Method and system for detecting the surface profile of a fabric based on laser measurement

Abstract

The application discloses a fabric surface profile detection method and system based on laser measurement, and relates to the technical field of fabric surface quality detection.The method comprises the following steps: placing a fabric to be detected on a fabric conveying platform, and sequentially arranging a laser profile scanning array and a laser speckle interference array in the conveying direction of the fabric conveying platform; scanning two-dimensional profile data of the fabric to be detected according to the laser profile scanning array, and extracting a static profile feature set; applying a dynamic excitation to the fabric to be detected, obtaining a speckle pattern sequence of a detection area based on the laser speckle interference array, and extracting a dynamic profile feature set; discriminating the profile defects of the fabric to be detected based on the static profile feature set and the dynamic profile feature set, and constructing a multi-dimensional profile feature vector of the fabric to be detected.The application is used to solve the problems of single-laser swing scanning in the prior art, such as measurement accuracy, and single-dimension measurement of only static geometric profile.

Description

Technical Field

[0001] This invention relates to the field of fabric surface quality inspection technology, and more specifically, to a method and system for fabric surface contour inspection based on laser measurement. Background Technology

[0002] Currently, laser-based surface contour detection has been widely used in many fields.

[0003] For example, Chinese invention patent application document CN103900489A discloses a method and corresponding device for measuring three-dimensional contours by line laser scanning. The method involves a line laser that swings under the control of a magnetostrictive micro-displacement controller to form a scanning surface for the object being detected. The swing angle ω is determined by the size of the object to ensure that the line laser completely scans the contour of the object being detected. Then, the height information of each contour point is obtained to reconstruct the three-dimensional surface contour of the object.

[0004] However, the above method has the following shortcomings: (1) Single laser oscillation scanning has measurement accuracy problems: The above method uses a magnetostrictive micro-displacement controller to drive a line laser to oscillate back and forth. When the laser oscillates back and forth, factors such as small vibrations and the conversion of oscillation in the forward and reverse directions cause the oscillation angular velocity to be a variable, making it impossible to obtain uniform oscillation. It is difficult to accurately measure and calculate the coordinate values ​​of each point in the curved surface space, resulting in random measurement and calculation errors. (2) Only static geometric contours are measured, and the measurement dimension is single: The above method only focuses on the static three-dimensional geometric contours of the object surface and does not involve the dynamic mechanical response measurement of the object under stress. For flexible materials (such as fabrics), static geometric data cannot reflect its elasticity, resilience, stiffness distribution and other mechanical performance information.

[0005] To address the aforementioned shortcomings, there is an urgent need for a fabric surface contour detection method that can comprehensively and multidimensionally detect complex fabric contours and jointly detect both static geometric contours and dynamic mechanical responses of fabrics. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a method and system for detecting fabric surface contours based on laser measurement, in order to solve the problems of measurement accuracy issues in single-laser oscillating scanning and the problem of measuring only static geometric contours with a single measurement dimension in the prior art.

[0007] To achieve the above objectives, this application provides the following technical solution: The first aspect of this invention provides a method for detecting the surface contour of a fabric based on laser measurement, comprising: placing the fabric to be detected on a fabric conveying platform, wherein a laser contour scanning array and a laser speckle interferometer array are sequentially arranged along the conveying direction of the fabric conveying platform; scanning the two-dimensional contour data of the corresponding detection area of ​​the fabric to be detected based on the laser contour scanning array, solving for the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extracting a static contour feature set based on the optimal three-dimensional coordinates; applying dynamic excitation to the fabric to be detected, obtaining a speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculating the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extracting a dynamic contour feature set based on the full-field displacement field and full-field strain field; judging contour defects of the fabric to be detected based on the static contour feature set and the dynamic contour feature set, and normalizing and mapping the static contour feature set and the dynamic contour feature set to construct a multi-dimensional contour feature vector of the fabric to be detected.

[0008] In a preferred embodiment, the fabric to be inspected is placed on a fabric conveying platform, comprising: the fabric conveying platform for placing the fabric to be inspected and conveying the fabric to be inspected according to a preset conveying direction and conveying speed; the laser profile scanning array is composed of a preset number of laser profile sensors, which are arranged sequentially above the conveying direction of the fabric conveying platform based on the surface normal of the fabric to be inspected; the laser speckle interferometer array includes a beam expander laser and a high-speed camera, the beam expander laser being used to form an interference speckle pattern on the surface of the fabric to be inspected, and the high-speed camera being used to continuously acquire the speckle pattern.

[0009] In a preferred embodiment, the optimal three-dimensional coordinates of each surface feature point within the detection area are determined based on two-dimensional contour data, including: the laser contour sensors of the laser contour scanning array simultaneously emit linear lasers from their respective azimuth angles to scan and acquire two-dimensional contour data of the detection area, wherein the two-dimensional contour data includes the lateral coordinates and height values ​​of each surface feature point; the spatial feature parameters of the laser contour sensors are acquired; the two-dimensional contour data are transformed into a coordinate system based on the spatial feature parameters to obtain the three-dimensional coordinates of each surface feature point in the detection area; and the three-dimensional coordinates of each surface feature point are solved simultaneously to obtain the optimal three-dimensional coordinates of the surface feature points.

[0010] In a preferred embodiment, extracting a static contour feature set based on optimal three-dimensional coordinates includes: obtaining a reference plane by fitting the optimal three-dimensional coordinates of each surface feature point in the detection area; obtaining the surface flatness of the fabric to be tested based on the deviation value between each surface feature point and the reference plane; obtaining the average contour height of each surface feature point in the detection area; obtaining the local undulation amplitude of the fabric to be tested based on the absolute value of the difference between the actual contour height of each surface feature point and the average contour height; obtaining the principal curvature of each surface feature point in the detection area; calculating the Gaussian curvature of each surface feature point in the detection area based on the principal curvature; and forming a static contour feature set of the fabric to be tested based on the surface flatness, local undulation amplitude, and Gaussian curvature.

[0011] In a preferred embodiment, a dynamic excitation is applied to the fabric to be tested, and a speckle pattern sequence of the test area is obtained based on a laser speckle interferometer array, including: acquiring an initial speckle pattern on the surface of the fabric to be tested based on the laser speckle interferometer array; applying a dynamic excitation to the fabric to be tested after the initial speckle pattern acquisition is completed, and acquiring a dynamic speckle pattern on the surface of the fabric to be tested during the dynamic excitation process; and forming a speckle pattern sequence based on the dynamic speckle pattern and the initial speckle pattern.

[0012] In a preferred embodiment, calculating the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence includes: selecting the initial position point of the initial speckle pattern to construct a reference window, and searching for candidate windows in the dynamic speckle pattern based on the reference window; calculating the regional similarity between the candidate window and the reference window, selecting the target position point in each dynamic speckle pattern based on the regional similarity, obtaining the displacement vector of the target position point based on the offset between the target position point and the initial position point; forming the full-field displacement field of the fabric surface to be detected based on the displacement vector of the target position point, and obtaining the full-field strain field of the fabric surface to be detected by spatial differentiation of the full-field displacement field.

[0013] In a preferred embodiment, a dynamic contour feature set is extracted based on the full-field displacement field and the full-field strain field, including: extracting the peak displacement vector of each target position point in the full-field displacement field over a time series, and forming a spatial distribution map based on the peak displacement vector; calculating the full-field average strain and strain standard deviation based on the strain vector of each surface feature point in the full-field strain field, and determining the strain concentration region based on the full-field average strain and strain standard deviation; generating a displacement response curve of the target position point based on the displacement vector of the target position point in the detection area at each time, and obtaining the dynamic response consistency of the fabric to be tested based on the displacement response curve and a preset reference curve; obtaining the peak displacement vector and average displacement vector of each target position point based on the displacement response curve, and obtaining the dynamic response recovery rate of the fabric to be tested based on the peak displacement vector and average displacement vector; and constructing a dynamic contour feature set of the fabric to be tested based on the spatial distribution map, the strain concentration region, the dynamic response consistency, and the dynamic response recovery rate.

[0014] In a preferred embodiment, contour defect discrimination of the fabric to be inspected is performed based on a static contour feature set and a dynamic contour feature set, including: performing binarized mapping on the contour features in both the static and dynamic contour feature sets to obtain the contour defect type of the fabric to be inspected; obtaining the judgment time and defect judgment location of the contour defect type, and obtaining the remaining distance from the defect judgment location to the preset defect processing station; calculating the estimated arrival time of the fabric to be inspected to the defect processing station based on the remaining distance, the judgment time, and the conveying speed of the fabric conveying platform; and marking the contour defect of the fabric to be inspected at the defect processing station according to the estimated arrival time.

[0015] In a preferred embodiment, the static contour feature set and the dynamic contour feature set are normalized and mapped to construct a multidimensional contour feature vector of the fabric to be detected, including: normalizing and mapping the static contour feature set to obtain a static contour feature vector; normalizing and mapping the dynamic contour feature set to obtain a dynamic contour feature vector; and constructing a multidimensional contour feature vector of the fabric to be detected based on the static contour feature vector and the dynamic contour feature vector.

[0016] A second aspect of the present invention provides a fabric surface contour detection system based on laser measurement, comprising: a fabric conveying module for placing the fabric to be inspected on a fabric conveying platform, wherein a laser contour scanning array and a laser speckle interferometer array are sequentially arranged along the conveying direction of the fabric conveying platform; a static contour feature module for scanning two-dimensional contour data of the corresponding detection area of ​​the fabric to be inspected based on the laser contour scanning array, solving for the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extracting a static contour feature set based on the optimal three-dimensional coordinates; a dynamic contour feature module for applying dynamic excitation to the fabric to be inspected, obtaining a speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculating the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extracting a dynamic contour feature set based on the full-field displacement field and full-field strain field; and a multi-dimensional contour feature module for judging contour defects of the fabric to be inspected based on the static contour feature set and the dynamic contour feature set, and performing normalization mapping on the static contour feature set and the dynamic contour feature set to construct a multi-dimensional contour feature vector of the fabric to be inspected.

[0017] The beneficial effects of this invention are: 1. This invention places the fabric to be inspected on a fabric conveying platform, with a laser contour scanning array and a laser speckle interference array arranged sequentially along the conveying direction of the fabric conveying platform. The arrangement of the laser contour scanning array solves the problem in the prior art where it is difficult to accurately measure and calculate the coordinate values ​​of each point in the curved surface space due to the use of a magnetostrictive micro-displacement controller to drive a line laser back and forth. The arrangement of the fabric conveying platform and the laser contour scanning array can accurately detect complex three-dimensional contours such as draping wrinkles, sidewall cracks, and three-dimensional textures on the fabric surface, improving the comprehensiveness and accuracy of physical surface contour detection.

[0018] 2. The fabric conveying platform of the present invention is equipped with a laser speckle interferometer array in the conveying direction, and scans the two-dimensional contour data of the detection area of ​​the fabric to be inspected based on the laser contour scanning array. The optimal three-dimensional coordinates of each surface feature point in the detection area are solved based on the two-dimensional contour data, and a static contour feature set is extracted based on the optimal three-dimensional coordinates. Dynamic excitation is applied to the fabric to be inspected, and the speckle pattern sequence of the detection area is obtained based on the laser speckle interferometer array. The full-field displacement field and full-field strain field of the detection area are calculated based on the speckle pattern sequence, and a dynamic contour feature set is extracted based on the full-field displacement field and full-field strain field. By using the static contour feature set and the dynamic contour feature set, the surface contour defects of the fabric to be inspected are judged and identified from both static and dynamic dimensions, thereby further improving the comprehensiveness and accuracy of fabric surface contour detection. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the process for a laser-based fabric surface contour detection method. Figure 2 This is a schematic diagram of a fabric surface contour detection system based on laser measurement. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0021] Example 1, please refer to Figure 1 , Figure 1 This invention presents a method for detecting fabric surface contours based on laser measurement, comprising the following steps: placing the fabric to be tested on a fabric conveying platform, wherein a laser contour scanning array and a laser speckle interferometer array are sequentially arranged along the conveying direction of the fabric conveying platform; scanning the two-dimensional contour data of the corresponding detection area of ​​the fabric to be tested using the laser contour scanning array, solving for the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extracting a static contour feature set based on the optimal three-dimensional coordinates; applying dynamic excitation to the fabric to be tested, obtaining the speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculating the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extracting a dynamic contour feature set based on the full-field displacement field and full-field strain field; judging contour defects in the fabric to be tested based on the static contour feature set and the dynamic contour feature set, and normalizing and mapping the static contour feature set and the dynamic contour feature set to construct a multi-dimensional contour feature vector of the fabric to be tested.

[0022] S1. The fabric to be inspected is placed on a fabric conveying platform. A laser profile scanning array and a laser speckle interferometer array are sequentially arranged along the conveying direction of the fabric conveying platform, including: S11. The fabric conveying platform is used to place the fabric to be inspected and to convey the fabric to be inspected according to the preset conveying direction and conveying speed. In this embodiment, the fabric conveying platform continuously conveys the fabric to be tested along the conveying direction at a constant conveying speed.

[0023] S12. The laser contour scanning array consists of a preset number of laser contour sensors. Based on the surface normal of the fabric to be detected, the laser contour sensors are arranged sequentially above the conveying direction of the fabric conveying platform. In this embodiment, the preset number of laser contour sensors in the laser contour scanning array is greater than or equal to 3, wherein the preset number is preferably 5. The laser profile sensor works based on the principle of laser triangular reflection. Specifically, the laser profile sensor projects a laser line onto the surface of the fabric to be inspected, a camera captures the reflected light, and the height of the fabric to be inspected is measured based on the height of the laser beam.

[0024] Arranging laser profile sensors sequentially above the fabric conveying platform along the surface normal of the fabric means that the angle between the laser emission direction of the laser profile sensor in the laser profile scanning array and the normal of the fabric surface is different for each sensor. In a preferred scheme, the pitch angles of the five laser profile sensors are +60°, +30°, 0°, -30°, and -60°, respectively, where the normal of the fabric surface is taken as 0°, and positive and negative values ​​represent the forward and backward tilt of the conveying direction. Furthermore, the scanning planes of each laser profile sensor are coplanar, ensuring that all laser profile sensors scan the same transverse detection area on the surface of the fabric to be tested.

[0025] It should be noted that this embodiment also introduces a synchronous trigger controller, which simultaneously sends trigger signals to the laser contour sensors in the laser contour scanning array, so that each laser contour sensor emits a linear laser from its own azimuth angle at the same time to scan the same transverse detection area on the surface of the fabric to be detected.

[0026] S13. The laser speckle interferometer array includes a beam expander laser and a high-speed camera. The beam expander laser is used to form an interference speckle pattern on the surface of the fabric to be inspected, and the high-speed camera is used to continuously acquire the speckle pattern. In this embodiment, the laser speckle interferometer array is arranged behind the laser profile scanning array. That is, the fabric to be tested on the fabric conveying platform is first scanned by the laser profile scanning array and then subjected to speckle interference by the laser speckle interferometer array. The laser speckle interferometer array includes a beam expander laser and a high-speed camera. The beam expander laser is used to generate a coherent expanded beam laser covering the corresponding detection area of ​​the fabric to be tested. After beam expansion, the coherent expanded beam laser uniformly illuminates the surface of the fabric to be tested, forming a random interference speckle pattern on the surface of the fabric to be tested. The high-speed camera is used to continuously acquire time-series images of the speckle pattern generated on the surface of the fabric to be tested, with an acquisition frame rate ≥500fps. In this embodiment, the acquisition frequency is preferably 1000fps.

[0027] S2. Based on the laser contour scanning array, scan the two-dimensional contour data of the detection area of ​​the fabric to be detected, solve the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extract the static contour feature set based on the optimal three-dimensional coordinates. S21. Based on the two-dimensional contour data, solve for the optimal three-dimensional coordinates of each surface feature point within the detection area, including: S211. The laser contour sensors of the laser contour scanning array emit linear lasers from their respective azimuth angles at the same time to scan and acquire two-dimensional contour data of the detection area. The two-dimensional contour data includes the lateral coordinates and height values ​​of each surface feature point in the fabric to be detected. Two-dimensional contour data is a sequence of two-dimensional coordinates of continuous discrete points corresponding to the same transverse detection area on the surface of the fabric to be detected, obtained by each laser contour sensor in its local coordinate system. Each surface feature point in the fabric to be detected is a continuous discrete point of the same transverse detection area on the surface of the fabric to be detected. Due to the different installation positions and azimuth orientations of each laser profile sensor, multiple sets of two-dimensional profile data for each surface feature point are located in the local coordinate system of their respective laser profile sensors. In the local coordinate system With the laser emission center of the laser contour sensor as the origin, the x-axis is along the transverse width direction of the fabric to be detected, and this component represents continuous discrete points within the same transverse detection area; the y-axis is along the fabric conveying direction, and since the laser contour sensor performs a single-frame instantaneous scan of the fabric, this component is 0; the z-axis is the laser emission direction, and this component represents the height value measured by the laser contour sensor. Due to the local coordinate system of the laser contour sensor... Since the y-axis component is 0, the two-dimensional contour data is actually two-dimensional data represented in the XZ coordinate system. The two-dimensional contour data can be represented as follows: ,in These represent the lateral coordinates of each surface feature point in the corresponding detection area of ​​the fabric to be inspected. The height value is measured by a laser profile sensor.

[0028] S212. Obtain the spatial feature parameters of the laser contour sensor, and perform coordinate system transformation on the two-dimensional contour data based on the spatial feature parameters to obtain the three-dimensional coordinates of each surface feature point in the fabric to be detected. In this embodiment, the spatial feature parameters of the laser contour sensor are used to transform the coordinate system of multiple sets of two-dimensional contour data of each surface feature point in the detection area of ​​the fabric to be detected, so that the multiple sets of two-dimensional contour data of each surface feature point are transformed from the local coordinate system. Unify to standard coordinate system Specifically: (1) Obtain the spatial characteristic parameters of the laser profile sensor. The spatial characteristic parameters of the laser profile sensor include spatial installation parameters and spatial position parameters. Spatial installation parameters include the pitch angle, yaw angle, and roll angle of the laser profile sensor, where the pitch angle refers to the angle in the local coordinate system of the laser profile sensor. In this context, the rotation angle of the laser profile sensor around the z-axis is defined by the width direction of the fabric being inspected. The rotation manifests as a pitch angle, specifically the tilt angle between the laser emission direction of the laser profile sensor and the vertical normal. The yaw angle refers to the angle within the local coordinate system of the laser profile sensor. In the diagram, the rotation angle of the laser profile sensor around the y-axis is defined, with the rotation axis being vertical. The rotation manifests as a left-right sway, and the yaw angle is specifically the tilt angle between the laser emission direction of the laser profile sensor and the vertical normal. The roll angle refers to the angle within the local coordinate system of the laser profile sensor. In the figure, the rotation angle of the laser profile sensor around the x-axis is the conveying direction of the fabric conveying platform. The rotation is a reverse rotation. The specific value of the roll angle is the left and right tilt angle value of the laser profile sensor around the width direction when it is installed, that is, the angle value caused by the deviation of one side of the laser profile sensor being higher and the other side being lower. Spatial position parameters include the lateral offset, longitudinal offset, and installation height of the laser profile sensor. The lateral offset refers to the lateral offset of the laser profile sensor relative to the center line of the fabric conveying platform, which is determined by the position of the mechanical mounting base of the laser profile sensor. The longitudinal offset refers to the longitudinal offset of the laser profile sensor along the conveying direction. Since multiple laser profile sensors are distributed sequentially along the conveying direction, the longitudinal offset of each laser profile sensor is different. The installation height refers to the installation height of the sensor relative to the reference zero plane, which is the vertical distance from the laser profile sensor to the surface of the fabric conveyor belt of the fabric conveying platform.

[0029] It should be noted that the spatial characteristic parameters of the laser profile sensor can be obtained through factory calibration and on-site installation calibration of the laser profile sensor.

[0030] (2) A rotation matrix is ​​constructed based on the pitch, yaw, and roll angles in the space installation parameters. Specifically, the pitch, yaw, and roll angles are combined in the order of yaw-pitch-roll to obtain the rotation matrix, which is 3. The orthogonal matrix of 3 is used to correct the deviation between the local coordinate system and the standard coordinate system caused by the installation tilt angle of the laser profile sensor. The physical meanings of the pitch angle, yaw angle and roll angle in the rotation matrix are as follows: By adjusting the pitch angle, each laser profile sensor can illuminate the side wall area of ​​the fabric folds to be detected from different tilt angles, thereby obtaining the coordinate data of the occluded area in subsequent steps; By adjusting the yaw angle and roll angle, it is ensured that the transverse profile lines collected by each laser profile sensor remain parallel in space and are located on the same height reference plane of the standard coordinate system.

[0031] It should be noted that the process of constructing the rotation matrix is ​​a well-known and mature method in the existing technology.

[0032] (3) Based on the lateral offset, longitudinal offset, and installation height in the spatial position parameters, a translation vector is constructed. The translation vector represents the absolute spatial coordinates of the local coordinate system origin of each laser contour sensor in the standard coordinate system. The translation vector of each laser profile sensor is specifically represented as follows: ,in The first The lateral offset, longitudinal offset, and installation height of each laser profile sensor.

[0033] (4) If the first The local two-dimensional coordinates obtained by the laser profile sensor are The coordinate system is transformed from the two-dimensional contour data using the following rigid body transformation equations to obtain: ; In the formula, For the first Local two-dimensional coordinates of the sensor Three-dimensional coordinates in the standard coordinate system These are the coordinate components of the three-dimensional coordinates; For the first Rotation matrix of a laser profile sensor; For the first Translation vector of a laser profile sensor.

[0034] It should be noted that through the coordinate transformation in this step, the contour data of each set of surface feature points are unified into the same three-dimensional spatial coordinate system, thereby achieving the unification of the two-dimensional contour data of surface feature points.

[0035] S213. Solve the simultaneous three-dimensional coordinates of each surface feature point to obtain the optimal three-dimensional coordinates of the surface feature points; After coordinate transformation, for the same surface feature point P on the fabric surface to be inspected, the two-dimensional contour data obtained by scanning in N laser contour sensors should all correspond to the same three-dimensional coordinates in the standard coordinate system after coordinate transformation. ; The observation equations of the N laser profile sensors for the surface feature point P are combined to form a system containing 3 unknowns. An overdetermined system of equations with 2N equations can be considered. Since 2N≥4>3 when N≥2, the number of equations exceeds the number of unknowns, thus constituting an overdetermined system. This system can be solved using the well-known least squares method to find a set of least squares solutions that minimizes the sum of squared errors of all equations. This least squares solution represents the optimal three-dimensional coordinates of the surface feature point P. The specific solution to the overdetermined system of equations can be achieved using numerical methods such as SVD decomposition, which are well-known in existing technology.

[0036] It should be noted that, due to the different observation angles of the various laser profilometers on the fabric being inspected, sidewall areas that are obscured during vertical scanning (such as the sides of folds or the inner walls of cracks) may still be observed by laser profilometers at other azimuth angles even if they are obscured in one laser profilometer. Therefore, through the fusion of the aforementioned multi-view data, the complete three-dimensional profile data of the fabric surface can be detected comprehensively and accurately.

[0037] S22. Extract the static contour feature set of the fabric to be detected based on the optimal three-dimensional coordinates, including: S221. A reference plane is obtained by fitting the optimal three-dimensional coordinates of each surface feature point in the detection area. The surface flatness of the fabric to be tested is obtained based on the deviation between each surface feature point and the reference plane. Based on all surface feature points of the corresponding detection area of ​​the fabric to be tested, a reference plane is obtained by fitting using the least squares method known in the prior art, and the vertical distance from each surface feature point to the reference plane is obtained. This vertical distance is used as the deviation value between the surface feature point and the reference plane. The specific calculation process for surface flatness is as follows: squaring the deviation value between each surface feature point and the reference plane, and summing all the squared results of the deviation values ​​to obtain the deviation sum of squares; dividing the deviation sum of squares by the total number of surface feature points to obtain the average value of the deviation sum of squares, and taking the second square root of this average value to obtain the surface flatness of the corresponding detection area of ​​the fabric to be tested.

[0038] It should be noted that surface smoothness is the root mean square value of the deviation between surface feature points and the reference plane. It is used to evaluate the overall smoothness of the surface. The smaller the surface smoothness value, the smoother the surface of the fabric to be tested; conversely, the larger the value, the greater the surface undulation of the fabric to be tested.

[0039] S222. Obtain the average contour height of each surface feature point within the detection area, and obtain the local undulation amplitude of the fabric to be detected based on the absolute value of the difference between the actual contour height of each surface feature point and the average contour height. The contour height values ​​of each surface feature point in the detection area are obtained along the conveying direction of the fabric conveying platform. The contour height value is actually the Z value of each surface feature point in the standard coordinate system. The average contour height of each surface feature point in the detection area is calculated and used as the reference centerline. The absolute value of the difference between each surface feature point in the detection area and the reference centerline is obtained, and the arithmetic mean of the absolute values ​​of the differences of all surface feature points in the detection area is calculated to obtain the local undulation amplitude of the fabric to be detected.

[0040] It should be noted that the local undulation amplitude is used to quantify the average degree of micro-undulation on the surface of the fabric to be tested. The larger the value of the local undulation amplitude, the more severe the micro-undulation.

[0041] S223. Obtain the principal curvature of each surface feature point in the detection area, and calculate the Gaussian curvature of each surface feature point in the detection area based on the principal curvature. The calculation process of Gaussian curvature is as follows: search for the K nearest neighbors of the surface feature point in standard three-dimensional coordinates to form a local neighborhood point set of the surface feature point. The value of K can be preset according to the point cloud density, with a preferred range of 20 to 50. In this embodiment, the preferred value is 30. Using this local neighborhood point set, a quadratic parametric surface, such as a quadratic paraboloid or a B-spline surface, is fitted using the least squares method known in the prior art to approximate the local surface shape near the surface feature point. According to the first and second basic forms of the local parametric surface known in the prior art, the two principal curvatures K1 and K2 of the surface at the surface feature point are calculated. Multiplying K1 and K2 together, the Gaussian curvature of each surface feature point in the detection area is obtained.

[0042] It should be noted that Gaussian curvature is used to accurately describe the curvature type of the local surface where the surface feature point is located. When Gaussian curvature > 0, it means that the local surface where the surface feature point is located is similar to a sphere, corresponding to independent protrusions or pits; when Gaussian curvature = 0, it means that the local surface where the surface feature point is located is flat; when Gaussian curvature < 0, it means that the local surface where the surface feature point is located is similar to a saddle surface, corresponding to a twisted area of ​​wrinkles.

[0043] S224. Based on surface flatness, local undulation amplitude, and Gaussian curvature, a static contour feature set of the fabric to be tested is formed.

[0044] S3. Apply dynamic excitation to the fabric to be inspected, obtain the speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculate the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extract the dynamic contour feature set based on the full-field displacement field and full-field strain field. S31. Apply dynamic excitation to the fabric to be inspected, and acquire the speckle pattern sequence of the inspection area based on a laser speckle interferometer array, including:

[0045] S311. Acquire the initial speckle pattern on the surface of the fabric to be tested based on a laser speckle interferometer array; Based on the beam expander laser in the laser speckle interferometer array, the expanded coherent laser is emitted. When the coherent laser illuminates the fabric to be tested with an optically rough surface, due to the random phase difference between the scattered light from the coherent light in each tiny area of ​​the fabric surface, these scattered lights interfere with each other in space, thus forming a randomly distributed bright and dark interference speckle pattern on the fabric surface. Each fabric detection area will produce a unique speckle pattern due to its unique fiber arrangement, surface texture and microstructure. When the fabric surface undergoes a small displacement or deformation, the speckle pattern will also undergo a corresponding displacement or distortion. Therefore, by tracking the changes in the speckle pattern, the displacement and deformation of the fabric surface to be tested can be inferred.

[0046] Before applying dynamic excitation to the fabric to be tested, the surface of the fabric to be tested, which has completed the acquisition of static contour feature set, is uniformly irradiated with laser based on the beam expander laser in the laser speckle interferometer array to obtain an initial speckle pattern, and the initial speckle pattern on the surface of the fabric to be tested is simultaneously acquired based on the high-speed camera in the laser speckle interferometer array.

[0047] S312. Apply dynamic excitation to the fabric to be tested after the initial speckle pattern acquisition is completed, and acquire the dynamic speckle pattern on the surface of the fabric to be tested during the dynamic excitation process. Applying dynamic excitation to the fabric to be tested after initial speckle pattern acquisition refers to applying controllable dynamic excitation to the fabric. The choice of excitation method depends on the fabric type and the detection target. For example, mechanical vibration excitation can be used: a small-amplitude sinusoidal vibration is applied to the fabric to be tested to detect the elastic modulus distribution and damping characteristics using a vibrator. The vibration frequency range can be [range missing]. Hz, the amplitude range can be mm; pulse tension excitation is adopted: an instantaneous tension pulse is applied to the fabric to be tested through the tension roller to detect the tensile strength and strain concentration area of ​​the fabric. The variation amplitude of the tension pulse is 1% to 10% of the fabric breaking strength. The fabric breaking tension is the maximum tensile force recorded by the instrument during the continuous stretching of the fabric to be tested until it is completely broken. The fabric breaking tension can be obtained based on the factory parameters or historical data of the fabric to be tested.

[0048] The acquisition of dynamic speckle patterns on the surface of the fabric to be tested during dynamic excitation refers to the continuous acquisition of dynamic speckle patterns on the surface of the fabric to be tested during dynamic excitation by a high-speed camera in a laser speckle interferometer array at a preset acquisition frequency. The preset acquisition frequency is ≥500fps, and in this embodiment, it is preferably 1000fps. It should be noted that the application time of dynamic excitation is precisely controlled by a synchronous trigger controller, and the synchronous trigger controller ensures that the acquisition start time of the high-speed camera corresponds precisely with the application time of dynamic excitation, ensuring that the application of dynamic excitation and the acquisition sequence of the high-speed camera are strictly synchronized.

[0049] S313. A speckle pattern sequence is formed based on the dynamic speckle pattern and the initial speckle pattern; The speckle pattern sequence is a time series of speckle patterns on the surface of the fabric to be tested. Specifically, the speckle pattern sequence can be represented as: ,in The initial speckle image, To induce a dynamic speckle pattern that is continuously collected over time on the surface of the fabric to be tested during the excitation process.

[0050] S32. Calculate the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, including: S321. Select the initial position points of the initial speckle pattern to construct a reference window, and search for candidate windows in the dynamic speckle pattern based on the reference window; In the initial speckle pattern, the initial position point of the initial speckle pattern is the target point selected for calculating the displacement. A square reference window is selected with the initial position of the initial speckle pattern as the center. The size parameter of the reference window can be from 21×21 pixels to 51×51 pixels. In this embodiment, the size parameter of the reference window is preferably 21×21 pixels. In each dynamic speckle pattern in the speckle pattern sequence, the reference window is moved pixel by pixel along the horizontal and vertical directions to obtain multiple candidate windows in the dynamic speckle pattern.

[0051] S322. Calculate the regional similarity between the candidate window and the reference window, select the target position point in each dynamic speckle pattern based on the regional similarity, and obtain the displacement vector of the target position point based on the offset between the target position point and the initial position point. The region similarity between the candidate window and the baseline window is calculated using the normally known normalized cross-correlation function. The specific formula for calculating region similarity is as follows: ; In the formula, The region similarity between the candidate window and the baseline window; For pixels within the reference window The grayscale value at that location; For the pixels within the candidate window The grayscale value at that location; and These are the average grayscale values ​​of all pixels within the baseline window and the candidate window, respectively. This is a two-dimensional double summation function, which means iterating through all pixels within the window.

[0052] Selecting the target location point in each dynamic speckle pattern based on regional similarity means that among all candidate windows in the dynamic speckle pattern, the center position of the candidate window with the greatest similarity to the reference window region is taken as the target location point in the dynamic speckle pattern. The physical meaning of this target location point is the best matching position of the reference window in the dynamic speckle pattern. The distance between the target position point and the initial position point is used as the offset of the target position point relative to the initial position point, and this offset is used as the displacement vector of the target position point.

[0053] S323. Based on the displacement vector of the target location point, form the full-field displacement field of the fabric surface to be tested, and perform spatial differentiation on the full-field displacement field to obtain the full-field strain field of the fabric surface to be tested. The full-field displacement field is composed of the displacement vectors of the target positions of all dynamic speckle patterns in the speckle pattern sequence. Specifically, the full-field displacement field can be expressed as: where... ,in For all target location points The displacement component in the X direction, For all target location points Displacement component in the Y direction; For each target location point, the following three strain components are calculated: the normal strain component in the X direction, the normal strain component in the Y direction, and the shear strain components in the X and Y directions. Based on the three strain components in all target locations, the full-field strain field of the fabric surface to be tested is formed. The calculation processes for the normal strain components in the X direction, the normal strain components in the Y direction, and the shear strain components in the X and Y directions are as follows: based on the displacement component in the X direction. Taking the partial derivative with respect to the X-coordinate yields the normal strain component in the X-direction. This component reflects the tensile or compressive deformation of the fabric in the width direction; positive values ​​indicate tension, and negative values ​​indicate compression. The displacement component in the Y-direction... Taking the partial derivative with respect to the Y-coordinate yields the normal strain component in the Y-direction. This component reflects the tensile or compressive deformation of the fabric under test in the conveying direction; positive values ​​indicate tension, and negative values ​​indicate compression. The displacement component in the X-direction... Taking the partial derivative with respect to the Y-coordinate, based on the displacement components in the Y direction. Take the partial derivative with respect to the X coordinate, and multiply the sum of the two partial derivatives by 0.5 to obtain the shear strain components in the X and Y directions. The shear strain components are used to reflect the shear deformation of the surface of the fabric to be tested, that is, the torsional deformation of the fabric to be tested when subjected to non-axial force.

[0054] S33. Extracting dynamic contour feature sets based on the full-field displacement field and full-field strain field, including: S331. Extract the peak displacement vector of each target location point in the full displacement field in the time series, and form a spatial distribution map based on the peak displacement vector; The X-direction displacement component and the Y-direction displacement component of the target position point (X,Y) at time t are squared respectively. The squared X-direction and Y-direction displacement components are added together, and the second square root of the sum is taken to obtain the resultant displacement amplitude of the target position point at time t. For each target position point, the maximum value of the resultant displacement amplitude corresponding to the target position point is found within the time window of the dynamic excitation. A spatial distribution map is formed based on the maximum value of the resultant displacement amplitude corresponding to each target position point.

[0055] S332. Calculate the average strain and standard deviation of the strain based on the strain vector of each surface feature point in the full-field strain field, and determine the strain concentration area based on the average strain and standard deviation of the strain. The strain threshold is generated based on the overall average strain and strain standard deviation. Specifically, the strain threshold can be expressed as: ,in For the average strain across the entire field, The strain standard deviation is defined as the strain concentration region, where the strain value exceeds the strain threshold. It should be noted that strain concentration is a precursor to potential damage or structural inhomogeneity in a fabric. In a uniform fabric, strain should be evenly distributed. When there are internal defects in a certain area (such as yarn breakage, fiber loss, or localized hardening), the strain response in that area will deviate significantly from the average level, forming strain concentration.

[0056] S333. Based on the displacement vector of each target location point in the detection area at each time, generate the displacement response curve of the target location point, and obtain the dynamic response consistency of the fabric to be tested based on the displacement response curve and the preset reference curve. Based on the displacement vectors of each target location point within the detection area at each time step, a displacement response curve of the target location point as a function of time during the continuous dynamic excitation is generated. The preset reference curve is specifically the displacement response curve of the selected geometric center point of the detection area. Based on the Pearson correlation coefficient known in the prior art, the similarity between the displacement response curve of each target location point within the detection area and the preset reference curve is calculated, and this similarity is used as the consistency of the dynamic response of the fabric to be tested.

[0057] It should be noted that the closer the dynamic response consistency is to 1, the more it indicates that the different parts of the fabric being tested fluctuate at the same amplitude after being subjected to force, indicating strong structural continuity. The lower the dynamic response consistency, the more it indicates that the energy conduction inside the fabric is blocked. For example, cracks, delamination, or severe warp and weft separation will prevent the vibration wave from being transmitted across the testing area, thus causing the motion waveforms of the two testing areas to be asynchronous.

[0058] S334. Obtain the peak value and average value of the displacement vector at each target location point based on the displacement response curve, and obtain the dynamic response recovery rate of the fabric to be tested based on the peak value and average value of the displacement vector. At the moment when the dynamic excitation is first applied, the peak displacement vector of the target position is recorded. After the dynamic excitation continues for a period of time (e.g., after the vibration excitation lasts for 100ms), the fabric under test under high-frequency vibration reaches a stable stage of dynamic equilibrium. The average displacement vector of the target position during this stable stage is taken, and the ratio of the average displacement vector to the peak displacement vector is taken as the dynamic response recovery rate of the fabric under test.

[0059] It should be noted that a higher dynamic response recovery rate indicates better fatigue and creep resistance of the fabric under test; a lower dynamic response recovery rate indicates more severe irreversible plastic deformation of the fabric under test.

[0060] S335. Based on the spatial distribution map, strain concentration area, dynamic response consistency, and dynamic response recovery rate, a dynamic contour feature set of the fabric to be tested is constructed.

[0061] S4. Determination of contour defects in the fabric to be inspected is performed based on static and dynamic contour feature sets. Based on static and dynamic contour feature sets, contour defects are identified in the fabric to be inspected, the contour defect types of the corresponding inspection areas of the fabric are obtained, and contour defects are marked on the fabric to be inspected according to the contour defect types. Specifically: S41. Based on the static contour feature set and the dynamic contour feature set, the contour defect of the fabric to be inspected is determined to obtain the contour defect type of the fabric to be inspected, including: S411. Binarize and map each static contour feature in the static contour feature set to determine the static contour state of the fabric to be detected.

[0062] Static contour states include static normal and static abnormal. All static contour features in the static contour feature set are compared to their respective preset normal ranges. If all static contour features in the set are within the normal range, the static contour state of the fabric under test is static normal. If any static contour feature in the set is outside the normal range, the static contour state of the fabric under test is static abnormal, and the type of abnormal static contour feature is recorded. It should be noted that the preset normal ranges for each static contour feature can be calibrated based on historical data of the fabric under test.

[0063] S412. Binarize and map each dynamic contour feature in the dynamic contour feature set to determine the dynamic contour state of the fabric to be detected. The dynamic profile status includes dynamic normal and dynamic abnormal. All dynamic profile features in the dynamic profile feature set are compared with their respective preset pass / fail thresholds. If all dynamic profile features in the set are within the normal range, the dynamic profile status of the fabric under test is dynamic normal. If any dynamic profile feature in the set is outside the normal range, the dynamic profile status of the fabric under test is dynamic abnormal, and the type of abnormal dynamic profile feature is recorded. It should be noted that the preset pass / fail thresholds for each dynamic profile feature can be calibrated based on historical data of the fabric under test.

[0064] S413. Determine the type of contour defect of the fabric to be inspected based on the static and dynamic contour states. Contour defect types include surface geometric defects, latent mechanical defects, and severe defects. Fabrics under inspection with a static contour state of static anomaly and a dynamic contour state of dynamic normality are classified as surface geometric defects; fabrics under inspection with a static contour state of static normality and a dynamic contour state of dynamic anomaly are classified as latent mechanical defects; and fabrics under inspection with a static contour state of static anomaly and a dynamic contour state of dynamic anomaly are classified as severe defects.

[0065] S42. Mark the contour defects of the fabric to be inspected according to the type of contour defect, including: S421. Introduce a defect handling station in the conveying direction of the fabric conveying platform, obtain the judgment time and defect judgment position when the contour defect type is detected, and obtain the remaining distance from the defect judgment position to the defect handling station. The defect processing station is set in the conveying direction of the fabric conveying platform, and after the laser profile scanning array and the laser speckle interferometer array. That is, the fabric to be inspected is first scanned and interfered with by the laser profile scanning array and the laser speckle interferometer array, and then the defect processing station is used to process the defect of the fabric to be inspected.

[0066] S422. Based on the remaining distance, the judgment time, and the conveying speed of the fabric conveying platform, calculate the estimated arrival time of the fabric to be inspected to the defect handling station. The specific calculation process for the estimated arrival time of the fabric to be inspected at the defect handling station is as follows: add the ratio of the remaining distance to the conveying speed to the judgment time to obtain the estimated arrival time of the fabric to be inspected at the defect handling station.

[0067] S423. Obtain the spatial coordinates of the defect detection area, generate an execution instruction based on the spatial coordinates, estimated arrival time and contour defect type, and mark the contour defect on the fabric to be inspected according to the execution instruction. The spatial coordinates include the transverse coordinates of the detection area of ​​the fabric with defects in the fabric width direction and the longitudinal coordinates of the fabric length direction, with the origin of the longitudinal coordinates being the starting point of the fabric conveying platform. The defect processing station marks the outline of the fabric to be inspected according to the execution instructions. Specifically, it automatically sprays identifiable color marks or inkjet codes on the inspection area of ​​the fabric to be inspected where there are defects. The color marks or inkjet codes correspond to the defect types of the inspection area of ​​the fabric to be inspected.

[0068] S5. Normalize and map the static contour feature set and the dynamic contour feature set to construct a multidimensional contour feature vector of the fabric to be detected. S51. Normalize the mapping based on the static contour feature set to obtain the contour static feature vector; The static contour feature set includes surface smoothness, local undulation amplitude, and Gaussian curvature. The relative deviation values ​​of these three static contour features and their corresponding preset reference values ​​are calculated by relative deviation normalization for surface smoothness and local undulation amplitude in the static contour feature set. The Gaussian curvature in the static contour feature set is normalized by the region proportion normalization method.

[0069] It should be noted that the corresponding preset reference values ​​for surface smoothness and local undulation amplitude can be obtained by batch calibration method based on production statistics. Specifically, a batch (which may contain 100 pieces of fabric) of the same type as the fabric to be tested and which is identified as qualified in quality is continuously inspected. The average value of various static contour parameters of this batch of fabrics is calculated, and the average value is taken as the corresponding preset reference value for surface smoothness and local undulation amplitude in the fabric to be tested.

[0070] S511. Using relative deviation normalization, the surface smoothness and local undulation amplitude in the static contour feature set are normalized respectively to obtain the relative deviation values ​​corresponding to the surface smoothness and local undulation amplitude, including: The calculation process for the relative deviation of surface flatness is as follows: calculate the absolute value of the difference between the surface flatness and its corresponding preset reference value, and multiply the ratio of the absolute value of the difference to the surface flatness by 100% to obtain the relative deviation of surface flatness. The calculation process for the relative deviation of the local fluctuation amplitude is as follows: calculate the absolute value of the difference between the local fluctuation amplitude and its corresponding preset reference value, and multiply the ratio of the absolute value of the difference to the local fluctuation amplitude by 100% to obtain the relative deviation of the local fluctuation amplitude. It should be noted that the normalization mapping process makes the static contour feature parameters of different amplitudes comparable laterally.

[0071] S512. Introduce the region proportion normalization method to normalize the Gaussian curvature in the static contour feature set to obtain the area proportion of convex and concave shapes and fold saddle point shapes. If relative deviation normalization is used to normalize Gaussian curvature, it cannot handle the physical meaning of the sign and zero crossings. Therefore, this embodiment uses the region proportion normalization method to normalize the Gaussian curvature in the static contour feature set. This means converting the Gaussian curvature into the area proportion of convex and concave shapes and the area proportion of fold saddle points. Specifically: (1) Obtain the total area of ​​surface feature points with Gaussian curvature greater than 0, the total area of ​​surface feature points with Gaussian curvature less than 0, and the effective total area of ​​all surface feature points in the detection area of ​​the fabric to be tested; (2) Divide the total area of ​​surface feature points with Gaussian curvature greater than 0 in the detection area by the effective total area to obtain the area fraction of the convex and concave morphological region in the detection area. Multiply the area fraction by 100% to obtain the area ratio of the convex and concave morphology. (3) Divide the total area of ​​surface feature points with Gaussian curvature less than 0 in the detection area by the effective total area to obtain the area fraction of the folded saddle point region in the detection area. Multiply the area fraction by 100% to obtain the morphological area ratio of the folded saddle point. (4) The area ratio of convex and concave shapes and the area ratio of fold saddle points are used as the normalized output of Gaussian curvature. The normalized output of Gaussian curvature can be specifically expressed as follows: ,in The area ratio of the convex and concave shape. The area ratio of the folded saddle point shape is 1, and the sum of the area ratios of the convex and concave shapes and the area ratio of the folded saddle point shape is 1.

[0072] S513. The contour static feature vector is composed of the relative deviation value of surface flatness, the relative deviation value of local undulation amplitude, the convex and concave shape corresponding to Gaussian curvature, and the area ratio of the fold saddle point shape. The geometric static contour index dimension reflects the spatial morphological characteristics of the fabric surface through the contour static feature vector.

[0073] S52. Normalize the mapping based on the dynamic contour feature set to obtain the contour dynamic feature vector; The dynamic profile feature set includes spatial distribution map, strain concentration region, dynamic response consistency, and dynamic response recovery rate; S521. Obtain the standard deviation and arithmetic mean of the combined displacement amplitude of each target location point in the spatial distribution map. Use the ratio of the standard deviation to the arithmetic mean as the uniformity index of the displacement response amplitude. Use the range normalization method to obtain the deviation of the uniformity index of the displacement response amplitude. Specifically, obtain the preset maximum uniformity index and minimum uniformity index. Divide the difference between the uniformity index of the displacement response amplitude and the minimum uniformity index by the difference between the maximum uniformity index and the minimum uniformity index. Multiply the quotient by 100% to obtain the deviation of the uniformity index of the displacement response amplitude. The deviation of the uniformity index reflects the percentage of uniformity deterioration.

[0074] It should be noted that the maximum uniformity index is the upper limit threshold of the uniformity index allowed by the system in actual engineering applications. The maximum uniformity index can be determined based on historical data statistics, process standards or user presets. In this embodiment, the maximum uniformity index can be 1.0, and the minimum uniformity index is 0 under ideal conditions.

[0075] S522. Obtain the similarity between the displacement response curve of the target position point and the preset reference curve in dynamic response consistency. Multiply the difference between the similarity and the similarity reference value by 100% to obtain the deviation of dynamic response consistency. The similarity reference value is 1, which means that the displacement response curve is completely related to the preset reference curve. For example, if the similarity between the displacement response curve of the target position point and the preset reference curve in dynamic response consistency is 0.95 and the similarity reference value is 1, then the deviation of dynamic response consistency is 5%, indicating that the integrity of the contour structure is lost by 5%.

[0076] S523. Obtain the maximum value and arithmetic mean of the strain field within the strain concentration region. Use the ratio of the maximum value to the arithmetic mean as the strain concentration index of the strain concentration region. Multiply the difference between this strain concentration index and the benchmark strain concentration index by 100% to obtain the deviation of the strain concentration index corresponding to the strain concentration region. The benchmark strain concentration index is 1, which means that the strain is uniform throughout the field and there is no concentration. For example, if the strain concentration index of the strain concentration region is 2.5 and the benchmark strain concentration index is 1, then the deviation of the strain concentration index corresponding to the strain concentration region is 150%, which intuitively shows that the strain concentration index of the strain concentration region is 1.5 times higher than that of the uniform state.

[0077] S524. Multiply the difference between the dynamic response recovery rate and the reference response recovery rate by 100% to obtain the deviation of the dynamic response recovery rate. The reference response recovery rate is 1, which means that the fabric under test can fully rebound. For example, if the measured dynamic response recovery rate is 0.85 and the reference response recovery rate is 1, then the deviation of the dynamic response recovery rate is 15%, which quantifies the relative proportion of irreversible plastic deformation in the test area as 15%.

[0078] S525. Based on the deviation of the uniformity index of displacement response amplitude in the spatial distribution map, the deviation of the strain concentration index corresponding to the strain concentration area, the deviation of dynamic response consistency, and the deviation of dynamic response recovery rate, a contour dynamic feature vector is formed.

[0079] The contour dynamic feature vector reflects the deformation characteristics of each detection area on the surface of the fabric under applied excitation conditions.

[0080] S53. Combine the static contour feature vector and the dynamic contour feature vector to construct a multidimensional contour feature vector of the fabric to be detected. By integrating the static and dynamic features of the fabric surface profile into a multi-dimensional evaluation index system using multi-dimensional contour feature vectors, the system includes a geometric contour index dimension based on the static contour feature set to reflect the spatial morphological features of the fabric surface, and an index based on the dynamic contour feature set to reflect the deformation characteristics of each region of the fabric surface under applied excitation conditions.

[0081] Example 2, Figure 2 This invention presents a fabric surface contour detection system based on laser measurement, comprising: a fabric conveying module for placing the fabric to be inspected on a fabric conveying platform, wherein a laser contour scanning array and a laser speckle interferometer array are sequentially arranged along the conveying direction of the fabric conveying platform; a static contour feature module for scanning the two-dimensional contour data of the corresponding detection area of ​​the fabric to be inspected based on the laser contour scanning array, solving for the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extracting a static contour feature set based on the optimal three-dimensional coordinates; a dynamic contour feature module for applying dynamic excitation to the fabric to be inspected, obtaining the speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculating the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extracting a dynamic contour feature set based on the full-field displacement field and full-field strain field; and a multi-dimensional contour feature module for judging contour defects of the fabric to be inspected based on the static contour feature set and the dynamic contour feature set, and normalizing and mapping the static contour feature set and the dynamic contour feature set to construct a multi-dimensional contour feature vector of the fabric to be inspected.

[0082] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0083] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0084] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0085] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0086] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0087] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting fabric surface contours based on laser measurement, characterized in that, Includes the following steps: The fabric to be tested is placed on a fabric conveying platform, and a laser contour scanning array and a laser speckle interferometer array are arranged sequentially along the conveying direction of the fabric conveying platform. The laser contour scanning array scans the two-dimensional contour data of the detection area of ​​the fabric to be inspected, and the optimal three-dimensional coordinates of each surface feature point in the detection area are solved based on the two-dimensional contour data. The static contour feature set is then extracted based on the optimal three-dimensional coordinates. Dynamic excitation is applied to the fabric to be inspected. The speckle pattern sequence of the detection area is obtained based on the laser speckle interferometer array. The full-field displacement field and full-field strain field of the detection area are calculated based on the speckle pattern sequence. The dynamic contour feature set is extracted based on the full-field displacement field and full-field strain field. Based on the static contour feature set and the dynamic contour feature set, contour defects of the fabric to be inspected are identified, and the static contour feature set and the dynamic contour feature set are normalized and mapped to construct a multi-dimensional contour feature vector of the fabric to be inspected.

2. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, The fabric to be inspected is placed on the fabric conveyor platform, including: The fabric conveying platform is used to place the fabric to be inspected and to convey the fabric to be inspected according to a preset conveying direction and conveying speed. The laser profile scanning array consists of a preset number of laser profile sensors, which are arranged sequentially above the conveying direction of the fabric conveying platform based on the surface normal of the fabric to be detected. The laser speckle interferometer array includes a beam expander laser and a high-speed camera. The beam expander laser is used to form an interference speckle pattern on the surface of the fabric to be inspected, and the high-speed camera is used to continuously acquire the speckle pattern.

3. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, The optimal three-dimensional coordinates of each surface feature point within the detection area are determined based on two-dimensional contour data, including: The laser contour sensors of the laser contour scanning array emit linear lasers from their respective azimuth angles at the same time to scan and acquire two-dimensional contour data of the detection area. The two-dimensional contour data includes the lateral coordinates and height values ​​of each surface feature point. The spatial feature parameters of the laser contour sensor are obtained, and the coordinate system is transformed based on the spatial feature parameters to obtain the three-dimensional coordinates of each surface feature point in the detection area. The optimal three-dimensional coordinates of the surface feature points are obtained by solving the simultaneous three-dimensional coordinates of each surface feature point.

4. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, Static contour feature set extracted based on optimal 3D coordinates, including: The reference plane is obtained by fitting the optimal three-dimensional coordinates of each surface feature point in the detection area, and the surface flatness of the fabric to be tested is obtained based on the deviation value between each surface feature point and the reference plane. The average contour height of each surface feature point within the detection area is obtained, and the local undulation amplitude of the fabric to be detected is obtained based on the absolute value of the difference between the actual contour height of each surface feature point and the average contour height. Obtain the principal curvature of each surface feature point in the detection area, and calculate the Gaussian curvature of each surface feature point in the detection area based on the principal curvature. The static contour feature set of the fabric to be tested is composed of the surface flatness, local undulation amplitude, and Gaussian curvature.

5. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, A dynamic excitation is applied to the fabric to be inspected, and a speckle pattern sequence of the inspection area is obtained based on a laser speckle interferometer array, including: The initial speckle pattern on the surface of the fabric to be tested is acquired based on a laser speckle interferometer array. A dynamic excitation is applied to the fabric to be tested after the initial speckle pattern acquisition is completed, and the dynamic speckle pattern on the surface of the fabric to be tested during the dynamic excitation process is acquired. A speckle pattern sequence is formed based on the dynamic speckle pattern and the initial speckle pattern.

6. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, The full-field displacement field and full-field strain field of the detection area are calculated based on the speckle pattern sequence, including: A reference window is constructed by selecting the initial position points of the initial speckle pattern, and candidate windows in the dynamic speckle pattern are retrieved based on the reference window. Calculate the regional similarity between the candidate window and the reference window, select the target position point in each dynamic speckle pattern based on the regional similarity, and obtain the displacement vector of the target position point based on the offset between the target position point and the initial position point; The full-field displacement field of the fabric surface to be tested is formed by the displacement vector of the target location point, and the full-field strain field of the fabric surface to be tested is obtained by spatial differentiation of the full-field displacement field.

7. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, Dynamic contour feature sets are extracted based on the full-field displacement field and the full-field strain field, including: Extract the peak displacement vector of each target location point in the full displacement field over time, and form a spatial distribution map based on the peak displacement vector; The average strain and strain standard deviation of the whole field are calculated based on the strain vector of each surface feature point in the whole field strain field, and the strain concentration area is determined based on the average strain and strain standard deviation of the whole field. Based on the displacement vector of the target location point in the detection area at each time, a displacement response curve of the target location point is generated. Based on the displacement response curve and the preset reference curve, the dynamic response consistency of the fabric to be tested is obtained. The peak value and mean value of the displacement vector at each target location are obtained based on the displacement response curve, and the dynamic response recovery rate of the fabric to be tested is obtained based on the peak value and mean value of the displacement vector. The dynamic contour feature set of the fabric to be tested is constructed based on the spatial distribution map, strain concentration area, dynamic response consistency, and dynamic response recovery rate.

8. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, Delineation defect discrimination of the fabric to be inspected is based on static and dynamic contour feature sets, including: The contour features in both the static and dynamic contour feature sets are binarized and mapped to obtain the contour defect type of the fabric to be detected. The system obtains the judgment time and location of the contour defect type, and the remaining distance from the defect judgment location to the preset defect handling station. Based on the remaining distance, judgment time, and the conveying speed of the fabric conveying platform, the system calculates the estimated arrival time of the fabric to be inspected to the defect handling station. The defect handling station marks the fabric to be inspected with contour defects based on the expected arrival time.

9. The fabric surface contour detection method based on laser measurement according to claim 1, characterized in that, Normalize and map the static and dynamic contour feature sets to construct a multidimensional contour feature vector for the fabric to be detected, including: Normalize and map the static contour feature set to obtain the static contour feature vector; Normalize the dynamic contour feature set to obtain the dynamic contour feature vector; A multidimensional contour feature vector of the fabric to be detected is constructed by combining the static and dynamic contour feature vectors.

10. A system for detecting fabric surface contours based on laser measurement as described in any one of claims 1-9, comprising: A fabric conveying module is used to place the fabric to be inspected on a fabric conveying platform, wherein a laser contour scanning array and a laser speckle interference array are arranged sequentially along the conveying direction of the fabric conveying platform. The static contour feature module is used to scan the two-dimensional contour data of the detection area of ​​the fabric to be inspected based on the laser contour scanning array, solve the optimal three-dimensional coordinates of each surface feature point in the detection area based on the two-dimensional contour data, and extract the static contour feature set based on the optimal three-dimensional coordinates. The dynamic contour feature module is used to apply dynamic excitation to the fabric to be inspected, obtain the speckle pattern sequence of the detection area based on the laser speckle interferometer array, calculate the full-field displacement field and full-field strain field of the detection area based on the speckle pattern sequence, and extract the dynamic contour feature set based on the full-field displacement field and full-field strain field. The multidimensional contour feature module is used to identify contour defects in the fabric to be inspected based on the static contour feature set and the dynamic contour feature set, and to normalize and map the static contour feature set and the dynamic contour feature set to construct the multidimensional contour feature vector of the fabric to be inspected.

Images