A method and apparatus for detecting surface anomalies in wind power generation equipment and blades.

By installing camera devices on the outside of the nacelle components of wind turbine generators and using blade anomaly detection models, the problem of detecting surface defects on the blades during wind turbine generator operation has been solved, enabling timely and accurate detection of surface defects on the blades and improving the safety and power generation efficiency of wind turbine generators.

CN123109970APending Publication Date: 2026-11-17YUANJIAN WIND POWER JIANGYINENVISION ENERGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611494822.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-23
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies make it difficult to detect blade surface defects in a timely and accurate manner while the wind turbine is in operation, resulting in a low defect detection frequency and affecting the safety and power generation efficiency of the wind turbine.

Method used

A camera device is installed on the outside of the nacelle assembly of a wind power generator. It is connected to the side wall of the nacelle assembly through an extension component. Different areas of the blade are photographed from different shooting positions without stopping the machine. The images are then analyzed using a blade anomaly detection model.

Benefits of technology

It enables timely and accurate detection of surface defects on wind turbine blades during operation, increases the detection frequency, and ensures the safety and power generation efficiency of wind turbine units.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123109970A_ABST
    Figure CN123109970A_ABST
Patent Text Reader

Abstract

This application relates to the field of wind turbine blade surface defect detection technology, and discloses a method and apparatus for detecting surface anomalies in wind turbine equipment and blades. The wind turbine includes a tower, an impeller assembly and a nacelle assembly disposed on the upper part of the tower, and a camera device disposed outside the nacelle assembly. The camera device is disposed on the side wall of the nacelle assembly via an extension component, which is used to avoid obstruction of the camera device by the nacelle assembly. The imaging area of ​​the camera device is the area where the blades in the impeller assembly pass through the lower part of the tower. The camera device is configured to capture images of different areas of the blades from different shooting positions while the turbine is running. The captured images are used to identify whether any anomalies have occurred on the blade surface. This achieves the effect of timely and accurate detection of surface defects on wind turbine blades during operation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of wind power equipment blade surface defect detection technology, and in particular to a method and apparatus for detecting surface anomalies in wind power equipment and blades. Background Technology

[0002] With the continuous rise in global energy consumption, wind turbines have been deployed and widely used on a large scale. As the core component for capturing wind energy, the operational health of the blades directly determines the safety and power generation efficiency of the wind turbine. Currently, the wind turbine is kept in a shutdown state so that technicians can conduct inspections to detect defects such as bulges and cracks on the blade surface. However, this method requires the turbine to be shut down, and in reality, wind turbine shutdowns are relatively infrequent, meaning the blades are in operation most of the time. Therefore, there is an urgent need for a method that can promptly and accurately detect surface defects on wind turbine blades while they are in operation. Summary of the Invention

[0003] This application provides a method for detecting surface anomalies in wind power generation equipment and blades, which can detect surface defects of wind turbine blades in operation in a timely and accurate manner.

[0004] In a first aspect, embodiments of this application provide a wind power generation device, including a tower, an impeller assembly and a nacelle assembly disposed on the upper part of the tower, and a camera device disposed outside the nacelle assembly; the camera device is disposed on the side wall of the nacelle assembly via an extension component; the camera device is configured to capture images of different areas of the blade using different shooting postures while the machine is running; wherein the captured images are used to identify whether any abnormalities have occurred on the surface area of ​​the blade.

[0005] In the above solution, by installing a camera device on the outside of the nacelle assembly of the wind turbine, specifically by connecting the camera device to the side wall of the nacelle assembly through an extension component, the camera device can avoid the nacelle assembly from obstructing the area it is photographing when taking pictures of the blades. In this way, the installed camera device can take pictures of different areas of the blades from different shooting positions without stopping the machine, and perform anomaly detection on various areas of the blade surface based on the collected pictures, thus achieving the effect of timely and accurate detection of surface defects of wind turbine blades in operation.

[0006] In one possible wind power generation device, the camera device captures the area where the blades of the rotor assembly pass through the lower part of the tower.

[0007] In one possible wind power generation device, the camera is mounted on the left and / or right side walls of the nacelle assembly via an extension member; or, the camera is mounted on the top of the nacelle assembly via an extension member.

[0008] In one possible wind power generation device, the wind power generation device further includes a processing device connected to the camera device; the processing device is configured to perform anomaly detection on the image of the blade surface area.

[0009] In one possible wind power generation device, the wind power generation device further includes a control component; the control component is configured to adjust the blade pitch angle according to the operating parameters of the wind power generation device; the camera device is configured to take pictures of various areas on the blade at different preset pitch angles, and the pictures collected for a blade at each preset pitch angle satisfy the complete surface area of ​​the blade.

[0010] In one possible wind power generation device, each preset pitch angle is obtained by adjusting the wind power generation device under low operating parameters; and / or at least one preset pitch angle is obtained by a pitch adjustment request.

[0011] In one possible wind power generation device, the camera is configured to capture two images of adjacent areas on the blade that overlap, and the overlap rate of the images increases from the blade root towards the blade tip.

[0012] In a possible wind power generation device, the shooting parameters of the camera device under different shooting postures are solidified through a calibration stage. The calibration stage involves the camera device gradually shooting different areas of the same blade at a single preset pitch angle in a unidirectional manner, and establishing the correspondence between the shooting parameters and the blade areas based on the images taken under different shooting postures.

[0013] In one possible wind power generation device, the camera is further configured to filter multiple images taken as the blades pass through the shooting area and send the images that meet the requirements to a processing device.

[0014] In a possible wind power generation device, the processing device is specifically configured to: detect abnormal areas on the blade surface of the image using a blade anomaly detection model; wherein the blade anomaly detection model is trained using historical images acquired by close-range acquisition when the wind power generation device is shut down.

[0015] In one possible wind power generation device, the blade anomaly detection model is optimized using images captured by the camera device.

[0016] In one possible wind power generation device, the camera device is further configured to receive an image acquisition command sent by the processing device when an abnormal area is detected, and adjust to the shooting pose corresponding to the abnormal area indicated by the image acquisition command to take pictures of different blades, thereby determining the target blade with the abnormal surface area.

[0017] Secondly, embodiments of this application provide a method for detecting surface anomalies on a blade, which receives images of different areas of a blade taken by a camera device in different shooting positions while the device is running without shutting down; wherein the camera device is mounted on the side wall of a nacelle assembly via an extension component, and the nacelle assembly is mounted on the upper part of the tower in a wind power generation device; and the image is used to detect abnormal areas on the blade surface through a blade anomaly detection model.

[0018] In the above solution, by installing a camera device on the outside of the nacelle assembly of the wind turbine, specifically by connecting the camera device to the side wall of the nacelle assembly through an extension component, the camera device can avoid the nacelle assembly from obstructing the area it is photographing when taking pictures of the blades. In this way, the installed camera device can take pictures of different areas of the blades from different shooting positions without stopping the machine. The pictures can be used to detect anomalies in various areas of the blade surface by using a blade anomaly detection model, thus achieving the effect of timely and accurate detection of surface defects of wind turbine blades in operation.

[0019] In one possible implementation, the blade anomaly detection model is trained using historical images acquired at close range when the wind power equipment is shut down, and optimized using images taken by the camera device.

[0020] In one possible implementation method, if an abnormal area is detected, an image acquisition command is sent to the camera device; wherein, the image acquisition command is used to instruct the camera device to take pictures of different leaves according to the shooting pose corresponding to the abnormal area, thereby determining the target leaf with the abnormal area.

[0021] In one possible implementation method, before detecting abnormal regions on the leaf surface of the image using the leaf anomaly detection model, the method further includes: performing image quality enhancement processing on the image to obtain an enhanced image; wherein the image quality enhancement processing includes at least one of the following: quality screening, multi-frame fusion restoration, and N-fold super-resolution enhancement.

[0022] In one possible implementation method, the detection of abnormal regions on the leaf surface of the image using a leaf anomaly detection model includes: dividing the image into multiple sub-images according to a preset segmentation method, and detecting abnormal regions on the leaf surface of any sub-image using the leaf anomaly detection model; any sub-image has at least an overlapping area with its adjacent sub-images; merging the multiple sub-images according to the reverse process of the preset segmentation method, thereby determining whether there are abnormal regions on the leaf surface indicated by the image.

[0023] Thirdly, embodiments of this application provide a blade surface anomaly detection device, which includes: a receiving unit for receiving images of different areas of a blade taken by a camera device in different shooting postures while the device is running without shutting down; wherein the camera device is disposed on the left and / or right side walls of a nacelle assembly via an extension component, the nacelle assembly is disposed on the upper part of the tower in a wind power generation device, the extension component is used to avoid the nacelle assembly from obstructing the camera device, and the shooting area of ​​the camera device is the part of the area where the blade passes under the tower; and a detection unit for detecting abnormal areas on the blade surface in the images using a blade anomaly detection model.

[0024] Fourthly, embodiments of this application provide a computing device, including: Memory, used to store program instructions; A processor is configured to invoke program instructions stored in the memory and execute any implementation method of the second aspect according to the obtained program.

[0025] Fifthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions for causing a computer to perform any implementation method as described in the second aspect.

[0026] In a sixth aspect, embodiments of this application provide a computer program product, the computer program product including computer-executable instructions, the computer-executable instructions being used to cause a computer to perform any implementation method as described in the second aspect. Attached Figure Description

[0027] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0028] Figure 1 A schematic diagram of a wind power generation device provided in an embodiment of this application; Figure 2A schematic diagram of yet another wind power generation device provided in the embodiments of this application; Figure 3 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 4 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 6 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 7 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 9 This is a schematic diagram of the installation of a camera device provided in an embodiment of this application; Figure 10 A schematic diagram of a blade surface anomaly detection method provided in an embodiment of this application; Figure 11 This is a schematic diagram of a blade surface anomaly detection device provided in an embodiment of this application; Figure 12 This is a schematic diagram of a computing device provided in an embodiment of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] like Figure 1The diagram shown is a schematic representation of a wind power generation device according to an embodiment of this application. For ease of description, the windward side of the blades in this embodiment is used as the front, and this schematic diagram shows the wind power generation device from the side and rear. The wind power generation device includes a tower 1, an impeller assembly mounted on the upper part of the tower 1, the impeller assembly including three blades 2, a nacelle assembly 3 mounted on the upper part of the tower 1, and a camera device 4 mounted outside the nacelle assembly. Optionally, a weather station 5 is also included, mounted on top of the nacelle assembly. The tower 1 serves as the main load-bearing structure, vertically supporting the nacelle and impeller assembly above, thus lifting the blades to higher altitudes to capture more wind energy. The nacelle assembly 3 is a shell mounted on the top of the tower 1, housing the core transmission and power generation equipment, such as the main shaft, gearbox, generator, brakes, yaw drive, and main control cabinet. Blade 2, typically a three-dimensional curved surface with twist, is the core component for energy capture. It converts wind energy into rotational mechanical energy through aerodynamic effects. Because it is subjected to wind loads, vibrations, icing, and other loads, the blade surface is prone to defects such as surface cracks, bulges, coating peeling, and leading-edge damage. Meteorological station 5 collects environmental meteorological parameters, providing input data to the unit's control system to achieve automatic control of the unit.

[0031] like Figure 2 The diagram shown is a schematic diagram of another wind power generation device provided in this application embodiment. This schematic diagram also shows the wind power generation device from the side and rear, but compared to... Figure 1 The observation distance is closer, so Figure 2 compared to Figure 1 It can more clearly show the installation location of camera device 4, for example, according to Figure 2 The camera device 4 is installed on the right side wall outside the cabin component 3.

[0032] like Figure 3 The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be installed on the right side wall outside the cabin assembly 3. Figure 3 The horizontal double-headed arrow located directly below the extension component 6 indicates that the extension component 6 can be extended and retracted in the horizontal direction. By adjusting the extension length of the extension component 6, the obstruction of the field of view of the camera device 4 by the cabin component 3 can be avoided.

[0033] like Figure 4The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be installed on the left side wall outside the cabin assembly. Figure 4 The horizontal double-headed arrow located directly below the extension component 6 indicates that the extension component 6 can be extended and retracted in the horizontal direction. By adjusting the extension length of the extension component 6, the obstruction of the field of view of the camera device 4 by the cabin component 3 can be avoided.

[0034] like Figure 5 The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be installed on the upper side wall outside the cabin assembly 3, that is, on the top of the cabin assembly. Figure 5 The vertical double-headed arrow located directly to the left of the extension component 6 indicates that the extension component 6 can be extended and retracted in the vertical direction. By adjusting the extension length of the extension component 6 to a certain height, the obstruction of the field of view of the camera device 4 by the cabin component 3 can be avoided.

[0035] like Figure 6 The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be installed on the lower side wall outside the cabin assembly 3, that is, at the bottom of the cabin assembly. Figure 6 The vertical double-headed arrow located directly to the left of the extension component 6 indicates that the extension component 6 can be extended and retracted in the vertical direction. By adjusting the extension length of the extension component 6 to a certain height, the obstruction of the field of view of the camera device 4 by the cabin component 3 can be avoided.

[0036] like Figure 7The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be set on the upper side wall outside the cabin assembly 3, that is, on the top of the cabin assembly. The extension member 6 is designed with a right-angle bend. The vertical part marked 61 of the extension member 6 can extend and retract vertically, and the horizontal part marked 62 of the extension member 6 can rotate 360° along the horizontal plane. During the rotation, the horizontal part marked 62 is extended and retracted. By adjusting the extension length in the vertical direction and the rotation direction and extension length in the horizontal direction of the extension member 6, the obstruction of the field of view of the camera device 4 by the cabin assembly 3 can be avoided.

[0037] for Figure 7 The installation diagram of the camera device shown illustrates that when the vertical part marked 61 and the horizontal part marked 62 in the extension component 6 are connected, the horizontal part marked 62 can be restricted to rotating only 360° in the horizontal plane, i.e., corresponding to... Figure 7 The right-angle bend shown can also be made so that the horizontal part marked 62 rotates relative to the vertical part marked 61, for example, by rotating the horizontal part marked 62 upwards or downwards by a certain angle relative to its horizontal plane. Figure 7 The right-angle bend shown can be updated to an acute-angle bend, an obtuse-angle bend, and a straight-angle bend. When a straight-angle bend is achieved, the resulting image can be referenced. Figure 5 .

[0038] like Figure 8 The figure shows an installation diagram of a camera device provided in an embodiment of this application. The figure shows the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, so that the camera device can be set on the lower side wall outside the cabin assembly 3, that is, at the bottom of the cabin assembly. The extension member 6 is designed with a right-angle bend. The vertical part marked 61 of the extension member 6 can extend and retract vertically, and the horizontal part marked 62 of the extension member 6 can rotate 360° along the horizontal plane. During the rotation, the horizontal part marked 62 is extended and retracted. By adjusting the extension length in the vertical direction and the rotation direction and extension length in the horizontal direction of the extension member 6, the obstruction of the field of view of the camera device 4 by the cabin assembly 3 can be avoided.

[0039] for Figure 8The installation diagram of the camera device shown illustrates that when the vertical part marked 61 and the horizontal part marked 62 in the extension component 6 are connected, the horizontal part marked 62 can be restricted to rotating only 360° in the horizontal plane, i.e., corresponding to... Figure 8 The right-angle bend shown can also be made so that the horizontal part marked 62 rotates relative to the vertical part marked 61, for example, by rotating the horizontal part marked 62 upwards or downwards by a certain angle relative to its horizontal plane. Figure 8 The right-angle bend shown can be updated to an acute-angle bend, an obtuse-angle bend, and a straight-angle bend. When a straight-angle bend is achieved, the resulting image can be referenced. Figure 6 .

[0040] For the above examples of six types of camera installation diagrams, the corresponding actual operation can be as follows: First, a through installation opening is set on the side wall of the cabin component, and the extension component is inserted through or fixed to the installation opening; for the extension component, one end is connected to the internal load-bearing structure of the cabin component or the side wall of the cabin component, and the other end is located outside the cabin component and the camera device is installed. In some embodiments, the extension component may include: (1) a flange seat that is sealed to the side wall of the cabin component; (2) a support arm or mounting cylinder extending from the flange seat to the outside of the cabin; (3) a vibration damping component set between the support arm and the cabin; (4) a waterproof, dustproof, and salt spray-proof sealing component set at the installation opening; (5) a power supply and communication cable channel set inside or on the side of the extension component; (6) a protective cover, wiper, heating defogging or anti-icing accessory set outside the camera device.

[0041] The above illustrations of the four camera mounting orientations simply illustrate the scenario of installing only one camera on the outside of the nacelle component. To improve the efficiency of detecting defects on the blade surface, it can be a combination of at least two of the four single mounting orientations. For example, one camera can be installed on the left and right walls of the outside of a nacelle component; another can be installed on the left, top, and bottom walls of the outside of a nacelle component; and yet another can be installed on the left, right, top, and bottom walls of the outside of a nacelle component.

[0042] for Figure 7 and Figure 8 The installation location of the camera device shown can also utilize an extension component that is a deformable, rotatable arc shape. For example... Figure 9The diagram shows an installation schematic of a camera device according to an embodiment of this application. The diagram illustrates the positional relationship between the cabin assembly 3 and the camera device 4. The cabin assembly 3 and the camera device 4 are connected by an extension member 6, allowing the camera device to be mounted on the lower side wall of the cabin assembly 3, i.e., at the bottom of the cabin assembly. The extension member 6 is designed as a deformable, rotatable arc shape. Rotatability means that the camera, currently facing the left side wall, can be rotated to face the right side wall. Deformation capability means that the extension member has a certain degree of extensibility and malleability. Extensibility means that the extension member can be compressed or extended according to the currently displayed arc curvature, and malleability means that the current arc curvature can be adjusted. By adjusting the extension height and orientation of the arc-shaped extension member, the obstruction of the camera device 4's field of view by the cabin assembly 3 can be avoided.

[0043] It should be noted that the extended components installed on the upper and lower side walls of the cabin components are not limited to the styles exemplified above. Other styles that can avoid the cabin components from obstructing the field of view of the camera device are also within the scope of protection of this application, and will not be exemplified here.

[0044] In conjunction with the wind power generation equipment described above, by connecting a camera device to the side wall outside the nacelle assembly via an extension component, and by controlling the shooting angle of the camera device to be unobstructed by the nacelle assembly, the camera device can be configured to take pictures of different areas of the blade from different shooting positions without stopping the machine. The pictures obtained can then be used to identify whether any abnormalities have occurred on the blade surface.

[0045] In some embodiments, the camera device captures the area where the blades of the impeller assembly pass through the lower part of the tower.

[0046] Referring to the examples described above, please continue to see... Figure 1 , Figure 1 It also includes Sky 7 and Earth 8, which are separated by a dashed horizontal line, according to Figure 1 The field of view of camera device 4, as shown in the image, is mostly directed towards the ground 8 and partially towards the sky 7. This indicates that the camera device 4 captures the area where the blade 2 passes under the tower 1, meaning its field of view is relatively low. This method avoids the problem of overexposure and insufficient image clarity when photographing blades against a bright or even directly lit sky background. By primarily using the ground as a background, the image clarity is improved, facilitating subsequent inspection of blade surface defects.

[0047] In some embodiments, the camera device is disposed on the left and / or right side walls of the cabin assembly via an extension member; or, the camera device is disposed on the top of the cabin assembly via an extension member.

[0048] Based on the examples described above, for Figure 5 As shown in the method of installing the camera device on the top of the nacelle, this application can adjust the extension height of the extension component 6 to make the height of the extension component 6 suitable. For example, at a suitable height, and by rotating the lens of the camera device to face the part of the blade that the blade passes under the tower, the camera device can completely capture a blade by using different shooting postures. Based on this method, the blade can be clearly and completely captured, which is beneficial for subsequent detection of abnormal areas on the blade surface.

[0049] for Figure 6 As shown in the method of installing the camera device at the bottom of the nacelle, this application can adjust the extension height of the extension component 6 to make the height of the extension component 6 suitable. For example, at a suitable height, and by rotating the lens of the camera device to face the part of the blade that the blade passes under the tower, the camera device can completely capture a blade by using different shooting postures. Based on this method, the blade can be clearly and completely captured, which is beneficial for subsequent detection of abnormal areas on the blade surface.

[0050] for Figure 3 , Figure 4 The practice of directly mounting the camera device on the left and right side walls of the cabin exterior using non-bending extension components, and Figure 7 , Figure 8 , Figure 9 By installing the camera device on the top or bottom of the nacelle using bending or deformable extension components, and pointing the camera device towards the left or right side wall of the nacelle, and by controlling the lens of the camera device to point towards the area where the blade passes under the tower, the camera device can also capture a complete image of a blade by using different shooting postures. Based on this method, the blade can be captured clearly and completely, which is beneficial for subsequent detection of abnormal areas on the blade surface.

[0051] In some embodiments, the wind power generation equipment further includes a processing device connected to a camera device; the processing device is configured to perform anomaly detection on the image of the blade surface area.

[0052] Referring to the example described above, a camera device installed on the external sidewall of the wind turbine nacelle via an extension component can be configured to capture images of different areas of the blade surface from various shooting positions without shutting down the turbine. The captured images can be used to detect defects on the blade surface. Specifically, this application can include a processing device connected to the camera device for detecting anomalies in the images of the blade surface areas. This processing device can be installed inside the nacelle assembly or located at a remote office location.

[0053] In some embodiments, the wind power generation equipment further includes a control component; the control component is configured to adjust the blade pitch angle according to the operating parameters of the wind power generation equipment; the camera device is configured to take pictures of various regions on the blade at different preset pitch angles, and the pictures collected for a blade at each preset pitch angle satisfy the complete surface area of ​​the blade.

[0054] Based on the example described above, blades generally exhibit three-dimensional curved surfaces with twist angles. If the camera device only detects defects on the surface of the blade at a single pitch angle, then the detected defects are only those on the blade surface captured by the camera device at that current pitch angle. Whether there are defects on other surfaces of the blade that are not captured is unknown, meaning the detection accuracy of blade surface defects is low.

[0055] To address this issue, this application proposes to install a control component in the wind power generation equipment. This control component can adjust the blade pitch angle according to the operating parameters of the wind power generation equipment. By adjusting the pitch angle, the blade surface entering the field of view of the camera device changes. Therefore, the camera device can be configured to capture images of various areas on the blade at different preset pitch angles. The images captured for a single blade at each preset pitch angle must cover the entire surface area of ​​the blade.

[0056] Operating parameters may include wind speed, output power, engine speed, and current pitch angle. Based on these parameters, the adjustable range of the pitch angle can be determined, that is, the adjustable range of the pitch angle can be determined without affecting the safety and power generation of the wind power generation equipment. For example, a larger pitch angle adjustment range is allowed under operating conditions such as low wind speed, idling, or standby, while the pitch angle cannot be adjusted under full power generation conditions.

[0057] Taking the example of a camera device capturing images of the blade surface near the blade root from a specific shooting position, by setting the camera device to capture images of this area at multiple different pitch angles, the multiple images of the blade at these different pitch angles can cover the entire blade surface in this area. This allows for the acquisition of all images of this area near the blade root, including the leading edge, windward side, leeward side, and rearward side. Other surface areas of the blade are processed in the same way, thus obtaining multiple images of the complete surface areas of each blade. Based on these multiple images of the complete surface areas of each blade, all defects on the complete surface of each blade can be identified in the future, improving the accuracy of blade surface defect identification.

[0058] In some embodiments, each preset pitch angle is obtained by adjusting the wind power generation equipment under low operating parameters; and / or at least one preset pitch angle is obtained through a pitch adjustment request.

[0059] Based on the examples described above, a large pitch angle adjustment range is allowed under operating conditions such as low wind speed, idling, or standby. In other words, this application can adjust the pitch angle over a large range under low operating parameters. For example, by rotating the blade three times at the current pitch angle, the camera can capture images of the entire surface of the blade. The four pitch angles, including the current pitch angle, can serve as an example of each preset pitch angle.

[0060] In other examples, a pitch adjustment request can be triggered by a scheduled periodic inspection of wind power equipment, with the pitch adjustment request indicating a preset pitch angle.

[0061] In other examples, when a suspected surface defect is identified based on an image of the blade captured by a camera, and the camera is required to re-image the blade surface area where the suspected defect is located, a pitch adjustment request can be issued to adjust the pitch angle. Thus, at a suitable pitch angle for capturing the blade surface defect from the front, the camera can re-capture the blade image, which can be better used to verify whether the suspected surface defect is a real defect.

[0062] In some embodiments, the camera device is configured to capture two images of adjacent areas on the leaf that overlap, and the overlap rate of the images tends to increase from the leaf root towards the leaf tip.

[0063] Based on the example described above, when the camera device is configured to capture images of different areas of the blade in different shooting positions, the camera device can capture one image for each of two adjacent areas on the blade. In order to avoid blind spots or gaps in the shooting of two adjacent areas when the camera device captures two images of adjacent areas of the blade, this application can configure the camera device to have overlapping images in the two images captured. This can avoid the problem of missing detection of blade surface defects due to blind spots or gaps in the shooting.

[0064] Furthermore, when the blades are installed on the upper part of the tower, the blade root is closer to the nacelle assembly, while the blade tip is farther away. Based on this blade installation method, and considering that the camera device in this embodiment is installed relatively close to the nacelle assembly (i.e., the camera device is closer to the blade root and farther from the blade tip), according to the imaging principle of perspective, when overlaying two images of adjacent areas on the blade in the aforementioned example, if a consistently low overlap rate is used, a blind spot will inevitably appear towards the blade tip. Conversely, if a consistently high overlap rate is used, there will be significant redundant acquisition towards the blade root, leading to a substantial increase in the number of shooting poses and acquisition time. To address this issue, this application proposes that when overlaying two images of adjacent areas on the blade in the aforementioned example, the overlap rate for the area closer to the blade root can be set lower, while the overlap rate for the area closer to the blade tip can be set higher. This improves image acquisition efficiency while avoiding blind spots.

[0065] In some embodiments, the overlap rate of the image near the leaf root area can be set to 15-25%, and the overlap rate of the image near the leaf tip area can be set to 35-50%.

[0066] In some embodiments, it is assumed that the leaf is divided into 10 regions from the leaf root to the leaf tip, labeled as region 1, region 2, region 3... region 9, region 10. The two images corresponding to region 1 and region 2 overlap, denoted as overlapping image 1; the two images corresponding to region 2 and region 3 overlap, denoted as overlapping image 2... and the two images corresponding to region 9 and region 10 overlap, denoted as overlapping image 10. The overlap rate of the images increases uniformly from the leaf root to the leaf tip. For example, the overlap rate of overlapping image 1 is 15%, overlapping image 2 is 17%, overlapping image 3 is 19%, overlapping image 4 is 21%, overlapping image 5 is 23%, overlapping image 6 is 37%, overlapping image 7 is 39%, overlapping image 8 is 41%, overlapping image 9 is 43%, and overlapping image 10 is 45%. Alternatively, the overlap rate of the images can increase unevenly from the leaf base to the leaf tip. For example, the overlap rate of overlapping image 1 is 17%, overlapping image 2 is 15%, overlapping image 3 is 19%, overlapping image 4 is 17%, overlapping image 5 is 23%, overlapping image 6 is 39%, overlapping image 7 is 35%, overlapping image 8 is 41%, overlapping image 9 is 41%, and overlapping image 10 is 45%.

[0067] In some embodiments, the shooting parameters of the camera device under different shooting poses are solidified through a calibration stage. The calibration stage involves the camera device gradually shooting different areas of the same blade at a single preset pitch angle in a unidirectional manner, and establishing the correspondence between the shooting parameters and the blade areas based on the images taken under different shooting poses.

[0068] Referring to the example described above, the camera device is configured to capture images of different areas of the wind turbine blades using different shooting poses while the wind turbine is running continuously. This process can be understood as the operational phase. During this operational phase, the camera device is configured to capture images of specific areas of the blades using specific shooting poses, directly utilizing the shooting parameters fixed during the calibration phase. This eliminates the need to determine the shooting poses in real-time during the operational phase to meet requirements such as overlapping images of adjacent areas.

[0069] In this application, taking the example described earlier where a blade is divided into 10 regions from the blade root to the blade tip, the calibration stage can be as follows: using region 1 located in the blade root direction as the propulsion reference, firstly, a camera is used to capture an image of the blade in region 1 at the current blade pitch angle, and the shooting parameters when the camera captures the corresponding image of region 1 are recorded, establishing a correspondence between these shooting parameters and region 1. Since region 1 is the first region captured, its shooting parameters do not need to be very strict when set. Next, a camera is used to capture an image of the blade in region 2 at the current blade pitch angle, where the overlap rate of the image of region 2 captured according to the camera parameters adjusted by the camera needs to be the same as that of region 1. The overlap rate of the images meets the preset image overlap rate requirement, thus recording the shooting parameters when the camera device captures the image corresponding to area 2, and establishing the correspondence between the shooting parameters and area 2; then the camera device is used to capture the image of the blade in area 3 at the current blade pitch angle, wherein the overlap rate of the image of area 3 captured according to the camera device adjusted to the preset image overlap rate requirement needs to meet the overlap rate of the image of area 2, thus recording the shooting parameters when the camera device captures the image corresponding to area 3, and establishing the correspondence between the shooting parameters and area 3; and so on, until the correspondence between the image of area 10 located at the blade tip and the shooting parameters when capturing the image of area 10 is determined.

[0070] The above examples illustrate the process of using region 1 at the leaf root as the reference point, and sequentially determining the correspondence between region 1, region 2, region 3... region 10 and their respective shooting parameters along the direction from the leaf root to the leaf tip. In another embodiment, region 10 at the leaf tip can also be used as the reference point, and sequentially determining the correspondence between region 10, region 9, region 8... region 1 and their respective shooting parameters along the direction from the leaf tip to the leaf root. The above demonstrates two methods of step-by-step shooting using a unidirectional advancement approach.

[0071] In some embodiments, the camera device is further configured to filter multiple images taken as the blade passes through the shooting area and send the images that meet the requirements to the processing device.

[0072] Based on the example described above, the camera device takes pictures of the wind turbine blades while the wind turbine is running continuously. During this continuous operation, the blades are generally rotating. When a blade rotates, there is a time lag between its appearance in the camera's field of view and its departure from that area. Furthermore, since the width of a blade is much smaller than the distance between adjacent blades, the camera will not capture images of the blade for an even longer period. Therefore, to improve the efficiency of the processing device in detecting defects on the blade surface, this application allows for the selection of multiple photos taken by the camera device as the blade moves through the field of view, and only the images containing the blade are processed. Instead of sending images without blades, the camera can send images to the processing unit. Furthermore, when blades pass through the camera's field of view, they are easily affected by factors such as wind, rain, snow, and dust, leading to various problems in the captured images, such as some blades being distorted, blurred, or not displayed. In such cases, the camera can choose to send images with blades to the processing unit, or images with high blade clarity and minimal distortion. This avoids wasting transmission resources by sending images without blades to the processing unit and the additional resources that the processing unit would need to spend on identifying images without blades when detecting defects.

[0073] In some embodiments, the processing device is specifically configured to: detect abnormal areas on the blade surface of the image using a blade anomaly detection model; wherein the blade anomaly detection model is trained using historical images acquired by close-range acquisition when the wind power generation equipment is shut down.

[0074] Based on the example described above, the camera device can capture images of the area where the blades pass under the tower. From these images, the camera device can select images containing blades and send them to the processing device. Upon receiving these images, the processing device can use its pre-trained blade anomaly detection model to detect abnormal areas on the blade surface. Regarding the training process of the blade anomaly detection model, considering that the images sent by the camera device have relatively low resolution and clarity, this application proposes to improve the detection effect by using a drone to inspect the blade surface on a date relatively recent to obtain images containing blades. This drone inspection method is an example of a close-range acquisition method, and the images containing blades provided by the drone are examples of historical images. These historical images acquired using the close-range acquisition method are labeled with defects, and these labeled defective historical images are used to train the blade anomaly detection model, thereby obtaining a trained blade anomaly detection model.

[0075] In some embodiments, the images captured by the camera device are optimized for the leaf anomaly detection model.

[0076] Based on the examples described above, embodiments of this application can train a leaf anomaly detection model by obtaining historical images of the leaf at close range. In order to improve the adaptability of the trained leaf anomaly detection model with the images captured by the camera device of this application, this application can also use the images captured by the camera device to fine-tune the trained leaf anomaly detection model. Thus, the leaf anomaly detection model trained in this way can better detect abnormal areas on the leaf surface in the images captured by the camera device.

[0077] Before using images captured by a camera device to fine-tune the leaf anomaly detection model trained on historical images obtained through close-range acquisition, this application can perform image quality enhancement processing on the images captured by the camera device as follows, for example: (1) Quality screening: Remove blurry, overexposed, underexposed, or insufficient leaf proportion images; (2) Multi-frame fusion restoration: Information fusion is performed on multiple frames of images acquired during the same shooting pose of the same leaf to suppress motion blur and low light noise caused by leaf movement, and to improve imaging degradation caused by backlight and rain / fog. (3) Super-resolution enhancement: The restored image is magnified and reconstructed to increase the number of pixels per unit physical length, so that small defects have identifiable imaging quality. Preferably, the magnification ratio is determined based on the ratio of the pixel density of the UAV acquisition method to that of the camera device acquisition method.

[0078] After the above image quality enhancement operations, this application can optimize and train the blade anomaly detection model based on the enhanced images and the defects marked in the images.

[0079] Furthermore, considering the relatively low resolution and clarity of the images captured by the camera device in this application, it is relatively easy to see that the number of defective images in the captured images is relatively low, meaning the number of defective images captured by the camera device is relatively scarce. Therefore, to further improve the effect of using the images captured by the camera device for optimization training of the trained leaf anomaly detection model, as described above, this application can synthesize the defects marked in the images captured by the near-field acquisition method into the images captured by the camera device of this application or into images obtained through image quality enhancement operations, to construct training samples for optimization training. The operation of synthesizing the defects marked in the images captured by the near-field acquisition method into the images captured by the camera device of this application or into enhanced images may include, for example: (1) Based on the defect areas marked in the images acquired by the close-range acquisition method, extract the defect area images and their marked outlines; (2) Perform one or more of the following transformations on the extracted defect area: scale transformation, rotation transformation, color transformation and blur transformation, so that its imaging characteristics match the image or enhanced image acquired by the camera device of this application; (3) The transformed defect area is composited into the blade surface area of ​​the image or enhanced image captured by the camera device of this application, and the composite edge is processed. (4) Generate corresponding defect annotations based on the synthesis position and annotation contour to obtain training samples.

[0080] In some embodiments, the camera device is further configured to receive an image acquisition command sent by the processing device when an abnormal area is detected, and to take images of different blades after adjusting to the shooting pose corresponding to the abnormal area indicated by the image acquisition command, thereby determining the target blade with the abnormal surface area.

[0081] Referring to the example described above, when the processing device detects an abnormal area on the blade surface using a blade anomaly detection model, the processing device can further determine, for example, which blade in the impeller assembly has the abnormal area. To this end, the processing device can, for example, send an image acquisition command to a camera device. This image acquisition command instructs the camera device to adjust its current shooting posture to the shooting posture when the abnormal area was initially photographed and to capture images of different blades in the impeller assembly. Correspondingly, the camera device can receive the image acquisition command and, upon receiving it, adjust its current shooting posture to the shooting posture when the abnormal area was initially photographed. It then uses the corresponding shooting parameters at that shooting posture to capture images of different blades in the impeller assembly. The captured blade images are submitted to the processing device for secondary defect identification to determine whether any blade has an anomaly in that area. If an anomaly is confirmed, the processing device determines which blade has the anomaly, or the captured blade images are used for manual verification. The blade with the surface anomaly identified through secondary identification or manual verification is one example of a target blade.

[0082] After verifying the abnormal areas, the camera device can return to its original shooting position to continue inspecting the blade surface. The defect images confirmed by the processing device or manual verification, along with their annotations of the abnormal areas, can be added to the training sample set for subsequent iterations of the blade anomaly detection model.

[0083] Having described a wind power generation device, and in particular a scheme for installing a camera device therein, embodiments of this application can also provide a method for detecting anomalies on the blade surface. For example... Figure 10 The diagram shown is a schematic of a blade surface anomaly detection method provided in an embodiment of this application. This method can be executed by the processing device described in the previous example, and includes the following steps: Step 1001: Receive images of different areas of the leaf taken by the camera device in different shooting positions while the device is running without stopping.

[0084] The camera device is mounted on the side wall of the nacelle assembly via an extension component, and the nacelle assembly is mounted on the upper part of the tower in the wind power generation equipment.

[0085] In this step, the structure of the wind power generation equipment can be referenced. Figure 1 The wind power generation equipment includes a tower 1, a nacelle assembly 3 located on the upper part of the tower 1, and an impeller assembly located on the upper part of the tower 1. The impeller assembly includes three blades 2. One end of an extension component 6 is connected to the side wall of the nacelle assembly 3, and the other end is connected to a camera device 4. Through the extension component 6, the nacelle assembly 3 can avoid obstructing the area captured by the camera device. In this way, the camera device can be configured to capture different areas of the blades from different shooting positions without stopping the machine.

[0086] After the camera device takes pictures of different areas of the blades in a non-stop state using different shooting positions, the camera device can send the captured pictures to the processing device; correspondingly, the processing device can receive the pictures sent by the camera device.

[0087] Step 1002: Detect abnormal areas on the leaf surface of the image using a leaf anomaly detection model.

[0088] After the processing device receives the image sent by the camera device, it can call its pre-trained blade anomaly detection model to detect abnormal areas on the blade surface of the image, thereby identifying the abnormal areas that appear on the blade surface.

[0089] In some embodiments, the blade anomaly detection model is trained using historical images acquired at close range when the wind power equipment is shut down, and optimized using images taken by a camera device.

[0090] For example, in conjunction with the example described above, regarding the training process of the blade anomaly detection model, considering that the images sent by the camera device are relatively low in resolution and clarity, in order to improve the detection effect of the blade anomaly detection model, this application proposes to use a drone to inspect the blade surface in a date relatively close to the current detection behavior to obtain images with blades. This drone inspection method is, for example, an example of a close-range acquisition method, and the images with blades provided by the drone are, for example, historical images.

[0091] Furthermore, in order to improve the compatibility between the trained leaf anomaly detection model and the images captured by the camera device of this application, this application can also use the images captured by the camera device to fine-tune the trained leaf anomaly detection model. Thus, the leaf anomaly detection model trained in this way can better detect abnormal areas on the leaf surface in the images captured by the camera device.

[0092] In some embodiments, if an abnormal area is detected, an image acquisition command is sent to the camera device; wherein, the image acquisition command is used to instruct the camera device to take pictures of different blades according to the shooting pose corresponding to the abnormal area, thereby determining the target blade with the abnormal area.

[0093] Referring to the example described above, when the processing device detects an abnormal area on the blade surface using a blade anomaly detection model, the processing device can further determine, for example, which blade in the impeller assembly has the abnormal area. To this end, the processing device can, for example, send an image acquisition command to a camera device. This image acquisition command instructs the camera device to adjust its current shooting posture to the shooting posture when the abnormal area was initially photographed and to capture images of different blades in the impeller assembly. Correspondingly, the camera device can receive the image acquisition command and, upon receiving it, adjust its current shooting posture to the shooting posture when the abnormal area was initially photographed. It then uses the corresponding shooting parameters at that shooting posture to capture images of different blades in the impeller assembly. The captured blade images are submitted to the processing device for secondary defect identification to determine whether any blade has an anomaly in that area. If an anomaly is confirmed, the processing device determines which blade has the anomaly, or the captured blade images are used for manual verification. The blade with the surface anomaly identified through secondary identification or manual verification is one example of a target blade.

[0094] In some embodiments, before detecting abnormal areas on the leaf surface of the image using the leaf anomaly detection model, the method further includes: performing image quality enhancement processing on the image to obtain an enhanced image; wherein the image quality enhancement processing includes at least one of the following: quality screening, multi-frame fusion restoration, and N-fold super-resolution enhancement.

[0095] For example, referring to the example described above, in order to improve the detection effect of the leaf anomaly detection model on abnormal areas on the surface of the leaf in the image, the processing device of this application can perform image quality enhancement processing on the received image to obtain an enhanced image. The image quality enhancement processing can include one or more of the following methods: (1) Quality screening: Remove blurry, overexposed, underexposed, or insufficient leaf proportion images; (2) Multi-frame fusion restoration: Information fusion is performed on multiple frames of images acquired during the same shooting pose of the same leaf to suppress motion blur and low light noise caused by leaf movement, and to improve imaging degradation caused by backlight and rain / fog. (3) Super-resolution enhancement: The restored image is magnified and reconstructed to increase the number of pixels per unit physical length, so that small defects have identifiable imaging quality. Preferably, the magnification ratio is determined based on the ratio of the pixel density of the UAV acquisition method to that of the camera device acquisition method.

[0096] In some embodiments, detecting abnormal regions on the blade surface of an image using a blade anomaly detection model includes: dividing the image into multiple sub-images according to a preset segmentation method, and detecting abnormal regions on the blade surface of any sub-image using the blade anomaly detection model; any sub-image has at least an overlapping area with its adjacent sub-images; merging the multiple sub-images according to the reverse process of the preset segmentation method, thereby determining whether there are abnormal regions on the blade surface indicated by the image.

[0097] Based on the examples described above, the processing device of this application can use a blade anomaly detection model to detect surface anomaly regions in an image. During the detection process, this application can provide the following two implementation methods, for example: Method 1: The image to be detected is input into the blade anomaly detection model as a whole. After the blade anomaly detection model detects the anomaly, it can output whether there are defects in the blade shown in the image. If there are defects, it will indicate the area where the defects occur.

[0098] Method 2 involves segmenting the image to be detected into multiple sub-images according to a preset segmentation method, with overlapping areas between adjacent sub-images. Each sub-image is then input as a whole into the blade anomaly detection model, which outputs the detection result for blade defects. Finally, after all sub-images have been detected by the blade anomaly detection model, the multiple sub-images can be merged and restored by reversing the previous segmentation process, thereby ultimately determining whether there are abnormal areas on the blade surface indicated by the image to be detected. This detection method can reduce the memory usage of the blade anomaly detection model per detection.

[0099] Based on the same concept, embodiments of this application also provide a blade surface anomaly detection device. For example... Figure 11 The diagram shown is a schematic of a blade surface anomaly detection device provided in an embodiment of this application. The device includes: The receiving unit 1101 is used to receive images of different areas of the blade taken by the camera device in different shooting positions while the device is running without stopping; wherein the camera device is set on the left side wall and / or right side wall of the nacelle assembly through an extension component, and the nacelle assembly is set on the upper part of the tower in the wind power generation equipment. The detection unit 1102 is used to detect abnormal areas on the surface of the blade in the image using a blade anomaly detection model.

[0100] Furthermore, for this device, the blade anomaly detection model is trained using historical images acquired at close range when the wind power generation equipment is shut down, and optimized using images taken by the camera device.

[0101] Furthermore, the device also includes a sending unit 1103; the sending unit 1103 is used to send an image acquisition command to the camera device if an abnormal area is detected; wherein, the image acquisition command is used to instruct the camera device to take pictures of different leaves according to the shooting pose corresponding to the abnormal area, thereby determining the target leaf with the abnormal area.

[0102] Furthermore, for this device, the detection unit 1102 is also used to perform image enhancement processing on the image to obtain an enhanced image; wherein the image quality enhancement processing includes at least one of the following: quality screening, multi-frame fusion restoration, and N-fold super-resolution enhancement.

[0103] Furthermore, for this device, the detection unit 1102 is specifically used to divide the image into multiple sub-images according to a preset segmentation method, and to detect abnormal areas on the blade surface of any sub-image using the blade anomaly detection model; any sub-image has at least an overlapping area with its adjacent sub-images; and to merge the multiple sub-images according to the reverse process of the preset segmentation method, thereby determining whether there are abnormal areas on the blade surface indicated by the image.

[0104] This application also provides a computing device, which may specifically be a desktop computer, portable computer, smartphone, tablet computer, personal digital assistant (PDA), etc. The computing device may include a central processing unit (CPU), memory, input / output devices, etc. Input devices may include a keyboard, mouse, touchscreen, etc., and output devices may include display devices, such as liquid crystal displays (LCDs) and cathode ray tubes (CRTs).

[0105] The memory may include read-only memory (ROM) and random access memory (RAM), and provides the processor with program instructions and data stored in the memory. In the embodiments of this application, the memory can be used to store program instructions for a blade surface anomaly detection method; The processor is used to call the program instructions stored in the memory and execute the blade surface anomaly detection method according to the obtained program.

[0106] like Figure 12 The diagram shown is a schematic representation of a computing device provided in an embodiment of this application. The computing device includes: The processor 1201, memory 1202, transceiver 1203, and bus interface 1204 are included; wherein the processor 1201, memory 1202, and transceiver 1203 are connected via bus 1205. The processor 1201 is used to read the program in the memory 1202 and execute the above-described blade surface anomaly detection method. Processor 1201 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP. It can also be a hardware chip. The aforementioned hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The aforementioned PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof.

[0107] The memory 1202 is used to store one or more executable programs and can store data used by the processor 1201 when performing operations.

[0108] Specifically, the program may include program code, which includes computer operation instructions. Memory 1202 may include volatile memory, such as random-access memory (RAM); memory 1202 may also include non-volatile memory, such as flash memory, hard disk drive (HDD), or solid-state drive (SSD); memory 1202 may also include combinations of the above types of memory.

[0109] Memory 1202 stores the following elements: executable modules or data structures, or subsets thereof, or extended sets thereof: Operation instructions: This includes various operation instructions used to perform various operations.

[0110] Operating system: includes various system programs used to implement various basic business functions and handle hardware-based tasks.

[0111] The 1205 bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 12 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0112] Bus interface 1204 can be a wired communication interface, a wireless bus interface, or a combination thereof. The wired bus interface can be, for example, an Ethernet interface. The Ethernet interface can be an optical interface, an electrical interface, or a combination thereof. The wireless bus interface can be a WLAN interface.

[0113] This application also provides a computer-readable storage medium storing computer-executable instructions for causing a computer to execute a blade surface anomaly detection method.

[0114] This application also provides a computer program product, which includes computer-executable instructions for causing a computer to execute a blade surface anomaly detection method.

[0115] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0116] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0117] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0118] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0119] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A wind power generation device, characterized in that, It includes a tower, an impeller assembly and a nacelle assembly disposed on the upper part of the tower, and a camera device disposed outside the nacelle assembly; The camera device is mounted on the side wall of the cabin assembly via an extension component; The camera device is configured to capture images of different areas of the blade from different shooting positions without stopping the machine; wherein the captured images are used to identify whether any abnormalities have occurred on the surface of the blade.

2. The wind power generation equipment as described in claim 1, characterized in that, The camera device captures the area where the blades of the impeller assembly pass through the lower part of the tower.

3. The wind power generation equipment as described in claim 1, characterized in that, The camera device is mounted on the left and / or right side walls of the cabin assembly via an extension component; or, The camera device is mounted on top of the cabin assembly via an extension component.

4. The wind power generation equipment as described in claim 1, characterized in that, The wind power generation equipment also includes a processing device connected to the camera device; The processing device is configured to perform anomaly detection on the surface area of ​​the blade in the image.

5. The wind power generation equipment as described in claim 1, characterized in that, The wind power generation equipment also includes control components; The control component is configured to adjust the blade pitch angle according to the operating parameters of the wind power generation equipment; The camera device is configured to capture images of various regions on the blade at different preset pitch angles, and the images captured for a blade at each preset pitch angle satisfy the complete surface area of ​​the blade.

6. The wind power generation equipment as described in claim 1, characterized in that, The camera device is configured to capture two images of adjacent areas on the leaf that overlap, and the overlap rate of the images increases from the leaf root towards the leaf tip.

7. The wind power generation equipment as described in claim 1, characterized in that, The shooting parameters of the camera device under different shooting postures are solidified through a calibration stage. The calibration stage is in which the camera device gradually shoots different areas of the same blade at a single preset pitch angle in a unidirectional advancing manner, and establishes the correspondence between the shooting parameters and the blade area based on the images taken under different shooting postures.

8. The wind power generation equipment as described in any one of claims 1-7, characterized in that, The camera device is also configured to filter multiple images taken as the blade passes through the shooting area and send images that meet the requirements to the processing device.

9. The wind power generation equipment as described in claim 4, characterized in that, The processing device is specifically configured as follows: The image is used to detect abnormal areas on the blade surface using a blade anomaly detection model; wherein, the blade anomaly detection model is trained using historical images acquired at close range when the wind power equipment is shut down.

10. The wind power generation equipment as described in claim 9, characterized in that, The blade anomaly detection model is optimized using images captured by the camera device.

11. The wind power generation equipment as described in claim 10, characterized in that, The camera device is also configured to receive an image acquisition command sent by the processing device when an abnormal area is detected, and adjust to the shooting pose corresponding to the abnormal area indicated by the image acquisition command to take pictures of different blades, thereby determining the target blade with an abnormal surface area.

12. A method for detecting surface anomalies on a blade, characterized in that, The method includes: The system receives images of different areas of the blade taken by a camera device in different shooting positions while the system is running without shutting down; wherein the camera device is mounted on the side wall of the nacelle assembly via an extension component, and the nacelle assembly is mounted on the upper part of the tower in the wind power generation equipment. The image is used to detect abnormal areas on the leaf surface using a leaf anomaly detection model.

13. The method as described in claim 12, characterized in that, The blade anomaly detection model is trained using historical images acquired at close range when the wind power equipment is shut down, and optimized using images taken by the camera device.

14. The method as described in claim 12, characterized in that, The method further includes: If an abnormal area is detected, an image acquisition command is sent to the camera device; wherein, the image acquisition command is used to instruct the camera device to take pictures of different leaves according to the shooting pose corresponding to the abnormal area, thereby determining the target leaf with the abnormal area.

15. The method as described in claim 12, characterized in that, Before detecting abnormal areas on the leaf surface of the image using the leaf anomaly detection model, the method further includes: The image is enhanced to obtain an enhanced image; wherein the image quality enhancement process includes at least one of the following: quality screening, multi-frame fusion restoration, and N-fold super-resolution enhancement.

16. The method according to any one of claims 12-15, characterized in that, The step of detecting abnormal areas on the leaf surface of the image using a leaf anomaly detection model includes: The image is divided into multiple sub-images according to a preset segmentation method, and the leaf anomaly detection model is used to detect abnormal areas on the leaf surface of any sub-image; any sub-image has at least an overlapping area with its adjacent sub-images. Multiple sub-images are merged according to the reverse process of the preset segmentation method to determine whether there are abnormal areas on the blade surface indicated by the image.

17. A computing device, characterized in that, include: Memory, used to store program instructions; A processor is configured to invoke program instructions stored in the memory and execute the method as described in any one of claims 12-16 according to the obtained program.

18. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the method as described in any one of claims 12-16.