Method for measuring dynamic breathing deformation of wind power blade

Abstract

The invention belongs to the technical field of deformation measurement, and particularly relates to a method for measuring dynamic breathing deformation of a wind power blade, which comprises the following steps of: constructing a testing device comprising a magnetic fixing frame, a light supplementing lamp and an industrial camera; a circle of section is selected in the blade to be measured, and a circle of targets are evenly arranged on the inner wall of the blade along the section line; carrying out industrial camera calibration; in a fatigue test state, image data acquisition is completed; template matching is carried out to recognize the target, the area where the center point of the target is located is determined, and the unique center point pixel coordinate of the area is extracted; calculating three-dimensional coordinates of each target center point in an industrial camera coordinate system through a space forward intersection algorithm; the target center coordinates under the industrial camera coordinate system are converted into a blade respiration measurement coordinate system, a dynamic track route of each target center point is generated, and blade respiration deformation is analyzed according to the track routes. According to the invention, the deformation degree of the blade can be better monitored, so that the occurrence of blade breakage and other conditions is reduced.

Description

Technical Field

[0001] The invention belongs to the technical field of deformation measurement, and in particular relates to a method for measuring the dynamic breathing deformation of a wind turbine blade. Background Art

[0002] Large blades are large blades used in wind turbines to capture wind energy and convert it into mechanical energy. Since large blades are exposed to the natural environment for a long time, they are eroded and damaged by natural factors such as wind, rain, snow, and sunlight, especially the influence of wind fields, which is very likely to cause cracks, fatigue, deformation and other problems inside the blades, thus affecting the structural safety and power generation efficiency of the blades, and even causing the blades to break, causing safety accidents.

[0003] Currently, most of the analysis of blade breathing deformation and the dynamic response of blades due to changes in aerodynamic shape, structural stiffness and other areas is carried out through finite element simulation. However, due to manufacturing deviations and the discreteness of material properties, it is difficult for the simulation model to match the actual model well in details. Therefore, dynamic measurement of breathing deformation of wind turbine blades under fatigue testing has become an important way to extract the dynamic deformation response characteristics of blades, providing feedback and closed-loop verification for simulation analysis, which is a key issue that needs to be solved urgently. Summary of the invention

[0004] In view of the above deficiencies in the prior art, an object of the present invention is to provide a method for measuring the dynamic breathing deformation of a wind turbine blade, which can better monitor the degree of blade deformation, thereby reducing the occurrence of blade breakage and the like.

[0005] To achieve the above objectives, the present invention provides a method for measuring the dynamic breathing deformation of a wind turbine blade, comprising the following steps: S1. Build a test device including a magnetic fixing bracket, a fill light, and an industrial camera; S2. Select a circle of cross sections inside the blade to be tested, evenly set a circle of targets on the inner wall of the blade along the cross section line, and install the test device at the hole of the baffle at the root of the blade; S3, use the chessboard grid calibration plate to calibrate the industrial camera; S4. Perform fatigue test on the blade. Under the fatigue test state, the target position is photographed by an industrial camera to complete image data acquisition; S5. For the collected image data, perform template matching to identify the target and determine the area where the target center point is located, extract the unique center point pixel coordinates of the area through the corner point detection algorithm, and then use the refined corner point function to further optimize the center point pixel coordinates; S6. According to the pixel coordinates of the same target center point in the stereo image, the three-dimensional coordinates of each target center point in the industrial camera coordinate system are calculated by using a spatial forward intersection algorithm; S7. Construct a blade breathing measurement coordinate system, convert the target center coordinates in the industrial camera coordinate system into the blade breathing measurement coordinate system, use the three-dimensional coordinates to generate a dynamic trajectory route for each target center point, and analyze the blade breathing deformation based on the trajectory route.

[0006] As a preferred embodiment of the present invention, in S1, the test device is arranged in such a way that the magnetic fixing frame includes a longitudinal rod, a transverse rod and a magnetic base, the longitudinal rod is arranged along the length direction of the blade, the end of the longitudinal rod is fixed on the magnetic base, the magnetic base is arranged on a test bench for fatigue testing, two industrial cameras are symmetrically installed on the transverse rod, the fill light is arranged in the front and lower part of the industrial camera, and the transverse rod is arranged on the longitudinal rod through a connecting device.

[0007] As a preferred solution of the present invention, in S1, the industrial camera has a USB or 1394 data cable to transmit data to the computer in real time, and the fill light is an LED light source.

[0008] As a preferred embodiment of the present invention, in S2, the number of targets is 12, and the selected cross-sectional position is 4 meters from the root of the blade to the tip of the blade in the length direction of the blade, and the targets are set in black and white; in S3, the checkerboard grid calibration plate is an alumina panel with an accuracy of 0.01mm, and the side length of each grid is 6cm.

[0009] As a preferred solution of the present invention, in S3, the process of industrial camera calibration is as follows: the image coordinates are determined by the pixel coordinates in the image data, and the object coordinates are determined by taking the fixed side length of the chessboard grid as the actual reference coordinates. The calibration formula is expressed as: (1); In the formula, is the image point coordinate of the image control point, that is, the position of the image control point in the image captured by the industrial camera; is the object coordinate of the image control point, that is, the position of the image control point in the object space coordinate system; are the coordinates of the image principal point; and It is the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system; It is a coordinate transformation function that converts the object coordinates to the camera coordinate system; represents a function used to correct the distortion of the image point; F is the internal parameter matrix of the industrial camera; assumed , represents the transformed coordinate point, then: (2); (3); (4); (5); is the focal length of the industrial camera in the X-axis and Y-axis directions; , , It is the radial distortion parameter, which is used to correct the radial distortion of the lens; , is the tangential distortion parameter, which is used to correct the tangential distortion of the lens; , , yes Coordinates in the camera coordinate system; , It is used to explain the function Parameters; r represents the distance between the image point and the imaging center; According to equations (1) to (5), the intrinsic parameters and extrinsic parameters of the two industrial cameras are calculated respectively. The intrinsic parameters include radial distortion parameters, tangential distortion parameters, focal length and coordinates of the image principal point. The extrinsic parameters include the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system. This realizes the calibration of the two industrial cameras and provides model parameters for the subsequent calculation of the three-dimensional coordinates of each target center point in the industrial camera coordinate system.

[0010] As a preferred solution of the present invention, in the S4, the acquisition frame rate of the industrial camera is 15 frames per second, and the image resolution is 25 million pixels; a soft trigger mode is used to control the industrial camera to perform synchronous data acquisition, and when the blade fatigue test reaches full load, a trigger command is initiated to perform acquisition, thereby realizing multiple batches of image data acquisition.

[0011] As a preferred embodiment of the present invention, in S5, performing template matching to identify the target and determining the area where the target center point is located is specifically to determine the area where the target center point is located by using a shape-based feature matching detection method for the target on the blade cross section. The shape-based feature matching detection method is: Set the template as the reference image for training, transform the scale during training, and rotate around the center of the template image, while retaining the template feature information in the process, expressed as: (6); (7); (8); (9); (10); Where Q and q are scale transformation parameters; cx and cy are the positions of the template pixel values; is a transformation matrix used to map points in the template image to points in the image block; scale represents a scale factor used to adjust the size of the template; angle represents the rotation angle, which is used to determine the degree of rotation of the template relative to the image block; , Respectively represent the number of columns and rows of the image; In the matching process, the template matching function is used to calculate the similarity between the template and the image block to track the target. Then, by tracking the center of the template image, the center is multiplied by the transformation matrix to identify the center position of the target, which is expressed as: (11); In the formula, is the final image center; is the transformation matrix from the original image to the final result; , Respectively represent the translation of the template image center in the X-axis and Y-axis directions; , They respectively represent the X-axis and Y-axis coordinates of the center of the template image in the target image; tracking other points is the same as tracking the center.

[0012] As a preferred solution of the present invention, in S5, the corner point refinement function is specifically to use the goodFeaturesToTrack function of opencv to find a corner point with the strongest features in a small window, and then use the cornerSubPix function of opencv to perform sub-pixel optimization on the corner point found by the goodFeaturesToTrack function.

[0013] As a preferred embodiment of the present invention, in S7, the process of constructing the leaf respiration measurement coordinate system is as follows: Calculate the center of gravity coordinates of all targets observed by the industrial camera, determine the origin of the blade breathing measurement coordinate system, and the center of gravity expression is: (12); In the formula, is the coordinate of the i-th target in the industrial camera coordinate system; n is the number of targets; is the coordinate of the center of gravity calculated by the forward intersection algorithm; Fit the plane to n targets and define the intermediate variables , , : (13); The fitting equation is obtained: (14); In the formula, a, b, and c are the coefficients of the plane fitting equation, which define the direction of the fitting plane in three-dimensional space; Make constraints , and the constrained least squares solution is: (15); In the formula, T represents transpose; Represents the normal vector of the blade section and is the Z-axis direction vector of the blade breathing measurement coordinate system; According to the relative rotation matrix of the industrial camera and the relative translation vector of the stereo camera , get the baseline vector of the industrial camera ; where t is represented by the centroided coordinates, and the projection matrix N from the spatial point to the fitting plane is: (16); Where I is the identity matrix; Then the projection vector of the baseline vector of the industrial camera on the fitting plane is: (17); In the formula, Represents the Euclidean norm of a vector; is the X-axis direction vector of the leaf breathing measurement coordinate system; Y-axis direction vector of leaf breathing measurement coordinate system ; Represents the cross product of two vectors.

[0014] As a preferred solution of the present invention, in S7, the process of converting the industrial camera coordinate system to the blade breathing measurement coordinate system is: A set of basis vectors that constitute the blade breathing measurement coordinate system and calculate the basis vectors of the industrial camera coordinate system arrive The transition matrix is ​​expressed as: (18); In the formula, It represents the projection coefficient of the first basis vector in the blade breathing measurement coordinate system in the direction of the first basis vector in the industrial camera coordinate system, and the rest are similar; The inverse matrix of the transition matrix is ​​multiplied by the target coordinates in the industrial camera coordinate system to obtain the target coordinates in the blade breathing measurement coordinate system: (19); In the formula, The three-dimensional coordinates of the target in the leaf breathing measurement coordinate system; is the three-dimensional coordinate of the target in the industrial camera coordinate system.

[0015] The algorithm involved in the present invention can be executed by an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the above algorithm calculation is implemented by executing the software by the processor.

[0016] The beneficial effects of the present invention are: The present invention can monitor the deformation degree of the blade under the blade fatigue test, and each target can be analyzed from three dimensions. For the magnetic fixing frame: it is connected to a test bench with almost no obvious vibration, which can reduce the influence of blade deformation on the camera to the greatest extent; its own weight structure can maintain good stability and reduce vibration propagation. For industrial cameras: high-pixel specifications can ensure high-precision data acquisition of blade deformation; the acquisition rate of about 15 frames per second can take into account the integrity of the blade deformation trajectory acquisition. For the target: its only black and white colors can better identify and track the target area; the only corner point in its central area can more easily track and generate the blade deformation trajectory. For the shape-based feature matching detection method: it can adapt well to the complex environment inside the blade and accurately identify the target. The matching method combining templates and features adopted by it can make up for the shortcomings of each algorithm and improve the accuracy of recognition. Therefore, the present invention can realize high-precision measurement of blade root breathing deformation, and can more accurately monitor the degree of blade deformation, thereby reducing the occurrence of blade breakage and major economic and safety accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic diagram of the process of the present invention; Figure 2 It is a schematic diagram of the structure of the testing device of the present invention; Figure 3 yes Figure 2 A right view structural diagram of ; Figure 4 It is a schematic diagram of the structure of the target setting.

[0018] In the figure: 1. vertical bar; 2. horizontal bar; 3. magnetic base; 4. industrial camera; 5. connecting device; 6. target. DETAILED DESCRIPTION

[0019] The embodiments of the present invention are further described below in conjunction with the accompanying drawings: like Figure 1 As shown, a method for measuring the dynamic breathing deformation of a wind turbine blade comprises the following steps: S1. Build a test device including a magnetic fixing frame, a fill light, and an industrial camera 4; S2. Select a circle of cross sections inside the blade to be tested (no need to cut, only select a circle of cross sections for setting the target 6), evenly set a circle of targets 6 on the inner wall of the blade along the cross section line, and install the test device at the hole of the baffle at the root of the blade; S3, calibrating the industrial camera 4 using a chessboard grid calibration plate; S4, performing fatigue test on the blade, and photographing the position of the target 6 by the industrial camera 4 in the fatigue test state to complete image data acquisition; S5, for the collected image data, perform template matching to identify the target 6 and determine the area where the center point of the target 6 is located, extract the unique center point pixel coordinates of the area through the corner point detection algorithm, and then use the refined corner point function to further optimize the center point pixel coordinates; S6, according to the pixel coordinates of the center point of the same target 6 in the stereo image, the three-dimensional coordinates of the center point of each target 6 in the coordinate system of the industrial camera 4 are calculated by a spatial forward intersection algorithm; S7. Construct a blade breathing measurement coordinate system, convert the target 6 center coordinates in the industrial camera 4 coordinate system into the blade breathing measurement coordinate system, use the three-dimensional coordinates to generate a dynamic trajectory route of each target 6 center point, and analyze the blade breathing deformation based on the trajectory route.

[0020] Corner detection algorithm (sub-pixel corner detection algorithm) and spatial forward intersection algorithm are existing technologies, and the existing algorithms and their formulas can be directly used for calculation.

[0021] like Figure 2 and Figure 3 As shown, in S1, the test device is set up in such a way that the magnetic fixing frame includes a longitudinal rod 1, a transverse rod 2 and a magnetic base 3. The longitudinal rod 1 is set along the length direction of the blade, and the end of the longitudinal rod 1 is fixed on the magnetic base 3. The magnetic base 3 is set on the experimental table for fatigue testing. Two industrial cameras 4 are symmetrically installed on the transverse rod 2, and a fill light (not shown in the figure) is set in the front and lower part of the industrial camera 4. The transverse rod 2 is set on the longitudinal rod 1 through a connecting device 5. The industrial camera 4 can be set on the transverse rod 2 through a known structure such as a fixing frame or a fixing seat. The experimental table for fatigue testing is a prior art, and the existing experimental device can be used.

[0022] The magnetic fixing frame is the main body, which can reduce the shaking of the industrial camera 4 when the blade fatigue test is fully loaded; the industrial camera 4 reduces the shaking through the magnetic fixing frame, so that it can remain stable under the full-load movement of the blade; the connection method of the connecting device 5 is a directional connection, that is, a locking connection that cannot significantly change the direction of the camera, and this method can improve the stability of the camera. The connecting device 5 can be set with a known structure.

[0023] The testing device is installed at the hole of the baffle plate at the root of the blade, and the target 6 is photographed by an industrial camera 4 extending into the interior of the blade.

[0024] In S1, the industrial camera 4 has a USB or 1394 data cable to transmit data to the computer in real time, and the fill light is an LED light source.

[0025] like Figure 4 As shown, in S2, the number of targets 6 is 12, and the selected cross-sectional position is 4 meters from the root of the blade to the tip of the blade in the length direction of the blade. The targets are set in black and white. Figure 4 The horizontal plate in is the web of the blade (the load transfer component of the wind turbine blade, used in the main beam structure of the wind turbine blade, responsible for supporting the main beam, bearing the blade bending load, and resisting the blade shear force); in S3, the chessboard grid calibration plate is an aluminum oxide panel with an accuracy of 0.01mm. The panel is opaque and non-reflective, and the side length of each grid is 6cm.

[0026] In S3, the calibration process of the industrial camera 4 is as follows: the image coordinates are determined by the pixel coordinates in the image data, and the object coordinates are determined by taking the fixed side length of the chessboard grid as the actual reference coordinates. The calibration formula is expressed as: (1); In the formula, is the image point coordinate of the image control point, that is, the position of the image control point in the image captured by the industrial camera 4; is the object coordinate of the image control point, that is, the position of the image control point in the object space coordinate system; are the coordinates of the image principal point; and It is the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system; It is a coordinate transformation function that converts the object coordinates to the camera coordinate system; represents a function used to correct the distortion of the image point; F is the internal parameter matrix of the industrial camera 4; assumed , represents the transformed coordinate point, then: (2); (3); (4); (5); is the focal length of the industrial camera 4 in the X-axis and Y-axis directions; , , It is the radial distortion parameter, which is used to correct the radial distortion of the lens; , is the tangential distortion parameter, which is used to correct the tangential distortion of the lens; , , yes Coordinates in the camera coordinate system; , It is used to explain the function Parameters; r represents the distance between the image point and the imaging center; According to equations (1) to (5), the intrinsic parameters and extrinsic parameters of the two industrial cameras 4 are calculated respectively. The intrinsic parameters include radial distortion parameters, tangential distortion parameters, focal length and coordinates of the image principal point. The extrinsic parameters include the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system. This realizes the calibration of the two industrial cameras 4 and provides model parameters for the subsequent calculation of the three-dimensional coordinates of the center point of each target 6 in the coordinate system of the industrial camera 4.

[0027] In S4, the acquisition frame rate of industrial camera 4 is 15 frames per second, and the image resolution is 25 million pixels. The soft trigger mode is used to control industrial camera 4 to perform synchronous data acquisition. When the blade fatigue test reaches full load, the trigger command is initiated to collect data, thereby realizing the acquisition of multiple batches of image data.

[0028] In S5, template matching is performed to identify the target 6 and determine the area where the center point of the target 6 is located. Specifically, the target 6 on the blade cross section is determined using a shape-based feature matching detection method to determine the location area where the center point of the target 6 is located. The shape-based feature matching detection method is: Set the template as the reference image for training, transform the scale during training, and rotate around the center of the template image, while retaining the template feature information in the process, expressed as: (6); (7); (8); (9); (10); Where Q and q are scale transformation parameters; cx and cy are the positions of the template pixel values; is a transformation matrix used to map points in the template image to points in the image block; scale represents a scale factor used to adjust the size of the template; angle represents the rotation angle, which is used to determine the degree of rotation of the template relative to the image block; , Respectively represent the number of columns and rows of the image; In the matching process, the template matching function is used to calculate the similarity between the template and the image block to track the target 6, and then the center of the template image is tracked and the center is multiplied by the transformation matrix to identify the center position of the target 6, which is expressed as: (11); In the formula, is the final image center; is the transformation matrix from the original image to the final result; , Respectively represent the translation of the template image center in the X-axis and Y-axis directions; , They respectively represent the X-axis and Y-axis coordinates of the center of the template image in the target image; tracking other points is the same as tracking the center.

[0029] In S5, the corner point refinement function is to use the goodFeaturesToTrack function of opencv to find a corner point with the strongest features in a small window, and then use the cornerSubPix function of opencv to perform sub-pixel optimization on the corner point found by the goodFeaturesToTrack function.

[0030] Among them, the goodFeaturesToTrack function detects corner points by calculating the first-order derivative information of image brightness changes, which specifically includes the following steps: Initialization: Select some candidate points from the input image as the initial set of feature points; Calculate brightness change: For each feature point, calculate the brightness change in the window around it. You can use a one-dimensional or two-dimensional Sobel operator to calculate. Maximum value screening: Select high-quality feature points based on the maximum value of brightness change, which can be controlled by setting a threshold; ‌Maximum response screening‌: Select the most prominent feature points based on their brightness change response, for example, select the first few feature points with the largest brightness change response.

[0031] The cornerSubPix function optimizes the position of the corner point by minimizing an error function. Specifically, it uses an iterative optimization method to find the corner point position that minimizes the error based on the gradient information around the corner point. This can be achieved by minimizing the following error function: , ; In the formula, is the pixel position around the corner point considered in the current iteration step; Indicates that during the optimization process, for each pixel around the corner point , the calculated error; Yes The gradient of the image at point The grayscale change rate at is used to describe the change of the image at this point; s is the corner point position to be optimized, that is, the corner point position with sub-pixel accuracy that we hope to find.

[0032] Ideally, the gradient on flat areas is zero, while the gradient on edges is in the direction of The direction is vertical, so should be 0, by minimizing all The square sum of is used to obtain a more accurate sub-pixel corner point position.

[0033] In S7, the process of constructing the leaf respiration measurement coordinate system is as follows: Calculate the center of gravity coordinates of all targets 6 observed by the industrial camera 4, determine the origin of the blade breathing measurement coordinate system, and the center of gravity expression is: (12); In the formula, is the coordinate of the i-th target 6 in the coordinate system of the industrial camera 4; n is the number of targets 6; is the coordinate of the center of gravity calculated by the forward intersection algorithm; Fit the n targets 6 with a plane and then define the intermediate variables , , : (13); The fitting equation is obtained: (14); In the formula, a, b, and c are the coefficients of the plane fitting equation, which define the direction of the fitting plane in three-dimensional space; Make constraints , and the constrained least squares solution is: (15); In the formula, T represents transpose; Represents the normal vector of the blade section and is the Z-axis direction vector of the blade breathing measurement coordinate system; According to the relative rotation matrix of industrial camera 4 and the relative translation vector of the stereo camera , get the 4 baseline vector of industrial camera ; where t is represented by the centroided coordinates, and the projection matrix N from the spatial point to the fitting plane is: (16); Where I is the identity matrix; Then the projection vector of the baseline vector of industrial camera 4 on the fitting plane is: (17); In the formula, Represents the Euclidean norm of a vector; is the X-axis direction vector of the leaf breathing measurement coordinate system; Y-axis direction vector of leaf breathing measurement coordinate system ; Represents the cross product of two vectors.

[0034] In S7, the process of converting the industrial camera 4 coordinate system to the blade breathing measurement coordinate system is as follows: A set of basis vectors that constitute the blade breathing measurement coordinate system, and calculate the basis vectors of the industrial camera 4 coordinate system arrive The transition matrix is ​​expressed as: (18); In the formula, It represents the projection coefficient of the first basis vector in the blade breathing measurement coordinate system in the direction of the first basis vector in the industrial camera 4 coordinate system, and the rest are similar; The inverse matrix of the transition matrix is ​​multiplied by the coordinates of target 6 in the industrial camera 4 coordinate system to obtain the coordinates of target 6 in the blade breathing measurement coordinate system: (19); In the formula, is the three-dimensional coordinate of target 6 in the leaf breathing measurement coordinate system; is the three-dimensional coordinate of the target 6 in the coordinate system of the industrial camera 4.

Claims

1. A method for measuring the dynamic breathing deformation of a wind turbine blade, characterized in that The following steps are involved: S1. Build a test device including a magnetic fixing bracket, a fill light, and an industrial camera; S2. Select a circle of cross sections inside the blade to be tested, evenly set a circle of targets on the inner wall of the blade along the cross section line, and install the test device at the hole of the baffle at the root of the blade; S3, use the chessboard grid calibration plate to calibrate the industrial camera; S4. Perform fatigue test on the blade. Under the fatigue test state, the target position is photographed by an industrial camera to complete image data acquisition; S5. For the collected image data, perform template matching to identify the target and determine the area where the target center point is located, extract the unique center point pixel coordinates of the area through the corner point detection algorithm, and then use the refined corner point function to further optimize the center point pixel coordinates; S6. According to the pixel coordinates of the same target center point in the stereo image, the three-dimensional coordinates of each target center point in the industrial camera coordinate system are calculated by using a spatial forward intersection algorithm; S7. Construct a blade breathing measurement coordinate system, convert the target center coordinates in the industrial camera coordinate system into the blade breathing measurement coordinate system, use the three-dimensional coordinates to generate a dynamic trajectory route for each target center point, and analyze the blade breathing deformation based on the trajectory route.

2. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the S1, the test device is set up in such a way that the magnetic fixing frame includes a longitudinal rod, a transverse rod and a magnetic base. The longitudinal rod is arranged along the length direction of the blade, and the end of the longitudinal rod is fixed on the magnetic base. The magnetic base is arranged on a test bench for fatigue testing. Two industrial cameras are symmetrically installed on the transverse rod, and the fill light is arranged in front and below the industrial camera. The transverse rod is arranged on the longitudinal rod through a connecting device.

3. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the S1, the industrial camera has a USB or 1394 data cable to transmit data to the computer in real time, and the fill light is an LED light source.

4. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the S2, the number of targets is 12, and the selected cross-sectional position is 4 meters from the root of the blade to the tip of the blade in the length direction of the blade. The targets are set in black and white. In S3, the checkerboard grid calibration plate is an alumina panel with an accuracy of 0.01mm, and the side length of each grid is 6cm.

5. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In S3, the process of industrial camera calibration is as follows: the image coordinates are determined by the pixel coordinates in the image data, and the object coordinates are determined by taking the fixed side length of the chessboard grid as the actual reference coordinates. The calibration formula is expressed as: （1）； In the formula, is the image point coordinate of the image control point, that is, the position of the image control point in the image captured by the industrial camera; is the object coordinate of the image control point, that is, the position of the image control point in the object space coordinate system; are the coordinates of the image principal point; and It is the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system; It is a coordinate transformation function that converts the object coordinates to the camera coordinate system; represents a function used to correct the distortion of the image point; F is the internal parameter matrix of the industrial camera; assumed , represents the transformed coordinate point, then: （2）； （3）； （4）； （5）； is the focal length of the industrial camera in the X-axis and Y-axis directions; , , It is the radial distortion parameter, which is used to correct the radial distortion of the lens; , is the tangential distortion parameter, which is used to correct the tangential distortion of the lens; , , yes Coordinates in the camera coordinate system; , It is used to explain the function Parameters; r represents the distance between the image point and the imaging center; According to equations (1) to (5), the intrinsic parameters and extrinsic parameters of the two industrial cameras are calculated respectively. The intrinsic parameters include radial distortion parameters, tangential distortion parameters, focal length and coordinates of the image principal point. The extrinsic parameters include the rotation matrix and translation vector from the object space coordinate system to the camera coordinate system. This realizes the calibration of the two industrial cameras and provides model parameters for the subsequent calculation of the three-dimensional coordinates of each target center point in the industrial camera coordinate system.

6. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the S4, the acquisition frame rate of the industrial camera is 15 frames per second, and the image resolution is 25 million pixels; the soft trigger mode is used to control the industrial camera to perform synchronous data acquisition. When the blade fatigue test reaches full load, the trigger command is initiated to collect data, thereby realizing multi-batch image data collection.

7. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the above S5, performing template matching to identify the target and determining the area where the target center point is located is specifically to determine the area where the target center point is located by using a shape-based feature matching detection method for the target on the blade cross section. The shape-based feature matching detection method is: Set the template as the reference image for training, transform the scale during training, and rotate around the center of the template image, while retaining the template feature information in the process, expressed as: （6）； （7）； （8）； （9）； （10）； Where Q and q are scale transformation parameters; cx and cy are the positions of the template pixel values; is a transformation matrix used to map points in the template image to points in the image block; scale represents a scale factor used to adjust the size of the template; angle represents the rotation angle, which is used to determine the degree of rotation of the template relative to the image block; , Respectively represent the number of columns and rows of the image; In the matching process, the template matching function is used to calculate the similarity between the template and the image block to track the target. Then, by tracking the center of the template image, the center is multiplied by the transformation matrix to identify the center position of the target, which is expressed as: （11）； In the formula, is the final image center; is the transformation matrix from the original image to the final result; , Respectively represent the translation of the template image center in the X-axis and Y-axis directions; , They respectively represent the X-axis and Y-axis coordinates of the center of the template image in the target image; tracking other points is the same as tracking the center.

8. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the above S5, the function for refining the corner point is specifically to use the goodFeaturesToTrack function of opencv to find a corner point with the strongest features in a small window, and then use the cornerSubPix function of opencv to perform sub-pixel optimization on the corner point found by the goodFeaturesToTrack function.

9. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 1, characterized in that: In the above-mentioned S7, the process of constructing the leaf respiration measurement coordinate system is as follows: Calculate the center of gravity coordinates of all targets observed by the industrial camera, determine the origin of the blade breathing measurement coordinate system, and the center of gravity expression is: （12）； In the formula, is the coordinate of the i-th target in the industrial camera coordinate system; n is the number of targets; is the coordinate of the center of gravity calculated by the forward intersection algorithm; Fit the plane to n targets and define the intermediate variables , , : （13）； The fitting equation is obtained: （14）； In the formula, a, b, and c are the coefficients of the plane fitting equation, which define the direction of the fitting plane in three-dimensional space; Make constraints , and the constrained least squares solution is: （15）； In the formula, T represents transpose; Represents the normal vector of the blade section and is the Z-axis direction vector of the blade breathing measurement coordinate system; According to the relative rotation matrix of the industrial camera and the relative translation vector of the stereo camera , get the baseline vector of the industrial camera ; where t is represented by the centroided coordinates, and the projection matrix N from the spatial point to the fitting plane is: （16）； Where I is the identity matrix; Then the projection vector of the baseline vector of the industrial camera on the fitting plane is: （17）； In the formula, Represents the Euclidean norm of a vector; is the X-axis direction vector of the leaf breathing measurement coordinate system; Y-axis direction vector of leaf breathing measurement coordinate system ; Represents the cross product of two vectors.

10. A method for measuring dynamic breathing deformation of wind turbine blades according to claim 9, characterized in that: In the above-mentioned S7, the process of converting the industrial camera coordinate system to the blade breathing measurement coordinate system is as follows: A set of basis vectors that constitute the blade breathing measurement coordinate system and calculate the basis vectors of the industrial camera coordinate system arrive The transition matrix is ​​expressed as: （18）； In the formula, It represents the projection coefficient of the first basis vector in the blade breathing measurement coordinate system in the direction of the first basis vector in the industrial camera coordinate system, and the rest are similar; The inverse matrix of the transition matrix is ​​multiplied by the target coordinates in the industrial camera coordinate system to obtain the target coordinates in the blade breathing measurement coordinate system: （19）； In the formula, The three-dimensional coordinates of the target in the leaf breathing measurement coordinate system; is the three-dimensional coordinate of the target in the industrial camera coordinate system.

Images