A method for rapid detection of surface defects of a wind turbine blade

Abstract

The application discloses a kind of wind power blade surface defect rapid detection methods, comprising the following steps: A1, wind power blade surface defect image acquisition: for wind turbine microcosmic inspection mode, four rotor unmanned aerial vehicle is used to the image acquisition of wind turbine blade surface, under the premise of fully considering unmanned aerial vehicle obstacle avoidance function, when unmanned aerial vehicle detects wind turbine blade, the distance between the two is generally maintained between 1 meter to 2 meters.The wind power blade surface defect rapid detection method solves the image motion blurring phenomenon generated when the unmanned aerial vehicle takes a photo of the wind power blade, reduces the influence of the image preprocessing caused by the deceptive targets such as stains, bird droppings and other deceptive targets on the wind power blade, and improves the clarity and fidelity of the scratch edge after image threshold segmentation.

Description

Technical Field

[0001] This invention belongs to the field of wind turbine blade inspection technology, specifically relating to a rapid detection method for surface defects in wind turbine blades. Background Technology

[0002] With social progress and the continuous development of science and technology, electricity has become an indispensable part of people's lives. However, due to the scarcity of traditional energy sources and the deteriorating ecological environment, the demand for developing new energy sources is becoming increasingly urgent. Wind energy is a renewable and green energy source that is relatively easy to develop and has a minimal impact on the environment and ecology, leading to its widespread development. The blades used in wind power generation are a crucial component of wind turbine units; any defects in the blades, and the severity of those defects, significantly affect the power generation efficiency of the wind turbine unit.

[0003] For the inspection of quality defects in wind turbine blades, domestic and foreign manufacturers typically employ manual visual inspection and ultrasonic testing. Manual visual inspection generally uses mobile cranes or suspended platforms for manual inspection with the naked eye, or inspection from a distance using telescopes, with defects marked and recorded. Ultrasonic testing, on the other hand, utilizes the characteristic that ultrasonic energy penetrates deep into the wind turbine blade material and is emitted at the edges of cross-sections as it travels from one section to another to inspect for defects. When the ultrasonic beam travels from the surface of the part through the probe into the metal interior, it generates reflected waves upon encountering defects and the bottom surface of the part, forming pulse waveforms on a fluorescent screen. The location and size of the defects are determined based on these pulse waveforms.

[0004] However, traditional manual visual inspection methods suffer from high labor costs, slow inspection speed, and low accuracy. Ultrasonic testing, on the other hand, is time-consuming and slow in inspecting wind turbine blade defects. Furthermore, the large number of wind turbine blades to be inspected leads to uncontrollable inspection costs and inconsistent results. Summary of the Invention

[0005] To address the problems mentioned in the background section, this invention provides a rapid detection method for surface defects in wind turbine blades.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a rapid detection method for surface defects of wind turbine blades, comprising the following steps:

[0007] A1. Image acquisition of surface defects of wind turbine blades: For the micro-inspection of wind turbine units, a quadcopter drone is used to acquire images of the surface of wind turbine blades. Under the premise of fully considering the obstacle avoidance function of the drone, the distance between the drone and the wind turbine blade is generally maintained between 1 meter and 2 meters when the drone inspects the wind turbine blades.

[0008] A2. Image Preprocessing of Wind Turbine Blade Surface Defects: The images of wind turbine blades acquired by the UAV are preprocessed. The image preprocessing process includes grayscale conversion; motion blur removal of the wind turbine blade defect images; removal of Gaussian noise and salt-and-pepper noise using an improved adaptive median filtering algorithm; and improvement of the morphology-based image enhancement algorithm to improve the overall quality of the preprocessed images.

[0009] A3. Image segmentation and extraction of surface defects on wind turbine blades: Image segmentation was performed on the surface defects of wind turbine blades after image preprocessing. A combination of bicubic interpolation operator and Otsu's method was used to binarize the image, and morphological opening operation was used to extract blade defects from the binarized image.

[0010] A4. Surface Defect Identification of Wind Turbine Blades: To accurately obtain the degree of damage and the type of damage on the surface of wind turbine blades, the connected components of the morphologically processed binary image are first divided. Then, the feature values ​​of the area, length of the defect, and degree of damage of the blade in the image are calculated to analyze the characteristics of the blade defects in the image.

[0011] In a preferred embodiment of the present invention, the preprocessing of the wind turbine blade surface defect image in A2 further includes the following steps:

[0012] B1. Convert the collected wind turbine blade defect images to grayscale, transforming the multi-channel RGB color images into single-channel grayscale images;

[0013] B2. Motion blur removal operation: Wiener filtering, a motion blur removal method that takes into account both the degradation transfer function and noise, is used to eliminate image blurring caused by the motion of the drone and the wind turbine themselves, as well as the relative motion between them.

[0014] B3. Image noise reduction: An improved adaptive median filtering method is used to remove Gaussian noise and salt-and-pepper noise caused by the diversity of the wind turbine environment, thus reducing noise.

[0015] B4. Image enhancement operation: In order to highlight some inconspicuous small cracks on the wind turbine blades and reduce the influence of external factors, the emphasize operator is used to enhance the image and calculate the pixel grayscale value after image enhancement.

[0016] In a preferred embodiment of the present invention, the Wiener filtering algorithm in the preprocessing of wind turbine blade surface defect images aims to find the estimated value f^(x) of the original image f(x) with blade defects, thereby minimizing the mean square error between the two; the mean square error formula is:

[0017] e 2(x)=|f(x)-f^(x)| 2

[0018] The minimum value of the error function in the above mean square error formula is calculated in the frequency domain as follows:

[0019]

[0020] In the formula, H(x,y) is the degenerate function; S n (x,y) represents the power spectrum of the noise; S f (x,y) represents the power spectrum of the undegraded image; G(x,y) represents the transformation of the degraded image.

[0021] In a preferred embodiment of the present invention, the pixel grayscale value g(x,y) after image enhancement based on the emphasize operator in the preprocessing of wind turbine blade surface defect images is:

[0022] g(x,y)=round{[p-mean]×k}+p

[0023] In the formula: p is the pixel grayscale value before image enhancement; mean is the pixel grayscale value after the image is filtered using the mean_image operator; k is the image enhancement coefficient.

[0024] In a preferred embodiment of the present invention, the image segmentation and extraction of surface defects on wind turbine blades in A3 further includes the following steps:

[0025] C1: The image of the blade defect after image enhancement is smoothed by cubic spline interpolation to ensure that the target image has smooth edges and high definition. The interpolated image is then cropped to ensure that the original image and the cropped image are the same size.

[0026] C2: Image segmentation is performed on the image after image smoothing using the Otsu's method;

[0027] C3: The morphological opening operation technique is selected to identify the target region of the leaf defect image after image segmentation, which can effectively extract the leaf defect from the background and noise of the target image.

[0028] In a preferred embodiment of the present invention, the optimal threshold calculated using the Otsu's inter-class variance algorithm in the segmentation and extraction of surface defects on wind turbine blades is:

[0029]

[0030] Let K take values ​​from 0 to L-1. Based on the inter-class variances for the corresponding K values, let the inter-class variance σ be such that... 2 The value of k corresponding to the maximum value of (k) is the optimal threshold to be calculated.

[0031] In a preferred embodiment of the present invention, the image recognition of surface defects of wind turbine blades in A4 includes the following steps:

[0032] D1: Perform connected component partitioning on the binarized image of leaf defects after morphological processing;

[0033] D2: Calculate the feature values ​​used to describe blade defects in the wind turbine blade image, including defect area A, defect elongation L, and blade damage degree.

[0034] In a preferred embodiment of the present invention, the defect area A of the blade in D2 is the total number of pixels contained in the binary image f(x,y) within the boundary of the target region, where the size of f(x,y) is MN, and the calculation formula is:

[0035]

[0036] In the formula:

[0037] In a preferred embodiment of the present invention, the blade defect elongation in D2 refers to the ratio of the smallest rectangular area surrounding the blade defect to the blade defect area A, and is defined as follows:

[0038]

[0039] In the formula: A is the area of ​​the blade defect, i.e., the total number of pixels in the region; W and L are the width and length pixels of the smallest rectangle surrounding the blade defect region, respectively; when the elongation is close to 1, the image of the blade defect region is close to a circle; when the elongation is greater than 1, the image of the blade defect region is close to a rectangle.

[0040] In a preferred embodiment of the present invention, the degree of blade damage in D2 refers to the ratio of the scratch defect area to the total area of ​​the identified target, defined as:

[0041]

[0042] In the formula: A is the area of ​​the blade defect, i.e., the total number of pixels in the area, and A1 is the number of pixels in the wind turbine blade specimen; the larger the ε value, the greater the degree of defect on the blade surface. Based on the definition of blade defect requirements and actual application requirements, the presence of scratches on the blade specimen can be determined by calculating the degree of blade damage.

[0043] This invention addresses the shortcomings of the prior art and has the following beneficial effects:

[0044] 1. This rapid detection method for surface defects on wind turbine blades solves the problem of motion blurring in images when taking pictures of wind turbine blades using drones, reduces the impact of deceptive targets such as dirt and bird droppings on wind turbine blades on image preprocessing, and improves the clarity and fidelity of scratch edges after image thresholding.

[0045] 2. The rapid detection method for surface defects of wind turbine blades uses the maximum inter-class variance method to segment the image after image smoothing. The threshold calculated by this method can ensure that the numerical variance between the target area and the background area of ​​the image reaches the maximum, thus achieving the goal of optimal image segmentation.

[0046] 3. This rapid detection method for surface defects of wind turbine blades addresses the Gaussian noise and salt-and-pepper noise generated by UAV imaging due to the diversity of the wind turbine environment. An improved adaptive median filtering method is used for noise reduction. To further improve the algorithm's running speed, the number of pixels involved in median sorting is reduced, which shortens the algorithm's running time while providing better protection for image details. Attached Figure Description

[0047] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0048] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0049] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments. These drawings are simplified schematic diagrams, which are only used to illustrate the basic structure of the present invention and therefore only show the components relevant to the present invention.

[0050] Example 1

[0051] A rapid detection method for surface defects in wind turbine blades includes the following steps:

[0052] A1. Wind Turbine Blade Surface Defect Image Acquisition: For microscopic inspection of wind turbine units, a quadcopter drone is used to acquire images of the wind turbine blade surface. While fully considering the drone's obstacle avoidance capabilities, the distance between the drone and the wind turbine blade is generally maintained between 1 and 2 meters during inspection. A2. Wind Turbine Blade Surface Defect Image Preprocessing: The wind turbine blade images acquired by the drone are preprocessed. The image preprocessing process includes grayscale conversion; motion blur removal of the wind turbine blade defect images; removal of Gaussian noise and salt-and-pepper noise using an improved adaptive median filtering algorithm; and enhancement of the image using an improved morphology-based image enhancement algorithm. The overall quality of the preprocessed image; A3, Image segmentation and extraction of surface defects on wind turbine blades: Image segmentation was performed on the surface defects of wind turbine blades after image preprocessing. A combination of bicubic interpolation operator and Otsu's method was used to binarize the image, and morphological opening operation was used to extract blade defects from the binarized image; A4, Surface defect identification of wind turbine blades: In order to accurately obtain the degree of damage and the type of damage defect on the surface of wind turbine blades, the connected components of the morphologically processed binary image were first divided, and then the feature values ​​of the area, defect elongation and degree of blade damage in the image were calculated, so as to analyze the characteristics of blade defects in the image.

[0053] This method solves the motion blur problem that occurs when using drones to photograph wind turbine blades, reduces the impact of deceptive targets such as dirt and bird droppings on wind turbine blades on image preprocessing, and improves the clarity and fidelity of scratch edges after image thresholding.

[0054] The preprocessing of wind turbine blade surface defect images in A2 also includes the following steps:

[0055] B1. Convert the collected wind turbine blade defect images to grayscale, transforming the multi-channel RGB color images into single-channel grayscale images;

[0056] B2. Motion blur removal operation: Wiener filtering, a motion blur removal method that takes into account both the degradation transfer function and noise, is used to eliminate image blurring caused by the motion of the drone and the wind turbine themselves, as well as the relative motion between them.

[0057] B3. Image noise reduction: An improved adaptive median filtering method is used to remove Gaussian noise and salt-and-pepper noise caused by the diversity of the wind turbine environment, thus reducing noise.

[0058] B4. Image enhancement operation: In order to highlight some inconspicuous small cracks on the wind turbine blades and reduce the influence of external factors, the emphasize operator is used to enhance the image and calculate the pixel grayscale value after image enhancement.

[0059] This rapid detection method for surface defects on wind turbine blades addresses the Gaussian noise and salt-and-pepper noise generated by UAV imaging due to the diversity of wind turbine environments. It employs an improved adaptive median filtering method for noise reduction. To further improve the algorithm's running speed, the number of pixels involved in median sorting is reduced, shortening the algorithm's running time while providing better protection for image details.

[0060] In the preprocessing of wind turbine blade surface defect images, the Wiener filtering algorithm aims to find the estimated value f^(x) of the original image f(x) with blade defects, thereby minimizing the root mean square error between the two; the root mean square error formula is:

[0061] e 2 (x)=|f(x)-f^(x)| 2

[0062] The minimum value of the error function in the above mean square error formula is calculated in the frequency domain as follows:

[0063]

[0064] In the formula, H(x,y) is the degenerate function; S n (x,y) represents the power spectrum of the noise; S f (x,y) represents the power spectrum of the undegraded image; G(x,y) represents the transformation of the degraded image.

[0065] The pixel grayscale value g(x,y) after image enhancement based on the emphasize operator in the preprocessing of wind turbine blade surface defect images is:

[0066] g(x,y)=round{[p-mean]×k}+p

[0067] In the formula: p is the pixel grayscale value before image enhancement; mean is the pixel grayscale value after the image is filtered using the mean_image operator; k is the image enhancement coefficient.

[0068] The image segmentation and extraction of surface defects on wind turbine blades in A3 also includes the following steps:

[0069] C1: The image of the blade defect after image enhancement is smoothed by cubic spline interpolation to ensure that the target image has smooth edges and high definition. The interpolated image is then cropped to ensure that the original image and the cropped image are the same size.

[0070] C2: Image segmentation is performed on the image after image smoothing using the Otsu's method;

[0071] C3: The morphological opening operation technique is selected to identify the target region of the leaf defect image after image segmentation, which can effectively extract the leaf defect from the background and noise of the target image.

[0072] This method for rapid detection of surface defects on wind turbine blades uses the maximum inter-class variance method to segment the image after image smoothing. The threshold calculated by this method can ensure that the numerical variance between the target area and the image background area reaches the maximum, thus achieving the goal of optimal image segmentation.

[0073] The optimal threshold calculated using the Otsu's inter-class variance algorithm in the image segmentation and extraction of surface defects on wind turbine blades is:

[0074]

[0075] Let K take values ​​from 0 to L-1. Based on the inter-class variances for the corresponding K values, let the inter-class variance σ be such that... 2 The value of k corresponding to the maximum value of (k) is the optimal threshold to be calculated.

[0076] Image recognition of surface defects on wind turbine blades in A4 includes the following steps:

[0077] D1: Perform connected component partitioning on the binarized image of leaf defects after morphological processing;

[0078] D2: Calculate the feature values ​​used to describe blade defects in the wind turbine blade image, including defect area A, defect elongation L, and blade damage degree.

[0079] In D2, the defect area A of the blade is the total number of pixels contained within the boundary of the target region in the binary image f(x,y), where the size of f(x,y) is MN, and the calculation formula is:

[0080]

[0081] In the formula:

[0082] In D2, the blade defect elongation refers to the ratio of the smallest rectangular area surrounding the blade defect to the blade defect area A, defined as:

[0083]

[0084] In the formula: A is the area of ​​the blade defect, i.e., the total number of pixels in the region; W and L are the width and length pixels of the smallest rectangle surrounding the blade defect region, respectively; when the elongation is close to 1, the image of the blade defect region is close to a circle; when the elongation is greater than 1, the image of the blade defect region is close to a rectangle.

[0085] In D2, the degree of blade damage refers to the ratio of the scratch defect area to the total area of ​​the identified target, defined as:

[0086]

[0087] In the formula: A is the area of ​​the blade defect, i.e., the total number of pixels in the area, and A1 is the number of pixels in the wind turbine blade specimen; the larger the ε value, the greater the degree of defect on the blade surface. Based on the definition of blade defect requirements and actual application requirements, the presence of scratches on the blade specimen can be determined by calculating the degree of blade damage.

[0088] Example 2

[0089] Based on Embodiment 1, a rapid detection method for surface defects of wind turbine blades includes a wind turbine blade image acquisition module, a wind turbine blade surface defect image preprocessing module, a wind turbine blade surface defect image segmentation and extraction module, and a wind turbine blade surface defect identification module.

[0090] The wind turbine blade image acquisition module is used for acquiring wind turbine blade images; the wind turbine blade surface defect image preprocessing module preprocesses the acquired wind turbine blade images; the wind turbine blade surface defect image segmentation and extraction module is used for segmenting and extracting wind turbine blade defect features from the preprocessed images; and the wind turbine blade surface defect recognition module is used for recognizing wind turbine blade defects.

[0091] The wind turbine blade image acquisition module uses a quadcopter drone to acquire images of the wind turbine blades, and the original image is denoted as f1(x,y).

[0092] The wind turbine blade surface defect image preprocessing module 2 preprocesses the acquired wind turbine blade images, specifically through the following steps:

[0093] E1. Perform grayscale processing on the acquired color RGB image f1(x,y) and calculate the brightness value of each pixel in the two-dimensional matrix of grayscale image f2(x,y).

[0094] Y = 0.299R + 0.587G + 0.114B

[0095] E2. Apply Wiener filtering to the grayscale image f2(x,y) of the wind turbine blade defect to remove motion blur, and find the estimated value f3(x,y) of the original defect image f2(x,y) such that the mean square error between the two is minimized.

[0096] e 2 (x) = |f2(x,y) - f3(x,y)| 2

[0097] Where in the frequency domain e 2 When (x) is at its minimum, f3(x,y) can be calculated;

[0098] Noise analysis was performed on the image f3(x,y) that had undergone motion blur removal. Due to the harsh and complex working environment of wind turbines, the acquired images will contain a large amount of Gaussian noise and salt-and-pepper noise.

[0099] Gaussian noise:

[0100]

[0101] In the formula: z represents the image gray value, μ represents the mean or expected value of z, and σ represents the standard deviation of z.

[0102] Salt and pepper noise:

[0103]

[0104] In the formula: if b > a, then the gray value b will be displayed as a small bright spot in the fault image; conversely, the value of a will be displayed as a small dark spot.

[0105] An improved adaptive median filtering method is used to denoise the wind turbine blade defect image f3(x,y), which can quickly obtain the denoised blade defect image f4(x,y).

[0106] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0107] In the description of this specification, the terms "one embodiment," "some embodiments," "embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example that are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0108] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make modifications, alterations, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A rapid detection method for surface defects in wind turbine blades, characterized in that, Includes the following steps: A1. Image acquisition of surface defects of wind turbine blades: For the micro-inspection of wind turbine units, a quadcopter drone is used to acquire images of the surface of wind turbine blades. While fully considering the obstacle avoidance function of the drone, the distance between the drone and the wind turbine blades is maintained between 1 and 2 meters when the drone is inspecting the wind turbine blades. A2. Image preprocessing of surface defects on wind turbine blades: Preprocessing the images of wind turbine blades acquired by the UAV equipment. The image preprocessing process includes grayscale conversion of the images. Perform motion blur removal processing on wind turbine blade defect images; An improved adaptive median filtering algorithm is used to remove Gaussian noise and salt-and-pepper noise; an improved morphology-based image enhancement algorithm is used to improve the overall quality of the preprocessed image. A3. Image segmentation and extraction of surface defects on wind turbine blades: Image segmentation was performed on the surface defects of wind turbine blades after image preprocessing. A combination of bicubic interpolation operator and Otsu's method was used to binarize the image, and morphological opening operation was used to extract blade defects from the binarized image. A4. Surface Defect Identification of Wind Turbine Blades: In order to accurately obtain the degree of damage and the type of damage on the surface of wind turbine blades, the connected components of the morphologically processed binary image are first divided. Then, the feature values ​​of the area, defect elongation and degree of blade damage in the image are calculated to analyze the characteristics of blade defects in the image. The preprocessing of wind turbine blade surface defect images in A2 also includes the following steps: B1. Convert the collected wind turbine blade defect images to grayscale, transforming the multi-channel RGB color images into single-channel grayscale images; B2. Motion blur removal operation: Wiener filtering, a motion blur removal method that takes into account both the degradation transfer function and noise, is used to eliminate image blurring caused by the motion of the drone and the wind turbine themselves, as well as the relative motion between them. B3. Image noise reduction: An improved adaptive median filtering method is used to remove Gaussian noise and salt-and-pepper noise caused by the diversity of the wind turbine environment, thus reducing noise. B4. Image enhancement operation: In order to highlight some inconspicuous small cracks on the wind turbine blades and reduce the influence of external factors, the emphasize operator is used to enhance the image and calculate the pixel grayscale value after image enhancement.

2. The method for rapid detection of surface defects in wind turbine blades according to claim 1, characterized in that, The Wiener filtering algorithm in the preprocessing of wind turbine blade surface defect images aims to find the original image containing blade defects. Valuation This minimizes the mean squared error between the two; the formula for mean squared error is: The minimum value of the error function in the above mean square error formula is calculated in the frequency domain as follows: In the formula It is a degenerate function; The power spectrum of the noise; The power spectrum of the undegraded image; Transformation of degraded images.

3. The method for rapid detection of surface defects in wind turbine blades according to claim 1, characterized in that, Image preprocessing of wind turbine blade surface defect images using the emphasize operator to enhance pixel grayscale values for: In the formula: It represents the pixel grayscale value before image enhancement; mean represents the pixel grayscale value after the image is filtered using the mean_image operator. It is the image enhancement factor.

4. The method for rapid detection of surface defects in wind turbine blades according to claim 1, characterized in that, The image segmentation and extraction of surface defects on wind turbine blades in A3 also includes the following steps: C1: The image of the blade defect after image enhancement is smoothed by cubic spline interpolation to ensure that the target image has smooth edges and high definition. The interpolated image is then cropped to ensure that the original image and the cropped image are the same size. C2: Image segmentation is performed on the image after image smoothing using the Otsu's method; C3: The morphological opening operation technique is selected to identify the target region of the leaf defect image after image segmentation, which can effectively extract the leaf defect from the background and noise of the target image.

5. The method for rapid detection of surface defects in wind turbine blades according to claim 4, characterized in that, The optimal threshold calculated using the Otsu's inter-class variance algorithm in the image segmentation and extraction of surface defects on wind turbine blades is: Let K take values ​​from 0 to L-1. Based on the inter-class variances for the corresponding K values, make the inter-class variances... The value of k corresponding to the maximum value is the optimal threshold to be calculated.

6. The method for rapid detection of surface defects in wind turbine blades according to claim 1, characterized in that, Image recognition of surface defects on wind turbine blades in A4 includes the following steps: D1: Perform connected component partitioning on the binarized image of leaf defects after morphological processing; D2: Calculate the feature values ​​used to describe blade defects in the wind turbine blade image, including defect area A, defect elongation L, and blade damage degree.

7. The method for rapid detection of surface defects in wind turbine blades according to claim 6, characterized in that, The blade defect area A in D2 is a binary image. This includes the total number of pixels within the target region boundary, where... The size is MN, and the calculation formula is: In the formula: .

8. The method for rapid detection of surface defects in wind turbine blades according to claim 6, characterized in that, In D2, the blade defect elongation refers to the ratio of the smallest rectangular area surrounding the blade defect to the blade defect area A, defined as: In the formula: A is the total number of pixels in the area of ​​the blade defect; W and L are the number of pixels in width and length of the smallest rectangle surrounding the blade defect area, respectively; when the elongation approaches 1, the image of the blade defect area is close to a circle. When the elongation is greater than 1, the image of the defect area of ​​the blade is close to a rectangle.

9. The method for rapid detection of surface defects in wind turbine blades according to claim 8, characterized in that, In D2, the degree of blade damage refers to the ratio of the scratch defect area to the total area of ​​the identified target, defined as: In the formula: A is the area of ​​the blade defect, i.e., the total number of pixels in that area. It is the number of pixels in the wind turbine blade specimen; among which The larger the value, the greater the degree of surface defect on the blade. Based on the definition of blade defect requirements and actual application requirements, the presence of scratches on the blade specimen can be determined by calculating the degree of blade damage.

Images