Method for detecting three-dimensional spatial weaving angle of non-planar composite material fiber

Abstract

This invention relates to a method for detecting the three-dimensional spatial braiding angle of fibers in non-planar composite materials, comprising the following steps: 1) establishing a unified global coordinate system for the industrial robotic arm, a 3D contour scanner, and a polarization camera to achieve a unified spatial reference for 3D point cloud data and polarization image data; 2) acquiring 3D geometric data of the surface of the component under test using the 3D contour scanner and extracting the local surface normal vector corresponding to each sampling point; 3) acquiring polarization images of the component surface using the polarization camera and calculating the two-dimensional linear polarization angle data; 4) establishing a spatial correspondence between the two-dimensional polarization image and the 3D point cloud in the global coordinate system of the robotic arm using the 3D geometric data obtained in step 2) and the two-dimensional linear polarization angle data obtained in step 3); finally calculating the three-dimensional spatial braiding angle of the fiber. This invention achieves non-destructive testing of the three-dimensional spatial braiding angle of fibers in the test area of ​​complex curved surface composite materials.

Description

Technical Field

[0001] This invention relates to a method for detecting fiber weaving angle, specifically a method for detecting the three-dimensional spatial weaving angle of non-planar composite fiber, belonging to the field of non-destructive testing technology for composite materials. Background Technology

[0002] With the widespread application of fiber-woven composite materials in the aerospace field, the internal fiber arrangement angle has become the most critical microstructural parameter determining the macroscopic mechanical properties, damage tolerance, and fatigue life of components. During the manufacturing process, the fibers are prone to severe local shear deformation and angular misalignment due to mold extrusion or weaving tension. Therefore, achieving high-precision, full-field non-destructive testing of the true fiber angles of non-planar composite materials is a pressing technical challenge that needs to be addressed in the fields of aerospace manufacturing and mechanical evaluation.

[0003] Currently, non-destructive testing techniques for fiber angles in composite materials suffer from two major limitations: 1) Two-dimensional polarization detection suffers from severe projection distortion errors. Most existing polarization camera-based detection methods assume the surface of the object being tested is an ideal plane. When dealing with non-planar components with complex curvature and torsional features, the surface normal vectors at different locations differ significantly. In this case, the two-dimensional linear polarization angle calculated by the polarization camera is merely the projection of the three-dimensional fiber onto the two-dimensional photosensitive target. The greater the local curvature, the greater the error between this projection angle and the actual three-dimensional fiber weaving angle, leading to the complete invalidation of the detection data in three-dimensional space. 2) Traditional 3D and vision fusion schemes consume enormous computing power and are difficult to register. Some existing technologies attempt to combine 3D contour scanning with 2D cameras, but their processing logic usually relies on extremely time-consuming 3D surface meshing or surface reconstruction algorithms. The massive discrete point cloud computing not only places extremely high demands on hardware computing power, making it difficult to meet the needs of efficient industrial testing; but also lacks a unified spatial and spatiotemporal linkage benchmark between heterogeneous multimodal sensors, making it difficult to achieve sub-pixel-level accurate matching between high-resolution optical features and 3D geometric features.

[0004] Therefore, in order to solve the above problems, it is indeed necessary to provide an innovative method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber to overcome the defects in the prior art. Summary of the Invention

[0005] The purpose of this invention is to provide a method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber. Based on the integration of light polarization characteristics and three-dimensional spatial morphology, it effectively reduces the projection error caused by polarization measurement on complex curved surfaces and realizes non-destructive testing of the three-dimensional spatial braiding angle of the fiber.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows: a method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber, comprising the following steps: 1) Fix the non-planar composite material to be tested on the test stand, calibrate the pose transformation relationship between the 3D contour scanner and the polarization camera relative to the end effector of the industrial robot arm, establish a unified global coordinate system for the robot arm, and obtain the pose information of the industrial robot arm, the 3D contour scanner and the polarization camera in the global coordinate system of the robot arm. 2) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm to carry a 3D contour scanner to scan the surface of the non-planar composite material to be tested, and obtain the original 3D point cloud data of the material surface; convert the original 3D point cloud data to the global coordinate system of the robotic arm, and extract the local surface normal vectors corresponding to each sampling point to obtain 3D geometric data containing 3D point coordinates and local surface normal vectors. 3) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm equipped with a polarization camera to move to the set observation pose and acquire polarization images of the surface of the non-planar composite material to be tested; process the polarization pixels of the polarization images, calculate the two-dimensional linear polarization angle of the component surface based on the polarization images, and obtain the corresponding two-dimensional linear polarization angle data. 4) Using the three-dimensional geometric data obtained in step 2) and the two-dimensional linear polarization angle data obtained in step 3), establish the spatial correspondence between the two-dimensional polarization image and the three-dimensional point cloud in the global coordinate system of the robotic arm; finally, calculate the three-dimensional spatial weaving angle of the fiber.

[0007] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further comprising: step 1) specifically comprising: 1-1), The non-planar composite material to be tested is mounted on the testing stand, and the six degrees of freedom of the material are constrained; (1-2) Based on the fixed installation relationship between the 3D contour scanner and the polarization camera relative to the end effector of the industrial robotic arm, the spatial pose transformation relationships between the contour scanner coordinate system, the polarization camera coordinate system, and the end flange coordinate system of the robotic arm are established respectively, and the pose transformation matrix T of the 3D contour scanner is calculated. ES And polarization camera pose transformation matrix T EC ; 1-3) Real-time reading of the end flange pose matrix output by the industrial robotic arm controller. Where B represents the global coordinate system of the robotic arm and E represents the coordinate system of the end flange; combined with the pose transformation matrix T of the 3D contour scanner obtained in steps 1-2), ES And polarization camera pose transformation matrix T EC Calculate the real-time pose of the 3D contour scanner and polarization camera in the global coordinate system of the robotic arm:

[0008]

[0009] in: Real-time pose of the 3D contour scanner in the global coordinate system of the robotic arm; Real-time pose of the polarization camera in the global coordinate system of the robotic arm.

[0010] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described in step 2): 2-1) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm to carry a three-dimensional contour scanner to perform non-contact scanning of the non-planar composite material to be tested, acquire the original three-dimensional point cloud data of the material surface in real time, and simultaneously record the real-time pose information of the end effector of the industrial robotic arm during the scanning process. 2-2), call the 3D contour scanner pose transformation matrix T obtained in step 1-2). ES In addition to the real-time pose information of the robotic arm recorded in step 2-1), coordinate transformation is performed on the original three-dimensional point cloud data collected in step 2-1), and the three-dimensional points in the scanner coordinate system are uniformly transformed to the global coordinate system of the robotic arm to obtain a three-dimensional point cloud dataset in a unified coordinate system. 2-3) Based on the 3D point cloud data obtained in step 2-2), calculate the local surface normal vector corresponding to each sampling point, and form 3D geometric data containing the 3D point coordinates and their corresponding local surface normal vectors. The data representation is as follows:

[0011] in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to the point; N represents the total number of point cloud sampling points.

[0012] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described in step 3): 3-1) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm carrying the polarization camera to move to the preset observation pose and keep it stationary; adjust the exposure time and gain of the polarization camera according to the preset exposure parameters to complete the preparation for polarization image acquisition; 3-2) Use a polarization camera to perform a single exposure and simultaneously acquire original polarization images in four polarization directions: 0°, 45°, 90° and 135°. Based on the light intensity information of the above four polarization images, calculate the Stokes parameters and calculate them pixel by pixel to generate a linear polarization degree matrix and a two-dimensional linear polarization angle matrix. 3-3), call the linear polarization degree matrix obtained in step 3-2) as the polarization confidence index to perform quality screening on pixel-level polarization data.

[0013] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described in step 3-2): 3-2-1) A single exposure was performed using a polarization camera to simultaneously capture polarization intensity images of four different polarization directions (0°, 45°, 90°, and 135°) within the same field of view, denoted as [images of polarized light intensity]. , , , ; 3-2-2), using the four polarization intensity images acquired in step 3-2-1) as input, and based on Stokes polarization theory, the polarization state parameters are calculated pixel by pixel; among them, The total light intensity component is ; Horizontal and vertical polarization difference components diagonal polarization difference component ; Subsequently, the two-dimensional linear polarization angle and degree of linear polarization under the current camera target surface are calculated pixel by pixel using the above components: where, the two-dimensional linear polarization angle ; linear polarization degree ; After the calculation is completed, the two-dimensional linear polarization angle matrix and linear polarization degree matrix are output.

[0014] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described as follows: the screening method in step 3-3) is as follows: A global statistical analysis of the linear polarization degree DoLP of the linear polarization degree matrix was performed, and its mean was calculated. with standard deviation And set the filtering threshold as follows:

[0015] in, This is an empirical adjustment coefficient; When a pixel satisfies: If the region corresponding to the pixel is determined to be a low polarization confidence region, it is removed; otherwise, its corresponding two-dimensional linear polarization angle data is retained.

[0016] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described in step 4): 4-1) Using the 3D point cloud data and local surface normal vector data obtained in step 2), and the 2D linear polarization angle matrix and linear polarization degree matrix obtained in step 3), the 3D point cloud data and polarized image data are converted to the global coordinate system of the robotic arm based on the fixed spatial pose relationship between the 3D contour scanner and the polarization camera. Using the intrinsic parameters of the polarization camera and its pose information relative to the global coordinate system of the robotic arm, a projection correspondence between the 2D image pixel coordinates and the 3D point cloud spatial coordinates is established to achieve one-to-one matching between 2D polarized pixels and 3D spatial points. Finally, a fused dataset containing spatial coordinates, local surface normal vectors, and polarization information is obtained. 4-2), based on the fused dataset obtained in step 4-1), the two-dimensional linear polarization angle AoLP is used as a directional representation of the fiber axis in the camera imaging plane, that is, for any detection point: Its corresponding local surface normal vector is: And calculate the three-dimensional weaving angle corresponding to that point; 4-3) Associate the 3D spatial weaving angle data of each point obtained in step 4-2) with the corresponding 3D point cloud data to form a 3D angle dataset containing spatial coordinates and angle information: The three-dimensional angle dataset is loaded using the three-dimensional visualization software ParaView, and the calculated three-dimensional spatial weaving angles are attached to the three-dimensional surface of the component in a color mapping manner to realize the visualization output of the fiber spatial weaving angle distribution results in the test area.

[0017] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention is further described as follows: the specific representation of the fusion dataset established in step 4-1) is as follows:

[0018] in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to that point; The corresponding two-dimensional linear polarization angle AoLP is obtained by pixel mapping; This corresponds to the linear polarization degree DoLP.

[0019] The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention can also be as follows: the specific method for the three-dimensional spatial braiding angle in step 4-2) is as follows: Calculate the two-dimensional fiber projection direction based on the two-dimensional polarization angle AoLP: By associating this direction with the local surface normal vector of the corresponding 3D point, the 3D direction vector of the fiber at that location can be obtained: ; Calculate the three-dimensional spatial weaving angle corresponding to this point based on the angle between the fiber's three-dimensional direction vector and the component's reference direction vector.

[0020] Wherein, the component reference direction vector The initial braiding reference axial vector selected along the non-planar component to be measured.

[0021] Compared with the prior art, the present invention has the following beneficial effects: 1. The method for detecting the three-dimensional spatial braiding angle of fibers in non-planar composite materials of the present invention integrates the polarization characteristics of light with the three-dimensional spatial morphology, effectively reducing the projection error caused by polarization measurement on complex curved surfaces, and realizing non-destructive testing of the three-dimensional spatial braiding angle of fibers in the test area of ​​complex curved surface composite materials.

[0022] 2. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber of the present invention relies on a unified global coordinate system of the robotic arm to accurately retrieve the point cloud normal vector and perform reverse projection compensation on the two-dimensional polarization angle, upgrading the traditional surface mapping detection to three-dimensional spatial orientation, which has extremely high engineering application value. Attached Figure Description

[0023] Figure 1 This is a flowchart of the method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber according to the present invention.

[0024] Figure 2 This is a diagram of the hardware device used in this invention.

[0025] Figure 3 It is the point cloud data map acquired by the three-dimensional contour scanner in step 2) of the present invention.

[0026] Figure 4 The polarization camera in step 3-2 of this invention acquires original polarization images in four polarization directions: 0°, 45°, 90°, and 135°.

[0027] Figure 5 The polarization degree image and linear polarization angle image are acquired by the polarization camera in step 3-2 of this invention.

[0028] Figure 6 This is a visualization output of the three-dimensional spatial weaving angle distribution result of step 4) of the present invention. Detailed Implementation

[0029] Please refer to the instruction manual appendix. Figure 1 As shown, this invention provides a method for detecting the three-dimensional spatial braiding angle of fibers in non-planar composite materials, which includes the following process steps: 1) Fix the non-planar composite material 2 to be tested onto the testing stand 1 (as per the instruction manual). Figure 2 As shown, the pose transformation relationship between the 3D contour scanner 7 and the polarization camera 6 relative to the end effector of the industrial robotic arm 4 is calibrated respectively, a unified global coordinate system {B} for the robotic arm is established, and the pose information of the industrial robotic arm, the 3D contour scanner and the polarization camera in the global coordinate system of the robotic arm is obtained, so as to provide a unified coordinate reference for the spatial correspondence between the 3D point cloud data in step 2) and the polarization image data in step 3).

[0030] This step is as follows: 1-1), the non-planar composite material 2 to be tested is installed on the test stand 1 and rigidly fixed with bolts. The six degrees of freedom of the material 2 are constrained to ensure that the component always maintains its spatial position during the three-dimensional contour scanning and polarization image acquisition process, so as to provide stable detection conditions for subsequent multi-source data fusion.

[0031] (1-2) Based on the fixed installation relationship of the 3D contour scanner 7 and the polarization camera 6 relative to the end effector of the industrial robotic arm 4, the spatial pose transformation relationships between the coordinate system {S} of the 3D contour scanner, the coordinate system {C} of the polarization camera, and the coordinate system {E} of the end flange of the robotic arm are established respectively. The pose transformation matrix T of the 3D contour scanner is calculated. ES And polarization camera pose transformation matrix T EC .

[0032] The robotic arm's global coordinate system {B} is used to represent the positional relationship of all spatial data during the detection process; the robotic arm's end flange coordinate system {E} is fixed to the industrial robotic arm's end effector and is used to describe the real-time pose changes of the end effector; the 3D contour scanner coordinate system {S} and the polarization camera coordinate system {C} are used to describe the acquisition coordinates of 3D point cloud data and 2D polarization image data, respectively.

[0033] The pose transformation matrix of the 3D contour scanner is denoted as:

[0034] The extrinsic parameter matrix of a polarization camera is denoted as:

[0035] Where R represents a 3×3 rotation matrix, used to describe the attitude transformation between two coordinate systems; This represents a 3×1 translation vector used to describe the positional relationship between the origins of two coordinate systems. During the detection process, the aforementioned extrinsic parameter matrix remains unchanged, serving as the fixed mounting parameters for the 3D contour scanner and polarization camera relative to the robotic arm's end effector.

[0036] 1-3) Real-time reading of the end flange pose matrix output by the industrial robotic arm controller. Where B represents the global coordinate system of the robotic arm and E represents the coordinate system of the end flange. Combined with the 3D contour scanner pose transformation matrix T obtained in steps 1-2), ES And polarization camera pose transformation matrix T EC Calculate the real-time pose of the 3D contour scanner and polarization camera in the global coordinate system of the robotic arm:

[0037]

[0038] in: Real-time pose of the 3D contour scanner in the global coordinate system of the robotic arm; Real-time pose of the polarization camera in the global coordinate system of the robotic arm.

[0039] Through the above coordinate transformation, both the 3D point cloud data and the polarization image data are uniformly expressed in the robot arm's base coordinate system {B}, providing a unified spatial reference for subsequent multi-source data fusion.

[0040] 2) Based on the global coordinate system {B} of the robotic arm established in step 1), control the industrial robotic arm 4 to carry the 3D contour scanner 7 to scan the surface of the non-planar composite material 1 to be tested, and obtain the original 3D point cloud data of the material surface (as shown in the appendix of the instruction manual). Figure 3 (As shown). The original 3D point cloud data is converted to the robot arm's global coordinate system {B}, and the local surface normal vectors corresponding to each sampling point are extracted to obtain 3D geometric data containing 3D point coordinates and local surface normal vectors.

[0041] This step specifically involves: 2-1), based on the global coordinate system of the robotic arm established in step 1), controlling the industrial robotic arm to carry a 3D contour scanner to perform non-contact scanning of the non-planar composite material to be tested, acquiring the original 3D point cloud data of the material surface in real time, and simultaneously recording the real-time pose information of the end effector of the industrial robotic arm during the scanning process, providing pose data for the 3D point cloud coordinate transformation.

[0042] 2-2), call the 3D contour scanner pose transformation matrix T obtained in step 1-2). ES In addition to the real-time pose information of the robotic arm recorded in step 2-1), the original three-dimensional point cloud data collected in step 2-1) is subjected to coordinate transformation, and the three-dimensional points in the scanner coordinate system are uniformly transformed to the global coordinate system of the robotic arm to obtain a three-dimensional point cloud dataset in a unified coordinate system, which provides geometric data for establishing a spatial correspondence between the subsequent polarization image and the three-dimensional point cloud.

[0043] 2-3), Based on the 3D point cloud data obtained in step 2-2), the local surface normal vectors corresponding to each sampling point are calculated using 3D point cloud processing software, forming 3D geometric data containing the 3D point coordinates and their corresponding local surface normal vectors. The data representation is as follows: ; in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to the point; N represents the total number of point cloud sampling points.

[0044] The three-dimensional geometric data is used as the data input for step 4) to establish the spatial correspondence between the two-dimensional polarization image and the three-dimensional point cloud and to restore the three-dimensional spatial weaving angle.

[0045] 3) Based on the global coordinate system {B} of the robotic arm established in step 1), control the industrial robotic arm 4 to move to the set observation pose with the polarization camera 6, and acquire the polarization image of the surface of the non-planar composite material 1 to be tested; process the polarization pixels of the polarization image, calculate the two-dimensional linear polarization angle of the component surface based on the polarization image, and obtain the corresponding two-dimensional linear polarization angle data.

[0046] This step is based on the fact that the fiber bundles on the surface of the fiber composite material have a significant anisotropic structure, and their reflection characteristics of incident light are related to the fiber axial direction. When natural light is incident on the fiber surface, the reflected light exhibits directional linear polarization characteristics due to the different electric field vibration responses of the longitudinal and transverse structures of the fiber during light propagation and reflection. The principal polarization direction of the reflected light corresponds to the projection direction of the fiber axial direction onto the imaging plane. Therefore, by acquiring light intensity information under different polarization directions using a polarization camera and calculating the two-dimensional linear polarization angle (AoLP), the two-dimensional projection direction of the fiber axial direction onto the camera's imaging plane can be characterized. This step utilizes the principle of polarization imaging to obtain the two-dimensional linear polarization angle information of the component surface, and uses the obtained two-dimensional linear polarization angle data as the data input for step 4) to establish the spatial correspondence and recover the three-dimensional spatial weaving angle.

[0047] This step specifically involves: 3-1) Based on the global coordinate system of the robotic arm established in step 1), controlling the industrial robotic arm equipped with a polarization camera to move to the preset observation pose and remain stationary. The operator determines the fiber weaving angle detection range and adjusts the exposure time and gain of the polarization camera according to the preset exposure parameters to complete the preparation for polarization image acquisition.

[0048] 3-2) Using a DoFP (Do-Focus Plane) polarization camera, a single exposure is performed to simultaneously acquire raw polarization images in four polarization directions: 0°, 45°, 90°, and 135°, as shown in the instruction manual. Figure 4As shown in a to d in the figure. Based on the light intensity information of the above four polarization images, Stokes parameters are calculated and calculated pixel by pixel to generate a linear polarization degree (DoLP) matrix and a two-dimensional linear polarization angle (AoLP) matrix; to achieve intuitive visualization of the detection results, the matrix data is processed by pseudo-color mapping, and a polarization degree image is rendered and output (as shown in the attached figure). Figure 5 (as shown in a) and the linear polarization angle image (as attached) Figure 5 (as shown in b in the text).

[0049] Specifically, 3-2-1), using a polarization camera, a single exposure is performed to simultaneously capture polarization intensity images of four different polarization directions (0°, 45°, 90°, and 135°) within the same field of view, denoted as [images of polarized light intensity]. , , , The use of a single-exposure, multi-channel acquisition mechanism ensures absolute consistency of images with different polarization states in both time and space, providing input data for subsequent polarization parameter calculations.

[0050] 3-2-2), using the four polarization intensity images acquired in step 3-2-1) as input, calculates the polarization state parameters pixel-by-pixel based on Stokes polarization theory. Among them, The total light intensity component is ; Horizontal and vertical polarization difference components ; diagonal polarization difference component .

[0051] Subsequently, the two-dimensional linear polarization angle and degree of linear polarization under the current camera target surface are calculated pixel by pixel using the above components: where, the two-dimensional linear polarization angle ; linear polarization degree .

[0052] After the calculation is completed, the two-dimensional linear polarization angle (AoLP) matrix and linear polarization degree (DoLP) matrix are output.

[0053] Among them, the two-dimensional linear polarization angle matrix is ​​used as the input for effective polarization data screening in step 3-3) and for recovering the three-dimensional spatial weaving angle in step 4), and the linear polarization degree matrix is ​​used as the basis for effective polarization region screening in step 3-3).

[0054] 3-3), call the linear polarization degree matrix obtained in step 3-2) as the polarization confidence index to perform quality screening on pixel-level polarization data.

[0055] The specific screening method involves: performing a global statistical analysis on the linear polarization degree (DoLP) matrix of the linear polarization degree matrix and calculating its mean. with standard deviation And set the filtering threshold as follows:

[0056] in, This is an empirical adjustment coefficient, with a value range of 0.5-1 depending on the captured image.

[0057] When a pixel satisfies: If the region corresponding to the pixel is determined to be a low polarization confidence region, it is removed; otherwise, its corresponding two-dimensional linear polarization angle data is retained.

[0058] The filtered AoLP data is output as high-confidence polarization angle information and used for the three-dimensional spatial weaving angle recovery in subsequent step 4).

[0059] 4) Using the three-dimensional geometric data obtained in step 2) and the two-dimensional linear polarization angle data obtained in step 3), establish the spatial correspondence between the two-dimensional polarization image and the three-dimensional point cloud in the global coordinate system of the robotic arm; finally, calculate the three-dimensional spatial weaving angle of the fiber.

[0060] Step 4) specifically involves: 4-1) Utilizing the 3D point cloud data and local surface normal vector data obtained in step 2), and the 2D linear polarization angle matrix and linear polarization degree matrix obtained in step 3), the 3D point cloud data and polarized image data are transformed into the robot's global coordinate system based on the fixed spatial pose relationship between the 3D contour scanner and the polarization camera. Using the polarization camera's intrinsic parameters and its pose information relative to the robot's global coordinate system, a projection correspondence between the 2D image pixel coordinates and the 3D point cloud spatial coordinates is established, achieving a one-to-one matching of 2D polarized pixels and 3D spatial points; ultimately, a fused dataset containing spatial coordinates, local surface normal vectors, and polarization information is obtained.

[0061] The specific representation of the fused dataset is as follows:

[0062] in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to that point; The corresponding two-dimensional linear polarization angle AoLP is obtained by pixel mapping; The corresponding linear polarization degree DoLP is used to characterize the confidence level of this data point.

[0063] The fused dataset serves as the input data basis for the three-dimensional spatial weaving angle recovery in step 4-2).

[0064] 4-2), based on the fused dataset obtained in step 4-1), the two-dimensional linear polarization angle AoLP is used as a directional representation of the fiber axis in the camera imaging plane, that is, for any detection point: Its corresponding local surface normal vector is: This is used to determine the local spatial attitude of the component surface at that point. Since the surface of a non-planar component has curvature variations, the two-dimensional polarization angle AoLP only represents the projection direction of the fiber direction onto the camera imaging plane. Therefore, the spatial attitude of this two-dimensional direction is corrected using the surface normal vector of the corresponding point, transforming the two-dimensional polarization direction to the corresponding three-dimensional curved surface position, thus realizing the recovery from the two-dimensional projection direction to the three-dimensional spatial direction.

[0065] Next, calculate the three-dimensional spatial weaving angle corresponding to this point. The specific calculation method is as follows: First, calculate the two-dimensional fiber projection direction based on the two-dimensional polarization angle AoLP: By associating this direction with the local surface normal vector of the corresponding 3D point, and eliminating the projection deviation caused by surface tilt through surface geometric constraints, the 3D direction vector of the fiber at this location is obtained: .

[0066] Then, based on the angle between the fiber's three-dimensional direction vector and the component's reference direction vector, calculate the corresponding three-dimensional spatial weaving angle at that point. ; Wherein, the component reference direction vector The initial braiding reference axial vector selected along the non-planar component to be measured.

[0067] The above calculations enable the conversion from two-dimensional polarization angle information to three-dimensional spatial weaving angle.

[0068] 4-3) Associate the 3D spatial weaving angle data of each point obtained in step 4-2) with the corresponding 3D point cloud data to form a 3D angle dataset containing spatial coordinates and angle information: The 3D angle dataset is loaded using the 3D visualization software ParaView. The calculated 3D spatial weaving angles are then applied to the 3D surface of the component using a color mapping method, thereby achieving a visual output of the 3D spatial weaving angle distribution results (as shown in the instruction manual). Figure 6 As shown in the figure, the three-dimensional spatial weaving angle distribution map of the area to be tested and the corresponding detection results are finally obtained.

[0069] The above-described specific embodiments are merely preferred embodiments of this invention and are not intended to limit this invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this invention should be included within the protection scope of this invention.

Claims

1. A method for detecting the three-dimensional spatial braiding angle of fibers in non-planar composite materials, characterized in that: The steps include the following: 1) Fix the non-planar composite material to be tested on the test stand, calibrate the pose transformation relationship between the 3D contour scanner and the polarization camera relative to the end effector of the industrial robot arm, establish a unified global coordinate system for the robot arm, and obtain the pose information of the industrial robot arm, the 3D contour scanner and the polarization camera in the global coordinate system of the robot arm. 2) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm to carry a 3D contour scanner to scan the surface of the non-planar composite material to be tested, and obtain the original 3D point cloud data of the material surface; convert the original 3D point cloud data to the global coordinate system of the robotic arm, and extract the local surface normal vectors corresponding to each sampling point to obtain 3D geometric data containing 3D point coordinates and local surface normal vectors. 3) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm equipped with a polarization camera to move to the set observation pose and acquire polarization images of the surface of the non-planar composite material to be tested; process the polarization pixels of the polarization images, calculate the two-dimensional linear polarization angle of the component surface based on the polarization images, and obtain the corresponding two-dimensional linear polarization angle data. 4) Using the three-dimensional geometric data obtained in step 2) and the two-dimensional linear polarization angle data obtained in step 3), establish the spatial correspondence between the two-dimensional polarization image and the three-dimensional point cloud in the global coordinate system of the robotic arm; finally, calculate the three-dimensional spatial weaving angle of the fiber.

2. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 1, characterized in that: Step 1) specifically refers to: 1-1), The non-planar composite material to be tested is mounted on the testing stand, and the six degrees of freedom of the material are constrained; (1-2) Based on the fixed installation relationship between the 3D contour scanner and the polarization camera relative to the end effector of the industrial robotic arm, the spatial pose transformation relationships between the contour scanner coordinate system, the polarization camera coordinate system, and the end flange coordinate system of the robotic arm are established respectively, and the pose transformation matrix T of the 3D contour scanner is calculated. ES And polarization camera pose transformation matrix T EC ; 1-3) Real-time reading of the end flange pose matrix output by the industrial robotic arm controller. Where B represents the global coordinate system of the robotic arm and E represents the coordinate system of the end flange; combined with the pose transformation matrix T of the 3D contour scanner obtained in steps 1-2), ES And polarization camera pose transformation matrix T EC Calculate the real-time pose of the 3D contour scanner and polarization camera in the global coordinate system of the robotic arm: in: Real-time pose of the 3D contour scanner in the global coordinate system of the robotic arm; Real-time pose of the polarization camera in the global coordinate system of the robotic arm.

3. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 2, characterized in that: Step 2) specifically refers to: 2-1) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm to carry a three-dimensional contour scanner to perform non-contact scanning of the non-planar composite material to be tested, acquire the original three-dimensional point cloud data of the material surface in real time, and simultaneously record the real-time pose information of the end effector of the industrial robotic arm during the scanning process. 2-2), call the 3D contour scanner pose transformation matrix T obtained in step 1-2). ES In addition to the real-time pose information of the robotic arm recorded in step 2-1), coordinate transformation is performed on the original three-dimensional point cloud data collected in step 2-1), and the three-dimensional points in the scanner coordinate system are uniformly transformed to the global coordinate system of the robotic arm to obtain a three-dimensional point cloud dataset in a unified coordinate system. 2-3) Based on the 3D point cloud data obtained in step 2-2), calculate the local surface normal vector corresponding to each sampling point, and form 3D geometric data containing the 3D point coordinates and their corresponding local surface normal vectors. The data representation is as follows: in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to the point; N represents the total number of point cloud sampling points.

4. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 1, characterized in that: Step 3) specifically refers to: 3-1) Based on the global coordinate system of the robotic arm established in step 1), control the industrial robotic arm carrying the polarization camera to move to the preset observation pose and keep it stationary; adjust the exposure time and gain of the polarization camera according to the preset exposure parameters to complete the preparation for polarization image acquisition; 3-2) Use a polarization camera to perform a single exposure and simultaneously acquire original polarization images in four polarization directions: 0°, 45°, 90° and 135°. Based on the light intensity information of the above four polarization images, calculate the Stokes parameters and calculate them pixel by pixel to generate a linear polarization degree matrix and a two-dimensional linear polarization angle matrix. 3-3), call the linear polarization degree matrix obtained in step 3-2) as the polarization confidence index to perform quality screening on pixel-level polarization data.

5. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 4, characterized in that: Step 3-2) specifically involves: 3-2-1) A single exposure was performed using a polarization camera to simultaneously capture polarization intensity images of four different polarization directions (0°, 45°, 90°, and 135°) within the same field of view, denoted as [images of polarized light intensity]. , , , ; 3-2-2), using the four polarization intensity images acquired in step 3-2-1) as input, and based on Stokes polarization theory, the polarization state parameters are calculated pixel by pixel; among them, The total light intensity component is ; Horizontal and vertical polarization difference components diagonal polarization difference component ; Subsequently, the two-dimensional linear polarization angle and degree of linear polarization under the current camera target surface are calculated pixel by pixel using the above components: where, the two-dimensional linear polarization angle ; linear polarization degree ; After the calculation is completed, the two-dimensional linear polarization angle matrix and linear polarization degree matrix are output.

6. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 4, characterized in that: The screening method in step 3-3) is as follows: A global statistical analysis of the linear polarization degree DoLP of the linear polarization degree matrix was performed, and its mean was calculated. with standard deviation And set the filtering threshold as follows: in, This is an empirical adjustment coefficient; When a pixel satisfies: If the region corresponding to the pixel is determined to be a low polarization confidence region, it is removed; otherwise, its corresponding two-dimensional linear polarization angle data is retained.

7. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 1, characterized in that: Step 4) specifically refers to: 4-1) Using the 3D point cloud data and local surface normal vector data obtained in step 2), and the 2D linear polarization angle matrix and linear polarization degree matrix obtained in step 3), the 3D point cloud data and polarized image data are converted to the global coordinate system of the robotic arm based on the fixed spatial pose relationship between the 3D contour scanner and the polarization camera. Using the intrinsic parameters of the polarization camera and its pose information relative to the global coordinate system of the robotic arm, a projection correspondence between the 2D image pixel coordinates and the 3D point cloud spatial coordinates is established to achieve one-to-one matching between 2D polarized pixels and 3D spatial points. Finally, a fused dataset containing spatial coordinates, local surface normal vectors, and polarization information is obtained. 4-2), based on the fused dataset obtained in step 4-1), the two-dimensional linear polarization angle AoLP is used as a directional representation of the fiber axis in the camera imaging plane, that is, for any detection point: Its corresponding local surface normal vector is: And calculate the three-dimensional weaving angle corresponding to that point; 4-3) Associate the 3D spatial weaving angle data of each point obtained in step 4-2) with the corresponding 3D point cloud data to form a 3D angle dataset containing spatial coordinates and angle information: ; The three-dimensional angle dataset is loaded using the three-dimensional visualization software ParaView, and the calculated three-dimensional spatial weaving angles are attached to the three-dimensional surface of the component in a color mapping manner to achieve a visual output of the three-dimensional spatial weaving angle distribution results.

8. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 7, characterized in that: The specific representation of the fused dataset established in step 4-1) is as follows: in: () represents the spatial coordinates of the iii-th point in the robot's global coordinate system; This represents the unit normal vector corresponding to that point; The corresponding two-dimensional linear polarization angle AoLP is obtained by pixel mapping; This corresponds to the linear polarization degree DoLP.

9. The method for detecting the three-dimensional spatial braiding angle of non-planar composite fiber as described in claim 7, characterized in that: The specific method for the three-dimensional spatial weaving angle in step 4-2) is as follows: Calculate the two-dimensional fiber projection direction based on the two-dimensional polarization angle AoLP: By associating this direction with the local surface normal vector of the corresponding 3D point, the 3D direction vector of the fiber at that location can be obtained: ; Calculate the three-dimensional spatial weaving angle corresponding to this point based on the angle between the fiber's three-dimensional direction vector and the component's reference direction vector. Wherein, the component reference direction vector The initial braiding reference axial vector selected along the non-planar component to be measured.

Images