A method and system for detecting the glaze thickness of porcelain insulator heads

By using a dual-axis rotating platform with a Z-axis height coarse adjustment mechanism and three-dimensional line laser scanning technology in the glaze thickness measurement of porcelain insulator heads, combined with CAD model fine registration and normal vector screening, the problems of low detection accuracy and poor efficiency in existing technologies have been solved, and efficient and accurate measurement of glaze thickness of porcelain insulator heads has been achieved.

CN123111153APending Publication Date: 2026-11-17BAODING TANGBEI ELECTRIC PORCELAIN ELEC APPLIANCE CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611301275.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-08-26
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies for detecting the glaze thickness of porcelain insulator heads suffer from low accuracy and poor efficiency. In particular, it is difficult to achieve efficient and accurate measurement in cases of complex curved surfaces and local micro-defects. Furthermore, multi-axis robotic arms have high hardware costs and complex calibration, which cannot meet the needs of large-scale, fast-paced quality inspection.

Method used

A dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism is used, combined with three-dimensional line laser scanning and CAD model fine registration. Candidate measurement points are screened by normal vectors, and rotational analytical values ​​are obtained by using an arctangent function with quadrant judgment. This ensures that the fixed measurement direction of the thickness gauge is aligned with the normal of the porcelain insulator surface, thereby achieving the selection of the optimal measurement point and the acquisition of thickness data.

Benefits of technology

It achieves high-precision and high-efficiency detection of the glaze thickness at the head of porcelain insulators, reduces hardware costs, simplifies calibration complexity, improves detection success rate and data reliability, and provides high-confidence quality inspection data support.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123111153A_ABST
    Figure CN123111153A_ABST
Patent Text Reader

Abstract

This invention relates to the field of thickness measurement technology, specifically to a method and system for measuring the glaze thickness of the head of a porcelain insulator. The method involves clamping an insulator on a dual-axis rotating platform equipped with a Z-axis height coarse adjustment mechanism and acquiring a point cloud to be measured. This cloud is then precisely registered with a standard CAD model to extract the head's measurement point set, and candidate measurement points are selected based on curvature and normal vectors. A fixed measurement direction is defined for the thickness gauge, and the rotational analytical value of the dual-axis rotating platform is obtained using an arctangent function with quadrant judgment combined with alignment constraints. Z-axis coarse adjustment ensures the head's rotation radius falls within the working distance range. Based on the rotational analytical value, the coordinate offset of the candidate measurement points after rotation is determined to obtain the optimal measurement point. Finally, the head glaze thickness is obtained based on the optimal measurement point and the rotational analytical value. This invention solves the problems of accurate alignment of curved surface normals and the high cost and complex calibration of existing robotic arms, achieving high-precision and high-efficiency automated fixed-point thickness measurement.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of thickness detection technology, specifically to a method and system for detecting the glaze thickness of the head of a porcelain insulator. Background Technology

[0002] Porcelain insulators are core insulation components of power transmission and transformation lines, widely used in overhead transmission lines, substations, and other power infrastructure applications. The glaze layer at the insulator's head serves as a crucial protective barrier, and its thickness directly affects the insulator's anti-flashover performance, aging resistance, and mechanical strength. If the glaze layer thickness is substandard, the insulator's insulation performance will significantly decrease, easily leading to serious power grid safety accidents such as flashover and breakdown under harsh operating conditions. Therefore, high-precision measurement of the glaze layer thickness is an indispensable and critical step in the factory inspection and routine maintenance sampling of porcelain insulators. In recent years, with the rapid expansion of the power grid, the demand for porcelain insulator testing has surged year by year, creating an urgent need for automated and efficient testing technologies.

[0003] Currently, the thickness measurement of the glaze layer at the head of porcelain insulators largely relies on six-axis or multi-axis industrial robotic arms to hold a thickness gauge and dynamically adjust its measurement posture. However, the hardware cost of multi-axis robotic arms and their associated flexible control systems is extremely high, hindering large-scale deployment. Secondly, the hand-eye calibration and spatial coordinate calibration process between the robotic arm's end effector and the thickness gauge is extremely complex, easily leading to accumulated errors during long-term operation in industrial settings. More importantly, existing automation solutions only provide simple trajectory tracking or coarse local plane fitting, failing to design a deep-level, precise normal matching algorithm and an optimal dynamic selection mechanism for measuring points specifically tailored to the complex curved surface features of porcelain insulator heads. This results in the robotic arm frequently changing posture when facing abrupt curvature changes or local surface defects, leading to low detection efficiency. Furthermore, it is highly susceptible to acquiring abnormal data or even causing the detection process to stall, making it difficult to balance detection efficiency and accuracy, and failing to meet the current demands for large-scale, fast-paced, and high-standard automated quality inspection across multiple scenarios. Summary of the Invention

[0004] To address the problems of low accuracy and poor efficiency in existing methods for detecting the glaze thickness at the head of porcelain insulators, the present invention aims to provide a method and system for detecting the glaze thickness at the head of porcelain insulators. The specific technical solution adopted is as follows: In a first aspect, one embodiment of the present invention provides a method for detecting the glaze layer thickness of a porcelain insulator head, the method comprising the following steps: Porcelain insulators are mounted on a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism. The test point cloud of the porcelain insulators is collected and precisely registered with its standard CAD model. The head test point set is extracted based on the geometric attributes of the standard CAD model. Candidate measurement points are selected based on the curvature and normal vector of each point in the head test point set. Define the fixed measurement direction of the thickness gauge, and use the arctangent function with quadrant judgment to obtain the rotational analytical value of the dual-axis rotary platform based on the alignment constraint between the normal vector of the candidate measurement point and the fixed measurement direction. The Z-axis height coarse adjustment mechanism is used to make the radius of rotation of the head surface of the porcelain insulator fall into the working distance range of the thickness gauge. Based on the rotational analysis value, the Z-axis coordinate and XY coordinate offset of each candidate measurement point after rotation is determined, and the optimal measurement point is obtained. The thickness of the glaze layer at the head of the porcelain insulator is obtained based on the optimal measurement point and the corresponding rotational analytical value.

[0005] Furthermore, the method for obtaining the head test point set is as follows: The range corresponding to the upper and lower limits of the Z-axis height of the head region in the standard CAD model of the porcelain insulator is obtained as the head height range; By using Z-axis pass-through filtering, the set of points in the standard point cloud after discretizing the standard CAD model that are within the head height range is taken as the standard head region point set; Based on the KD-tree nearest neighbor search algorithm, in the point cloud to be tested after fine registration with the standard CAD model, a point set that matches the standard head region point set is found and used as the head test point set.

[0006] Furthermore, the method for obtaining the candidate measurement points is as follows: Based on the normal vectors of each point in the head test point set, calculate the neighborhood variance of the normal vector of each point in the preset neighborhood. The candidate measurement point set consists of points in the head measurement point set whose curvature is less than or equal to a preset curvature threshold and whose neighborhood variance of the normal vector is less than or equal to a preset normal vector variance threshold.

[0007] Furthermore, the method for obtaining the rotation resolution value is as follows: The opposite direction of the fixed measurement direction of the thickness gauge is defined as the target normal vector; a dual-axis rotation matrix is ​​constructed based on the rotational motion of the dual-axis rotation platform around the X and Y axes; Obtain the directional components of the normal vector of each candidate measurement point along the X, Y, and Z axes; Based on the alignment constraint that the normal vector of each candidate measurement point is equal to the target normal vector after transformation by the dual-axis rotation matrix, the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotation platform around the X-axis corresponding to each candidate measurement point; the negative value of the direction component of the X-axis, the square root of the sum of the squares of the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotation platform around the Y-axis corresponding to each candidate measurement point.

[0008] Furthermore, the method for obtaining the optimal measurement point is as follows: Based on the rotation analysis value corresponding to each candidate measurement point, obtain the spatial coordinates of each candidate measurement point after rotation; Candidate measurement points whose rotated Z-axis coordinates fall within the working distance range of the thickness gauge, and whose absolute values ​​of the rotated X-axis and Y-axis are both less than or equal to the preset spot radius, are selected as valid measurement points. When the number of valid measurement points is less than the preset threshold, a system alarm is triggered. When the number of valid measurement points is greater than or equal to the preset number threshold, the sum of the absolute values ​​of the X and Y coordinates of each valid measurement point after rotation is calculated and used as the XY coordinate offset value of each valid measurement point. The valid measurement points are sorted in ascending order of their XY coordinate offset values ​​to obtain a sequence of valid measurement points. The first specified number of valid measurement points in the valid measurement point sequence are selected as the optimal measurement points.

[0009] Furthermore, the method for obtaining the spatial coordinates of the candidate measurement point after rotation is as follows: For any candidate measurement point, the initial three-dimensional coordinates of the candidate measurement point are used as the initial coordinate column vector; Multiply the biaxial rotation matrix determined by the rotation analysis value corresponding to the candidate measurement point with the initial coordinate column vector, and then add it with the preset translation offset vector to obtain the rotated target coordinate column vector; The X-axis, Y-axis, and Z-axis coordinates are extracted from the target coordinate column vector and used as the spatial coordinates of the candidate measurement point after rotation.

[0010] Furthermore, the method for obtaining the thickness of the glaze layer on the head is as follows: For any optimal measurement point, the dual-axis rotary platform is driven to rotate to the target pose by the rotation resolution value corresponding to the optimal measurement point, and the thickness gauge is controlled to perform a preset number of thickness measurements on the optimal measurement point. The range of thickness measurements is taken as the measurement change value of the optimal measurement point; if the measurement change value is less than or equal to the preset change threshold, the average value of the thickness measurements is taken as the head glaze thickness corresponding to the optimal measurement point. If the measured change value is greater than the preset change threshold, then a valid measurement point is selected according to the preset conditions and used as a candidate measurement point for the optimal measurement point. When the number of candidate measurement points exceeds the reference threshold, the thickness of each candidate measurement point is remeasured. If the measurement change value of the current candidate measurement point is less than or equal to the preset change threshold, the average thickness measurement value of the current candidate measurement points is taken as the head glaze thickness corresponding to the optimal measurement point and the remeasurement is stopped. If the measurement change values ​​of all candidate measurement points are greater than the preset change threshold, the optimal measurement point is marked as the failure to obtain the head glaze thickness. When the number of candidate measurement points is less than or equal to the reference number threshold, the optimal measurement point is marked as a failure to obtain the head glaze thickness.

[0011] Furthermore, the preset condition is that the spatial distance between the effective measurement point and the optimal measurement point is less than or equal to a preset distance threshold.

[0012] Furthermore, the fixed measurement direction of the thickness gauge is the negative Z-axis direction, which is vertically downward.

[0013] Secondly, another embodiment of the present invention provides a porcelain insulator head glaze thickness detection system, the system comprising: a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of any of the above methods.

[0014] The present invention has the following beneficial effects: This invention first clamps a porcelain insulator onto a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism. By providing a physical displacement compensation margin in the Z-axis direction, it ensures from the physical source that the insulator can match the optimal working space of the fixed thickness gauge. Furthermore, by using a dual-axis platform instead of a traditional multi-axis robotic arm, it significantly reduces hardware costs and calibration complexity. The collected point cloud of the porcelain insulator to be measured is precisely registered with its standard CAD model, effectively eliminating coordinate system offsets caused by clamping deviations. Then, based on the geometric attributes of the standard CAD model, the head set of points to be measured is extracted, realizing the core measurement. The precise positioning of the region effectively concentrated the system's computing power. To ensure the accuracy of subsequent normal alignment and the stability of the thickness gauge's measurement signal, candidate measurement points were screened based on the curvature and normal vectors of each point in the head-side measurement point set, precisely eliminating inferior points unsuitable for measurement, such as those with large edge curvature and local surface defects. To ensure that the measurement points on the porcelain insulator surface could accurately align with the thickness gauge's spot in a vertical orientation, a fixed measurement direction for the thickness gauge was defined. Using an arctangent function with quadrant judgment, the alignment constraints between the normal vectors of the candidate measurement points and the fixed measurement direction were used to obtain the double... The rotation resolution value of the axis rotation platform effectively avoids the deadlock risk of the calculation denominator being zero due to the normal vector component approaching zero, ensuring that the spatial alignment command can be stably and uniquely calculated. In order to ensure that the measurement point after rotation can be completely covered in the real physical space and at the optimal detection depth, the Z-axis height coarse adjustment mechanism is used to make the radius of rotation of the head surface of the porcelain insulator fall into the working distance range of the thickness gauge. Based on the rotation resolution value, the Z-axis coordinate and XY coordinate offset of each candidate measurement point after rotation are determined. The alignment accuracy and relative detection distance of each candidate measurement point with the center of the thickness gauge spot after rotation are analyzed, so as to accurately obtain the optimal measurement point and completely eliminate invalid points that are in the detection blind zone or deviate from the spot, even if the normal is aligned, thus maximizing the detection success rate. In order to obtain high-confidence thickness data and realize quality inspection closed loop, the thickness of the glaze layer at the head of the porcelain insulator is obtained based on the optimal measurement point and the corresponding rotation resolution value. This not only effectively reduces the random interference of local minor physical defects, but also provides high-precision, high-efficiency and highly automated reliable data support for the factory inspection and quality control of power grid insulation components. Attached Figure Description

[0015] Figure 1 A schematic flowchart illustrating a method for detecting the glaze layer thickness at the head of a porcelain insulator according to an embodiment of the present invention; Figure 2 This is a structural diagram of a porcelain insulator head glaze thickness detection system provided in one embodiment of the present invention. Detailed Implementation

[0016] The following description, in conjunction with the accompanying drawings, details the specific scheme of the method and system for detecting the glaze layer thickness of a porcelain insulator head provided by the present invention.

[0017] Example 1: This invention proposes a method for detecting the glaze layer thickness at the head of a porcelain insulator. Please refer to [link / reference]. Figure 1 The diagram illustrates a schematic flowchart of a method for detecting the glaze thickness of a porcelain insulator head according to an embodiment of the present invention. The method includes the following steps: Step S1: Clamp the porcelain insulator on a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism, collect the test point cloud of the porcelain insulator, finely register it with its standard CAD model, extract the head test point set based on the geometric attributes of the standard CAD model, and screen candidate measurement points based on the curvature and normal vector of each point in the head test point set.

[0018] Specifically, this embodiment uses a porcelain insulator as an example for analysis; all subsequent references to the porcelain insulator refer to this porcelain insulator. It is known that the ultrasonic thickness gauge, fixedly mounted directly above the dual-axis rotating platform, typically measures in a fixed vertically downward direction, and its measuring spot diameter is 2mm (effective measuring range is relatively small). To precisely adjust the surface normal of any measuring point on the porcelain insulator head to coincide with the thickness gauge's measuring direction, the dual-axis rotating platform is set as a precision bearing platform capable of independently rotating around the X and Y axes, providing sufficient degrees of freedom for attitude adjustment. Simultaneously, considering that the dual-axis rotating platform only provides pure rotational motion kinematically, and the rotation process does not change the absolute distance from the porcelain insulator surface point to the platform's rotation center, if the radius of rotation after the porcelain insulator is clamped exceeds the effective working range of the fixed thickness gauge... If the distance range is too wide, the thickness gauge will not function properly. Therefore, a Z-axis height coarse adjustment mechanism is configured for the dual-axis rotary platform to provide a physical displacement compensation margin in the Z-axis direction. This ensures that porcelain insulators of different models or sizes can effectively fall into the optimal measuring distance range of the thickness gauge during rotation testing. Then, the porcelain insulator is clamped on the dual-axis rotary platform equipped with the Z-axis height coarse adjustment mechanism, and the head center of the porcelain insulator is approximately placed at the rotation center of the dual-axis rotary platform by adjusting the clamps. This ensures from a physical source that the normals of enough curved points can pass near the rotation center after rotation, which is beneficial for the measurement point after rotation to accurately fall within the 2mm diameter detection spot range of the thickness gauge.

[0019] To obtain three-dimensional morphological data of the porcelain insulator surface, a three-dimensional line laser scanner was used to acquire the test point cloud of the porcelain insulator. During the acquisition process, the dual-axis rotating platform remained stationary to ensure the integrity of the point cloud. To filter out platform background noise and accurately identify specific structural parts of the porcelain insulator, a standard CAD model corresponding to the model of the porcelain insulator to be tested was imported. The least squares method was used to precisely register the test point cloud with the standard CAD model, eliminating coordinate system offsets caused by clamping deviations and improving the physical accuracy of subsequent spatial coordinate calculations of measurement points and platform rotation control. To narrow the data processing scope and eliminate the skirt and steel foot parts that do not need to be measured, the head test point set was extracted based on the geometric properties of the standard CAD model to achieve precise positioning of the head region of the porcelain insulator. This facilitates the centralized computing power of the system to perform high-quality surface feature analysis of the core test area. Considering that the head surface may have local defects or areas with large curvature at the edges, in order to ensure the accuracy of subsequent normal alignment and the stability of the thickness gauge measurement signal, candidate measurement points were screened based on the curvature and normal vector of each point in the head test point set, and inferior points unsuitable for measurement were eliminated. The least squares method is well-known and will not be elaborated further.

[0020] Preferably, in one feasible embodiment, the method for obtaining the head test point set is as follows: First, by parsing the geometric attribute table of the imported porcelain insulator standard CAD model, the range corresponding to the upper and lower limits of the Z-axis height of the head region in the standard CAD model of the porcelain insulator is obtained as the head height range, providing a clear physical boundary basis for subsequent point cloud segmentation; then, by Z-axis pass-through filtering, the standard point cloud after the standard CAD model is discretized (i.e., the data obtained by sampling the surface point cloud of the continuous CAD surface model according to a preset sampling rate, for example, setting the preset sampling rate to be...) A sampling rate of 20 points / mm is used to fully represent the detailed features of the curved surface of the porcelain insulator head. The preset sampling rate can be set according to actual conditions and is not limited here. A set of points within the head height range is used as the standard head region point set, filtering out redundant point clouds belonging to the porcelain insulator skirts and steel feet. Finally, to accurately extract the true head structure from the actual collected test point cloud, a KD-tree nearest neighbor search algorithm is used to find a point set matching the standard head region point set in the test point cloud after precise registration with the standard CAD model. This set is then used as the head test point set, achieving automated segmentation and extraction of the head point cloud. The pass-through filtering and KD-tree nearest neighbor search algorithms are well-known and will not be elaborated further.

[0021] Preferably, in one feasible method of this embodiment, the method for obtaining candidate measurement points is as follows: Considering that the original point cloud to be measured may have local environmental noise during acquisition, leading to distortion in the calculation of local normal vectors, in order to evaluate the confidence level of the normal vector calculation, the variance of the normal vector neighborhood of each point in the head test point set is calculated based on the normal vector of each point (i.e., for any point, the normal vectors of all nearest neighbor points in the preset neighborhood of that point are extracted, and the variances of their X, Y, and Z components are calculated and then the mean is taken to characterize the dispersion of the normal vector in the local area of ​​that point). In this embodiment, the preset neighborhood is set as a spherical region with a radius of 2mm centered on each point, ensuring that the number of nearest neighbor points participating in the variance calculation is sufficient to reflect the true local surface features. Simultaneously, it matches the 2mm spot size of the ultrasonic thickness gauge. The implementer can set a preset neighborhood based on the point cloud density and the measurement spot size, which is not limited here. Among them, the smaller the variance of the normal vector neighborhood, the smoother the local neighborhood of the corresponding point, and the more accurate the normal vector calculation result of the corresponding point. In addition, considering that the curvature of the edge or severely scratched area of ​​the porcelain insulator is large, it is easy to cause ultrasonic scattering, resulting in extremely unstable thickness measurement signal. In order to eliminate poor measurement points that cannot provide stable echo signals, the head of the test points is concentrated on points with curvature less than or equal to a preset curvature threshold and normal vector neighborhood variance less than or equal to a preset normal vector variance threshold, forming a candidate measurement point set to ensure that the selected points have both normal reliability and surface smoothness. In this embodiment, the preset curvature threshold is set to be... Ensure that you avoid the transition edges of the umbrella skirt and areas with severe defects (normal head curvature is usually within...). - The implementer can set a preset curvature threshold according to the standard curvature parameters of different types of porcelain insulators, which is not limited here; set the preset normal vector variance threshold to 0.05 to ensure that the preset neighborhood of the selected point has high consistency and avoid errors in the subsequent rotation matrix solution caused by normal vector deviation. The implementer can set the preset normal vector variance threshold according to the acquisition accuracy of the scanner and the point cloud density, which is not limited here.

[0022] Step S2: Define the fixed measurement direction of the thickness gauge, and use the arctangent function with quadrant judgment to obtain the rotational analytical value of the dual-axis rotary platform based on the alignment constraint between the normal vector of the candidate measurement point and the fixed measurement direction.

[0023] Specifically, considering that ultrasonic thickness gauges are usually fixedly installed in automated inspection systems to ensure the stability of the coordinate reference, in order to ensure that the test point on the curved surface of the porcelain insulator can align with the spot of the thickness gauge in a vertical orientation, the fixed measurement direction of the thickness gauge, namely the negative Z-axis direction vertically downward, is defined as a fixed reference reference in the spatial coordinate system. Then, using the arctangent function with quadrant judgment, the potential program crash caused by the calculation denominator being zero due to the Z component of the normal vector approaching zero is avoided. Based on the alignment constraint between the normal vector of the candidate measurement point and the fixed measurement direction, the rotational analytical value of the dual-axis rotary platform is obtained, ensuring that the corresponding dual-axis motor motion angle command can be stably and uniquely calculated.

[0024] Preferably, in one feasible method of this embodiment, the method for obtaining the rotation resolution value is as follows: Considering that the thickness gauge emits ultrasonic signals vertically from top to bottom, in order to clarify the expected spatial attitude target of the point to be measured after rotation, the opposite direction of the fixed measurement direction of the thickness gauge (i.e., the positive direction of the vertically upward Z-axis) is defined as the target normal vector. That is, the normal vector of the point to be measured after rotation must be completely vertically upward to meet the thickness gauge, which is beneficial to directly construct the equality constraint between the initial normal vector and the target state. Considering that the rotation of the porcelain insulator with the dual-axis rotating platform is a rigid body transformation process, in order to establish a rigorous mathematical relationship between the attitude changes of the spatial surface points before and after rotation, a dual-axis rotation matrix is ​​constructed based on the rotational motion of the dual-axis rotating platform around the X-axis and Y-axis. That is, the pitch matrix of the dual-axis rotating platform rotating around the Y-axis is multiplied by the roll matrix of the dual-axis rotating platform rotating around the X-axis (i.e., the physical rotation order of the system is strictly limited to first rotating around the X-axis and then rotating around the Y-axis), which accurately reflects the global spatial attitude transformation relationship under the dual-axis combined motion; the methods for obtaining the roll matrix and pitch matrix are well known and will not be described in detail. To apply the aforementioned matrix transformation to point cloud data and obtain the directional components of the normal vector of each candidate measurement point along the X, Y, and Z axes, i.e., to extract and normalize the three-dimensional normal vector coordinates of the candidate measurement points. The normalization method here is as follows: divide the X-axis, Y-axis, and Z-axis components of the original normal vector by the magnitude of the original normal vector (i.e., the square root of the sum of the squares of the three direction components) to obtain a unit normal vector with the same direction but a length of 1. Considering the uniqueness of the inverse kinematics solution of the dual-axis rotating platform, in order to prevent the singularity deadlock problem caused by the divisor in the conventional trigonometric function approaching zero (e.g., when the normal of the measurement point is almost horizontal) during the solution process, based on the alignment constraint that the normal vector of each candidate measurement point is equal to the target normal vector after transformation by the dual-axis rotation matrix, the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotating platform around the X-axis corresponding to each candidate measurement point; the negative value of the direction component of the X-axis, and the square root of the sum of the squares of the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotating platform around the Y-axis corresponding to each candidate measurement point. The specific implementation process is as follows: The target normal vector is set as... Establish spatial alignment constraint equations Where R is the constructed biaxial rotation matrix; Let R be the transpose column vector of the initial normal vector of the candidate measurement points. It should be noted that R represents the pure angle transformation. Orthogonal rotation matrices are used to maintain the matching of vector dimension operations and ensure that the rotation of the normal vector is not affected by spatial translation. Furthermore, during the establishment of spatial alignment constraint equations, [the following is used]: Convert to The direction column vector, making it directly intersect with... Linear multiplication of the biaxial rotation matrices is performed without introducing homogeneous terms for translation compensation. To solve the aforementioned spatial alignment constraint equations, the equations are expanded and the rotation variables of each axis are separated. The analytical value of the biaxial rotation platform's rotation around the X-axis is... ;in, The arctangent function with quadrant determination can be determined according to... and The symbol automatically determines the correct quadrant position, effectively avoiding... The risk of division by zero; the analytical value of the rotation of the dual-axis rotary platform around the Y-axis. Similarly, it employs an arctangent function with quadrant-based judgment, effectively avoiding the situation where... and This also includes the division-by-zero collapse that occurs when the normal vector is 0 (i.e., completely horizontal). Therefore, the analytical equations described above can stably and efficiently determine the rotation angle required to accurately align the normal vector of the measured point with the thickness gauge. It should be noted that the analytical equations are derived based on a specific mechanical configuration of the dual-axis rotary platform, first rotating around the X-axis and then around the Y-axis. In practical industrial applications, if the rotation order of the physical structure of the dual-axis rotary platform is reversed, or if the theoretical extreme pose is caused by the completely horizontal normal vector, the dual-axis rotation matrix can be reconstructed using quaternion representation (a well-known method, not elaborated here) and the attitude inverse solution can be performed to ensure the continuity and deadlock-free nature of the spatial pose solution.

[0025] Step S3: The Z-axis height coarse adjustment mechanism is used to make the radius of rotation of the head surface of the porcelain insulator fall into the working distance range of the thickness gauge. Based on the rotation resolution value, the Z-axis coordinate and XY coordinate offset of each candidate measurement point after rotation are determined, and the optimal measurement point is obtained.

[0026] Specifically, considering that the dual-axis rotating platform only provides angle transformation without changing the absolute spatial distance from the surface point of the porcelain insulator to the rotation center of the platform, in order to ensure that the measured point after rotation can meet the distance measurement requirements of the fixedly installed ultrasonic thickness gauge in terms of vertical height, a Z-axis height coarse adjustment mechanism is used to make the radius of rotation of the curved surface of the porcelain insulator head fall into the working distance range of the thickness gauge. This avoids the situation where the thickness gauge cannot obtain an effective echo signal due to improper clamping height, ensuring the feasibility of subsequent measurements. Then, based on the rotation resolution value, the Z-axis coordinate and XY coordinate offset of each candidate measured point after rotation are determined. The alignment accuracy and relative detection distance of each candidate measured point with the center of the thickness gauge spot after rotation are analyzed to accurately obtain the optimal measured point and maximize the detection success rate and detection accuracy.

[0027] Preferably, in one feasible embodiment of this method, the optimal measurement point is obtained as follows: to determine the actual physical position of each candidate measurement point after the platform executes the corresponding rotation command, and then based on the rotation resolution value corresponding to each candidate measurement point, the spatial coordinates of each candidate measurement point after rotation are obtained, providing coordinate data support for subsequent field-of-view constraint judgment; wherein, the specific method for obtaining the spatial coordinates of the candidate measurement point after rotation is as follows: for any candidate measurement point, the initial three-dimensional coordinates of the candidate measurement point are used as the initial coordinate column vector, for example, the extracted three-dimensional coordinates are used as the initial coordinate column vector. Convert to The spatial coordinate column vector is used to efficiently perform rigid body transformation calculations of spatial coordinates using linear algebra methods and to compensate for the spatial offset between the origin of the standard CAD model and the absolute physical reference system of the thickness gauge. The biaxial rotation matrix determined by the rotation analysis value corresponding to the candidate measurement point is multiplied by the initial coordinate column vector and then added to a preset translation offset vector to obtain the rotated target coordinate column vector, achieving accurate mapping from the platform's initial coordinate system to the rotated attitude coordinate system. In this embodiment, the preset translation offset vector is set as a three-dimensional column vector containing the coarse adjustment displacement of the Z-axis height coarse adjustment mechanism, used to align the actual physical system's translation displacement in the Z-axis direction, ensuring that the rotated target coordinate column vector accurately reflects the physical position of the measurement point in the absolute detection coordinate system of the thickness gauge. To separate the specific dimensional values ​​that can be used for distance measurement and spot constraint judgment from the matrix operation results, the X-axis, Y-axis, and Z-axis coordinate values ​​are extracted from the target coordinate column vector as the rotated spatial coordinates of the candidate measurement point. Considering the upper and lower limits of the optimal working distance of the thickness gauge and the fact that the diameter of the ultrasonic measurement spot emitted by it is usually extremely small, in order to ensure absolute physical safety during the rotation process, a spatial bounding box is constructed using a standard CAD model or point cloud of the porcelain insulator. The corresponding rotation resolution value is extracted and falls within the physical limit range of the dual-axis rotation platform. The bounding box does not physically interfere with the thickness gauge when performing rotation sweep based on the rotation resolution value. At the same time, the Z-axis coordinate after rotation falls within the working distance range of the thickness gauge, and the absolute values ​​of the X and Y coordinates after rotation are both less than or equal to the preset spot radius. Candidate measurement points are considered as valid measurement points, and invalid points that, although the surface normals can be aligned, are physically located in the thickness gauge's detection blind zone or deviate from the spot coverage range are completely eliminated. In this embodiment, the preset spot radius is set to 1 mm by matching the physical parameters of the actual ultrasonic thickness gauge probe to ensure that the measurement point after rotation can accurately fall within the detection spot range of the 2 mm diameter thickness gauge without missing the target. The implementer can set the preset spot radius according to the specific ultrasonic thickness gauge probe specifications selected, which is not limited here. Considering the stability requirements of batch automated testing, and to avoid low error tolerance or unrepresentativeness in subsequent measurements due to insufficient available measurement points, a preset threshold of 5 is set. This ensures sufficient high-quality alternative points on the porcelain insulator surface to support multi-point comprehensive evaluation and retesting mechanisms. Implementers can set the preset threshold according to the measurement point density and insulator size required by the testing specifications; no specific limitation is imposed here. When the number of effective measurement points is less than the preset threshold, it indicates a serious deviation in the current porcelain insulator clamping or extremely poor overall surface quality, failing to provide sufficient effective measurement area. To prevent the testing system from falling into an invalid measurement loop, a system alarm is triggered, allowing for manual intervention to adjust the tooling or directly reject defective products. When the number of effective measurement points is large... When the number of valid measurement points is equal to or greater than a preset threshold, it indicates that there are sufficient compliant measurement points in the current state. To select the optimal measurement point from among many compliant points, the sum of the absolute values ​​of the X and Y coordinates of each valid measurement point after rotation is calculated as the XY coordinate offset value of each valid measurement point. This accurately reflects the Manhattan distance of each valid measurement point from the theoretical central axis of the thickness gauge on the horizontal plane after rotation. The smaller the XY coordinate offset value, the higher the physical alignment accuracy of the corresponding valid measurement point. To quickly extract the target set with the best alignment with the thickness gauge, the valid measurement points are sorted in ascending order of their XY coordinate offset values ​​to obtain a sequence of valid measurement points. Then, the first specified number of valid measurement points in the sequence are taken as the optimal measurement points. In this embodiment, the specified number is set to 5 to ensure detection efficiency while fully reflecting the overall glaze thickness distribution of the porcelain insulator head. Implementers can set the specified number according to the requirements of the power industry quality inspection standards for the number of sampling inspection points for porcelain insulators, which is not limited here.

[0028] Step S4: Based on the optimal measurement point and the corresponding rotational analytical value, obtain the thickness of the glaze layer at the head of the porcelain insulator.

[0029] Specifically, considering that ultrasonic thickness measurement is easily affected by random interference from minor local physical defects on the surface of porcelain insulators during actual automated detection, in order to obtain high-confidence head glaze thickness data, and then obtain the head glaze thickness of porcelain insulators based on the optimal measurement point and the corresponding rotation resolution value, high-precision non-destructive fixed-point thickness measurement can be achieved, providing reliable data support for the final product qualification judgment.

[0030] Preferably, in one feasible method of this embodiment, the method for obtaining the thickness of the glaze layer on the head is as follows: for any optimal measurement point, the dual-axis rotation platform is driven to rotate to the target pose (i.e., the spatial pose where the normal of the optimal measurement point is vertically upward and exactly within the effective spot of the thickness gauge) by the rotation resolution value corresponding to the optimal measurement point, and the thickness gauge is controlled to perform a preset number of thickness measurements on the optimal measurement point to ensure that the single-point measurement result has statistical significance and to reduce the interference of environmental vibration or random instrument error; in this embodiment, the preset number is set to 5 to ensure that the sample size is sufficient to reflect the short-term fluctuation of the signal. The implementer can set the preset number according to the sampling frequency of the thickness gauge and the overall detection efficiency requirements, which is not limited here. To quantitatively evaluate the stability of the thickness measurement signal in the local area of ​​the optimal measurement point, the range of multiple thickness measurements is used as the measurement variation value of the optimal measurement point, directly reflecting the maximum dispersion of the results of multiple consecutive measurements. Considering that a qualified glaze surface should be dense and uniform, the echo fluctuation of the thickness measurement signal should be within the minimum range allowed by the instrument's measurement accuracy. In order to accurately remove abnormal jump data caused by surface micro-scratches, bubbles, or stains, a preset variation threshold of 0.005mm is set to ensure that the adopted data truly reflects the thickness of the glaze body. The implementer can set the preset variation threshold according to the nominal accuracy of the thickness gauge itself and the glaze process factory standard, which is not limited here. If the measured change value is less than or equal to the preset change threshold, it indicates that the surface condition of the optimal measurement point is good and the ultrasonic echo signal is stable and reliable. In order to output the final representative measurement value, the average value of the thickness measurement is taken as the head glaze thickness corresponding to the optimal measurement point. If the measured change value is greater than the preset change threshold, it indicates that the optimal measurement point is likely to have local physical defects that cause measurement distortion. In order to improve the detection error tolerance and avoid directly judging the detection failure of the area where the optimal measurement point is located due to occasional extremely small local defects, a valid measurement point is selected as a candidate measurement point for the optimal measurement point through preset conditions. That is, a high-quality alternative point is found in the neighborhood of the optimal measurement point for supplementary verification. In this embodiment, the preset condition is that the spatial distance between the valid measurement point and the optimal measurement point is less than or equal to the preset distance threshold. The preset distance threshold is set to 10mm to ensure that the selected candidate measurement points can fully represent the thickness distribution characteristics of the local area of ​​the porcelain insulator where the optimal measurement point is located in physical space. The implementer can set the preset distance threshold according to the head size parameters and curvature gradient of the porcelain insulator, which is not limited here. When the number of candidate measurement points is greater than the reference quantity threshold, it indicates that there are sufficient available alternative measurement points around the optimal measurement point. In order to complete the re-measurement verification quickly and efficiently and avoid falling into an infinite loop of point selection and re-measurement, re-thickness measurement is performed on the candidate measurement points one by one. It should be noted that the specific execution process of the re-thickness measurement here is: drive the dual-axis rotation platform to rotate to the new target pose according to the rotation analysis value corresponding to the current candidate measurement point, and perform a preset number of thickness measurements on the current candidate measurement point to obtain the corresponding measurement variation value and the average thickness measurement value. If the measurement variation value of the current candidate measurement point is less than or equal to the preset variation threshold, the average thickness measurement value of the current candidate measurement point is taken as the head glaze layer thickness corresponding to the optimal measurement point, and the re-measurement is stopped. If the measurement variation values of all candidate measurement points are greater than the preset variation threshold, it indicates that the overall surface state of the local area of the optimal measurement point is extremely poor or there are large-area process defects, and valid thickness measurement data can no longer be obtained, then the optimal measurement point is marked as failed to obtain the head glaze layer thickness. When the number of candidate measurement points is less than or equal to the reference quantity threshold, it indicates that the point cloud around the optimal measurement point is severely sparse or the curvature changes drastically, and there is no alternative point that can meet the platform alignment constraint, then the optimal measurement point is marked as failed to obtain the head glaze layer thickness. In this embodiment, the reference quantity threshold is set to 0, which ensures that the rescue mechanism can be triggered as long as there is at least one compliant valid measurement point in its local area. The implementer can set the reference quantity threshold according to the comprehensive consideration of the detection strictness and the re-measurement time cost, which is not limited herein.

[0031] In order to realize the global automatic evaluation and data traceability of the quality of porcelain insulator products, all optimal measurement points are traversed, and the finally obtained head glaze layer thickness or acquisition failure mark of each optimal measurement point is summarized and recorded. If the head glaze layer thickness is successfully obtained for all optimal measurement points, and all the head glaze layer thicknesses are greater than or equal to the preset thickness qualification standard, it is comprehensively determined that the head glaze layer thickness detection of the porcelain insulator is qualified. If any optimal measurement point is marked as failed to obtain the head glaze layer thickness, or the obtained head glaze layer thickness is less than the preset thickness qualification standard, it is comprehensively determined that the detection of the porcelain insulator is unqualified. To facilitate the archiving management of quality inspection results and subsequent process optimization analysis, the system packages the porcelain insulator model, the original spatial coordinates of each optimal measurement point, the corresponding rotation analysis values, the finally measured head glaze layer thickness, and the global qualification judgment result to generate a structured detection report, which is stored in the database, completing the efficient and closed-loop automatic detection of the head glaze layer thickness of the porcelain insulator, and providing complete data support for the factory inspection and quality control of power grid insulation components.

[0032] In summary, this embodiment involves clamping an insulator onto a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism and acquiring a point cloud to be measured. This cloud is then precisely registered with a standard CAD model to extract the head's measurement point set. Candidate measurement points are selected based on curvature and normal vectors. A fixed measurement direction for the thickness gauge is defined, and the rotational analytical value of the dual-axis rotary platform is obtained using an arctangent function with quadrant judgment combined with alignment constraints. Z-axis coarse adjustment ensures the head's rotation radius falls within the working distance range. Based on the rotational analytical value, the coordinate offset of the candidate measurement points after rotation is determined to obtain the optimal measurement point. Finally, the head's glaze thickness is obtained based on the optimal measurement point and the rotational analytical value. This invention solves the problems of precise alignment of curved surface normals and the high cost and complex calibration of existing robotic arms, achieving high-precision and high-efficiency automated fixed-point thickness measurement.

[0033] Example 2: This invention also proposes a system for detecting the glaze thickness of porcelain insulator heads; please refer to [link / reference]. Figure 2 The diagram shows a structural diagram of a porcelain insulator head glaze thickness detection system according to an embodiment of the present invention. The system includes: a candidate measurement point acquisition module 10, a rotation resolution value acquisition module 20, an optimal measurement point acquisition module 30, and a head glaze thickness acquisition module 40.

[0034] The candidate measurement point acquisition module 10 is used to clamp porcelain insulators on a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism, collect the test point cloud of the porcelain insulators, finely register it with its standard CAD model, extract the head test point set based on the geometric attributes of the standard CAD model, and filter candidate measurement points based on the curvature and normal vector of each point in the head test point set.

[0035] The rotation resolution acquisition module 20 is used to define the fixed measurement direction of the thickness gauge. Using the arctangent function with quadrant judgment, the rotation resolution of the dual-axis rotation platform is obtained based on the alignment constraint between the normal vector of the candidate measurement point and the fixed measurement direction.

[0036] The optimal measurement point acquisition module 30 is used to make the radius of rotation of the head surface of the porcelain insulator fall into the working distance range of the thickness gauge through the Z-axis height coarse adjustment mechanism, and to determine the Z-axis coordinate and XY coordinate offset of each candidate measurement point after rotation based on the rotation analysis value, so as to obtain the optimal measurement point.

[0037] The head glaze thickness acquisition module 40 is used to acquire the head glaze thickness of the porcelain insulator based on the optimal measurement point and the corresponding rotation resolution value.

[0038] It should be noted that the porcelain insulator head glaze thickness detection system and the porcelain insulator head glaze thickness detection method provided in the above embodiments belong to the same concept. The specific implementation process can be found in the method embodiments, and will not be repeated here.

Claims

1. A method for detecting the thickness of a glaze layer on a head of a porcelain insulator, characterized by, The method includes the following steps: Porcelain insulators are mounted on a dual-axis rotary platform equipped with a Z-axis height coarse adjustment mechanism. The test point cloud of the porcelain insulators is collected and precisely registered with its standard CAD model. The head test point set is extracted based on the geometric attributes of the standard CAD model. Candidate measurement points are selected based on the curvature and normal vector of each point in the head test point set. Define the fixed measurement direction of the thickness gauge, and use the arctangent function with quadrant judgment to obtain the rotational analytical value of the dual-axis rotary platform based on the alignment constraint between the normal vector of the candidate measurement point and the fixed measurement direction. The Z-axis height coarse adjustment mechanism is used to make the radius of rotation of the head surface of the porcelain insulator fall into the working distance range of the thickness gauge. Based on the rotational analysis value, the Z-axis coordinate and XY coordinate offset of each candidate measurement point after rotation is determined, and the optimal measurement point is obtained. The thickness of the glaze layer at the head of the porcelain insulator is obtained based on the optimal measurement point and the corresponding rotational analytical value.

2. The method of claim 1, wherein the thickness of the glaze layer is measured by using a laser sensor. The method for obtaining the head test point set is as follows: The range corresponding to the upper and lower limits of the Z-axis height of the head region in the standard CAD model of the porcelain insulator is obtained as the head height range; By using Z-axis pass-through filtering, the set of points in the standard point cloud after discretizing the standard CAD model that are within the head height range is taken as the standard head region point set; Based on the KD-tree nearest neighbor search algorithm, in the point cloud to be tested after fine registration with the standard CAD model, a point set that matches the standard head region point set is found and used as the head test point set.

3. The method of claim 1, wherein the thickness of the glaze layer is measured by using a laser sensor. The method for obtaining the candidate measurement points is as follows: Based on the normal vectors of each point in the head test point set, calculate the neighborhood variance of the normal vector of each point in the preset neighborhood. The candidate measurement point set consists of points in the head measurement point set whose curvature is less than or equal to a preset curvature threshold and whose neighborhood variance of the normal vector is less than or equal to a preset normal vector variance threshold.

4. The method of claim 1, wherein the thickness of the glaze layer is measured by using a laser sensor. The method for obtaining the rotation resolution value is as follows: The opposite direction of the fixed measurement direction of the thickness gauge is defined as the target normal vector; a dual-axis rotation matrix is ​​constructed based on the rotational motion of the dual-axis rotation platform around the X and Y axes; Obtain the directional components of the normal vector of each candidate measurement point along the X, Y, and Z axes; Based on the alignment constraint that the normal vector of each candidate measurement point is equal to the target normal vector after transformation by the dual-axis rotation matrix, the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotation platform around the X-axis corresponding to each candidate measurement point; the negative value of the direction component of the X-axis, the square root of the sum of the squares of the direction components of the Y-axis and Z-axis are substituted into the arctangent function with quadrant judgment to obtain the rotation analysis value of the dual-axis rotation platform around the Y-axis corresponding to each candidate measurement point.

5. The method of claim 1, wherein the thickness of the glaze layer is measured by using a laser sensor. The method for obtaining the optimal measurement point is as follows: Based on the rotation analysis value corresponding to each candidate measurement point, obtain the spatial coordinates of each candidate measurement point after rotation; Candidate measurement points whose rotated Z-axis coordinates fall within the working distance range of the thickness gauge, and whose absolute values ​​of the rotated X-axis and Y-axis are both less than or equal to the preset spot radius, are selected as valid measurement points. When the number of valid measurement points is less than the preset threshold, a system alarm is triggered. When the number of valid measurement points is greater than or equal to the preset number threshold, the sum of the absolute values ​​of the X and Y coordinates of each valid measurement point after rotation is calculated and used as the XY coordinate offset value of each valid measurement point. The valid measurement points are sorted in ascending order of their XY coordinate offset values ​​to obtain a sequence of valid measurement points. The first specified number of valid measurement points in the valid measurement point sequence are selected as the optimal measurement points.

6. A method of detecting the thickness of the glaze layer on the head of a porcelain insulator according to claim 5, characterized in that, The method for obtaining the spatial coordinates of the candidate measurement point after rotation is as follows: For any candidate measurement point, the initial three-dimensional coordinates of the candidate measurement point are used as the initial coordinate column vector; Multiply the biaxial rotation matrix determined by the rotation analysis value corresponding to the candidate measurement point with the initial coordinate column vector, and then add it with the preset translation offset vector to obtain the rotated target coordinate column vector; The X-axis, Y-axis, and Z-axis coordinates are extracted from the target coordinate column vector and used as the spatial coordinates of the candidate measurement point after rotation.

7. A method of detecting the thickness of the glaze layer on the head of a porcelain insulator according to claim 5, characterized in that, The method for obtaining the thickness of the glaze layer on the head is as follows: For any optimal measurement point, the dual-axis rotary platform is driven to rotate to the target pose by the rotation resolution value corresponding to the optimal measurement point, and the thickness gauge is controlled to perform a preset number of thickness measurements on the optimal measurement point. The range of thickness measurements is taken as the measurement change value of the optimal measurement point; if the measurement change value is less than or equal to the preset change threshold, the average value of the thickness measurements is taken as the head glaze thickness corresponding to the optimal measurement point. If the measured change value is greater than the preset change threshold, then a valid measurement point is selected according to the preset conditions and used as a candidate measurement point for the optimal measurement point. When the number of candidate measurement points exceeds the reference threshold, the thickness of each candidate measurement point is remeasured. If the measurement change value of the current candidate measurement point is less than or equal to the preset change threshold, the average thickness measurement value of the current candidate measurement points is taken as the head glaze thickness corresponding to the optimal measurement point and the remeasurement is stopped. If the measurement change values ​​of all candidate measurement points are greater than the preset change threshold, the optimal measurement point is marked as the failure to obtain the head glaze thickness. When the number of candidate measurement points is less than or equal to the reference number threshold, the optimal measurement point is marked as a failure to obtain the head glaze thickness.

8. The method of claim 7, wherein the step of detecting the thickness of the glaze layer on the head of the porcelain insulator is characterized by: The preset condition is that the spatial distance between the effective measurement point and the optimal measurement point is less than or equal to a preset distance threshold.

9. The method for detecting the glaze layer thickness at the head of a porcelain insulator as described in claim 1, characterized in that, The thickness gauge is fixed to measure in the negative Z-axis direction, which is vertically downward.

10. A system for detecting the glaze thickness of a porcelain insulator head, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method for detecting the glaze layer thickness of a porcelain insulator head as described in any one of claims 1-9.