Methods, devices, equipment, and media for inspecting wind turbines using unmanned aerial vehicles (UAVs).

Abstract

This invention relates to the field of unmanned aerial vehicle (UAV) equipment technology, and discloses a method, apparatus, equipment, and medium for inspecting wind turbines based on UAVs. First, with the UAV reaching its initial cruise position and its battery level within a preset range, an image of the top of the wind turbine is acquired. Then, based on the top image of the wind turbine, the UAV is controlled to fly to the center point of the wind turbine blades and acquire an image of the swept surface. Next, the actual error distance of the swept surface is determined based on the swept surface image. If the actual error distance of the swept surface meets the preset conditions for the swept surface, the reference flight direction of the wind turbine components is determined. Then, based on the component safety distance, the component inspection image overlap rate, and the number of component inspection shots, a sequence of single-step movement distances for each component and the total movement distance of the component are determined. Finally, images are acquired based on the component reference flight direction, the sequence of single-step movement distances, and the total movement distance to obtain a set of component inspection images.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) equipment technology, and in particular to a method, apparatus, equipment and medium for inspecting wind turbines based on UAVs. Background Technology

[0002] With the development of drone technology, using drones to inspect wind power equipment has become commonplace.

[0003] While manually operated drone patrols have partially solved high-altitude safety issues, they are highly complex, requiring professional pilots, and the flight paths rely on manual planning. This makes dynamic adjustments difficult under complex weather conditions, potentially leading to blurred images or missed detections. Therefore, a novel drone-based method for wind turbine inspection is needed. Summary of the Invention

[0004] The embodiments described in this specification aim to at least partially solve one of the technical problems in the related art. To this end, the embodiments described in this specification propose a method, apparatus, equipment, and medium for inspecting wind turbines based on unmanned aerial vehicles (UAVs).

[0005] This specification provides a method for inspecting wind turbines based on unmanned aerial vehicles (UAVs), the method comprising:

[0006] When the drone reaches the initial cruise position and the battery level is within a preset range, an image of the top of the wind turbine is acquired, wherein the initial cruise position includes the coordinates and safe altitude of the wind turbine.

[0007] Based on the image of the top of the wind turbine, the drone is controlled to fly to the center point of the wind turbine blades and collect images of the swept surface of the wind turbine.

[0008] The actual error distance of the swept surface is determined based on the image of the swept surface of the wind turbine.

[0009] If the actual error distance of the swept surface meets the preset conditions of the swept surface, the reference flight direction of the wind turbine component is determined.

[0010] Based on the component safety distance, component inspection image overlap rate, and component inspection shooting number, determine the component single-step movement distance sequence set and the component total movement distance;

[0011] Image acquisition is performed based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance to obtain a component inspection image set.

[0012] This specification provides a wind turbine inspection device based on a drone, the device comprising:

[0013] The wind turbine top image acquisition module is used to acquire an image of the top of the wind turbine when the drone flies to the initial cruise position and the power is within a preset range. The initial cruise position includes the coordinates and safe altitude of the wind turbine.

[0014] The wind turbine swept surface image acquisition module is used to control the UAV to fly to the center point of the wind turbine blades and acquire the wind turbine swept surface image based on the top image of the wind turbine;

[0015] The actual error distance determination module for swept surface is used to determine the actual error distance of the swept surface based on the wind turbine swept surface image.

[0016] The component reference flight direction determination module is used to determine the component reference flight direction of the wind turbine when the actual error distance of the swept surface meets the preset conditions of the swept surface.

[0017] The component movement data determination module is used to determine the set of single-step movement distance sequences and the total movement distance of the component based on the component safety distance, the overlap rate of component inspection images, and the number of component inspection shots.

[0018] The component inspection image acquisition module is used to acquire images based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance to obtain a component inspection image set.

[0019] This specification provides a computer device, a memory, and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described in any of the above embodiments.

[0020] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in any of the above embodiments.

[0021] This specification provides a computer program product that includes instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0022] In the above-described implementation method, firstly, when the UAV reaches the initial cruise position and its battery level is within a preset range, an image of the top of the wind turbine is acquired. The initial cruise position includes the coordinates and safe altitude of the wind turbine. Then, based on the image of the top of the wind turbine, the UAV is controlled to fly to the center point of the wind turbine blades and acquire an image of the swept surface. Next, the actual error distance of the swept surface is determined based on the swept surface image. If the actual error distance of the swept surface meets the preset conditions for the swept surface, the reference flight direction of the wind turbine components is determined. Then, based on the component safe distance, the component inspection image overlap rate, and the number of component inspection shots, the set of single-step movement distance sequences and the total movement distance of the components are determined. Finally, based on the component reference flight direction, the set of single-step movement distance sequences, and the total movement distance, images are acquired to obtain a set of component inspection images. This implementation process dynamically optimizes the UAV's flight trajectory and shooting parameters by real-time fusion of environmental data from wind turbine images and multimodal sensor information, combined with visual image recognition algorithms. While reducing manual intervention, it improves the detection accuracy and efficiency in complex scenarios, thereby effectively resolving the contradiction between cost and performance. Attached Figure Description

[0023] Figure 1 A flowchart of a method for inspecting wind turbines based on unmanned aerial vehicles (UAVs) provided for embodiments of this specification.

[0024] Figure 2 A schematic diagram illustrating the process of controlling a drone to fly to the center point of the wind turbine blades and acquire images of the swept surface of the wind turbine, as provided in the embodiments of this specification.

[0025] Figure 3 A flowchart illustrating the determination of the actual error distance of the swept surface provided for the embodiments of this specification;

[0026] Figure 4a A flowchart illustrating the determination of the blade's vertical foot coordinates for embodiments described in this specification;

[0027] Figure 4b A schematic diagram of several key points corresponding to the top image of the fan provided for the embodiments of this specification;

[0028] Figure 5 A flowchart illustrating the determination of the actual top error distance provided for the implementation of this specification;

[0029] Figure 6a A flowchart illustrating the process of correcting the actual error distance at the top, provided for the implementation of this specification;

[0030] Figure 6b A top view diagram after correcting the actual top error distance, provided for the implementation of this specification;

[0031] Figure 7 A flowchart illustrating the process of determining the axial flight direction of the wind turbine, provided for the implementation of this specification;

[0032] Figure 8 A flowchart illustrating the process of acquiring images of the windward side of a fan, provided for the implementation of this specification.

[0033] Figure 9 A flowchart illustrating the determination of the actual error distance on the windward side, provided for the implementation of this specification;

[0034] Figure 10a A flowchart illustrating the process of correcting the actual error distance on the windward side, provided for the implementation of this specification;

[0035] Figure 10b A schematic diagram of the windward surface after correcting the actual error distance of the windward surface, provided for the implementation of this specification.

[0036] Figure 11 A flowchart illustrating the process of determining the reference flight direction of a component of a wind turbine, provided for the implementation of this specification.

[0037] Figure 12 A flowchart illustrating the process of determining the set of single-step movement distance sequences for blades and the total movement distance of components, provided for embodiments of this specification;

[0038] Figure 13 A flowchart illustrating the process of determining the set of single-step movement distance sequences and the total movement distance of the tower, provided for the implementation of this specification;

[0039] Figure 14a A schematic diagram illustrating the implementation of blade inspection in this specification.

[0040] Figure 14b A schematic diagram of the inspection of wind turbine blades and towers provided for the implementation of this specification;

[0041] Figure 15 A flowchart illustrating the implementation of tower inspection as provided in this specification.

[0042] Figure 16 A flowchart illustrating the process of acquiring images of the leeward side of a wind turbine, provided for the implementation of this specification;

[0043] Figure 17 A flowchart illustrating the determination of the actual error distance on the leeward side, provided for the implementation of this specification.

[0044] Figure 18 A flowchart illustrating the process of correcting the actual error distance on the leeward side, provided for the implementation of this specification.

[0045] Figure 19A schematic diagram of a drone-based wind turbine inspection device provided for embodiments of this specification;

[0046] Figure 20 A schematic diagram of the internal structure of the computer device provided in the embodiments of the instruction manual. Detailed Implementation

[0047] Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0048] In related technologies, manual suspended platform inspections rely on maintenance personnel climbing the towers of wind turbines (over 100 meters high) to visually inspect components such as blades and nacelles at close range using handheld devices (such as thermal imagers). However, this method requires 2-3 hours to inspect each wind turbine, and the annual high-altitude accident rate is 0.3‰, posing a risk of falls. Furthermore, manual inspections rely on limited observation, resulting in a high rate of missed inspections (up to 20%), especially for micro-cracks (e.g., cracks 2mm or smaller).

[0049] In related technologies, ground-based telescope / high-magnification microscope inspection methods achieve blade inspection by remotely capturing images of wind turbine blades using high-resolution equipment. However, these methods are limited by light and weather conditions, with single-shot coverage rates below 30%. Furthermore, the rate of missed detection for minute cracks is as high as 42%, and the effective inspection window for offshore wind farms is typically less than 100 days per year. The data from these methods is also highly subjective and lacks traceability.

[0050] In related technologies, fixed sensor online monitoring deploys vibration / temperature sensors in critical components of wind turbines, such as gearboxes and bearings, and transmits data in real time via a SCADA system. However, fixed sensor online monitoring only covers internal components of the wind turbine and cannot detect external damage to the turbine blades (such as lightning strikes or corrosion). Furthermore, it is incompatible with manual recording and drone data protocols, resulting in a fault prediction accuracy rate of less than 60%.

[0051] The aforementioned existing technologies suffer from drawbacks such as efficiency bottlenecks (e.g., reliance on downtime and low coverage), safety risks (e.g., high-altitude operations), data fragmentation (e.g., difficulty in integrating multi-source information), and blind spots in blade diagnosis (e.g., difficulty in detecting external damage). These drawbacks result in an average annual damage rate of up to 9.5% for wind turbine blades, leading to persistently high operation and maintenance costs.

[0052] While manually operated drone patrols have partially solved high-altitude safety issues, significant bottlenecks remain. Firstly, they are highly complex: requiring professional pilots, flight paths are manually planned, and dynamic adjustments are difficult under complex weather conditions (such as turbulence or sudden wind speed changes), easily leading to blurred images or missed detections. Secondly, they are costly: professional drones require equipment such as lidar and thermal imagers, and a single flight duration is ≤25 minutes, with frequent battery replacements increasing overall maintenance costs.

[0053] Based on the above analysis, this specification provides a method for inspecting wind turbines using a drone. First, with the drone reaching its initial cruise position and the battery level within a preset range, an image of the top of the wind turbine is acquired. The initial cruise position includes the wind turbine's coordinates and safe altitude. Then, based on the top image, the drone is controlled to fly to the center point of the wind turbine blades and acquire an image of the turbine's swept surface. Next, the actual error distance of the swept surface is determined based on the swept surface image. If the actual error distance meets the preset conditions for the swept surface, the reference flight direction of the wind turbine components is determined. Then, based on the component's safe distance, the overlap rate of the component inspection images, and the number of component inspection shots, a set of sequenced single-step movement distances for each component and the total movement distance of the component are determined. Finally, images are acquired based on the component's reference flight direction, the set of sequenced single-step movement distances, and the total movement distance to obtain a set of component inspection images. This implementation process dynamically optimizes the drone's flight trajectory and shooting parameters by real-time fusion of environmental data from the wind turbine images and multimodal sensor information, combined with visual image recognition algorithms. While reducing human intervention, it improves the detection accuracy and efficiency in complex scenarios, thereby effectively resolving the contradiction between cost and performance.

[0054] This specification provides a method for inspecting wind turbines based on unmanned aerial vehicles (UAVs). Please refer to [link / reference]. Figure 1 The drone-based wind turbine inspection method may include the following steps:

[0055] S110. When the drone arrives at the initial cruise position and the battery level is within the preset range, acquire an image of the top of the wind turbine.

[0056] The initial cruise position includes the coordinates of the wind turbine and the safe altitude.

[0057] Specifically, before the inspection begins, the tower height of the wind turbine is obtained through preliminary measurements or other methods. blade length And the coordinates of the wind turbine. Because a safe distance is required between the drone and the wind turbine to ensure the drone's safety, a safety distance setting is provided between the top-mounted drone and the wind turbine. Safety height Control the drone to take off vertically and ascend to a safe altitude. Afterwards, the drone flies to the coordinates of the wind turbine and then adjusts its nose to a preset direction (such as due north) in order to begin the inspection.

[0058] After the drone reaches its initial cruise position, it needs to determine if its battery level is within a preset range (e.g., 25% to 100%) to ensure it has sufficient power to return and meet inspection requirements. If the battery level is outside the preset range, it indicates insufficient power for subsequent inspections, resulting in mission failure and requiring the drone to return. If the battery level is within the preset range, it needs to determine if adaptive inspection parameters have been loaded. If not, these parameters must be loaded. If loaded, the drone's top camera gimbal angle is adjusted to 0°, and image acquisition is performed to obtain an image of the wind turbine's top.

[0059] S120: Based on the top image of the wind turbine, control the drone to fly to the center point of the wind turbine blades and collect images of the swept surface of the wind turbine.

[0060] S130. Determine the actual error distance of the swept surface based on the image of the swept surface of the wind turbine.

[0061] S140. If the actual error distance of the swept surface meets the preset conditions of the swept surface, determine the reference flight direction of the wind turbine components.

[0062] S150. Based on the component safety distance, component inspection image overlap rate, and component inspection shooting number, determine the component single-step movement distance sequence set and the component total movement distance.

[0063] S160. Based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance, image acquisition is performed to obtain a component inspection image set.

[0064] Specifically, a wind turbine has two swept surfaces. If at least one of them is not inspected, image recognition and calculation analysis are performed on the top image of the turbine to determine whether the drone is within a preset center range on the top of the wind turbine. If the drone is within the preset center range on the top of the wind turbine, a control algorithm is used to guide the drone to fly to the center point of the wind turbine blades, and then the drone's nose is adjusted to acquire images, thus obtaining images of the swept surface of the turbine.

[0065] Next, image recognition and calculation are performed on the swept surface images of the wind turbine to determine the actual error distance of the swept surface. If the actual error distance of the swept surface meets the preset conditions, it indicates that the UAV is at the actual blade center point. Then, the swept surface images of the wind turbine are analyzed and calculated to determine the reference flight direction of the wind turbine components. It is also necessary to calculate the component safety distance, component inspection image overlap rate, and the number of component inspection shots to determine the set of single-step movement distance sequences and the total movement distance of the components, in order to control the flight distance of the UAV. Subsequently, the control algorithm is used to control the UAV to fly along the component reference flight direction according to the set of single-step movement distance sequences, and to acquire images after completing one single-step movement distance to obtain the final component inspection image set. Finally, the control algorithm is used to control the UAV to fly in the opposite direction of the component reference flight direction according to the total movement distance of the components, returning to the actual blade center point for subsequent inspection or return flight.

[0066] It should be noted that since wind turbines have two swept surfaces, after determining the component inspection image set for one swept surface through the above process, the drone is controlled to fly back to the initial cruise position. Then, the battery level is reassessed to see if it is within the preset range. If the battery level is not within the preset range, the drone is controlled to return to base. If the battery level is within the preset range, subsequent operations are performed to obtain the component inspection image set for the other swept surface, and then the drone returns to base.

[0067] In the above implementation, firstly, when the UAV reaches the initial cruise position and its battery level is within a preset range, an image of the top of the wind turbine is acquired. The initial cruise position includes the coordinates and safe altitude of the wind turbine. Then, based on the image of the top of the wind turbine, the UAV is controlled to fly to the center point of the wind turbine blades and acquire an image of the swept surface. Next, the actual error distance of the swept surface is determined based on the swept surface image. If the actual error distance of the swept surface meets the preset conditions for the swept surface, the reference flight direction of the wind turbine components is determined. Then, based on the component safe distance, the overlap rate of component inspection images, and the number of component inspection shots, the set of sequence of single-step movement distances for the components and the total movement distance of the components are determined. Finally, image acquisition is performed based on the component reference flight direction, the set of sequence of single-step movement distances for the components, and the total movement distance of the components to obtain a set of component inspection images. This implementation process dynamically optimizes the UAV's flight trajectory and shooting parameters by real-time fusion of environmental data and multimodal sensor information from the wind turbine images, combined with visual image recognition algorithms. While reducing manual intervention, it improves the detection accuracy and efficiency in complex scenarios, thereby effectively resolving the contradiction between cost and performance.

[0068] In some implementations, please refer to Figure 2The top image of the wind turbine corresponds to a center point of the top image. Based on the top image of the wind turbine, controlling the drone to fly to the center point of the wind turbine blades and acquire the swept surface image of the wind turbine can include the following steps:

[0069] S210. Based on the image of the top of the wind turbine, the outline of the first target area is determined.

[0070] S220. Calculate the vertical coordinates of the blade based on the contour of the first target area.

[0071] S230. Based on the contour of the first target area, the vertical coordinates of the blade, and the center point of the top image, determine the actual error distance at the top.

[0072] S240. When the actual error distance at the top meets the preset conditions at the top, the axial flight direction of the wind turbine is determined based on the vertical coordinate of the blade.

[0073] S250, based on the wind turbine's axial flight direction, horizontal safety distance, and blade safety height, controls the UAV to fly to the center point of the wind turbine blades and collect images of the wind turbine's swept surface.

[0074] Specifically, an image recognition algorithm is used to identify the top image of the wind turbine to determine the outline of the first target area. Next, based on the obtained first target area outline, a geometric calculation method is used to determine the blade vertical coordinates. After determining the blade vertical coordinates, the image distance is converted to the actual distance by combining the first target area outline and the center point of the top image, thus determining the actual error distance at the top. When the actual error distance at the top meets the preset conditions, i.e., when the actual error is within the allowable range, it indicates that the UAV's position is within the preset center range of the top of the wind turbine. Then, by analyzing the positional relationship of the blade vertical coordinates, geometric principles and a heading planning algorithm are used to determine the axial flight direction of the wind turbine. Finally, based on the determined axial flight direction of the wind turbine, and combining the two key parameters of horizontal safety distance and blade safety height, the UAV is controlled to fly to the center point of the wind turbine blades and acquire images of the swept surface of the wind turbine.

[0075] In the above implementation, the outline of the first target area is determined based on the image of the top of the wind turbine. The vertical coordinates of the blades are calculated based on the outline of the first target area. The actual error distance of the top is determined based on the outline of the first target area, the vertical coordinates of the blades, and the center point of the top image. If the actual error distance of the top meets the preset conditions of the top, the axial flight direction of the wind turbine is determined based on the vertical coordinates of the blades. Based on the axial flight direction of the wind turbine, the horizontal safety distance, and the safe height of the blades, the UAV is controlled to fly to the center point of the wind turbine blades and collect the swept surface image of the wind turbine, providing a data basis for subsequent blade and tower inspection.

[0076] In some implementations, please refer to Figure 3 The swept surface image of the wind turbine corresponds to a center point of the swept surface image. Determining the actual error distance of the swept surface based on the swept surface image of the wind turbine can include the following steps:

[0077] S310. Based on the swept surface image of the wind turbine, the outline of the second target area is determined.

[0078] S320. Based on the contour of the second target region, determine the coordinates of the centroid of the wind turbine's central structure.

[0079] S330. Based on the contour of the second target area, the coordinates of the centroid of the wind turbine center structure, and the center point of the swept surface image, determine the actual error distance of the swept surface.

[0080] Specifically, an image recognition algorithm is used to identify the swept surface image of the wind turbine to determine the outline of the second target region. Next, the centroid coordinates of the wind turbine's central structure are determined using a centroid calculation formula. After determining the centroid coordinates of the wind turbine's central structure, a proportional conversion between the image distance and the actual distance is performed using the second target region outline and the center point of the swept surface image to determine the actual error distance of the swept surface.

[0081] In the above implementation, the outline of the second target area is determined based on the wind turbine swept surface image. Based on the outline of the second target area, the centroid coordinates of the wind turbine center structure are determined. Based on the outline of the second target area, the centroid coordinates of the wind turbine center structure, and the center point of the swept surface image, the actual error distance of the swept surface is determined, providing a data basis for subsequent UAV swept surface correction.

[0082] In some implementations, please refer to Figure 4a The first target region contour includes a first blade segmentation contour, a second blade segmentation contour, and a nacelle segmentation contour. Calculations based on the first target region contour to determine the blade's vertical foot coordinates may include the following steps:

[0083] S410. Calculate and determine the centroid coordinates of the first blade based on the segmented profile of the first blade.

[0084] S420. Calculate the centroid coordinates of the second blade based on the segmented profile of the second blade.

[0085] S430. Calculate the coordinates of the cabin's centroid based on the cabin's segmented profile.

[0086] S440. Based on the coordinates of the centroid of the first blade, the centroid of the second blade, and the centroid of the nacelle, the vertical foot coordinates of the blade are determined.

[0087] Specifically, image recognition algorithms are used to identify the segmentation contours in the image of the top of the wind turbine. If the image of the top of the wind turbine is determined to include the segmentation contours of the first and second blades corresponding to the two largest blades, as well as the nacelle segmentation contour, the centroid coordinates of the first blade are determined using a centroid calculation formula. The centroid coordinates of the second blade segmentation contour are also determined using the same formula. Finally, the centroid coordinates of the nacelle segmentation contour are determined using a vertical projection method. The distance between the nacelle centroid coordinates and the line connecting the first and second blade centroid coordinates is then calculated to determine the blade's vertical foot coordinates. If the image of the top of the wind turbine is determined to not include at least one of the segmentation contours of the first and second blades, or the nacelle segmentation contour, the UAV is controlled to return to its home position. For an example, please refer to [link to example]. Figure 4b , Figure 4b The green dot in the diagram represents the blade's center of mass, the red dot represents the nacelle's center of mass, and the dark blue dot located at the blade's center of mass represents the blade's foot.

[0088] For example, based on the segmentation contour of the first leaf Calculations were performed to determine the coordinates of the centroid of the first blade. Based on the segmentation profile of the second leaf. Calculations were performed to determine the coordinates of the centroid of the second blade. Based on cabin segmentation profile Calculations were performed to determine the coordinates of the cabin's center of mass. The centroid coordinates of the computer module are calculated using the following formula. coordinates of the centroid of the first blade Coordinates of the centroid of the second blade The perpendicular coordinates of the blades connecting the two lines :

[0089]

[0090]

[0091] in, For parameters.

[0092] In the above embodiments, the centroid coordinates of the first blade are determined based on the segmented profile of the first blade, the centroid coordinates of the second blade are determined based on the segmented profile of the second blade, the centroid coordinates of the nacelle are determined based on the segmented profile of the nacelle, and the vertical foot coordinates of the blade are determined based on the centroid coordinates of the first blade, the second blade, and the nacelle. This provides a data basis for subsequent correction of the top of the wind turbine by the UAV.

[0093] In some implementations, please refer to Figure 5 Determining the actual error distance at the top, based on the contour of the first target region, the vertical coordinates of the blade, and the center point of the top image, may include the following steps:

[0094] S510. Calculate and determine the error distance of the top image based on the vertical foot coordinates of the blade and the center point of the top image.

[0095] S520. Based on the area of ​​the cabin segmentation profile and the actual area of ​​the cabin top, determine the top scale factor.

[0096] S530. Based on the top image error distance and the top scale coefficient, determine the actual top error distance.

[0097] Specifically, the center point of the top image of the wind turbine is calculated using the top image to determine the center point of the top image. Then, based on the geometric projection relationship, the distance between the blade's vertical coordinates and the center point of the top image is calculated, and this distance is used as the top image error distance. The area of ​​the nacelle segmentation contour is calculated using the contour area calculation formula to determine the area of ​​the nacelle segmentation contour. Next, the calculated area of ​​the nacelle segmentation contour is compared with the actual area of ​​the top of the nacelle, and the top scale coefficient is determined by converting the area ratio of the two. Finally, the actual error distance of the top is determined by calculating based on the top image error distance and the top scale coefficient. The actual area of ​​the top of the nacelle is an adaptive inspection parameter. For an example, please refer to [link to example]. Figure 4b , Figure 4b The light blue dot is the center point of the top image.

[0098] For example, the center point of the top image is = The vertical coordinates of the blade are... The area of ​​the cabin's segmented outline is The actual area of ​​the cabin roof is The distance between the center point of the top image and the actual error at the top is calculated using the following formula:

[0099]

[0100]

[0101]

[0102]

[0103]

[0104] in, Width of the image at the top of the wind turbine. The height of the top image of the wind turbine. The distance of the top image error. This is the top scaling factor. This represents the actual error distance at the top.

[0105] In the above implementation, the top image error distance is determined by calculation based on the blade vertical foot coordinates and the center point of the top image. The top scale coefficient is determined based on the area of ​​the nacelle segmentation contour and the actual area of ​​the top of the nacelle. The actual top error distance is determined based on the top image error distance and the top scale coefficient, providing a data basis for subsequent correction of the top of the wind turbine on the UAV.

[0106] In some implementations, please refer to Figure 6a The method may also include the following steps:

[0107] S610. If the actual error distance at the top does not meet the preset conditions at the top, the corrected flight direction is determined by calculation based on the vertical coordinates of the blade and the center point of the top image.

[0108] S620: After correcting the flight direction and controlling the actual error distance of the top of the UAV, it collects the corrected image of the top of the wind turbine.

[0109] Among them, the corrected image of the top of the wind turbine corresponds to the center point of the corrected image.

[0110] Specifically, if the actual error distance at the top does not meet the preset conditions, it indicates that the drone's position is not within the preset center range of the wind turbine's top, indicating an offset. Therefore, the drone's position needs to be corrected to improve the accuracy and reliability of the overall inspection process and avoid inaccurate inspection data or omissions of key parts due to positional deviations. Using the drone's preset direction as a reference, the angle between the drone's current position and the blade's vertical coordinates is calculated based on the geometric relationship between the blade's vertical coordinates and the center point of the top image, determining the corrected flight direction. It should be noted that clockwise is considered positive. Then, the control algorithm guides the drone to fly along the corrected flight direction to the actual error distance at the top, reaching the desired position. Image acquisition then occurs, obtaining a corrected image of the wind turbine's top. The center point is calculated using the corrected image of the wind turbine's top, determining the center point of the corrected image. The preset condition at the top can be less than the center point correction threshold. .

[0111] For example, based on the blade foot coordinates and the center point of the top image = Calculations are performed to determine the correct flight direction. :

[0112]

[0113] S630: Based on the corrected image of the top of the wind turbine, identify and determine the outline of the target area for correction.

[0114] S640. Calculate and determine the vertical coordinates of the correction blade based on the contour of the target area.

[0115] Among them, the contour of the target area to be corrected corresponds to the coordinates of the center of mass of the cabin to be corrected.

[0116] Specifically, image recognition algorithms are used to identify the segmentation contours in the corrected image of the wind turbine top, thus determining the segmentation contours within the image. If the corrected image of the wind turbine top includes the segmentation contours corresponding to the two largest blades and the corrected nacelle segmentation contour, the centroid coordinates of each blade are determined using the centroid calculation formula. The centroid coordinates of the corrected nacelle are then determined using the same formula. Finally, the distance between the line connecting the corrected nacelle centroid coordinates and the centroid coordinates of the two blades is calculated using the vertical projection method, thereby determining the vertical foot coordinates of the corrected blades. If the corrected image of the wind turbine top does not include at least one of the segmentation contours corresponding to the two largest blades or the corrected nacelle segmentation contour, the UAV is controlled to return to its home position.

[0117] S650. Based on the contour of the target area to be corrected, the vertical coordinates of the corrected blade, and the center point of the corrected image, determine the actual error distance at the top after correction.

[0118] Specifically, based on geometric projection relationships, the distance between the vertical coordinates of the corrected blade and the center point of the corrected image is calculated, yielding the distance value between the vertical foot and the image center point, which is then used as the error distance for the corrected top image. The area of ​​the corrected nacelle segmentation contour is calculated using the contour area calculation formula to determine its area. Next, the calculated area of ​​the corrected nacelle segmentation contour is compared with the actual area of ​​the nacelle top, and a ratio conversion is performed to determine the corrected top scaling factor. Finally, based on the corrected top image error distance and the corrected top scaling factor, the actual error distance of the corrected top is determined.

[0119] S660. Based on the comparison between the corrected actual top error distance and the top preset conditions, if the corrected actual top error distance does not meet the top preset conditions and the number of iterations does not meet the preset number of iterations, the above process is iteratively executed until the corrected actual top error distance meets the top preset conditions and the corrected nacelle centroid coordinates are used as the nacelle centroid coordinates, and the corrected blade vertical foot coordinates are used as the blade vertical foot coordinates.

[0120] Specifically, the corrected actual top error distance is compared with the preset top conditions. If the corrected actual top error distance does not meet the preset top conditions, the number of position corrections already performed by the UAV is determined to establish the iteration count. If the number of iterations does not meet the preset iteration count, the above process is repeated until the corrected actual top error distance meets the preset top conditions. At this point, the corrected nacelle centroid coordinates are used as the nacelle centroid coordinates, and the corrected blade vertical foot coordinates are used as the blade vertical foot coordinates. If the number of iterations meets the preset iteration count, the UAV is controlled to return to base. The preset iteration count can be the maximum number of corrections in a single inspection phase. .

[0121] For example, please refer to Figure 6b , Figure 6b The image is corrected so that the actual error distance at the top meets the preset conditions at the top.

[0122] In the above embodiments, by correcting the position of the UAV, more accurate UAV flight movement and acquisition of inspection images can be achieved.

[0123] In some implementations, please refer to Figure 7 The axial flight direction of the wind turbine includes the flight direction on the windward side and the flight direction on the leeward side. Assuming the actual error distance at the top meets the preset conditions at the top, determining the axial flight direction of the wind turbine based on the vertical coordinates of the blades can include the following steps:

[0124] S710 determines the flight direction of the windward side based on the coordinates of the nacelle's center of mass and the vertical coordinates of the blades.

[0125] The S720 determines the flight direction on the leeward side based on the vertical coordinates of the blades and the coordinates of the nacelle's center of mass.

[0126] Specifically, assuming the actual error distance at the top meets the preset conditions, the angle between the UAV's current position and the nacelle is calculated based on the geometric relationship between the nacelle's center of mass coordinates and the blade's vertical foot coordinates, using the UAV's preset direction as a reference, thus determining the flight direction of the windward side. It should be noted that clockwise is considered positive.

[0127] Using the UAV's preset direction as a reference, the angle between the nacelle and the UAV's current position is calculated based on the geometric relationship between the blade's vertical foot coordinates and the nacelle's center of mass coordinates, thus determining the leeward flight direction. It should be noted that clockwise is recorded as positive.

[0128] It should be noted that the inspection sequence can be preset. Depending on the actual situation, you can choose to inspect the windward side first and then the leeward side, or inspect the leeward side first and then the windward side.

[0129] In the above implementation, the direction in which the drone needs to fly is determined, thereby enabling the drone to be controlled so that the wind turbine inspection can be completed subsequently.

[0130] In some implementations, please refer to Figure 8 The swept surface image of the wind turbine includes an image of the windward side of the wind turbine. Based on the axial flight direction of the wind turbine rotor, the horizontal safety distance, and the safe height of the blades, the process of controlling the UAV to fly to the center point of the wind turbine blades and acquire the swept surface image of the wind turbine may include the following steps:

[0131] S810 controls the UAV to fly horizontally at a safe distance based on the windward flight direction, and then flies downwards to a safe height above the blades, so that the UAV reaches the center point of the wind turbine blades.

[0132] S820 adjusts the nose direction of the UAV based on the windward flight direction and collects images of the windward side of the wind turbine.

[0133] Among them, the windward side image of the wind turbine corresponds to the center point of the windward side image.

[0134] Specifically, blade length Safe distance from wind turbines The sum of the distances between them constitutes the safe height of the blade. To ensure the safety of the UAV during inspection, a control algorithm is used to control the UAV to fly horizontally at a safe distance along the windward flight direction, ensuring that the UAV can effectively avoid the wind turbine blades while maintaining a suitable inspection range. After reaching the safe horizontal distance, the UAV is then controlled to fly downwards at the safe blade height, reaching the center point of the wind turbine blade. Upon reaching the center point, the UAV adjusts its nose direction based on the windward flight direction to ensure the nose is aligned with the windward side of the wind turbine before image acquisition, obtaining an image of the windward side of the wind turbine. The center point is calculated using this image of the windward side of the wind turbine, determining the center point of the windward side image. The horizontal safe distance is an adaptive inspection parameter.

[0135] In the above implementation, the drone is controlled to fly horizontally at a safe distance based on the windward flight direction, and then flies downward to a safe height relative to the blade, so that the drone reaches the center point of the wind turbine blade. Based on the windward flight direction, the drone's nose is adjusted to face the windward side and images of the wind turbine are collected, providing a data basis for subsequent blade and tower inspections.

[0136] In some implementations, please refer to Figure 9 The second target region contour includes the hub segmentation contour. Based on the second target region contour, determining the centroid coordinates of the wind turbine's central structure can include the following steps:

[0137] S910. Calculate the hub centroid coordinates based on the hub segmentation profile.

[0138] Specifically, during the inspection of the windward side, image recognition algorithms are used to identify the segmented contours in the wind turbine's windward image, thus determining the segmented contours within the image. If the windward image is determined to include the segmented contours of the three blades, the tower, and the hub, the hub's centroid coordinates are determined using a centroid calculation formula. If the windward image is determined not to include at least one of the segmented contours of the three blades, the tower, or the hub, the drone is controlled to return to its home position.

[0139] Determining the actual error distance of the swept surface based on the contour of the second target region, the centroid coordinates of the wind turbine's central structure, and the center point of the swept surface image can include the following steps:

[0140] S920: Based on the hub centroid coordinates and the center point of the windward image, the error distance of the windward image is determined.

[0141] S930. Based on the area of ​​the wheel hub segmentation profile and the actual area of ​​the wheel hub, determine the windward proportion coefficient.

[0142] S940. Based on the image error distance of the windward side and the windward side ratio coefficient, determine the actual error distance of the windward side.

[0143] Specifically, the distance between the hub's centroid coordinates and the center point of the windward image is calculated based on geometric projection relationships. This distance is then used as the windward image error distance. The hub segmentation contour is calculated using the contour area calculation formula to determine its area. Next, the calculated area of ​​the hub segmentation contour is compared with the actual area of ​​the hub. By converting the area ratio between the two, the windward scale factor is determined. Finally, the actual windward error distance is calculated based on the windward image error distance and the windward scale factor. The actual hub area is an adaptive inspection parameter.

[0144] For example, the center point of the windward image is = The coordinates of the wheel hub's center of mass are The area of ​​the wheel hub's segmented profile is... The actual area of ​​the wheel hub is The error distance between the center point of the windward image and the actual windward surface is calculated using the following formula:

[0145]

[0146]

[0147]

[0148]

[0149]

[0150] in, The width of the image of the windward side of the fan. The image height of the windward side of the wind turbine. For the image error distance of the windward side, This is the windward surface ratio coefficient. This represents the actual error distance on the windward side.

[0151] In the above implementation, calculations are performed based on the hub segmentation profile to determine the hub centroid coordinates. Then, calculations are performed based on the hub centroid coordinates and the center point of the windward image to determine the windward image error distance. Finally, the windward scale coefficient is determined based on the area of ​​the hub segmentation profile and the actual hub area. Finally, the actual windward error distance is determined based on the windward image error distance and the windward scale coefficient, providing a data foundation for subsequent windward correction of the UAV.

[0152] In some implementations, please refer to Figure 10a The method may also include the following steps:

[0153] S1010. If the actual error distance of the windward surface does not meet the preset conditions of the swept surface, the windward correction flight direction is determined by calculation based on the hub centroid coordinates and the center point of the windward surface image.

[0154] S1020. After controlling the actual error distance of the UAV's flight direction on the windward side based on the windward correction, collect the corrected image of the windward side.

[0155] Among them, the corrected windward image corresponds to the center point of the corrected windward image.

[0156] Specifically, if the actual error distance on the windward side does not meet the preset conditions for the swept surface, it indicates that the drone's position is not within the preset center range of the wind turbine's windward side, indicating an offset. Therefore, the drone's position needs to be corrected to improve the accuracy and reliability of the overall inspection process and avoid inaccurate inspection data or omissions of key parts due to positional deviations. Using the drone's preset direction as a reference, the angle between the drone's current position and the hub's center coordinates is calculated based on the geometric relationship between the hub's center coordinates and the center point of the windward side image, determining the windward correction flight direction. It should be noted that clockwise is considered positive. Then, using a control algorithm, the drone flies along the windward correction flight direction to the actual error distance on the windward side, reaching the desired position. Image acquisition then occurs, obtaining the corrected windward side image. The center point is calculated using the corrected windward side image to determine the center point of the windward correction image. The preset condition for the swept surface can be less than the center point correction threshold. .

[0157] For example, based on the hub centroid coordinates ) and the center point of the windward image = Calculations are performed to determine the correct flight direction against the wind. :

[0158]

[0159] S1030. Based on the corrected windward image, identify and determine the contour of the windward correction target area.

[0160] S1040. Based on the contour of the target area for windward correction, the coordinates of the center of gravity of the corrected wheel hub, and the center point of the windward correction image, determine the actual windward error distance after correction.

[0161] Among them, the contour of the target area for windward correction corresponds to the coordinates of the center of gravity of the wheel hub.

[0162] Specifically, by using image recognition algorithms to identify the corrected windward image, the contour of the windward correction target area can be determined. Given that the windward correction target area contour includes the segmented contours of the three blades and the corrected wheel hub segmented contour, the centroid coordinates of the corrected wheel hub are determined using the centroid calculation formula.

[0163] Based on the geometric projection relationship, the distance between the center coordinates of the corrected wheel hub and the center point of the corrected frontal image is calculated, yielding the distance between the wheel hub and the image center point, which is then used as the frontal correction image error distance. The area of ​​the corrected wheel hub segmentation contour is calculated using the contour area calculation formula to determine its area. Next, the calculated area of ​​the corrected wheel hub segmentation contour is compared with the actual area of ​​the wheel hub, and a conversion between their areas is performed to determine the frontal correction ratio coefficient. Finally, based on the frontal correction image error distance and the frontal correction ratio coefficient, the actual frontal error distance after correction is determined.

[0164] S1050. Based on the comparison between the corrected actual windward error distance and the preset conditions of the swept surface, if the corrected actual windward error distance does not meet the preset conditions of the swept surface and the number of iterations meets the preset number of windward surface iterations, the above process is iteratively executed until the corrected actual windward error distance meets the preset conditions of the swept surface and the corrected windward surface image is used as the windward surface image of the wind turbine.

[0165] Specifically, the corrected actual windward error distance is compared with the preset conditions for the swept surface. If the corrected actual windward error distance does not meet the preset conditions for the swept surface, the number of position corrections already performed by the UAV is determined, and the number of iterations is calculated. If the number of iterations does not meet the preset number of windward iterations, the above process can be repeated until the corrected actual windward error distance meets the preset conditions for the swept surface. At this point, the corrected windward image is used as the windward image of the wind turbine. If the number of iterations meets the preset number of windward iterations, the UAV is controlled to return to base. The preset number of windward iterations can be the maximum number of corrections in a single inspection phase. .

[0166] For example, please refer to Figure 10b , Figure 10b The image is designed to ensure that the actual windward error distance after correction meets the preset conditions of the swept surface.

[0167] In the above embodiments, by correcting the position of the UAV, more accurate UAV flight movement and acquisition of inspection images can be achieved.

[0168] In some implementations, please refer to Figure 11 The second target area also includes the third blade segmentation profile, the fourth blade segmentation profile, the fifth blade segmentation profile, and the tower segmentation profile. Determining the reference flight direction of the wind turbine components, assuming the actual error distance of the swept surface meets the preset conditions of the swept surface, may include the following steps:

[0169] S1110. Calculate the centroid coordinates of the third blade based on the segmented profile of the third blade.

[0170] S1120. Calculate the centroid coordinates of the fourth blade based on the segmented profile of the fourth blade.

[0171] S1130. Calculate the centroid coordinates of the fifth blade based on the segmented profile of the fifth blade.

[0172] S1140. Calculate the coordinates of the tower centroid based on the tower segmentation profile.

[0173] S1150. Based on the coordinates of the centroid of the third blade and the coordinates of the centroid of the wind turbine center structure, the flight direction of the third blade is determined.

[0174] S1160. Based on the coordinates of the fourth blade's center of mass and the coordinates of the wind turbine's central structure's center of mass, determine the fourth blade's flight direction.

[0175] S1170. Based on the coordinates of the fifth blade's center of mass and the coordinates of the wind turbine's central structure's center of mass, determine the flight direction of the fifth blade.

[0176] S1180. Based on the coordinates of the tower's center of mass and the coordinates of the wind turbine's center of mass, the flight direction of the tower is determined.

[0177] Specifically, image recognition algorithms are used to identify the segmented contours in the windward image of the wind turbine. Given that the windward image includes the segmented contours of the third, fourth, and fifth blades, the tower, and the rotor's central structure, the centroid coordinates of the third blade are determined using a centroid calculation formula. Similarly, the centroid coordinates of the fourth, fifth, and finally the tower are determined using the same formula.

[0178] Using the drone's preset direction as a reference, the required flight angle of the drone towards the third blade is calculated based on the geometric relationship between the coordinates of the third blade's center of mass and the coordinates of the wind turbine's central structure, thus determining the flight direction of the third blade. Similarly, the required flight angle of the drone towards the fourth blade is calculated based on the geometric relationship between the coordinates of the fourth blade's center of mass and the coordinates of the wind turbine's central structure, thus determining the flight direction of the fourth blade. Finally, the required flight angle of the drone towards the tower is calculated based on the geometric relationship between the coordinates of the tower's center of mass and the coordinates of the wind turbine's central structure, thus determining the tower's flight direction. It should be noted that clockwise is considered positive.

[0179] It should be noted that during the inspection of the windward side, the coordinates of the center of mass of the wind turbine are the coordinates of the hub's center of mass, while during the inspection of the leeward side, the coordinates of the center of mass of the wind turbine are the coordinates of the nacelle's center of mass.

[0180] In the above embodiments, the inspection direction of the blades and tower is determined so that the UAV can acquire inspection images more accurately.

[0181] In some implementations, please refer to Figure 12 The component safety distance includes blade length, the component inspection image overlap rate includes blade inspection image overlap rate, and the component inspection shooting count includes the number of blade inspection shooting counts. Based on the component safety distance, component inspection image overlap rate, and component inspection shooting count, the set of single-step movement distance sequences and the total movement distance of the component are determined, which may include the following steps:

[0182] S1210. Based on the blade length, the overlap rate of the blade inspection images, and the number of times the blade is photographed during inspection, determine the coverage length of a single photograph of the blade.

[0183] S1220: Based on the coverage length of a single shot and the overlap rate of the blade inspection images, calculate and determine the initial step distance and subsequent step length of the blade.

[0184] S1230. Based on the first step distance of the blade, the subsequent step length of the blade, and the number of times the blade is inspected and photographed, determine the set of single-step movement distance sequences of the blade.

[0185] S1240. Based on the blade length and the coverage length of a single image taken by the blade, determine the total moving distance of the blade.

[0186] Specifically, during blade inspection, a blade inspection image overlap rate (e.g., 20%) is set to ensure complete blade coverage by the inspection image. The single-shot coverage length of the blade is determined by calculating the blade length effectively covered by a single inspection image using the blade length, the blade inspection image overlap rate, and the number of blade inspection shots. The initial step distance of the blade is obtained by dividing the single-shot coverage length by an integer 2. Subsequent step lengths of the blade are determined by calculating the single-shot coverage length and the blade inspection image overlap rate. The number of times the subsequent step length is repeated is determined based on the number of blade inspection shots to ensure that the distribution of shooting points throughout the blade inspection process meets the requirements of complete image coverage and the set overlap rate. Thus, the initial step distance and the subsequent step lengths repeated a specific number of times constitute a set of single-step movement distance sequences for the blade. This set of single-step movement distance sequences guides the setting of the distance for each single step as the UAV moves from the actual blade center point towards the blade tip region. Based on the blade length and the single-shot coverage length, the total blade movement distance is determined using appropriate calculation methods. The total blade travel distance is the distance from the blade tip back to the actual blade center point after a single blade has been inspected.

[0187] For example, the blade length is The overlap rate of blade inspection images The number of times the blades were inspected and photographed was: The calculation is performed using the following formula:

[0188] Calculate the coverage length of a single image taken of the blade. :

[0189]

[0190] Calculate the set of single-step movement distance sequences of the blade :

[0191] First step distance of the blade (actual center point of the blade → center of the first shooting point)

[0192]

[0193] Subsequent step length of the leaf (distance moved between shooting points)

[0194]

[0195] Set of single-step movement distance sequences for blades:

[0196]

[0197] Calculate the total travel distance of the blades :

[0198]

[0199] It should be noted that since the length of each blade of a wind turbine is consistent, the set of single-step movement distance sequences and the total movement distance of each blade during the upwind inspection process are consistent with the set of single-step movement distance sequences and the total movement distance of each blade during the leeward inspection process.

[0200] In the above implementation, the inspection movement distance is dynamically planned by determining the set of single-step movement distance sequences of the blades and the total movement distance of the blades.

[0201] In some implementations, please refer to Figure 13 The component safety distance includes the tower safety height, the component inspection image overlap rate includes the tower inspection image overlap rate, and the component inspection shooting count includes the tower inspection shooting count. Based on the component safety distance, component inspection image overlap rate, and component inspection shooting count, the set of single-step movement distance sequences and the total movement distance of the component are determined, which may include the following steps:

[0202] S1310. Based on the tower's safe height, the overlap rate of tower inspection images, and the number of tower inspection shots, determine the coverage length of a single tower shot.

[0203] S1320. Based on the coverage length of a single tower shot and the overlap rate of tower inspection images, calculate and determine the first step distance and subsequent step length of the tower.

[0204] S1330. Based on the first step distance of the tower, the subsequent step length of the tower, and the number of times the tower is inspected and photographed, determine the set of single-step movement distance sequences of the tower.

[0205] S1340. Based on the tower's safe height and the coverage length of a single shot taken by the tower, determine the total moving distance of the tower.

[0206] Specifically, during tower inspection, to ensure complete tower coverage by the inspection image, a tower inspection image overlap rate (e.g., 20%) is set. The tower's safe height, the tower inspection image overlap rate, and the number of tower inspection shots are used to calculate the single-shot coverage length, representing the effective tower length covered by a single inspection image. The first-step distance is obtained by dividing the single-shot coverage length by an integer 2. Subsequent step lengths are determined using the single-shot coverage length and the tower inspection image overlap rate. The number of subsequent step lengths is determined by the number of tower inspection shots to ensure that the distribution of shooting points throughout the tower inspection process meets the requirements of complete image coverage and the set overlap rate. Thus, the first-step distance and the subsequent step lengths repeated a specific number of times constitute a sequence of single-step movement distances for the blade. This sequence of single-step movement distances guides the UAV's movement from the actual blade center point towards the bottom of the tower, setting the distance for each single step. Based on the tower's safe height and the coverage length of a single tower shot, the total tower movement distance is determined using appropriate calculation methods. This total tower movement distance is used to determine the distance traveled from the bottom of the tower back to the actual blade center point after the tower inspection is completed.

[0207] For example, the height of the wind turbine tower is The tower needs to be inspected and photographed a number of times. The minimum safe height of the tower is The overlap rate of the inspection images is Based on the height of the wind turbine tower The tower needs to be inspected and photographed a number of times. Determine the safe height of the tower. Calculate using the following formula:

[0208] Calculate the coverage length of a single shot taken by the tower. :

[0209]

[0210] Calculate the set of single-step movement distance sequences of the tower. :

[0211] First step distance of the tower (actual blade center point → center of the first shooting point)

[0212]

[0213] Tower follow-up step length (distance moved between shooting points)

[0214]

[0215] Tower single-step movement distance sequence set

[0216]

[0217] Calculate the total moving distance of the tower. :

[0218]

[0219] It should be noted that, since the safe height of the tower is consistent, the set of single-step movement distance sequences and the total movement distance of the tower during the windward inspection process are consistent with the set of single-step movement distance sequences and the total movement distance of the tower during the leeward inspection process.

[0220] In the above embodiments, the inspection movement distance is dynamically planned by determining the set of single-step movement distance sequences of the tower and the total movement distance of the tower.

[0221] In some implementations, please refer to Figure 14a Image acquisition based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance yields a component inspection image set, which may include the following steps:

[0222] S1410: Based on the flight direction of the third blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and collect images to obtain the third blade inspection image set.

[0223] S1420: Control the UAV to fly to the actual blade center point based on the total blade movement distance.

[0224] Specifically, based on the UAV's flight control system, and combining the flight direction of the third blade with the sequence of single-step movement distances of the blades, a control algorithm ensures that the UAV starts from the actual blade center point and performs inspection flights along the flight direction of the third blade. During this inspection flight, the UAV advances step-by-step according to the sequence of single-step movement distances, and after each single-step movement, it acquires image data in real time, obtaining the corresponding inspection image of the third blade. After each flight and image acquisition, a single-step movement distance inspection task is completed, ultimately forming a complete set of third blade inspection images. After completing the acquisition of the third blade inspection images, the UAV adjusts its flight direction and reverses the total blade movement distance based on the flight control system and control algorithm, returning to the actual blade center point to begin acquiring inspection images for the next blade.

[0225] For example, the flight direction of the third blade is The set of single-step movement distance sequences of the blade is The total distance the blades traveled was Control the drone from the actual center point of the blade towards... Directional flight, initial flight distance is After flight, the third blade inspection image was acquired. Then it continued towards... Fly in direction, each flight distance is ,repeat Each time, the third blade inspection image is acquired after each flight. After completing the flight of the single-step movement distance sequence of the blades, the third blade inspection image set can be obtained. Finally, the UAV is controlled to fly in the opposite direction. Return to the actual center point of the blade.

[0226] It should be noted that during the inspection of the windward side, a third blade inspection is performed according to the above procedure, and during the inspection of the leeward side, a third blade inspection is performed according to the above procedure.

[0227] S1430: Based on the flight direction of the fourth blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and collect images to obtain the inspection image set of the fourth blade.

[0228] S1440: Control the UAV to fly to the actual center point of the blade based on the total blade movement distance.

[0229] Specifically, based on the UAV's flight control system, and combining the flight direction of the fourth blade with the sequence of single-step movement distances of the blade, a control algorithm ensures that the UAV starts from the actual blade center point and performs inspection flights along the fourth blade's flight direction. During this inspection flight, the UAV advances step-by-step according to the sequence of single-step movement distances, and after each single-step movement, it acquires image data in real time, obtaining the corresponding inspection image of the fourth blade. After each flight and image acquisition, a single-step movement distance inspection task is completed, ultimately forming a complete set of fourth blade inspection images. After completing the fourth blade inspection image acquisition, the UAV adjusts its flight direction and reverses the total blade movement distance based on the flight control system and control algorithm, returning to the actual blade center point to begin acquiring inspection images for the next blade.

[0230] S1450: Based on the flight direction of the fifth blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and collect images to obtain the inspection image set of the fifth blade.

[0231] S1460: Control the UAV to fly to the actual center point of the blade based on the total blade movement distance.

[0232] Specifically, based on the UAV's flight control system, and combining the flight direction of the fifth blade with the sequence of single-step movement distances of the blade, a control algorithm ensures that the UAV starts from the actual blade center point and performs inspection flights along the fifth blade's flight direction. During this inspection flight, the UAV advances step-by-step according to the sequence of single-step movement distances, and after each single-step movement, it acquires image data in real time, obtaining the corresponding inspection image of the fifth blade. After each flight and image acquisition, a single-step movement distance inspection task is completed, ultimately forming a complete set of fifth blade inspection images. After acquiring the fifth blade inspection images, the UAV adjusts its flight direction and reverses the total blade movement distance based on the flight control system and control algorithm, returning to the actual blade center point to begin acquiring inspection images for the next blade.

[0233] For example, please refer to Figure 14b , Figure 14b Images of wind turbine blades and towers for inspection.

[0234] In the above implementation, by inspecting each blade of the wind turbine, the inspection images of each blade are collected so that subsequent operation and maintenance inspections of each blade can be carried out.

[0235] In some implementations, please refer to Figure 15 Image acquisition based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance yields a component inspection image set, which may include the following steps:

[0236] S1510: Based on the tower flight direction and the tower single-step movement distance sequence set, the UAV is controlled to fly from the actual blade center point and collect images to obtain the tower inspection image set.

[0237] S1520: Control the UAV to fly to the actual blade center point based on the total moving distance of the tower.

[0238] Specifically, based on the UAV's flight control system, and combining the tower's flight direction with a sequence of single-step tower movement distances, a control algorithm ensures that the UAV starts from the actual blade center point and performs inspection flights along the tower's flight direction. During these inspection flights, the UAV advances step-by-step according to the tower's single-step movement distance sequence, and after each single-step movement, it acquires image data in real time, obtaining the corresponding tower inspection image. After each flight and image acquisition, a single-step movement distance inspection task is completed, ultimately forming a complete tower inspection image set. After completing the tower inspection image acquisition, the UAV adjusts its flight direction and reverses the total tower movement distance based on the flight control system and control algorithm, returning to the actual blade center point.

[0239] For example, please refer to Figure 14b , Figure 14b Images are provided for inspecting the blades and tower of a wind turbine. It should be noted that during the inspection of the windward side, a tower inspection is performed once following the same procedure; during the inspection of the leeward side, a tower inspection is performed once following the same procedure.

[0240] In the above implementation, the tower of the wind turbine is inspected to collect tower inspection images, so as to facilitate subsequent tower operation and maintenance inspection.

[0241] In some implementations, please refer to Figure 16 The swept surface image of the wind turbine includes an image of the leeward side of the wind turbine. Based on the axial flight direction of the wind turbine rotor, the horizontal safety distance, and the safe height of the blades, controlling the UAV to fly to the center point of the wind turbine blades and acquire the swept surface image of the wind turbine may include the following steps:

[0242] S1610: Based on the flight direction on the leeward side, control the drone to fly horizontally at a safe distance, and then fly downwards to a safe height above the blades, so that the drone reaches the center point of the wind turbine blades.

[0243] S1620: Adjust the nose direction of the UAV based on the flight direction on the leeward side and collect images of the leeward side of the wind turbine.

[0244] Among them, the leeward side image of the wind turbine corresponds to the center point of the leeward side image.

[0245] Specifically, blade length Safe distance from wind turbines The sum of the distances between them constitutes the blade's safe height. To ensure the safety of the UAV during inspection, a control algorithm is used to control the UAV to fly horizontally along the leeward flight direction at a safe distance, ensuring that the UAV can effectively avoid the wind turbine blades while maintaining a suitable inspection range. After reaching the safe horizontal distance, the UAV is then controlled to fly downwards to the blade's safe height, reaching the center point of the wind turbine blade. Upon reaching the center point, the UAV adjusts its nose direction based on the leeward flight direction to ensure the nose is aligned with the windward side of the wind turbine before image acquisition, obtaining an image of the leeward side of the wind turbine. The center point is calculated using this leeward side image to determine the center point of the leeward side image. The horizontal safe distance is an adaptive inspection parameter.

[0246] In the above implementation, the drone is controlled to fly horizontally at a safe distance based on the leeward flight direction, and then flies downward to a safe height relative to the blade, so that the drone reaches the center point of the wind turbine blade. Based on the leeward flight direction, the drone's nose is adjusted to face the wind turbine and images of the leeward side are collected, providing a data basis for subsequent blade and tower inspections.

[0247] In some implementations, please refer to Figure 17 The second target region outline includes the nacelle segmentation outline. Based on the second target region outline, determining the centroid coordinates of the wind turbine's central structure can include the following steps:

[0248] S1710. Calculate the coordinates of the cabin's centroid based on the cabin's segmented profile.

[0249] Specifically, during the inspection of the leeward side, image recognition algorithms are used to identify the segmented contours in the wind turbine's leeward side image, thus determining the segmented contours within the image. If the leeward side image is determined to include the segmented contours of the three blades, the tower, and the nacelle, the nacelle's centroid coordinates are calculated using a centroid calculation formula. If the leeward side image is determined not to include at least one of the segmented contours of the three blades, the tower, or the nacelle, the drone is controlled to return to base.

[0250] Determining the actual error distance of the swept surface based on the contour of the second target region, the coordinates of the centroid of the wind turbine's central structure, and the center point of the swept surface image may include the following steps:

[0251] S1720. Based on the coordinates of the cabin's center of mass and the center point of the leeward image, the error distance of the leeward image is determined.

[0252] S1730. Based on the area of ​​the cabin segmentation profile and the actual area of ​​the cabin, determine the leeward proportion coefficient.

[0253] S1740. Based on the leeward side image error distance and the leeward side scale coefficient, determine the actual leeward side error distance.

[0254] Specifically, based on the geometric projection relationship, the distance between the centroid coordinates of the cabin and the center point of the leeward image is calculated to obtain the distance between the cabin and the image center point, and this distance is used as the leeward image error distance. The cabin segmentation contour is calculated using the contour area calculation formula to determine the area of ​​the cabin segmentation contour. Next, the calculated area of ​​the cabin segmentation contour is compared with the actual area of ​​the cabin, and the leeward scale coefficient is determined by converting the area ratio between the two. Finally, the actual leeward error distance is calculated based on the leeward image error distance and the leeward scale coefficient. The actual cabin area is an adaptive inspection parameter.

[0255] For example, the center point of the leeward image is = The coordinates of the cabin's center of mass are... The area of ​​the cabin's segmented outline is The actual cabin area is The error distance between the center point of the leeward image and the actual leeward side is calculated using the following formula:

[0256]

[0257]

[0258]

[0259]

[0260]

[0261] in, The width of the image on the leeward side of the wind turbine. The height of the image on the leeward side of the wind turbine. The leeward side image error distance, This is the leeward side ratio coefficient. This represents the actual error distance on the leeward side.

[0262] In the above implementation, calculations are performed based on the segmented outline of the cabin to determine the coordinates of the cabin's centroid. Calculations are then performed based on the coordinates of the cabin's centroid and the center point of the leeward image to determine the leeward image error distance. Based on the area of ​​the segmented outline of the cabin and the actual area of ​​the cabin, the leeward scale coefficient is determined. Finally, based on the leeward image error distance and the leeward scale coefficient, the actual leeward error distance is determined, providing a data basis for subsequent windward correction of the UAV.

[0263] In some implementations, please refer to Figure 18 The method may also include the following steps:

[0264] S1810. If the actual error distance on the leeward side does not meet the preset conditions of the swept surface, the leeward correction flight direction is determined by calculation based on the coordinates of the cabin centroid and the center point of the leeward side image.

[0265] S1820. After controlling the actual error distance of the UAV's flight leeward side based on the leeward correction flight direction control, the corrected leeward side image is collected.

[0266] Among them, the corrected leeward image corresponds to the center point of the leeward corrected image.

[0267] Specifically, if the actual error distance on the leeward side does not meet the preset conditions for the swept surface, it indicates that the UAV's position is not within the preset center range of the wind turbine's leeward side, indicating an offset. Therefore, the UAV's position needs to be corrected to improve the accuracy and reliability of the overall inspection process and avoid inaccurate inspection data or omissions of key parts due to positional deviations. Using the UAV's preset direction as a reference, the angle between the UAV's current position and the nacelle's centroid coordinates is calculated based on the geometric relationship between the nacelle's centroid coordinates and the center point of the leeward side image, determining the leeward correction flight direction. It should be noted that clockwise is considered positive. Then, using a control algorithm, the UAV is controlled to fly along the leeward correction flight direction, covering the actual error distance on the leeward side, until it reaches the desired position. Image acquisition then occurs, obtaining the corrected leeward side image. The center point is calculated using the corrected leeward side image to determine the center point of the leeward correction image. The preset condition for the swept surface can be less than the center point correction threshold. .

[0268] For example, based on the coordinates of the cabin's center of mass and the center point of the leeward side image = Calculations are performed to determine the leeward direction for correct flight. :

[0269]

[0270] S1830. Based on the corrected leeward image, identify and determine the contour of the leeward correction target area.

[0271] S1840. Based on the contour of the leeward correction target area, the coordinates of the center of mass of the corrected cabin, and the center point of the leeward correction image, determine the actual leeward error distance after correction.

[0272] Among them, the outline of the leeward correction target area corresponds to the coordinates of the center of mass of the correction nacelle;

[0273] Specifically, by using image recognition algorithms to identify the corrected windward image, the outline of the leeward correction target area can be determined. Given that the leeward correction target area outline includes the segmented outlines of the three blades and the nacelle segmented outline, the centroid calculation formula is used to calculate the nacelle segmented outline and determine the centroid coordinates of the nacelle.

[0274] Based on the geometric projection relationship, the distance between the centroid coordinates of the corrected nacelle and the center point of the leeward corrected image is calculated, yielding the distance between the nacelle and the image center point, which is then used as the leeward corrected image error distance. The area of ​​the corrected nacelle segmentation contour is calculated using the contour area calculation formula to determine its area. Next, the calculated area of ​​the corrected nacelle segmentation contour is compared with the actual area of ​​the nacelle, and a leeward correction ratio is calculated to determine the leeward correction scale factor. Finally, based on the leeward corrected image error distance and the leeward correction scale factor, the actual leeward error distance after correction is determined.

[0275] S1850. Based on the comparison between the corrected leeward actual error distance and the sweeping surface preset conditions, if the corrected leeward actual error distance does not meet the sweeping surface preset conditions and the number of iterations meets the leeward surface iteration preset number, the above process is iteratively executed until the corrected leeward actual error distance meets the sweeping surface preset conditions and the corrected leeward surface image is used as the wind turbine leeward surface image.

[0276] Specifically, the actual leeward error distance after correction is compared with the preset conditions of the swept surface. If the actual leeward error distance after correction does not meet the preset conditions of the swept surface, the number of position corrections already performed by the UAV is determined to determine the number of iterations. If the number of iterations does not meet the preset number of leeward surface iterations, the above process can be repeated until the actual leeward error distance after correction meets the preset conditions of the swept surface. At this point, the corrected leeward surface image is used as the wind turbine leeward surface image. If the number of iterations meets the preset number of leeward surface iterations, the UAV is controlled to return to base. The preset number of leeward surface iterations can be the maximum number of corrections in a single inspection phase. .

[0277] In the above embodiments, by correcting the position of the UAV, more accurate UAV flight movement and acquisition of inspection images can be achieved.

[0278] In some implementations, the focal length of the UAV camera can be adjusted based on the acquired images and calculated parameters, thereby achieving more accurate image acquisition.

[0279] This specification provides an embodiment of a wind turbine inspection device 1900 based on a drone. Please refer to [link / reference]. Figure 19 The wind turbine inspection device 1900 based on UAV includes: a wind turbine top image acquisition module 1910, a wind turbine swept surface image acquisition module 1920, a swept surface actual error distance determination module 1930, a component reference flight direction determination module 1940, a component movement data determination module 1950, and a component inspection image acquisition module 1960.

[0280] The wind turbine top image acquisition module 1910 is used to acquire an image of the top of the wind turbine when the drone flies to the initial cruise position and the power is within a preset range. The initial cruise position includes the coordinates and safe altitude of the wind turbine.

[0281] The wind turbine swept surface image acquisition module 1920 is used to control the UAV to fly to the center point of the wind turbine blades of the wind turbine generator and acquire the wind turbine swept surface image based on the top image of the wind turbine.

[0282] The actual error distance determination module 1930 for swept surface is used to determine the actual error distance of the swept surface based on the wind turbine swept surface image.

[0283] The component reference flight direction determination module 1940 is used to determine the component reference flight direction of the wind turbine when the actual error distance of the swept surface meets the preset conditions of the swept surface.

[0284] The component movement data determination module 1950 is used to determine the set of single-step movement distance sequences and the total movement distance of the component based on the component safety distance, the overlap rate of component inspection images, and the number of component inspection shots.

[0285] The component inspection image acquisition module 1960 is used to acquire images based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance to obtain a component inspection image set.

[0286] For a detailed description of the drone-based wind turbine inspection device, please refer to the description of the drone-based wind turbine inspection method above, which will not be repeated here.

[0287] In some embodiments, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 20 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for inspecting wind turbines based on unmanned aerial vehicles (UAVs). The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0288] Those skilled in the art will understand that Figure 20 The structures shown are merely block diagrams of some structures related to the solutions disclosed in this specification, and do not constitute a limitation on the computer device to which the solutions disclosed in this specification are applied. Specifically, the computer device may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements.

[0289] In some embodiments, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method steps described above.

[0290] This specification provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method in any of the above embodiments.

[0291] One embodiment of this specification provides a computer program product including instructions that, when executed by a processor of a computer device, enable the computer device to perform the steps of the method described in any of the above embodiments.

[0292] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

Claims

1. A method for inspecting wind turbines based on unmanned aerial vehicles (UAVs), characterized in that, The method includes: When the drone reaches the initial cruise position and the battery level is within a preset range, an image of the top of the wind turbine is acquired, wherein the initial cruise position includes the coordinates and safe altitude of the wind turbine. Based on the image of the top of the wind turbine, the drone is controlled to fly to the center point of the wind turbine blades and collect images of the swept surface of the wind turbine. The actual error distance of the swept surface is determined based on the image of the swept surface of the wind turbine. If the actual error distance of the swept surface meets the preset conditions of the swept surface, the reference flight direction of the wind turbine component is determined. Based on the component safety distance, component inspection image overlap rate, and component inspection shooting number, determine the component single-step movement distance sequence set and the component total movement distance; Image acquisition is performed based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance to obtain a component inspection image set.

2. The method according to claim 1, characterized in that, The top image of the wind turbine corresponds to a center point of the top image. Based on the top image of the wind turbine, controlling the drone to fly to the center point of the wind turbine blades and acquire an image of the swept surface of the wind turbine includes: Based on the image of the top of the wind turbine, the outline of the first target area is determined; The blade's vertical coordinates are determined based on the contour of the first target region. Based on the outline of the first target region, the vertical coordinates of the blade, and the center point of the top image, the actual error distance at the top is determined. If the actual error distance at the top meets the preset conditions at the top, the axial flight direction of the wind turbine is determined based on the vertical coordinates of the blade. Based on the axial flight direction of the wind turbine, the horizontal safety distance, and the blade safety height, the UAV is controlled to fly to the center point of the wind turbine blade and collect images of the swept surface of the wind turbine.

3. The method according to claim 1, characterized in that, The swept surface image of the wind turbine corresponds to a center point of the swept surface image. Determining the actual error distance of the swept surface based on the swept surface image of the wind turbine includes: Based on the image of the swept surface of the wind turbine, the outline of the second target region is determined; Based on the contour of the second target region, determine the coordinates of the centroid of the wind turbine's central structure; Based on the contour of the second target region, the centroid coordinates of the wind turbine center structure, and the center point of the swept surface image, the actual error distance of the swept surface is determined.

4. The method according to claim 2, characterized in that, The first target region contour includes a first blade segmentation contour, a second blade segmentation contour, and a nacelle segmentation contour. The step of calculating and determining the blade vertical coordinates based on the first target region contour includes: The centroid coordinates of the first blade are determined by calculation based on the segmented profile of the first blade. The centroid coordinates of the second blade are determined by calculation based on the segmented profile of the second blade. The coordinates of the cabin's centroid are determined by calculation based on the cabin's segmented profile. The vertical foot coordinates of the blade are determined by calculation based on the centroid coordinates of the first blade, the second blade, and the nacelle.

5. The method according to claim 4, characterized in that, The step of determining the actual error distance at the top based on the contour of the first target region, the vertical coordinates of the blade, and the center point of the top image includes: The error distance of the top image is determined by calculating based on the vertical foot coordinates of the blade and the center point of the top image. Based on the area of ​​the cabin segmentation outline and the actual area of ​​the cabin top, determine the top scaling factor; The actual error distance at the top is determined based on the top image error distance and the top scaling factor.

6. The method according to claim 2, characterized in that, The method further includes: If the actual error distance at the top does not meet the preset conditions at the top, the corrected flight direction is determined based on the vertical coordinates of the blade and the center point of the top image. After controlling the UAV to fly at the actual error distance of the top based on the corrected flight direction, the UAV is captured as a corrected image of the top of the wind turbine, wherein the corrected image of the top of the wind turbine corresponds to the center point of the corrected image. Based on the corrected image of the top of the wind turbine, the outline of the correction target area is determined, wherein the outline of the correction target area corresponds to the coordinates of the centroid of the correction nacelle. The vertical coordinates of the correction blade are determined based on the contour of the target area. Based on the contour of the target area to be corrected, the vertical coordinates of the corrected blade, and the center point of the corrected image, the actual error distance at the top after correction is determined. Based on the comparison between the corrected actual top error distance and the top preset condition, if the corrected actual top error distance does not meet the top preset condition and the number of iterations does not meet the preset number of iterations, the above process is iteratively executed until the corrected actual top error distance meets the top preset condition and the corrected nacelle centroid coordinates are used as the nacelle centroid coordinates, and the corrected blade vertical foot coordinates are used as the blade vertical foot coordinates.

7. The method according to claim 2 or 6, characterized in that, The axial flight direction of the wind turbine includes a windward flight direction and a leeward flight direction. Determining the axial flight direction of the wind turbine based on the blade foot coordinates, assuming the actual error distance at the top meets the preset conditions at the top, includes: The flight direction of the windward surface is determined based on the coordinates of the nacelle's center of mass and the vertical coordinates of the blades. The flight direction on the leeward side is determined based on the vertical foot coordinates of the blade and the center of mass coordinates of the nacelle.

8. The method according to claim 7, characterized in that, The swept surface image of the wind turbine includes an image of the windward side of the wind turbine. The process of controlling the UAV to fly to the center point of the wind turbine blades and acquire the swept surface image, based on the axial flight direction of the wind turbine rotor, the horizontal safety distance, and the blade safety height, includes: Based on the windward flight direction, the drone is controlled to fly horizontally at the horizontal safe distance, and then flies downward at the blade safe height, so that the drone flies to the center point of the wind turbine blade. The UAV's nose is adjusted based on the windward flight direction, and an image of the wind turbine's windward surface is acquired, wherein the wind turbine's windward surface image corresponds to a center point of the windward surface image.

9. The method according to claim 8, characterized in that, The second target region contour includes the hub segmentation contour. Based on the second target region contour, the coordinates of the centroid of the wind turbine's central structure are determined, including: The coordinates of the wheel hub's centroid are determined based on the wheel hub's segmented profile. Based on the contour of the second target region, the centroid coordinates of the wind turbine's central structure, and the center point of the swept surface image, the actual error distance of the swept surface is determined, including: The error distance of the windward image is determined by calculating based on the coordinates of the wheel hub centroid and the center point of the windward image. The windward proportion coefficient is determined based on the area of ​​the wheel hub segmentation profile and the actual area of ​​the wheel hub. Based on the image error distance of the windward side and the windward side ratio coefficient, the actual error distance of the windward side is determined.

10. The method according to claim 9, characterized in that, The method further includes: If the actual error distance of the windward surface does not meet the preset conditions of the swept surface, the windward correction flight direction is determined based on the hub centroid coordinates and the center point of the windward surface image. After controlling the UAV to fly at the actual error distance of the windward side based on the windward correction flight direction, the corrected windward side image is acquired, wherein the corrected windward side image corresponds to the center point of the windward correction image. Based on the corrected windward image, the contour of the windward correction target area is determined, wherein the contour of the windward correction target area corresponds to the coordinates of the center of gravity of the corrected wheel hub. Based on the outline of the target area for windward correction, the coordinates of the center of gravity of the corrected wheel hub, and the center point of the windward correction image, the actual windward error distance after correction is determined. Based on the comparison between the corrected actual windward error distance and the preset conditions of the swept surface, if the corrected actual windward error distance does not meet the preset conditions of the swept surface and the number of iterations meets the preset number of windward surface iterations, the above process is iteratively executed until the corrected actual windward error distance meets the preset conditions of the swept surface and the corrected windward surface image is used as the windward surface image of the wind turbine.

11. The method according to claim 3, characterized in that, The second target area also includes the third blade segmentation profile, the fourth blade segmentation profile, the fifth blade segmentation profile, and the tower segmentation profile. Determining the reference flight direction of the wind turbine components when the actual error distance of the swept surface meets the preset conditions of the swept surface includes: The centroid coordinates of the third blade are determined by calculating based on the segmented contour of the third blade. The centroid coordinates of the fourth blade are determined by calculating based on the segmented contour of the fourth blade. The centroid coordinates of the fifth blade are determined by calculation based on the segmented contour of the fifth blade. The coordinates of the tower's centroid are determined based on the tower's segmented profile. The flight direction of the third blade is determined based on the coordinates of the centroid of the third blade and the coordinates of the centroid of the wind turbine center structure. The flight direction of the fourth blade is determined based on the coordinates of the centroid of the fourth blade and the coordinates of the centroid of the wind turbine's central structure. The flight direction of the fifth blade is determined based on the coordinates of the centroid of the fifth blade and the coordinates of the centroid of the wind turbine's central structure. The flight direction of the tower is determined based on the coordinates of the tower's center of mass and the coordinates of the wind turbine's central structure.

12. The method according to claim 1, characterized in that, The component safety distance includes blade length, the component inspection image overlap rate includes blade inspection image overlap rate, and the component inspection shooting count includes the number of blade inspection shooting counts. The determination of the component single-step movement distance sequence set and the component total movement distance based on the component safety distance, component inspection image overlap rate, and component inspection shooting count includes: The coverage length of a single image of a leaf is determined based on the leaf length, the overlap rate of the leaf inspection images, and the number of times the leaf inspection images are taken. The first step distance and subsequent step length of the blade are determined by calculating based on the coverage length of a single shot and the overlap rate of the blade inspection images. Based on the first step distance of the blade, the subsequent step length of the blade, and the number of times the blade is inspected and photographed, a set of single-step movement distance sequences of the blade is determined. The total moving distance of the blade is determined based on the blade length and the coverage length of a single shot taken by the blade.

13. The method according to claim 1, characterized in that, The component safety distance includes the tower safety height, the component inspection image overlap rate includes the tower inspection image overlap rate, and the component inspection shooting count includes the tower inspection shooting count. The determination of the component single-step movement distance sequence set and the component total movement distance based on the component safety distance, component inspection image overlap rate, and component inspection shooting count includes: Based on the tower safety height, the tower inspection image overlap rate, and the number of tower inspection shots, the coverage length of a single tower shot is determined. The first step distance and subsequent step length of the tower are determined by calculating the coverage length of a single shot of the tower and the overlap rate of the tower inspection images. Based on the first step distance of the tower, the subsequent step length of the tower, and the number of times the tower is inspected and photographed, a set of single-step movement distance sequences of the tower is determined; The total moving distance of the tower is determined based on the tower's safe height and the length of coverage of a single shot taken by the tower.

14. The method according to claim 12, characterized in that, The image acquisition based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance yields a component inspection image set, including: Based on the flight direction of the third blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and perform image acquisition to obtain the third blade inspection image set; The UAV is controlled to fly to the actual center point of the blade based on the total moving distance of the blade. Based on the flight direction of the fourth blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and perform image acquisition to obtain the fourth blade inspection image set; The UAV is controlled to fly to the actual center point of the blade based on the total moving distance of the blade. Based on the flight direction of the fifth blade and the sequence of single-step movement distance of the blade, the UAV is controlled to fly from the actual center point of the blade and collect images to obtain the fifth blade inspection image set; The UAV is controlled to fly to the actual center point of the blade based on the total moving distance of the blade.

15. The method according to claim 13, characterized in that, The image acquisition based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance yields a component inspection image set, including: Based on the tower's flight direction and the set of single-step movement distance sequences of the tower, the UAV is controlled to fly from the actual blade center point and acquire images to obtain a tower inspection image set; The UAV is controlled to fly to the center point of the actual blade based on the total moving distance of the tower.

16. The method according to claim 7, characterized in that, The swept surface image of the wind turbine includes an image of the leeward side of the wind turbine. The process of controlling the UAV to fly to the center point of the wind turbine blades and acquire the swept surface image, based on the axial flight direction of the wind turbine rotor, the horizontal safety distance, and the blade safety height, includes: Based on the flight direction on the leeward side, the drone is controlled to fly horizontally at the horizontal safe distance, and then flies downward at the blade safe height, so that the drone flies to the center point of the wind turbine blade. The drone's nose is adjusted based on the leeward flight direction, and an image of the wind turbine's leeward side is acquired, wherein the wind turbine's leeward side image corresponds to a center point of the leeward side image.

17. The method according to claim 16, characterized in that, The second target region outline includes the nacelle segmentation outline. Based on the second target region outline, the coordinates of the centroid of the wind turbine's central structure are determined, including: The coordinates of the cabin's centroid are determined based on the cabin's segmented contour. Based on the contour of the second target region, the centroid coordinates of the wind turbine's central structure, and the center point of the swept surface image, the actual error distance of the swept surface is determined, including: The error distance of the leeward image is determined by calculating based on the coordinates of the cabin's centroid and the center point of the leeward image. The leeward scale factor is determined based on the area of ​​the cabin segmentation outline and the actual area of ​​the cabin. The actual error distance of the leeward side is determined based on the leeward side image error distance and the leeward side scale coefficient.

18. The method according to claim 17, characterized in that, The method further includes: If the actual error distance on the leeward side does not meet the preset conditions of the swept surface, the leeward correction flight direction is determined by calculation based on the coordinates of the cabin centroid and the center point of the leeward side image. After controlling the UAV to fly on the leeward side based on the actual error distance of the leeward side, the corrected leeward side image is acquired, wherein the corrected leeward side image corresponds to the center point of the leeward correction image. Based on the corrected leeward image, the contour of the leeward correction target area is determined, wherein the contour of the leeward correction target area corresponds to the coordinates of the center of mass of the correction nacelle. Based on the outline of the leeward correction target area, the coordinates of the centroid of the corrected nacelle, and the center point of the leeward correction image, the actual leeward error distance after correction is determined. Based on the comparison between the corrected leeward actual error distance and the sweeping surface preset conditions, if the corrected leeward actual error distance does not meet the sweeping surface preset conditions and the number of iterations meets the leeward surface iteration preset number, the above process is iteratively executed until the corrected leeward actual error distance meets the sweeping surface preset conditions and the corrected leeward surface image is used as the wind turbine leeward surface image.

19. A wind turbine inspection device based on unmanned aerial vehicles (UAVs), characterized in that, The device includes: The wind turbine top image acquisition module is used to acquire an image of the top of the wind turbine when the drone flies to the initial cruise position and the power is within a preset range. The initial cruise position includes the coordinates and safe altitude of the wind turbine. The wind turbine swept surface image acquisition module is used to control the UAV to fly to the center point of the wind turbine blades and acquire the wind turbine swept surface image based on the top image of the wind turbine; The actual error distance determination module for swept surface is used to determine the actual error distance of the swept surface based on the wind turbine swept surface image. The component reference flight direction determination module is used to determine the component reference flight direction of the wind turbine when the actual error distance of the swept surface meets the preset conditions of the swept surface. The component movement data determination module is used to determine the set of single-step movement distance sequences and the total movement distance of the component based on the component safety distance, the overlap rate of component inspection images, and the number of component inspection shots. The component inspection image acquisition module is used to acquire images based on the component's reference flight direction, the component's single-step movement distance sequence set, and the component's total movement distance to obtain a component inspection image set.

20. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 18.

21. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 18.

Images