Method for imaging wind turbine rotor blades

Abstract

The present invention describes a wind turbine rotor blade imaging array (1) comprising: a multi-axis gimbal (10) mounted on the outside of a wind turbine (2) and configured to adjust orientation in response to one or more received settings (10_α, 10_β, 10_γ); a camera (11) mounted on the multi-axis gimbal (10) and arranged to capture an image (11i) of a rotor blade (20); an image analysis unit (110) configured to analyze the captured images (11i); and a camera orientation controller (100) configured to calculate updated gimbal settings (10_α, 10_β, 10_γ) based on an image analysis output (110_out).

Description

Technology Field

[0001] The present invention describes a method for imaging wind turbine rotor blades and a wind turbine rotor blade imaging array. Background Technology

[0002] Wind turbine rotor blades are exposed to harsh environmental conditions and can suffer impact damage from hail, sand, or other airborne particles. To prevent more serious damage, it is important to repair minor damage to the exterior of the rotor blades. For example, water trapped in small cracks or fissures can expand upon freezing, exacerbating the damage. Additionally, even minor damage to the rotor blade sheath can make it a target for unintended lightning strikes, potentially leading to very serious or fatal damage. For these reasons, it is crucial to regularly inspect the exterior surface of the rotor blades to identify any surface anomalies.

[0003] However, due to the current dimensions of rotor blades (lengths can easily exceed 70m), the possibility of identifying defects in a single image of the entire rotor blade captured using a conventional camera is excluded. This is because the camera must be positioned too far away to include the entire rotor blade within its field of view, and the resolution is too low to identify damage of approximately a few millimeters.

[0004] In one approach, a person wearing a rope harness can descend from the hub and visually inspect the rotor blades for damage. However, this carries a high risk of injury. Additionally, this method is time-consuming and involves significant downtime, so this type of on-site inspection can be costly.

[0005] Therefore, various camera-based automatic inspection technologies have been proposed to detect damage to the exterior of rotor blades.

[0006] For example, the camera can be mounted on a horizontal track to slide back and forth while capturing images of the rotor blades, and can also be pitched during the imaging sequence so that all aspects of the rotor blade surface can be imaged.

[0007] US 2019 / 0370999 A1 provides a method, system, and equipment for automatically detecting and tracking a blade, used for tracking a blade of a wind turbine via an unmanned aerial vehicle. The unmanned aerial vehicle is provided with a cradle head, the cradle head controls the shooting angle of a camera, and comprises the following steps: acquiring a blade video file through the camera; detecting at least one frame of a blade image in the blade video file; extracting side edges of a blade region in at least one frame of blade images; tracking and detecting side edges of a blade region in temporally adjacent multi-frame blade images in the blade video file according to the side edges of the blade region; determining a center point of a blade region in each frame of a blade image; and adjusting the cradle head.

[0008] EP 3770424 A1 describes an automated turbine blade monitoring system that increases the frequency of blade inspections and ultimately extends the lifespan of wind turbine blades by providing on-demand photos to avoid massive labor costs associated with on-site inspections. Since blade condition data can be collected remotely without sending technicians to the turbine, the average time between visits to the unit can be improved.

[0009] In known technologies, some degree of error is unavoidable, and accumulated errors can lead to significant discrepancies between the actual position of a defect and local coordinates. For example, when using images obtained from drone-based systems, errors can be much larger due to the inherent instability of the aircraft and limitations in the accuracy of satellite-based altitude readings.

[0010] Image data obtained from drones may contain noise. Since rotor blade surfaces are typically uniform and lack significant features, it may not be possible to correct for noise using feature matching techniques between consecutive images. Additionally, images may be captured under poor or varied lighting conditions. Consequently, known techniques may not be able to image the entire rotor blade with the desired level of accuracy. This can lead to significant costs due to delays when field maintenance personnel fail to detect defects at reported locations and must re-locate them.

[0011] Accordingly, the present invention aims to provide an improved imaging method for wind turbine rotor blades.

[0012] This objective is achieved by the wind turbine rotor blade imaging array of claim 1 and the wind turbine rotor blade imaging method of claim 6.

[0013] According to the present invention, a wind turbine rotor blade imaging array comprises: a multi-axis gimbal mounted on the outside of a wind turbine and configured to adjust orientation in response to one or more received gimbal settings; a camera mounted on the multi-axis gimbal and arranged to capture images of the rotor blades; an image analysis unit configured to analyze the captured images; and a camera orientation controller configured to calculate updated gimbal settings based on the image analysis output. The image analysis unit generates output data that the camera orientation controller can use as a basis for calculating the updated gimbal settings.

[0014] The imaging array can be used to perform an image acquisition sequence that controls the camera to collect or capture multiple images of the currently displayed rotor blade surface, that is, the rotor blade surface that the camera can "see." For example, if the orientation of the rotor blade is set so that the pressure side faces the camera, the image acquisition sequence can generate 50 images of the pressure side.

[0015] By applying this teaching, for example, since the curvature of the outermost part of the rotor blade can cause it to go out of the camera's initial field of view, it becomes possible to automatically capture accurate images of pre-bent long rotor blades using a fixed camera.

[0016] In the imaging array of the present invention, the orientation of the optical axis of the camera with respect to two or more axes in three-dimensional space is directly determined by gimbal settings, namely the pitch, roll, and yaw settings of a 3-axis gimbal or the pitch and roll gimbal settings of a 2-axis gimbal. Since it can be assumed that the camera is rigidly mounted on the support arm of the gimbal, rotating all axes of the gimbal has the effect of rotating the camera by the same amount. Therefore, the terms "camera orientation controller" and "gimbal controller" can be considered synonyms.

[0017] An advantage of the imaging array of the present invention is that it can be implemented at a relatively low cost. It can be implemented without any dedicated hardware components and can be assembled from off-the-shelf components. The image analysis unit and the camera orientation controller can be fully or partially implemented as software modules running on the processor of a higher-level controller. Implementing it primarily in the form of software modules has the advantage that the steps of the method of the present invention can be installed and executed by updating applications already installed on existing systems with relatively little effort. An additional advantage of the imaging array of the present invention is that it does not require a device such as a drone or an operator with drone experience.

[0018] According to the present invention, a method for imaging a wind turbine rotor blade comprises the steps of: mounting a multi-axis gimbal of such an imaging array on the outside of a wind turbine; and operating a camera of the imaging array to capture images of the rotor blade, wherein the step of capturing images is preceded by at least the step of adjusting the multi-axis gimbal based on an image analysis output.

[0019] A multi-axis gimbal can be mounted on any suitable fixed structure outside the wind turbine. For example, the gimbal can be attached to a holder fixed to the tower at an appropriate height above the ground. In a particularly simple approach, the gimbal can be mounted on the lower area of ​​the tower. By bringing the rotor blades to an appropriate angular position, for example, the "4 o'clock" or "8 o'clock" position, the camera is then positioned to capture images along the full length of the rotor blades. Alternatively, for ground-based wind turbines, the gimbal can be mounted on a suitably sturdy support, such as a tripod, at a point near the wind turbine tower, for example, on the base.

[0020] In the context of the present invention, it should be understood that images captured by a camera may display only the area of ​​the rotor blade (where the imaged surface area fills the image) or also display the background (e.g., sky, ground, sea, wind turbine tower, parts of the hub). Long rotor blades are difficult to image with conventional camera setups because their very thin tips tape into a flat, narrow area, making them difficult to distinguish from the background. These problems are exacerbated in the case of pre-bent rotor blades with a distinct curvature in the direction of the updraft. The purpose of the updraft curvature is to prevent the rotor blade tips from colliding with the tower during the operation of the wind turbine, and the rotor blades are made to "straighten out" due to wind loads. However, the curvature of the rotor blades while stationary for the imaging procedure can add to the difficulty of obtaining a complete set of images.

[0021] The method of the present invention overcomes this problem by using a multi-axis gimbal to support the camera and adjusting the gimbal axes so that the camera is in an optimal position to capture the next image. The gimbal axes are adjusted based on image analysis of at least one previously captured image. For example, according to the image analysis, for the next image, the optical axis of the camera may need to be adjusted further outward (toward the tip) as well as further to one side. Each captured image can be processed by an image analysis unit, and the image analysis output can be used by a camera orientation controller to update the gimbal settings in preparation for a subsequent image capture.

[0022] In a particularly preferred embodiment of the present invention, the image analysis unit notifies the gimbal controller to "proceed" when the image is considered satisfactory in relation to, for example, resolution, brightness, contrast, color, and detailed position of the captured photograph, and notifies the gimbal controller to "re-take" when the image is considered unsatisfactory. That is, the steps of capturing an image of the rotor blade area, analyzing the captured image, and adjusting the gimbal settings are repeated as needed until a satisfactory image is obtained. In this way, the relevant steps can be repeated in a loop until the image of the corresponding rotor blade area is considered satisfactory.

[0023] As described above, the multi-axis gimbal may be a 2-axis gimbal configured to rotate around two orthogonal axes to allow for pitch and roll, pitch and yaw, or roll and yaw adjustments. However, in a particularly preferred embodiment of the present invention, the multi-axis gimbal is a 3-axis gimbal capable of rotating around about three orthogonal axes, allowing for pitch, roll, and yaw adjustments. Accordingly, gimbal settings calculated by the camera orientation controller may include pitch angles, roll angles, and yaw angles. These angles may be referred to as Euler angles, Tait-Bryan angles, Cardan angles, etc.

[0024] Particularly advantageous embodiments and features of the present invention are provided by the dependent claims as disclosed in the description below. Features of different claims may be appropriately combined to provide additional embodiments not described herein.

[0025] Wind turbines generally include several, typically three, rotor blades. Hereinafter, without limiting the invention in any way, it may be assumed that during an image acquisition sequence, the camera of the imaging array of the invention is controlled to collect or capture multiple images of a single rotor blade, for example, 50 images of the rotor blade faces (e.g., pressure side, suction side, leading edge).

[0026] In a particularly preferred embodiment of the present invention, the imaging array also includes a distance measuring device configured to measure the distance between a camera and an imaged area of ​​a rotor blade. The distance measuring device may include a LiDAR system, a laser distance measuring device, or any suitable distance tracking means. The distance measuring device may continuously set the distance from the camera to the imaged surface, and this distance may be reported at the moment the image is captured.

[0027] The camera orientation controller can adjust the gimbal (and camera) orientation by calculating relevant axis settings along with distance meter data capable of quantifying the distance to at least a previously imaged surface area.

[0028] The advantage of the method of the present invention is that a complete set of images can be obtained more quickly, and because adjacent images can be obtained at similar depths of field, the images exhibit advantageously low levels of noise. Therefore, less processing is required on the images before the image stitching procedure, so a more accurate result can ultimately be obtained.

[0029] The method of the present invention may be carried out as follows: The aerodynamic rotor is stopped at a position where the rotor blade to be inspected is best viewed, for example, a position where the rotor blade is at the "4 o'clock" position. The gimbal is controlled so that the camera is directed toward one end of the rotor blade. Then, the camera is controlled to begin capturing images transmitted to an image analysis unit, and a rangefinder transmits the measurements to a camera orientation controller. The output of the image analysis and rangefinder information is used by the camera orientation controller to calculate settings for the gimbal axes—in the image below—so that the camera's optical axis is directed toward a square area on the rotor blade (preferably, an area slightly overlapping with a previously correctly captured image). To this end, the image analysis module is configured to detect the rotor blade, that is, to distinguish the rotor blade from the background within the image (sky, clouds, etc.). For example, the image analysis unit may be configured to perform appropriate image processing steps, such as edge detection, to distinguish the rotor blade from the image background.

[0030] The updated gimbal settings ensure that the camera maintains an optimal orientation so that it can always "see" the next part of the rotor blade to be imaged. The gimbal settings allow the next image to follow the previous one; the camera settings ensure that the image covers as much of the rotor blade as possible (e.g., including both the "left" and "right" of the surface visible in the image). Controlling the camera in this way allows for the rapid acquisition of a complete set of high-quality images, thereby minimizing the essential downtime of the wind turbine. The advantageously high quality of the image set simplifies the next stage, image stitching, and helps accurately identify damage to the rotor blade surface.

[0031] Any appropriate rule may be used to describe the orientation of the camera’s rotating XYZ reference frame relative to a fixed xyz reference frame. The orientation of the camera’s XYZ axes can be described using Euler angles, Tate-Bryan angles, Cardan angles, etc., as shown above. Since the camera is rigidly mounted on the gimbal, the field of view and distance tracking unit effectively measures and reports the orientation of the camera’s optical axis.

[0032] The gimbal is preferably mounted on a rigid body, for example, an adapter secured by a strap or track attached around a wind turbine tower, or a rigid tripod arranged on the ground near the base of the tower.

[0033] The imaging array of the present invention preferably also includes a camera controller configured to adjust parameters such as the focal length and shutter speed of the camera. The camera may be assumed to have a variable focal length, for example, through a quasi-focal lens ("zoom lens"). Any one of the camera parameters may be adjusted or modified according to the image analysis output. Prior to the image capture step, the camera orientation controller may provide updated gimbal settings, and the camera controller may provide updated camera settings. Once the gimbal is adjusted, the camera controller then causes the camera to capture images. These steps are repeated until a sufficient number of images are collected, for example, for a desired area of ​​the rotor blade, one side of the rotor blade, or the entire rotor blade surface.

[0034] The camera orientation controller and camera controller may be implemented as a combined control unit or be dependent on a higher-level control unit. Similarly, control over all components of the imaging system may be managed by the user through an appropriate user interface; for example, the user may decide to adjust camera parameters (e.g., image capture rate, image resolution, etc.).

[0035] As mentioned above, a 2-axis gimbal can be configured to rotate around two orthogonal axes to allow, for example, a camera mounted on the gimbal to "pitch" (e.g., along the longitudinal axis of the rotor blade) and "roll" (e.g., in the left-right direction across the rotor blade). This type of gimbal may be sufficient for imaging long rotor blades that are essentially straight. However, as mentioned above, long rotor blades may be in a "pre-bent" state with significant curvature in the direction of the updraft. Long rotor blades may also be "twisted," meaning the chord plane is not flat and is twisted to some extent with respect to the longitudinal axis of the rotor blade. Thus, in a particularly preferred embodiment of the present invention, the multi-axis gimbal is a 3-axis gimbal configured to rotate around three orthogonal axes, that is, the camera mounted on the gimbal can be pitched, rolled, and yawed. Being able to rotate the camera around three orthogonal axes means that long rotor blades that are pre-bent and twisted can be optimally captured over their entire length.

[0036] A fixed reference frame is defined before starting the imaging sequence. This can be done by designating a specific location for the gimbal—that is, a location whose position within the reference frame is known—and defining the gimbal settings based on this location. Alternatively, it may be desirable to calibrate the imaging array with respect to the reference frame before the imaging sequence.

[0037] Preferably, the reference coordinate system (or "reference frame") is defined in relation to a stationary body, such as a wind turbine tower. This reference frame may be a Cartesian coordinate system with X, Y, and Z as the primary axes. These three axes can be assigned in any convenient manner. For example, the origin of this reference coordinate system may be defined to be located at the center of the wind turbine hub, that is, at the point where the longitudinal axes of the rotor blades intersect. Similarly, the reference coordinate system may be established from the perspective of the rotor blade being imaged, with the origin located at the bottom of the rotor blade. In this reference frame, the Z-axis may be assigned to coincide with the longitudinal axis of the rotor blade.

[0038] In a particularly preferred embodiment of the present invention, the initial calibration procedure comprises the steps of: defining a fixed reference frame; capturing an initial image by orienting the optical axis of the camera toward the origin of the reference frame; and recording distance measurements for the initial image. The parameters established in this manner—namely, the reference frame origin, initial field of view, and initial distance—become the reference for defining all subsequent images. The distance measurements for the initial image may be measured and recorded manually, for example, or set in an automated manner.

[0039] One or more of the gimbal axes may change orientation in each subsequent imaging step. Image data and position data are synchronized at each image capture step; that is, gimbal axis settings and distance meter measurements are recorded at the time of image capture. Using the system of the present invention, the orientation of the camera coordinate system relative to a fixed reference frame can be set for each image. With this knowledge, each image can be transformed or projected into a common coordinate system (e.g., a fixed reference frame) through a process called homography mapping, which is part of the stitching procedure. Once all images are "transformed" into a common coordinate system, distances extended across multiple images can be measured.

[0040] Since the camera is mounted on a gimbal and is ultimately firmly mounted on a fixed support, continuous images can be captured in similar camera setups. Accordingly, the present invention offers an improvement over known "noisy" approaches, such as automated drone technology, where it is difficult to synchronize distance measurements with captured images because the drone is generally always moving: since image data and distance meter data can be generated and transmitted separately, in a "worst-case" scenario from the perspective of image stitching, the images and distance measurements are "out of sync," and distance measurements are incorrectly assigned to the images due to the delay between the two types of data and drone movement within a short time interval.

[0041] On the other hand, the method of the present invention can provide a sequence of images each having very accurate orientation information, so the accuracy of the image stitching procedure is improved accordingly.

[0042] Since gimbal settings can be associated with the rotor blade's reference frame, in subsequent image analysis, every point in the image can be associated with an "actual" point on the rotor blade surface. That is, the position of a pixel in the image can be mapped to the intrinsic coordinates of the rotor blade's reference frame. This allows for the identification of defects on the rotor blade surface by applying appropriate image processing algorithms, such as those capable of detecting color anomalies, edge anomalies, etc. Since these algorithms are known to those skilled in the art, they do not need to be described in detail here. Any anomalies or "findings" detected by these algorithms can be reported to the user along with the coordinates of the rotor blade's reference frame. For example, the user may receive a message such as "Possible surface defect at a point 25m from the base end, on the suction side, at a point 3cm from the leading edge." The length, width, or area of ​​the defect may also be indicated in such defect reports.

[0043] Impact defects (holes, cracks, peeling of the outer paint layer, etc.) are often found near the leading edges of the outer blade regions. This is because as the rotational speed of the rotor blade increases in the tip direction, the impact velocity of the particles increases accordingly. Defects caused by high loads and bending moments may appear as cracks or wrinkles in the inner regions of the rotor blade. The method of the present invention can identify the location of any such defects with very high accuracy.

[0044] The object of the present invention is also achieved by a computer program product having a computer program comprising program units that can be directly loaded into the memory of a control unit of an image array and perform the steps of the method of the present invention when the program is executed by the control unit.

[0045] Other objects and features of the present invention will become apparent from the following detailed description in conjunction with the accompanying drawings. However, it should be understood that the drawings are provided for illustrative purposes only and do not define the limitations of the invention. Brief explanation of the drawing

[0046] FIG. 1 illustrates one embodiment of a wind turbine rotor blade imaging array of the present invention; FIG. 2 is a simplified block diagram of one embodiment of a wind turbine rotor blade imaging array of the present invention; FIG. 3 illustrates a stage during the implementation of the method of the present invention; FIG. 4 illustrates a stage during the implementation of the method of the present invention; FIG. 5 is a flowchart illustrating the steps of the method of the present invention; FIGS. 6 and FIGS. 7 each illustrate an image array of the prior art. Specific details for implementing the invention

[0047] Identical numbers in the drawings represent identical objects overall. Objects in the drawings are not necessarily drawn to actual scale.

[0048] FIG. 1 is a simplified drawing of an embodiment of the imaging array (1) of the present invention, FIG. 2 is a block diagram of the imaging array (1) of the present invention, and FIG. 5 is a flowchart illustrating the steps of the method of the present invention. In FIG. 1, a camera (11) is mounted on a multi-axis gimbal attached to a wind turbine tower (22). The camera (11) is mounted on a multi-axis gimbal (10). A distance measuring device (12) measures the distance (12_d) between the camera (11) and the imaged surface of the rotor blade.

[0049] FIG. 2 is a simplified block diagram of an imaging array of the present invention. This diagram shows a gimbal controller (100) that calculates updated settings for a 3-axis gimbal (10) and a camera (11), and it should be understood that the gimbal controller is mounted on the 3-axis gimbal (10) to enable rotation around three orthogonal axes in reference frame XYZ. Although a 3-axis gimbal is described herein, it should be understood that the camera can be mounted on a 2-axis gimbal. Such an embodiment may be sufficient, for example, to image a rotor blade that has not been previously bent.

[0050] A fixed reference frame is defined before starting the imaging sequence. For example, a 3D coordinate system can be defined with the origin at the center of the hub (20H) with one of the three orthogonal axes aligned parallel to the longitudinal axis of the wind turbine tower (the point indicates the origin of the geometric center of the hub (20H) at the intersection of the longitudinal axes of the three rotor blades). Similarly, a 3D coordinate system can be defined with the base of the rotor blade (the other point indicates the base of the rotor blade being imaged) as the origin with one of the three orthogonal axes aligned with the longitudinal axis of the rotor blade (20). Once the reference frame is selected, the imaging array (1) can be established by guiding the optical axis (11A) of the camera (11) toward the origin and capturing an initial image. At the same time, the distance to the imaged surface is measured and the gimbal settings are recorded. These parameters (reference frame origin, gimbal settings, and initial distance) serve as the reference for defining all subsequent images.

[0051] In FIG. 1, the rotor blade (20) being imaged is shown in the "four o'clock" position. In this position, the camera (11) can image the entire side (e.g., the pressure side). Then, when the rotor blade (20) is rotated to the "eight o'clock" position, the camera (11) can image the other side (e.g., the suction side). The gimbal settings are updated during the imaged process, for example, during the process in which the camera (11) captures a sequence of images from the base end (20R) of the rotor blade (20) to the tip end (20T). The imaged sequence can start at the base end (20R) with an initial set of gimbal settings (10_α, 10_β, 10_γ). Each image (11i) is analyzed to determine not only the image quality but also the overall progress. As reported by the distance measuring device (12), the distance (12_d) between the camera and the rotor blade (20) can be used to update the pitch setting (10_α).

[0052] For example, if the image is considered satisfactory in terms of resolution, brightness, contrast, color, and detailed position of the captured photo, one or more of the gimbal settings (10_α, 10_β, 10_γ) can be updated to adjust the camera orientation in preparation for the next image capture step. If the image is determined to be unsatisfactory, the gimbal settings (10_α, 10_β, 10_γ) can be updated to fine-tune the camera orientation to capture a more satisfactory image.

[0053] Since the camera (11) is mounted on the gimbal (10), the field of view of the camera (11) (at the moment the image (11i) is captured) is directly related to the gimbal settings (10_α, 10_β, 10_γ), so the pitch, roll, and yaw settings (10_α, 10_β, 10_γ) of the gimbal (10) can be understood as the field of view of the camera (11). The pitch, roll, and yaw settings (10_α, 10_β, 10_γ) of the gimbal (10) will determine the orientation of the camera's optical axis (11A) at any time. The method of the present invention provides a method to optimally aim the camera (11) at the rotor blade (20) by continuously adjusting the pitch, roll, and yaw settings (10_α, 10_β, 10_γ) of the gimbal (10) so that the entire length can be consistently captured with high image quality.

[0054] The flowchart of FIG. 5 illustrates the steps of the method of the present invention performed by the exemplary embodiment illustrated in FIG. 1 and FIG. 2. In the first step (51), the gimbal (10) is initialized so that the camera (11) aims at one end of the rotor blade, for example, the base end (20R). In the next step (52), the camera captures an image (11i), and the captured image is transmitted to an image analysis unit (110) and processed at the stage (53). In one embodiment, the image analysis unit (110) notifies the gimbal controller to "proceed" if the image is considered satisfactory in step (56), and notifies it to "retake" if the image is considered unsatisfactory. Before retaking the image (11i) at the same pitch setting (10_α), the yaw setting (10_β) and / or roll setting (10_γ) are appropriately adjusted at the stage (54). The amount of adjustment required can be indicated by the image analysis unit (110), which can appropriately notify the gimbal controller (100) how much the camera needs to be adjusted for a specific axis of the next image. Before capturing a subsequent image at stage (52), gimbal settings (10_α, 10_β, 10_γ) are transmitted to the gimbal (10) so that the orientation of the camera (11) can be adjusted in preparation for the image capture stage (52). This process is repeated until the entire rotor blade (20) is imaged. After each satisfactory image, the pitch setting can be updated at stage (55) to turn the camera further outward toward the rotor blade tip. Unsatisfactory images can be retaken by adjusting the yaw setting (10_β) and / or roll setting (10_γ) at stage (54) as described above. In step (57), the completion of the imaging procedure is detected by the image analysis unit (110), for example, the image analysis unit may be able to recognize the rotor blade tip in the final satisfactory image.

[0055] Of course, the method of the present invention can be performed by analyzing each image to determine pitch, roll, and yaw settings (10_α, 10_β, 10_γ) for subsequent images, thereby reducing the possibility of unsatisfactory images.

[0056] However, as previously mentioned, it may be faster to simply adjust the pitch setting (10_α) and correct the yaw and / or roll as needed, because these corrections may be more relevant to the thin airfoil portion of the rotor blade, where the camera orientation adjustments may be more necessary due to the curved shape.

[0057] FIG. 3 illustrates one stage of the method of the present invention. Any of the images described below may be considered as an image (11i) captured in stage (52) of the flowchart of FIG. 5. An image (30) is analyzed and deemed satisfactory. The gimbal settings are updated so that the camera is pitched further outward, and a subsequent image (31) is captured. However, the image analysis considers this subsequent image (31) to be unoptimal because it does not include both edges of the rotor blade (20) (of course, whether the image is satisfactory may be determined according to other criteria). Therefore, the gimbal controller (100) adjusts the gimbal settings to adjust the yaw and / or roll settings (without changing the pitch settings) and captures another image (32). The image analysis approves this image (32) because it shows the rotor blade from edge to edge. The gimbal settings can be adjusted once again to pitch the camera further outward, allowing for additional image capture in the direction of the rotor blade tip as exemplified in FIG. 4. Here, the image (40) is analyzed and deemed satisfactory. The gimbal settings are updated so that the camera is pitched further outward and a subsequent image (41) is captured. However, because the rotor blade (20) is pre-bent and significantly bent in the direction of the updraft in an unloaded state, the narrow tip end is no longer within the camera's field of view, and image analysis determines that this subsequent image (41) is not optimal. Therefore, the gimbal controller adjusts the gimbal settings to adjust the yaw and / or roll settings (without changing the pitch settings) and captures another image (42). Image analysis approves this image (42) because the camera's orientation has been changed to capture a narrow airfoil from edge to edge. By adjusting the gimbal settings once again to pitch the camera further outward, an image (43) of the rotor blade tip (20T) can be captured.

[0058] FIG. 6 illustrates an embodiment of a method according to the prior art. Here, a camera-equipped drone (60) is used to capture images of the rotor blades (20). Due to the difficulty of maintaining a stable drone position even under weak wind conditions, and also due to inaccuracies inherent in satellite-based geographic position readings, significant "noise" may occur in the images. Due to the noise, image stitching becomes difficult and the results may be poor. If the quality is unsatisfactory, the procedure must be repeated, which increases downtime and results in loss of revenue.

[0059] FIG. 7 illustrates an additional prior art in which a fixed camera (70) is arranged to view any part of the rotor blade (20). However, as illustrated herein, for pre-bent rotor blades, it may be difficult and time-consuming to adjust camera settings to capture the entire rotor blade surface.

[0060] Although the present invention has been disclosed in the form of preferred embodiments and variations thereof, it will be understood that numerous additional modifications and variations may be made without departing from the scope of the invention.

[0061] For clarity, it should be understood that throughout this application, the use of the singular (“a” or “an”) does not exclude the plural, and “comprising” does not exclude other steps or elements. The reference to “unit” or “module” does not exclude the use of more than one unit or module.

Claims

Claim 1 A wind turbine rotor blade imaging arrangement (1), comprising: a multi-axis gimbal (10) mounted on the outside of a wind turbine (2) and configured to adjust orientation in response to one or more received gimbal settings (10_α, 10_β, 10_γ) — said gimbal settings include one or more angles among pitch angle (10_α), roll angle (10_β), and yaw angle (10_γ); a camera (11) mounted on said multi-axis gimbal (10) and arranged to capture images (11i) of a rotor blade (20); an image analysis unit (110) configured to analyze said captured images (11i) and generate an image analysis output (110_out); and- a wind turbine rotor blade imaging array (1), comprising a camera orientation controller (100) configured to calculate updated gimbal settings (10_α, 10_β, 10_γ) based on the image analysis output (110_out). Claim 2 In claim 1, the wind turbine rotor blade imaging array (1) is configured to repeat capturing an image (11i) of the rotor blade (20) or the rotor blade region, analyzing the captured image (11i), and adjusting the gimbal settings (10_α, 10_β, 10_γ) based on the image analysis output (110_out) until the image (11i) of the rotor blade (20) or the rotor blade region is considered satisfactory. Claim 3 In claim 1, the multi-axis gimbal (10) is configured to rotate around a first axis (10Y) and around a second axis (10Z) orthogonal to the first axis (10Y), and the multi-axis gimbal (10) is configured to rotate around a third axis (10X) orthogonal to the first axis (10Y) and the second axis (10Z), wind turbine rotor blade imaging array (1). Claim 4 A wind turbine rotor blade imaging array (1), wherein, in claim 1, a rangefinder (12) configured to measure the distance (12_d) between the camera (11) and the imaged area of ​​the rotor blade (20), and further comprising a camera controller configured to adjust the focal length of the camera (11) based on one or more of the measured distance (12_d) and the image analysis output (110_out). Claim 5 In claim 1, the wind turbine rotor blade imaging array (1), wherein the multi-axis gimbal (10) is mounted on the outside of the wind turbine tower. Claim 6 A method for imaging a wind turbine rotor blade (20), comprising: - a step of providing a multi-axis gimbal (10) of a wind turbine rotor blade imaging array (1) according to any one of claims 1 to 5 to the outside of a wind turbine (2); - a step of operating a camera (11) of the wind turbine rotor blade imaging array (1) to capture images (11i) of the rotor blade (20), wherein at least one gimbal setting (10_α, 10_β, 10_γ) of the multi-axis gimbal (10) is adjusted based on the image analysis output (110_out), the step of capturing the images (11i) is preceded by the step of adjusting one or more gimbal settings (10_α, 10_β, 10_γ) of the multi-axis gimbal (10). Claim 7 A method for imaging a wind turbine rotor blade according to claim 6, comprising the steps of analyzing the images (11i), and calculating gimbal settings (10_α, 10_β, 10_γ) to guide the optical axis (11A) of the camera (11) to a target area on the rotor blade (20). Claim 8 A method for imaging a wind turbine rotor blade according to claim 6, wherein the steps of capturing an image (11i) of a rotor blade region, analyzing the captured image (11i), and adjusting gimbal settings (10_α, 10_β, 10_γ) based on the image analysis output (110_out) are repeated until the image (11i) of the rotor blade region is considered satisfactory. Claim 9 A method for imaging a wind turbine rotor blade according to claim 6, wherein the method comprises the step of determining an imaged fraction of the rotor blade. Claim 10 In claim 6, the completion of the imaging procedure is detected by the image analysis unit (110), a method for imaging a wind turbine rotor blade. Claim 11 A method for imaging a wind turbine rotor blade, wherein, in claim 6, the initial calibration procedure comprises: - defining a fixed reference frame; - guiding the optical axis (11A) of the camera (11) toward the origin of the reference frame and capturing an initial image (11i); and - recording a distance measurement for the initial image (11i). Claim 12 A method for imaging a wind turbine rotor blade according to claim 6, wherein the imaging procedure comprises: - bringing the rotor blade (20) to a first angle position before imaging one side of the rotor blade (20); and - bringing the rotor blade (20) to a second angle position before imaging the other side of the rotor blade (20). Claim 13 A method for imaging a wind turbine rotor blade according to claim 12, comprising the step of pitching the rotor blade (20) around the longitudinal axis (20L) of the rotor blade during the imaging procedure. Claim 14 A method for imaging a wind turbine rotor blade according to claim 6, comprising the step of analyzing the captured images (11i) to identify a finding (F) for the rotor blade (20) and to determine the coordinates of the finding (F) in a reference frame of the rotor blade (20). Claim 15 A computer program product for performing the steps of the method according to claim 6, wherein the computer program product performs the steps of the method when loaded into the memory of a programmable device configured to control one or more components of the wind turbine rotor blade imaging array (1).

Images