the maximum single inspection range and the maximum single wind turbine quantity of the UAV as constraint conditions, divide the topological structure diagram, and obtain several connected subgraphs, that is, several inspection areas.

[0062] At the same time, if the number of obtained inspection areas exceeds the total number of UAVs, the division result needs to be adjusted by an optimization algorithm to balance the inspection task volume of each area and the workload of the UAV.

[0063] Specifically, in this embodiment, the geographic information data of the wind farm is obtained, including the latitude and longitude coordinates, altitude, topography, etc. of each wind turbine, as well as the distribution map data of the wind turbine. These data can be obtained through a wind farm management system, a GIS platform or a field survey. For example, a wind farm has 100 wind turbines, and the precise location information and surrounding terrain data of each unit can be obtained. These data are converted into a data format suitable for graph theory algorithm processing, including an adjacency matrix or an adjacency list. The adjacency matrix is ​​a two-dimensional array used to represent the connection relationship between nodes in the graph. For example, if there is a path connection between unit 1 and unit 2, the corresponding position element value in the adjacency matrix is ​​1, otherwise it is 0. Construct a topological structure diagram of the wind farm, that is, the wind turbines are regarded as nodes in the graph, and the distances between the units are regarded as edges in the graph, thereby abstracting the wind farm into a graph theory model. According to the maximum range and load capacity parameters of the drone, the maximum inspection range and the maximum number of wind turbines that can be inspected by a single drone in a flight mission are calculated. For example, the maximum flight distance is 200 kilometers, the maximum load is 10 kilograms, and the UAV is flying at a cruising speed, and the inspection time of each unit is 30 minutes, then the maximum inspection time of a single flight mission is about 5 hours. Considering the round-trip flight time and standby time, a single UAV can inspect up to 8 wind turbines at a time. Using graph theory algorithms, including depth-first search or breadth-first search, the topological structure graph of the wind farm is divided into several connected subgraphs, each of which represents an inspection area. By setting appropriate threshold conditions, the scale and density of each inspection area are controlled to ensure that it meets the operational capacity limit of a single UAV. For example, the wind farm is divided into 5 inspection areas, the number of units in each inspection area does not exceed 20, and the distance between units does not exceed 10 kilometers. This can avoid a single mission exceeding the maximum range and inspection time limit of the UAV. If the number of inspection areas obtained by division exceeds the total number of UAVs, the division results need to be adjusted through the optimization algorithm to balance the inspection task volume of each area and the workload of the UAV. Two smaller areas can be merged into one, or a larger area can be split into two smaller areas to ensure that each drone is responsible for a roughly equal amount of inspection tasks.

[0064] Step 2: Construct the UAV flight trajectory for each inspection area based on the wind turbine structural parameters and terrain data, and preset the circling radius and camera angle.

[0065] Furthermore, the UAV flight trajectory of each inspection area is constructed according to the structural parameters of the wind turbine and the terrain data, including:

[0066] Construct a three-dimensional scene model based on wind turbine structural parameters and terrain data;

[0067] The artificial potential field algorithm is used for path planning in a three-dimensional scene model to obtain a flight trajectory;

[0068] Based on collision risk detection, the flight trajectory is locally optimized, and based on the particle swarm optimization algorithm, the flight trajectory is globally optimized to obtain the final flight trajectory.

[0069] Among them, using the artificial potential field algorithm for path planning in a three-dimensional scene model to obtain a flight trajectory includes:

[0070] Based on the three-dimensional scene model, the positions of wind turbines and terrain obstacles are obtained;

[0071] The position of the wind turbine is used as the target point in the artificial potential field algorithm, and the position of the terrain obstacle is used as the obstacle point in the artificial potential field algorithm;

[0072] By dynamically adjusting the distances between the UAV and the target point and the obstacle point, the flight trajectory of the UAV is planned.

[0073] Based on collision risk detection, the flight trajectory is locally optimized: According to the structural parameters of the wind turbine, collision detection and safety distance judgment are carried out on the flight trajectory in the three-dimensional scene model, and local optimization and adjustment are made to the paths with collision risks or too close distances.

[0074] Based on the particle swarm optimization algorithm, the flight trajectory is globally optimized: The particle swarm optimization algorithm is used to globally optimize the flight trajectory, with the shortest flight distance as the optimization goal, to obtain an optimal flight trajectory that takes into account coverage, safety, and efficiency.

[0075] Specifically, in this embodiment, structural parameters such as the precise position coordinates, height, and blade length of the wind turbines in each inspection area are obtained, as well as the terrain elevation data around the wind turbines, to construct a three-dimensional scene model. For example, the position coordinates (longitude, latitude, and altitude), tower height, blade length of 5 wind turbines in area A of a wind farm, and the elevation data (DEM data) of the surrounding terrain are obtained, and a three-dimensional scene model is constructed using these data in software. The model includes the three-dimensional models of the wind turbines and the terrain surface. This can visually display the actual situation of the wind farm and provide a basis for subsequent path planning. The positions of the wind turbines are used as target points in the artificial potential field algorithm, and the positions of the terrain obstacles are used as obstacle points in the artificial potential field algorithm to improve the artificial potential field function, so that the drone is subject to the gravitational force of the target point and the repulsive force of the obstacles during flight. For example, the positions of the 5 wind turbines in area A are set as target points to generate gravitational force on the drone; the positions of obstacles such as mountains and buildings in the terrain that are higher than the set threshold are set as obstacle points to generate repulsive force on the drone. The improved artificial potential field function can dynamically adjust the magnitudes of the gravitational and repulsive forces according to the distances between the drone and the target point and the obstacle point, guiding the drone to fly towards the target point and avoid obstacles. This method can effectively handle the path planning problem in complex environments. Through the improved artificial potential field algorithm for path planning, a flight path that covers all wind turbines and avoids terrain obstacles is obtained, and this path can be used as the basis for subsequent optimization. According to the structural parameters such as the height and blade length of the wind turbines, collision detection and safety distance judgment are performed on the flight path in the three-dimensional scene, and local optimization adjustments are made to the paths with collision risks or too-close distances. For example, in the scene of area A, check whether the initial path collides with the blades or tower of the wind turbine, or whether it meets the set safety distance requirements. If a collision risk or too-close distance is found, local adjustments are made to the path, such as changing the waypoint position or flight height, to ensure the flight safety of the drone. The particle swarm optimization algorithm is used to globally optimize the flight path, with the shortest flight distance as the optimization goal, to obtain an optimal flight trajectory that takes into account coverage, safety, and efficiency. For example, the initial path in area A is used as the input, and the particle swarm optimization algorithm is used to globally optimize the path, with the total flight distance as the objective function, while considering the safety distance and obstacle avoidance constraints, and finally an optimal flight trajectory that covers all wind turbines, avoids obstacles, and has the shortest flight distance is obtained.

[0076] Furthermore, the method for presetting the circumradius and the camera position angle is as follows:

[0077] Obtain the camera parameters and lidar parameters carried by the drone;

[0078] Simulate the flight in the 3D scene model based on the final flight trajectory, and calculate the data acquisition efficiency and quality at different circumradius and camera position angles according to the camera parameters and lidar parameters;

[0079] Based on the data acquisition efficiency and quality, and combined with flight safety, determine the optimal circumradius and the optimal camera position angle.

[0080] Specifically, in this embodiment, the simulated flight of the UAV is carried out according to the 3D scene model and the final flight trajectory, and the data acquisition efficiency and quality at different circumradius and camera position angles during the simulated flight are calculated according to the camera parameters (such as resolution, focal length, pixel size) and lidar parameters (such as scanning frequency, scanning angle, point cloud density) carried by the UAV. For example, assuming that the tower height of a wind turbine is 100 meters and the blade length is 80 meters, the UAV can be set to fly around at a position 150 meters away from the center of the tower and 120 meters in height, and the camera position angle is adjusted to simulate the data acquisition effect at different perspectives. The camera parameters and lidar parameters carried by the UAV will directly affect the resolution and accuracy of data acquisition. Assuming that a UAV is equipped with a camera with a resolution of 4K and the point cloud density of the lidar is 1000 points per square meter, the image resolution and point cloud density of key parts such as blades and towers can be calculated at different circumradius and camera position angles. Considering flight safety, data acquisition efficiency and data quality comprehensively, it is necessary to determine the optimal circumradius and the optimal camera position angle. A smaller circumradius can improve the data acquisition resolution, but may increase the flight risk. A larger circumradius can improve flight safety, but will reduce the data acquisition efficiency. A suitable camera position angle can ensure the complete data acquisition of key parts and avoid occlusion and perspective deviation.

[0081] Step 3: Control the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, circumradius and camera position angle, and collect wind turbine images and point cloud data during the circumferential flight.

[0082] Further, controlling the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, circumradius and camera position angle includes:

[0083] Discretize the flight trajectory into a number of waypoints, where the waypoints contain position, attitude, and speed information. Among them, the position and attitude information are calculated according to the circumradius and camera position angle, and the speed information is a preset speed;

[0084] After controlling the UAV to take off, obtain the real-time position, real-time attitude, and real-time speed of the UAV, and compare the real-time position, real-time attitude, and real-time speed with the position, attitude, and speed information in the next waypoint to obtain the real-time heading adjustment amount and the real-time speed adjustment amount;

[0085] Continuously correct the heading and speed of the UAV through the real-time heading adjustment amount and real-time speed adjustment amount until the flight trajectory ends.

[0086] Specifically, in this embodiment, the trajectory is discretized into a series of waypoints, and each waypoint contains information such as position, attitude, and speed. For example, the surrounding trajectory can be divided into 100 waypoints, with a distance of 10 meters between each waypoint. The first waypoint may be located 10 meters directly in front of the wind turbine, at a height of 50 meters, with a horizontal attitude towards the turbine and a speed of 5 m / s. The setting of each waypoint requires precise calculation to ensure that the UAV can stably and efficiently collect data during flight. After the UAV takes off, the control system obtains the real-time position and attitude information of the UAV through GPS and the inertial measurement unit (IMU), compares this information, and calculates the position deviation and attitude deviation. If the deviation value exceeds the preset threshold, the adjustment amounts of the UAV's heading and speed are calculated according to the deviation direction and magnitude, and the UAV corrects its heading and speed according to the adjustment amounts to make the actual flight trajectory coincide with the planned trajectory.

[0087] In this embodiment, according to the preset surrounding radius R and the position angle θ of the wind turbine, the position and attitude information of the next waypoint are dynamically calculated, and the information of the next waypoint is sent to the UAV. The UAV adjusts its heading and speed according to the waypoint information and continues to perform the surrounding flight task to ensure accurate arrival at the predetermined position until the UAV completes the full-round surrounding flight of the wind turbine.

[0088] In this embodiment, the images of the wind turbine are collected during the surrounding flight by an on-board camera, which are multi-angle and high-resolution images; the point cloud data is scanned by a lidar, which are the three-dimensional space coordinates of the wind turbine.

[0089] Step 4: Based on the images and point cloud data of the wind turbine, construct a three-dimensional model of the wind turbine, combine the terrain data to construct a three-dimensional map, and fuse the flight trajectory into the three-dimensional map to complete the data collection of the target offshore wind farm.

[0090] Further, constructing a three-dimensional model of the wind turbine based on the images and point cloud data of the wind turbine includes:

[0091] Preprocess the images and point cloud data of the wind turbine respectively, and then fuse them;

[0092] Extract the component features of the wind turbine, as well as the geometric contours and position information of each component, based on the fused images and point cloud data of the wind turbine, and construct a three-dimensional model of the wind turbine.

[0093] Among them, preprocessing the images of the wind turbine includes: stitching multiple images of the wind turbine into a panoramic image of the wind turbine, and then performing distortion correction processing on the panoramic image of the wind turbine to obtain the processed panoramic image.

[0094] After preprocessing the point cloud data, it includes: removing noise from the point cloud data, smoothing the point cloud, and obtaining the processed point cloud data.

[0095] Fusion includes: registering and mapping the processed panoramic image and the processed point cloud data to obtain a panoramic image containing the point cloud data.

[0096] Specifically, in this embodiment, a feature matching algorithm, such as Scale-Invariant Feature Transform (SIFT), is used to identify the same feature points between different images, so as to accurately align and splice the images; a distortion correction algorithm, such as bar distortion correction, is used to eliminate the distortion caused by lens curvature and splicing technology by adjusting the pixel positions in the image.

[0097] In this embodiment, the noise removal process of the point cloud data is implemented by statistical analysis methods, such as calculating the local density of each point and removing isolated points with a density lower than the threshold; the smoothing of the point cloud is performed using the Moving Least Squares (MLS) method to smooth the point cloud and improve the quality of the point cloud data.

[0098] In this embodiment, the fusion process is carried out through feature point matching, corresponding specific points in the image to the corresponding points in the point cloud, so as to realize the fusion of the image and the point cloud. And during the fusion process, the Iterative Closest Point (ICP) algorithm is used to optimize the registration accuracy and ensure the best alignment between the image and the point cloud.

[0099] In this embodiment, based on the fused wind turbine image and point cloud data, the component features of the wind turbine, as well as the geometric profiles and position information of each component, are extracted to construct a three-dimensional model of the wind turbine. Specifically, it includes:

[0100] First, use the fused data to extract the key component features of the wind turbine, such as the damage or wear condition of the blades. Through an edge detection algorithm, such as the Canny algorithm, identify the edges of the blades, and then analyze their shapes and integrity. Then use a region segmentation algorithm to identify and isolate different components in the image, such as the nacelle and the tower, for further analysis and evaluation. Then, based on the extracted features and point cloud data, use a surface reconstruction algorithm to construct a three-dimensional model of the wind turbine. Finally, post-process the three-dimensional model, including model simplification, texture mapping, etc., to obtain a lightweight three-dimensional model suitable for visualization and analysis, and import the three-dimensional model of the wind turbine into a virtual map, integrate the flight trajectory and environmental parameters into the virtual map, and enhance the visual effect of the model through advanced graphics rendering technology to improve the overall realism of the scene.

[0101] Extract the geometric parameters and semantic attributes of the key components of the wind turbine simultaneously, encode and compress the 3D models and attribute information of the key components, and send the data to the ground station via wireless or wired transmission to achieve remote monitoring and analysis. After receiving the data at the ground station, decode and decompress the data, store the 3D models and attribute information in the database, and provide visualization and analysis tools to support the condition monitoring, fault diagnosis, and maintenance decision-making of the key components of the wind turbine.

[0102] To further optimize the above technical solution, this embodiment also provides an offshore wind turbine data acquisition system, including:

[0103] An inspection area division module, configured to divide the target offshore wind farm into several inspection areas, and obtain the structural parameters and terrain data of the wind turbines in each inspection area;

[0104] A flight information construction module, configured to construct the UAV flight trajectories of each inspection area respectively according to the wind turbine structural parameters and terrain data, and preset the circumferential radius and position angles;

[0105] An inspection data acquisition module, configured to control the UAV to fly around the wind turbines in each inspection area based on the flight trajectories, circumferential radius, and position angles, and collect the wind turbine images and point cloud data during the circumferential flight;

[0106] An inspection data processing module, configured to construct a 3D model of the wind turbine based on the wind turbine images and point cloud data, construct a 3D map in combination with the terrain data, and fuse the flight trajectories into the 3D map to complete the data acquisition of the target offshore wind farm.

[0107] Further, the inspection area division module dividing the target offshore wind farm into several inspection areas includes:

[0108] Obtain the geographical information data and wind turbine distribution map of the target offshore wind farm, and construct a topological structure diagram, where the topological structure diagram uses the wind turbines as nodes and the distances between the wind turbines as edges;

[0109] Obtain the maximum flight range and payload capacity of the UAV, and calculate the maximum single inspection range and the maximum single quantity of wind turbines in the UAV inspection task;

[0110] Adopt graph theory algorithms, use the maximum single inspection range and the maximum single quantity of wind turbines of the UAV as constraint conditions, divide the topological structure diagram, and obtain several connected subgraphs, that is, several inspection areas.

[0111] Further, the flight information construction module constructing the UAV flight trajectories of each inspection area respectively according to the wind turbine structural parameters and terrain data includes:

[0112] Construct a three-dimensional scene model based on the structural parameters of the wind turbine and the terrain data;

[0113] Use the artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain the flight trajectory;

[0114] Based on the collision risk detection, locally optimize the flight trajectory, and globally optimize the flight trajectory based on the particle swarm optimization algorithm to obtain the final flight trajectory.

[0115] Among them, using the artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain the flight trajectory includes:

[0116] Obtain the positions of the wind turbines and terrain obstacles based on the three-dimensional scene model;

[0117] Take the positions of the wind turbines as the target points in the artificial potential field algorithm, and take the positions of the terrain obstacles as the obstacle points in the artificial potential field algorithm;

[0118] By dynamically adjusting the distances between the UAV and the target points and obstacle points, plan the flight trajectory of the UAV.

[0119] Furthermore, the method for the flight information construction module to preset the circumferential radius and the camera position angle is as follows:

[0120] Obtain the camera parameters and lidar parameters carried by the UAV;

[0121] Simulate the flight in the three-dimensional scene model based on the final flight trajectory, and calculate the data acquisition efficiency and quality at different circumferential radii and camera position angles according to the camera parameters and lidar parameters;

[0122] Based on the data acquisition efficiency and quality, combined with flight safety, determine the optimal circumferential radius and the optimal camera position angle.

[0123] Furthermore, the inspection data acquisition module controls the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, the circumferential radius, and the camera position angle, including:

[0124] Discretize the flight trajectory into a number of waypoints, and the waypoints contain position, attitude, and speed information. Among them, the position and attitude information are calculated through the circumferential radius and the camera position angle, and the speed information is the preset speed;

[0125] After controlling the UAV to take off, obtain the real-time position, real-time attitude, and real-time speed of the UAV, and compare the real-time position, real-time attitude, and real-time speed with the position, attitude, and speed information in the next waypoint to obtain the real-time heading adjustment amount and the real-time speed adjustment amount;

[0126] Continuously correct the heading and speed of the UAV through the real-time heading adjustment amount and the real-time speed adjustment amount until the flight trajectory ends.

[0127] Furthermore, the inspection data processing module constructs a three-dimensional model of the wind turbine based on the wind turbine images and point cloud data, including:

[0128] After preprocessing the wind turbine images and point cloud data respectively, fusion is carried out;

[0129] Based on the fused wind turbine images and point cloud data, extract the component features of the wind turbine, as well as the geometric contours and position information of each component, and construct a three-dimensional model of the wind turbine.

[0130] Among them, after preprocessing the wind turbine images and point cloud data respectively, and then carrying out fusion includes:

[0131] Stitch multiple wind turbine images into a panoramic image of the wind turbine, and then perform distortion correction processing on the panoramic image of the wind turbine to obtain the processed panoramic image;

[0132] Perform denoising and smoothing of the point cloud data on the point cloud data to obtain the processed point cloud data;

[0133] Register and map the processed panoramic image with the processed point cloud data to obtain a panoramic image containing the point cloud data.

[0134] The embodiments described above are only descriptions of the preferred embodiments of the present invention, and do not limit the scope of the present invention. Without departing from the design spirit of the present invention, various deformations and improvements made by those of ordinary skill in the art to the technical solutions of the present invention shall fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for collecting data of an offshore wind turbine, characterized in that: include: Divide the target offshore wind farm into several inspection areas, and obtain the structural parameters and terrain data of the wind turbines in each inspection area; Construct the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine and the terrain data, and preset the circling radius and the camera angle; Based on the flight trajectory, circling radius and aircraft position angle, the drone is controlled to circumvent the wind turbines in each inspection area, and images and point cloud data of the wind turbines are collected during the circling flight; Building a three-dimensional model of the wind turbine based on the wind turbine image and point cloud data, building a three-dimensional map in combination with the terrain data, integrating the flight trajectory into the three-dimensional map, and completing data collection of the target offshore wind farm; Wherein, constructing the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine set and the terrain data includes: Constructing a three-dimensional scene model according to the structural parameters of the wind turbine and the terrain data; Using an artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain a flight trajectory; Locally optimizing the flight trajectory based on collision risk detection, and globally optimizing the flight trajectory based on a particle swarm optimization algorithm to obtain a final flight trajectory; Using an artificial potential field algorithm to perform path planning in the three-dimensional scene model, obtaining a flight trajectory includes: Acquire the positions of wind turbines and terrain obstacles based on the three-dimensional scene model; Using the position of the wind turbine as the target point in the artificial potential field algorithm, and using the position of the terrain obstacle as the obstacle point in the artificial potential field algorithm; The flight trajectory of the UAV is planned by dynamically adjusting the distance between the UAV and the target point and the obstacle point.

2. The offshore wind turbine data collection method according to claim 1, characterized in that: The target offshore wind farm is divided into several inspection areas including: Acquire geographic information data and a wind turbine distribution map of the target offshore wind farm, and construct a topological structure map, wherein the topological structure map uses wind turbines as nodes and distances between wind turbines as edges; Obtain the maximum range and load capacity of the drone, and calculate the maximum single inspection range and the maximum number of wind turbines in a single inspection mission; A graph theory algorithm is used to divide the topological structure diagram by taking the single maximum inspection range of the UAV and the single maximum number of wind turbines as constraints to obtain a number of connected subgraphs, i.e., a number of inspection areas.

3. The offshore wind turbine data collection method according to claim 1, characterized in that: The method for presetting the surround radius and camera angle is: Get the camera parameters and lidar parameters carried by the drone; Simulating flight in the three-dimensional scene model based on the final flight trajectory, and calculating data collection efficiency and quality under different circling radii and camera angles according to the camera parameters and the laser radar parameters; Based on the data collection efficiency and quality, combined with flight safety, the optimal circling radius and the optimal aircraft position angle are determined.

4. The offshore wind turbine data collection method according to claim 1, characterized in that: Controlling the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, the circling radius and the aircraft position angle includes: Discretize the flight trajectory into a number of waypoints, each of which contains position, attitude, and speed information, wherein the position and attitude information are calculated by the orbiting radius and the aircraft position angle, and the speed information is a preset speed; After controlling the drone to take off, the real-time position, real-time attitude, and real-time speed of the drone are obtained, and the real-time position, real-time attitude, and real-time speed are compared with the position, attitude, and speed information of the next waypoint to obtain the real-time heading adjustment amount and the real-time speed adjustment amount; The heading and speed of the UAV are continuously corrected by the real-time heading adjustment amount and the real-time speed adjustment amount until the flight trajectory ends.

5. The offshore wind turbine data collection method according to claim 1, characterized in that: Constructing a three-dimensional model of a wind turbine generator system based on the wind turbine generator system image and point cloud data includes: Preprocessing the wind turbine image and point cloud data respectively and then fusing them; Based on the fused wind turbine image and point cloud data, the component features of the wind turbine and the geometric outline and position information of each component are extracted to construct a three-dimensional model of the wind turbine.

6. The offshore wind turbine data collection method according to claim 5, characterized in that: After preprocessing the wind turbine image and point cloud data respectively, the fusion includes: splicing a plurality of wind turbine group images into a panoramic image of the wind turbine group, and then performing distortion correction processing on the panoramic image of the wind turbine group to obtain a processed panoramic image; The point cloud data is processed to remove noise and smooth the point cloud to obtain the processed point cloud data; The processed panoramic image is registered and mapped with the processed point cloud data to obtain a panoramic image containing the point cloud data.

7. An offshore wind turbine data acquisition system, applied to the offshore wind turbine data acquisition method according to any one of claims 1 to 6, characterized in that: include: The inspection area division module is used to divide the target offshore wind farm into several inspection areas and obtain the structural parameters and terrain data of the wind turbines in each inspection area; A flight information construction module is used to construct the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine set and the terrain data, and to preset the circling radius and the aircraft position angle; The inspection data acquisition module is used to control the UAV to fly around the wind turbines in each inspection area based on the flight trajectory, the circling radius and the aircraft position angle, and to collect images and point cloud data of the wind turbines during the circling flight; An inspection data processing module is used to construct a three-dimensional model of the wind turbine based on the wind turbine image and point cloud data, and to construct a three-dimensional map in combination with the terrain data, and to integrate the flight trajectory into the three-dimensional map to complete the data collection of the target offshore wind farm; Wherein, constructing the UAV flight trajectory of each inspection area according to the structural parameters of the wind turbine set and the terrain data includes: Constructing a three-dimensional scene model according to the structural parameters of the wind turbine and the terrain data; Using an artificial potential field algorithm to perform path planning in the three-dimensional scene model to obtain a flight trajectory; Locally optimizing the flight trajectory based on collision risk detection, and globally optimizing the flight trajectory based on a particle swarm optimization algorithm to obtain a final flight trajectory; Using an artificial potential field algorithm to perform path planning in the three-dimensional scene model, obtaining a flight trajectory includes: Acquire the positions of wind turbines and terrain obstacles based on the three-dimensional scene model; Using the position of the wind turbine as the target point in the artificial potential field algorithm, and using the position of the terrain obstacle as the obstacle point in the artificial potential field algorithm; The flight trajectory of the UAV is planned by dynamically adjusting the distance between the UAV and the target point and the obstacle point.

Images