Precise detection method and system for sealing element

Abstract

The embodiment of the invention provides a precision detection method and system for a sealing element, and belongs to the technical field of precision detection, and the method comprises the steps: obtaining the angle data of a mobile detection head; the obtained placing rod is determined as a rotating node; calculating an initial connection weight corresponding to the rotating node; calculating according to the angle data to obtain a first matching parameter; calculating a distance parameter between the sample input vector and the output layer node through the initial connection weight; determining a neighborhood range of the output layer according to the distance parameter; outputting clustering result data of the rotating node by taking the first matching parameter as an adjustment amount; and the first motor controller and the second motor controller are used for controlling the placing rod and the movable detection head to detect the sealing element. The movable detection head rotates around the fixed clamp, so that the surface appearance defect of the sealing element is detected, the detection efficiency is greatly improved, the problem that the sealing element made of an elastomer material is easy to deform is solved, and the detection efficiency of the sealing element is improved.

Description

Technical Field

[0001] This invention relates to the field of precision testing technology, and in particular to a precision testing method and system for sealing components, a computer device, and a storage medium. Background Technology

[0002] As a critical basic component, seals are widely used in aerospace, automotive, energy equipment, and medical devices. Their sealing performance directly affects the safety, reliability, and lifespan of equipment. Precision testing is the core link in ensuring the quality of seals. Existing seal testing can generally be divided into contact measurement and non-contact measurement. Contact measurement refers to manual testing and simple functional testing using tools such as calipers and plug gauges, which has low accuracy and poor efficiency. On the other hand, there is also testing through automated equipment, such as laser scanning and CT scans, but the testing cost is relatively high and the testing procedures are complex. Summary of the Invention

[0003] In view of the above problems, embodiments of the present invention are proposed to provide a precision inspection method for a seal, a precision inspection system for a seal, a computer device, and a storage medium that overcome or at least partially solve the above problems.

[0004] To address the aforementioned problems, this invention discloses a precision inspection method for seals, applied to a fixed clamp and a movable inspection head. The fixed clamp includes multiple placement rods arranged in a dispersed manner, which are used to place seals. The movable inspection head is used to inspect the seals, comprising:

[0005] The angle data of the moving detection head is obtained;

[0006] The placement rod was identified as a rotation node.

[0007] Calculate the initial connection weights corresponding to the rotated nodes;

[0008] The first matching parameter is calculated based on the angle data;

[0009] The distance parameter between the sample input vector and the output layer node is calculated using the initial connection weights;

[0010] The neighborhood range of the output layer is determined based on the distance parameter.

[0011] Using the first matching parameter as the adjustment amount, calculate the weights of the output layer nodes and the neighborhood range that meet the minimum distance adjustment. If the number of iterations meets the preset conditions, output the clustering result data of the rotated nodes.

[0012] The clustering results data are input to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head. The first motor controller and the second motor controller control the placement rod and the moving detection head to perform the detection operation of the seal.

[0013] Preferably, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the method includes:

[0014] The roll angle, pitch angle, and yaw angle of the moving detection head are calculated using the accelerometer, magnetometer, and gyroscope.

[0015] Calculate the swing characteristic parameters of the placement rod;

[0016] The Kalman filter is modified by the swing characteristic parameters to obtain the modified Kalman filter;

[0017] The roll angle, pitch angle, and yaw angle of the motor of the moving detection head are filtered by a modified Kalman filter to obtain optimized roll angle, pitch angle, and yaw angle.

[0018] The optimized roll angle, pitch angle, and yaw angle are transmitted to the second motor controller of the moving detection head for data comparison, and the operating angle of the second motor controller of the moving detection head is adjusted.

[0019] Preferably, determining the neighborhood range of the output layer based on the distance parameter includes:

[0020] Generate multiple candidate solutions corresponding to the distance parameters;

[0021] Obtain the preset destruction operator and preset repair operator for the path of the rotated node;

[0022] Calculate the parameter value of each operator in the search cycle, and select the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value;

[0023] The optimal solution is determined as the neighborhood range of the output layer.

[0024] Preferably, the step of calculating the first matching parameter based on the angle data includes:

[0025] Obtain the angle data sequence of the moving detection head;

[0026] A smoothed angle sequence is obtained by performing a smoothed window filter on the angle data sequence.

[0027] Calculate the rate of change of the angles in the smoothed angle sequence, and obtain the smoothness index vector from the rate of change of the angles.

[0028] The similarity between the smoothness index vector and the preset vectors in the preset detection pattern library;

[0029] Extract the weight parameters of the preset vector with the highest similarity;

[0030] The first matching parameter is calculated using the weight parameters and similarity.

[0031] Preferably, the step of calculating and adjusting the weights of the output layer nodes and their neighborhood ranges that satisfy the minimum distance using the first matching parameter as an adjustment amount, and outputting the clustering result data of the rotated nodes if the number of iterations meets a preset condition, includes:

[0032] The output layer node that meets the minimum distance requirement is determined as the best matching unit;

[0033] The neighborhood function is determined based on the distance parameter between the best matching unit and its neighboring nodes; this neighborhood function sets the update magnitude of each node within the neighborhood relative to the best matching unit. The neighborhood range is the topological neighborhood centered on the best matching unit.

[0034] For the best matching unit and its neighboring nodes, set the first matching parameter as the adjustment amount, adjust the weight vector of the BMU and its neighboring nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

[0035] Preferably, the step of inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller, includes:

[0036] The clustering results are mapped to the first motion parameters and the second running parameters;

[0037] The first motor controller is controlled by the first motion parameters to rotate the placement rod and adjust it to the target angle.

[0038] The second motor controller, controlled by the second motion parameters, makes the moving detection head move along the path trajectory.

[0039] This invention discloses a precision inspection system for seals, applied to a fixed clamp and a moving inspection head. The fixed clamp includes multiple placement rods arranged in a dispersed manner for placing seals, and the moving inspection head is used to inspect the seals. The system includes:

[0040] The first acquisition module is used to acquire the angle data of the moving detection head;

[0041] The second acquisition module is used to acquire the placement rod as a rotation node;

[0042] The first calculation module is used to calculate the initial connection weights corresponding to the rotated nodes;

[0043] The second calculation module is used to calculate the first matching parameter based on the angle data;

[0044] The third calculation module is used to calculate the distance parameters between the sample input vector and the output layer nodes using the initial connection weights;

[0045] The determination module is used to determine the neighborhood range of the output layer based on the distance parameter;

[0046] The node weight module is used to calculate and adjust the weights of the output layer nodes and their neighborhood ranges that meet the minimum distance using the first matching parameter as the adjustment amount. If the number of iterations meets the preset conditions, the clustering result data of the rotated nodes is output.

[0047] The detection operation module is used to input clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and to control the placement rod and the moving detection head to perform the detection operation of the seal through the first motor controller and the second motor controller.

[0048] Preferably, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the system includes:

[0049] The fourth calculation module is used to calculate the roll angle, pitch angle and yaw angle of the moving detection head using the accelerometer, magnetometer and gyroscope;

[0050] The fifth calculation module is used to calculate the swing characteristic parameters of the placement rod;

[0051] The correction module is used to correct the Kalman filter using the swing characteristic parameters to obtain the corrected Kalman filter;

[0052] The filtering operation module is used to filter the roll angle, pitch angle and yaw angle of the motor of the moving detection head through a modified Kalman filter to obtain optimized roll angle, pitch angle and yaw angle;

[0053] The adjustment module is used to transmit the optimized roll angle, pitch angle and yaw angle to the second motor controller of the moving detection head for data comparison and to adjust the operating angle of the second motor controller of the moving detection head.

[0054] This invention also discloses a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described steps for precise detection of the seal.

[0055] This invention also discloses a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described steps for precise detection of the seal.

[0056] The embodiments of the present invention have the following advantages:

[0057] In this embodiment of the invention, the precision detection method for the seal is applied to a fixed fixture and a moving detection head. The fixed fixture includes multiple placement rods arranged in a dispersed manner, which are used to place the seal. The moving detection head is used to detect the seal. The method includes: acquiring angle data of the moving detection head; identifying the placement rod as a rotating node; calculating the initial connection weight corresponding to the rotating node; calculating a first matching parameter based on the angle data; calculating the distance parameter between the sample input vector and the output layer node using the initial connection weight; determining the neighborhood range of the output layer based on the distance parameter; using the first matching parameter as an adjustment amount, calculating and adjusting the node weights of the output layer node and the neighborhood range that satisfy the minimum distance; if the number of iterations meets a preset condition, outputting the clustering result data of the rotating node; inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller. The movable inspection head rotates around the fixed fixture to inspect the surface defects of the seal, which greatly improves the inspection efficiency, solves the problems of easy deformation of elastomeric material seals and the error introduced by contact measurement, reduces equipment costs, increases inspection speed, meets the full inspection requirements of large-scale production lines, saves energy consumption, and improves the inspection efficiency of seals. Attached Figure Description

[0058] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0059] Figure 1 This is a flowchart illustrating the steps of a precision testing method for a sealing component according to an embodiment of the present invention.

[0060] Figure 2 This is a schematic diagram of the cooperation between a fixed clamp and a movable detection head according to an embodiment of the present invention;

[0061] Figure 3 This is a structural block diagram of an embodiment of a precision testing system for a sealing component according to an embodiment of the present invention;

[0062] Figure 4 This is an internal structural diagram of a computer device according to one embodiment. Detailed Implementation

[0063] To make the technical problems, technical solutions, and beneficial effects solved by the embodiments of the present invention clearer, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely illustrative of the present invention and are not intended to limit the present invention.

[0064] In this embodiment of the invention, multiple fixed clamps are used to support the seal, and the starting point of the moving detection head is set at the center of the fixed clamp. The moving detection head rotates around the fixed clamp to detect surface defects of the seal, which greatly improves the detection efficiency, solves the problems of easy deformation of elastomeric material seals and the error introduced by contact measurement, reduces equipment costs, increases detection speed, and meets the full inspection requirements of large-scale production lines. Furthermore, by adjusting the moving path of the moving detection head and the angle of the fixed clamp, energy consumption is saved and the detection efficiency of the seal is improved.

[0065] Reference Figure 1 This diagram illustrates a flowchart of a precision inspection method for a seal according to an embodiment of the present invention. The method is applied to a fixed fixture and a moving inspection head. The fixed fixture includes multiple placement rods arranged in a dispersed manner, which are used to place the seal. The moving inspection head is used to inspect the seal. Specifically, the method may include the following steps:

[0066] Step 101: Obtain the angle data of the moving detection head;

[0067] In this embodiment of the invention, the method is applied to a detection system including a fixed clamp and a moving detection head, such as... Figure 3 As shown, the starting point of the moving detection head is set at the center of the fixed clamp. The moving detection head rotates around the fixed clamp. The seal can be placed and fixed on the placement rod. The placement rod can realize the self-rotation operation as shown at point A, and the movement in four directions (up, down, left, and right) as shown at point B. The moving detection head can be equipped with an image acquisition device or an infrared scanner, etc. This embodiment of the invention does not impose too many restrictions on this. It should be noted that the seal can include rubber rings, silicone gaskets, etc. This embodiment of the invention does not impose too many restrictions on this. In addition, there can be multiple fixed clamps. The moving detection head is used in conjunction with multiple fixed clamps, that is, one moving detection head can detect the seals on multiple fixed clamps.

[0068] In one specific example, when the seal can be placed and fixed on the placement rod, the moving detection head can take an image of the appearance of the corresponding seal while the placement rod rotates. After the image is taken, the placement rod can be moved downward to allow the seal to slide off, thus completing the detection process. After all the seals on the placement rods of the fixed fixture have been detected, the detection head is moved to the next fixed fixture, and the appearance image of the corresponding seal is taken again. The detection process is repeated until all the fixed fixtures have been detected.

[0069] In a preferred embodiment of the present invention, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; of course, multiple other types of sensors may also be provided, and the present invention does not impose excessive limitations on this. The method includes:

[0070] The roll angle, pitch angle, and yaw angle of the moving detection head are calculated using the accelerometer, magnetometer, and gyroscope.

[0071] Calculate the swing characteristic parameters of the placement rod;

[0072] The Kalman filter is modified by the swing characteristic parameters to obtain the modified Kalman filter;

[0073] The roll angle, pitch angle, and yaw angle of the motor of the moving detection head are filtered by a modified Kalman filter to obtain optimized roll angle, pitch angle, and yaw angle.

[0074] The optimized roll angle, pitch angle, and yaw angle are transmitted to the second motor controller of the moving detection head for data comparison, and the operating angle of the second motor controller of the moving detection head is adjusted.

[0075] In this embodiment of the invention, the oscillation characteristic parameters may include the oscillation angle amplitude and the oscillation stiffness coefficient. The Kalman filter is corrected by the oscillation angle amplitude and the oscillation stiffness coefficient to obtain the corrected Kalman filter. The optimized roll angle, pitch angle and yaw angle are then used to adjust the operating angle.

[0076] Specifically, the swing characteristic parameters may also include swing angular velocity or angular acceleration, swing frequency, equivalent damping coefficient, etc., but the embodiments of the present invention do not impose too many restrictions on this.

[0077] Among them, the swing angle amplitude refers to the maximum angle at which the end of the placement rod deviates from its target equilibrium position, which directly reflects the severity of the swing; while the swing stiffness coefficient refers to the equivalent stiffness data after simplifying the "placement rod-seal" system into a spring-mass system, which can be calculated by analyzing the ratio of the applied small torque to the resulting angle change.

[0078] In one specific example, a high-resolution encoder or a miniature IMU (inertial measurement unit) is installed on the motor shaft or base of each placement rod. When the placement rod performs rotation or movement operations, its angle and angular velocity data are read in real time, and the real-time or historical motion data collected above is analyzed.

[0079] The swing angle range can be obtained by calculating the maximum deviation between the actual angle curve and the target angle line (such as 90 degrees).

[0080] The equivalent stiffness coefficient can be obtained by applying a known small step torque command during system initialization or idle, measuring the resulting steady-state angle change, and calculating the ratio of torque to angle change.

[0081] One of the core components of a Kalman filter is the process noise covariance matrix Q, which represents the uncertainty of the system model. In this embodiment of the invention, the oscillation characteristic parameter is used to dynamically adjust this Q matrix.

[0082] Specifically, a large swing amplitude or a low stiffness coefficient indicates that the current state of the system is more susceptible to disturbances, and it is necessary to increase the noise variance values ​​related to position and velocity in the Q matrix.

[0083] The modified Kalman filter is Qx = Qj × (1 + α × C + β / D); where Qj represents the calibration value of the process noise covariance matrix, Qx represents the output value of the modified Kalman filter, C represents the swing angle amplitude, D represents the equivalent stiffness coefficient, and α and β represent the preset weighting coefficients, respectively.

[0084] In this embodiment of the invention, the original roll angle, pitch angle, and yaw angle calculated by the IMU of the detection head itself are input into the corrected Kalman filter for fusion filtering;

[0085] When the detection head is scanning a rubber ring with large swing, the Q matrix is ​​increased, which allows the filter to respond more quickly to the small wobbles transmitted by the jig swing. The output is a more accurate angle, which is compared with the target path angle of the second motor controller to generate more timely and accurate control commands. This allows the detection head motor to make compensatory movements, thereby maintaining the relative stability between the lens and the surface of the seal at the moment of image acquisition and improving the accuracy of the detection.

[0086] In this embodiment of the invention, the placement rod and the detection head are considered as a single unit. The dynamic characteristics of the fixture end are fed back in real-time to the state estimation and control algorithm of the detection head end through swing characteristic parameters. This enables the detection system to adapt to different workpiece characteristics. Whether it is soft rubber or harder silicone, the system can automatically adjust its filtering and control strategies to achieve the best balance between high precision and high responsiveness, improving the detection success rate and reliability under various complex conditions.

[0087] It should be noted that the reference process noise covariance matrix The calibration can be performed by driving the placement rod to perform a series of standard small-amplitude movements in a system-unloaded (no seals) and stationary state. Motor encoder data is collected as approximate true values ​​and angle data calculated by the IMU. Using a standard Kalman filter algorithm (with initial Q set to an empirical value), the optimal filtering error is derived through maximum likelihood estimation or covariance matching techniques. Values, which are fixed as the system baseline; weighting coefficients It can be a preset empirical coefficient.

[0088] Step 102: The placement rod is determined to be a rotation node;

[0089] Step 103: Calculate the initial connection weights corresponding to the rotated nodes;

[0090] Specifically, calculating the initial connection weights corresponding to the rotated nodes includes the following steps:

[0091] Constructing Node Feature Vectors: For each placement rod (i.e., rotation node), extract its multi-dimensional state features to form an n-dimensional feature vector. The n-dimensional eigenvector Exemplary features may include:

[0092] : Current position angle, obtained through the encoder of the first motor, normalized to [0, 1].

[0093] : Seal type code, for example, rubber O-ring = 0.1, silicone flat gasket = 0.3, metal-coated ring = 0.8, based on its rigidity preset.

[0094] The current swing angular velocity amplitude is calculated by the IMU at the base of the rod and then normalized.

[0095] Node traversal priority: Based on the preset detection path plan, assign an ordinal value and normalize it.

[0096] The current distance to the moving detection head is calculated and normalized according to the system coordinate system. .

[0097] Furthermore, the SOFM network can be initialized by first setting the output layer to one. A two-dimensional grid (e.g., 5x5). Each grid node This corresponds to a weight vector with the same dimension as the input feature vector. , where the superscript (0) indicates the initial state.

[0098] Furthermore, a linear initialization method is used to determine the initial connection weights. First, the principal components of the eigenvector sets of all rotated nodes are calculated. Then, the node coordinates of the output layer mesh are... A linear mapping is applied to the subspace spanned by the first two principal components. The specific formula is:

[0099] ;

[0100] in: It is all input feature vectors The mean vector, and These are the first and second principal component vectors, respectively. It corresponds to the standard deviation of the principal components and is used to control the distribution range of the initial weights in the feature space. It is the side length of the output layer grid, which can ensure that the initial weights cover the main distribution area of ​​the input data and accelerate network convergence.

[0101] Step 104: Calculate the first matching parameter based on the angle data;

[0102] In this embodiment of the invention, each placement rod can be converted into a rotation node, and an initial connection weight is preset for the rotation node. The initial connection weight is used to calculate the distance parameter between the sample input vector and the output layer node; then the first matching parameter is calculated through angle data.

[0103] Specifically, in this embodiment of the invention, calculating the first matching parameter based on the angle data includes:

[0104] Obtain the angle data sequence of the moving detection head;

[0105] A smoothed angle sequence is obtained by performing a smoothed window filter on the angle data sequence.

[0106] Calculate the rate of change of the angles in the smoothed angle sequence, and obtain the smoothness index vector from the rate of change of the angles.

[0107] The similarity between the smoothness index vector and the preset vectors in the preset detection pattern library;

[0108] Extract the weight parameters of the preset vector with the highest similarity;

[0109] The first matching parameter is calculated using the weight parameters and similarity.

[0110] In this embodiment of the invention, raw data is read in real time from the IMU (gyroscope, accelerometer, etc.) of the moving detection head, and the roll angle, pitch angle, and yaw angle corresponding to a series of timestamps are calculated to form an angle data sequence that changes over time.

[0111] Applying a moving average filter or a low-pass digital filter, such as a Butterworth filter, to the angle data sequence aims to preserve the low-frequency trends caused by real physical motion while filtering out high-frequency sensor measurement noise. This results in a smoothed sequence, i.e., a smoothed angle sequence.

[0112] Perform a difference operation on the smoothed angle sequence to calculate the angle change rate, i.e., angular velocity, at each time step; extract features such as the average angular velocity value and the standard deviation of angular velocity from the angle change rate data, and construct the corresponding smoothness index vector using the aforementioned features such as the average angular velocity value and the standard deviation of angular velocity;

[0113] In this embodiment of the invention, a preset detection mode library is also provided. This preset detection mode library contains the correspondence between multiple mode parameters, preset vectors, and weight parameters. For example, the preset detection mode library may include the correspondence between a flexible large workpiece mode, preset vector 1, and weight parameter 1; a rigid small workpiece mode, preset vector 2, and weight parameter 2; and a medium rigidity mode, preset vector 3, and weight parameter 3. Of course, it may also include other correspondences between mode parameters, preset vectors, and weight parameters. This embodiment of the invention does not impose too many restrictions on this.

[0114] Furthermore, the cosine similarity or Euclidean distance similarity between the smoothness index vector and the preset vectors in the preset detection mode library can be calculated. The preset detection mode with the highest similarity can be selected, and its corresponding weight parameters can be extracted. The weight parameters are preset by the system calibration method, which represents the urgency of adjusting the subsequent SOFM network weights when under this detection mode.

[0115] The first matching parameter can be calculated as follows:

[0116] E = λ × S; where E represents the first matching parameter, λ represents the weight parameter, and S represents the similarity. In this embodiment of the invention, the first matching parameter dynamically modulates the learning rate, enabling the entire system to adaptively change the convergence speed and amplitude of the control strategy based on the real-time perceived physical motion characteristics.

[0117] Specifically, the preset detection pattern library is established through calibration using historical best detection data and contains M entries. Each entry (pattern m) stores:

[0118] Preset vector The standard eigenvector representing smooth motion obtained by degradation calculation under this model can be obtained from the average angular velocity scalar and the angular velocity standard deviation scalar under this model.

[0119] Weight parameters The emergency adjustment weights for this mode are derived by inversely calculating the required learning rate adjustment for the network when this mode achieves the highest detection accuracy in historical data. The value typically ranges from [0.5, 2.0], with larger values ​​indicating more aggressive weight adjustments when matching this mode.

[0120] Specifically, similarity The calculation uses a Gaussian kernel function:

[0121] ;

[0122] in: It is the smoothness index vector obtained from the current calculation; It is a bandwidth parameter, set through cross-validation, for example, set to 20% of the average distance of the preset vectors between all modes.

[0123] Step 105: Calculate the distance parameter between the sample input vector and the output layer node using the initial connection weights;

[0124] In practical applications of this invention, in a preferred SOFM network, a rotating node refers to sample data consisting of each placement rod and its state (such as current position, angle, swing characteristic parameters, and the order in which nodes are traversed).

[0125] The output layer nodes represent the placement rod control strategy or placement rod configuration mode, and the control strategy for moving the detection head.

[0126] First, a sample input vector can be constructed. Then, the weight vector of the output layer node can be obtained. Finally, the distance parameter between the two can be calculated. Specifically, each placement rod can be set with 5 features: current position angle, seal type code, swing angular velocity, angular acceleration, and node traversal order. The sample input vector is constructed using the placement rod number, current position angle, seal type code, swing angular velocity, and angular acceleration. The weight vector of the output layer node is obtained, and the Euclidean distance parameter between the two is calculated. It should be noted that the node traversal order represents the movement path of the moving detection head.

[0127] Step 106: Determine the neighborhood range of the output layer based on the distance parameter;

[0128] In a preferred embodiment of the present invention, determining the neighborhood range of the output layer based on the distance parameter includes:

[0129] Generate multiple candidate solutions corresponding to the distance parameters;

[0130] Obtain the preset destruction operator and preset repair operator for the path of the rotated node;

[0131] Calculate the parameter value of each operator in the search cycle, and select the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value;

[0132] The optimal solution is determined as the neighborhood range of the output layer.

[0133] In this embodiment of the invention, the specific cell corresponding to the distance parameter is first determined. The specific cell may include a node located at the center of the grid, that is, the cell with the smallest distance parameter is the center of the grid. Multiple candidate solutions are generated, such as candidate solution 1 is a circular neighborhood centered on the specific cell with radius r1; candidate solution 2 is a circular neighborhood centered on the specific cell with radius r2; candidate solution 3 is a 3x3 rectangular neighborhood.

[0134] Furthermore, preset destruction operators and preset repair operators for the paths of rotated nodes can be obtained; for example, the preset destruction operator may include randomly removing 20% ​​of the nodes, removing the 30% of the nodes farthest from a specific cell, and removing the node with the largest distance parameter; while the preset repair operator may include adding the node closest to the specific cell, adding the node that minimizes the average distance, and maintaining the integrity of the topology.

[0135] The parameter value may include the average distance parameter of nodes in the neighborhood, the neighborhood size, the topological density, etc., and the embodiments of the present invention do not impose too many restrictions on this.

[0136] The parameter values ​​of each operator in the search cycle are calculated, and the optimal solution among multiple candidate solutions corresponding to the path of the rotating node is selected based on the parameter values. The neighborhood range that meets the conditions is obtained by comparing the average distance parameter, neighborhood size, and topological density. Compared with the traditional SOFM which uses a fixed decay neighborhood, this embodiment of the invention considers multiple factors such as distance parameter, topology, and iteration stage. It avoids local optima by destroying and repairing operators, finds a globally better neighborhood configuration, and automatically selects the most suitable control strategy to update the range.

[0137] Step 107: Using the first matching parameter as the adjustment amount, calculate the weights of the output layer nodes and the neighborhood range that satisfy the minimum distance. If the number of iterations meets the preset conditions, output the clustering result data of the rotated nodes.

[0138] In this embodiment of the invention, the step of calculating and adjusting the weights of the output layer nodes and their neighborhood ranges that satisfy the minimum distance using the first matching parameter as an adjustment amount, and outputting the clustering result data of the rotated nodes if the number of iterations meets a preset condition, includes:

[0139] The output layer node that meets the minimum distance requirement is determined as the best matching unit;

[0140] The neighborhood function is determined based on the distance parameter between the best matching unit and the neighboring nodes. This neighborhood function sets the update magnitude of each node in the neighborhood relative to the best matching unit, and the neighborhood range is the topological neighborhood centered on the best matching unit.

[0141] For the best matching unit and its neighboring nodes, set the first matching parameter as the adjustment amount, adjust the weight vector of the best matching unit and its neighboring nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

[0142] The preset condition can be any value set by those skilled in the art according to actual needs, and the embodiments of the present invention do not impose too many restrictions on it; for example, the preset condition can be that the number of iterations is greater than 500.

[0143] Based on the motion state of the current detection head quantified by the first matching parameter, the SOFM network is guided to automatically classify the complex states of all current placement rods (rotating nodes) and the movement path of the moving detection head into a few optimal preset control configuration modes at the most appropriate learning frequency.

[0144] In a preferred embodiment, the optimal matching unit can be the specific unit described above, that is, the output layer node that satisfies the minimum distance is determined as the optimal matching unit.

[0145] Centered on the best matching unit, a neighborhood range is defined on the topology of the output layer, such as a two-dimensional grid. The weights are iteratively updated using the first matching parameter as the adjustment amount, as shown in the following specific method.

[0146] W = E(wq + ηhX); where W represents the adjusted weight vector, E represents the first matching parameter, wq represents the weight vector before adjustment, η represents the learning rate, h represents the neighborhood function, X represents the difference vector between the current input state and the node's old weights, and ηhX represents the direction of weight adjustment.

[0147] Learning rate in embodiments of the present invention Exponential decay can be used, and it is subject to the first matching parameter. modulation.

[0148] ;

[0149] in, This is the initial learning rate, for example, 0.1. It is the current iteration number. It is the decay time constant, for example, set to 1 / 3 of the total number of iterations. It is the dynamic first matching parameter.

[0150] Neighborhood function of the present invention A Gaussian function can be used, and its width... It decays over time.

[0151] ;

[0152] in, These are the coordinates of the best matching unit (BMU). This is the topological distance (such as Manhattan distance) on the output layer mesh. Neighborhood radius. The attenuation formula is:

[0153] ;

[0154] The initial value is set to the output layer mesh radius (e.g., for a 5x5 mesh). ), It is the attenuation constant (usually with) (Related).

[0155] The weight update formula of this invention is as follows:

[0156] ;in, It is the feature vector of the rotated node input at the t-th iteration.

[0157] Furthermore, the preset conditions include: (a) the total number of iterations reaches a preset upper limit. (e.g., 1000 times); (b) the average weight change is less than the threshold. (For example The calculation formula is: N refers to the number of node coordinates.

[0158] Step 108: Input the clustering result data into the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and control the placement rod and the moving detection head to perform the detection operation of the seal through the first motor controller and the second motor controller.

[0159] Furthermore, in this embodiment of the invention, after obtaining the clustering result data, the clustering result data is normalized, and then the normalized data is input to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, respectively. The first motor controller and the second motor controller control the placement rod and the moving detection head to perform the detection operation of the seal.

[0160] In a specific example of this invention, the step of inputting clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller, includes:

[0161] The clustering results are mapped to the first motion parameters and the second running parameters;

[0162] The first motor controller is controlled by the first motion parameters to rotate the placement rod and adjust it to the target angle.

[0163] The second motor controller, controlled by the second motion parameters, makes the moving detection head move along the path trajectory.

[0164] In this embodiment of the invention, the mapping from physical state to sample input vector is performed as described in step 103. The physical state (angle, load type, dynamic swing, spatial position, path sequence) of each placement rod is measured and encoded by sensors, and then converted into normalized digital features. to Concatenate them into a sample input vector The key to this process is normalization, which requires determining the physical range of each feature, such as angles from 0 to 360 degrees, rigid encoding from 0 to 1, and using linear or nonlinear scaling to map it to the interval [0,1] or [-1,1] to ensure that features of different dimensions are comparable in the SOFM network.

[0165] In addition, the clustering results are mapped to motion control parameters, and the clustering results are output after the SOFM network is trained. Each cluster center... This represents an optimized "fixture-detection head" collaborative working mode. The mapping process is as follows:

[0166] First, the motion parameters can be decoded, and the cluster center weight vector can be obtained. Each dimension, through inverse normalization, is reduced to a specific physical quantity. For example, The first component (corresponding to) After inverse normalization, it is mapped back to the target angle value. (e.g., 120 degrees). The second component is mapped to the expected seal type and is used to select the corresponding force control parameters. The third and fourth components can be used to derive the smoothness constraints of the detection head movement.

[0167] Furthermore, control commands can be generated, including the following steps: First motion parameters (for the first motor controller): mainly a series of target angle sequences. It is sent directly to the first motor controller to control the rotation of the placement rod.

[0168] The second motion parameter (used by the second motor controller): Spatial path features decoded from the cluster center vector (implicit in the combination relationship of each vector component), combined with the detection task, generate a smooth three-dimensional path trajectory point sequence in the task space through B-spline curve interpolation. The data is then converted into angle commands for each joint motor and sent to the second motor controller to drive the moving detection head. This embodiment of the invention can establish a mapping lookup table between "cluster center vectors and optimal motion control parameter sets". By finding the nearest neighbor cluster center, the corresponding control parameters can be quickly obtained, achieving real-time control.

[0169] In this embodiment of the invention, the precision detection method for the seal is applied to a fixed fixture and a moving detection head. The fixed fixture includes multiple placement rods arranged in a dispersed manner, which are used to place the seal. The moving detection head is used to detect the seal. The method includes: acquiring angle data of the moving detection head; identifying the placement rod as a rotating node; calculating the initial connection weight corresponding to the rotating node; calculating a first matching parameter based on the angle data; calculating the distance parameter between the sample input vector and the output layer node using the initial connection weight; determining the neighborhood range of the output layer based on the distance parameter; using the first matching parameter as an adjustment amount, calculating and adjusting the node weights of the output layer node and the neighborhood range that satisfy the minimum distance; if the number of iterations meets a preset condition, outputting the clustering result data of the rotating node; inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller. The movable inspection head rotates around the fixed fixture to inspect the surface defects of the seal, which greatly improves the inspection efficiency, solves the problems of easy deformation of elastomeric material seals and the error introduced by contact measurement, reduces equipment costs, increases inspection speed, meets the full inspection requirements of large-scale production lines, saves energy consumption, and improves the inspection efficiency of seals.

[0170] It should be noted that, for the sake of simplicity, the method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0171] Reference Figure 3 This diagram illustrates a structural block diagram of a precision inspection system for a sealing component according to an embodiment of the present invention. The system is applied to a fixed clamp and a movable inspection head. The fixed clamp includes multiple placement rods arranged in a dispersed manner for placing the sealing component. The movable inspection head is used to inspect the sealing component and may specifically include the following modules:

[0172] The first acquisition module 301 is used to acquire the angle data of the moving detection head;

[0173] The second acquisition module 302 is used to acquire the placement rod as a rotation node;

[0174] The first calculation module 303 is used to calculate the initial connection weights corresponding to the rotated nodes;

[0175] The second calculation module 304 is used to calculate the first matching parameter based on the angle data;

[0176] The third calculation module 305 is used to calculate the distance parameter between the sample input vector and the output layer node through the initial connection weights;

[0177] The determining module 306 is used to determine the neighborhood range of the output layer based on the distance parameter;

[0178] The node weight module 307 is used to calculate and adjust the node weights of the output layer nodes and their neighborhood ranges that meet the minimum distance with the first matching parameter as the adjustment amount. If the number of iterations meets the preset conditions, the clustering result data of the rotated nodes is output.

[0179] The detection operation module 308 is used to input clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and to control the placement rod and the moving detection head to perform the detection operation of the seal through the first motor controller and the second motor controller.

[0180] Preferably, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the system includes:

[0181] The fourth calculation module is used to calculate the roll angle, pitch angle and yaw angle of the moving detection head using the accelerometer, magnetometer and gyroscope;

[0182] The fifth calculation module is used to calculate the swing characteristic parameters of the placement rod;

[0183] The correction module is used to correct the Kalman filter using the swing characteristic parameters to obtain the corrected Kalman filter;

[0184] The filtering operation module is used to filter the roll angle, pitch angle and yaw angle of the motor of the moving detection head through a modified Kalman filter to obtain optimized roll angle, pitch angle and yaw angle;

[0185] The adjustment module is used to transmit the optimized roll angle, pitch angle and yaw angle to the second motor controller of the moving detection head for data comparison and to adjust the operating angle of the second motor controller of the moving detection head.

[0186] Preferably, the determining module includes:

[0187] The generation submodule is used to generate multiple candidate solutions corresponding to the distance parameters;

[0188] The acquisition submodule is used to acquire the preset destruction operator and preset repair operator of the path of the rotated node;

[0189] The filtering submodule is used to calculate the parameter value of each operator in the search cycle, and filter out the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value;

[0190] A determination submodule is used to determine the optimal solution as the neighborhood range of the output layer.

[0191] Preferably, the second calculation module includes:

[0192] The sequence submodule is used to acquire the angle data sequence of the moving detection head;

[0193] The window filtering submodule is used to perform smooth window filtering on the angle data sequence to obtain a smooth angle sequence.

[0194] The smoothness index vector submodule is used to calculate the rate of change of the angle of the smooth angle sequence and obtain the smoothness index vector through the rate of change of the angle.

[0195] The similarity submodule is used to compare the smoothness index vector with the preset vectors in the preset detection pattern library.

[0196] The extraction submodule is used to extract the weight parameters of the preset vector with the highest similarity.

[0197] The calculation submodule is used to calculate the first matching parameter using the weight parameter and similarity.

[0198] Preferably, the node weight module includes:

[0199] The determination submodule is used to identify the output layer nodes that meet the minimum distance requirement as the best matching units;

[0200] The neighborhood function submodule determines the neighborhood function based on the distance parameter between the best matching unit and its neighboring nodes. This neighborhood function sets the update magnitude of each node within the neighborhood relative to the best matching unit. The neighborhood range is a topological neighborhood centered on the best matching unit.

[0201] The clustering result data submodule is used to set the first matching parameter as an adjustment amount for the best matching unit and its neighborhood nodes, adjust the weight vector of the BMU and its neighborhood nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

[0202] Preferably, the detection operation module includes:

[0203] The mapping submodule is used to map the clustering result data to the first motion parameters and the second running parameters;

[0204] The first control submodule is used to control the first motor controller to rotate the placement rod and adjust it to the target angle through the first motion parameters;

[0205] The second control submodule is used to control the second motor controller through the second motion parameters so that the moving detection head moves along the path trajectory.

[0206] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0207] Specific limitations regarding the precision testing system for seals can be found in the limitations of the precision testing methods for seals described above, and will not be repeated here. Each module in the aforementioned precision testing system for seals can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0208] The precision testing system for seals provided above can be used to perform the precision testing method for seals provided in any of the above embodiments, and has the corresponding functions and beneficial effects.

[0209] In one embodiment, a computer device is provided, which may be a control terminal corresponding to an atomizing air pump, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for precision detection of seals. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0210] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0211] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0212] The angle data of the moving detection head is obtained;

[0213] The placement rod was identified as a rotation node.

[0214] Calculate the initial connection weights corresponding to the rotated nodes;

[0215] The first matching parameter is calculated based on the angle data;

[0216] The distance parameter between the sample input vector and the output layer node is calculated using the initial connection weights;

[0217] The neighborhood range of the output layer is determined based on the distance parameter.

[0218] Using the first matching parameter as the adjustment amount, calculate the weights of the output layer nodes and the neighborhood range that meet the minimum distance adjustment. If the number of iterations meets the preset conditions, output the clustering result data of the rotated nodes.

[0219] The clustering results data are input to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head. The first motor controller and the second motor controller control the placement rod and the moving detection head to perform the detection operation of the seal.

[0220] Preferably, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the method includes:

[0221] The roll angle, pitch angle, and yaw angle of the moving detection head are calculated using the accelerometer, magnetometer, and gyroscope.

[0222] Calculate the swing characteristic parameters of the placement rod;

[0223] The Kalman filter is modified by the swing characteristic parameters to obtain the modified Kalman filter;

[0224] The roll angle, pitch angle, and yaw angle of the motor of the moving detection head are filtered by a modified Kalman filter to obtain optimized roll angle, pitch angle, and yaw angle.

[0225] The optimized roll angle, pitch angle, and yaw angle are transmitted to the second motor controller of the moving detection head for data comparison, and the operating angle of the second motor controller of the moving detection head is adjusted.

[0226] Preferably, determining the neighborhood range of the output layer based on the distance parameter includes:

[0227] Generate multiple candidate solutions corresponding to the distance parameters;

[0228] Obtain the preset destruction operator and preset repair operator for the path of the rotated node;

[0229] Calculate the parameter value of each operator in the search cycle, and select the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value;

[0230] The optimal solution is determined as the neighborhood range of the output layer.

[0231] Preferably, the step of calculating the first matching parameter based on the angle data includes:

[0232] Obtain the angle data sequence of the moving detection head;

[0233] A smoothed angle sequence is obtained by performing a smoothed window filter on the angle data sequence.

[0234] Calculate the rate of change of the angles in the smoothed angle sequence, and obtain the smoothness index vector from the rate of change of the angles.

[0235] The similarity between the smoothness index vector and the preset vectors in the preset detection pattern library;

[0236] Extract the weight parameters of the preset vector with the highest similarity;

[0237] The first matching parameter is calculated using the weight parameters and similarity.

[0238] Preferably, the step of calculating and adjusting the weights of the output layer nodes and their neighborhood ranges that satisfy the minimum distance using the first matching parameter as an adjustment amount, and outputting the clustering result data of the rotated nodes if the number of iterations meets a preset condition, includes:

[0239] The output layer node that meets the minimum distance requirement is determined as the best matching unit;

[0240] The neighborhood function is determined based on the distance parameter between the best matching unit and its neighboring nodes; this neighborhood function sets the update magnitude of each node within the neighborhood relative to the best matching unit. The neighborhood range is the topological neighborhood centered on the best matching unit.

[0241] For the best matching unit and its neighboring nodes, set the first matching parameter as the adjustment amount, adjust the weight vector of the BMU and its neighboring nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

[0242] Preferably, the step of inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller, includes:

[0243] The clustering results are mapped to the first motion parameters and the second running parameters;

[0244] The first motor controller is controlled by the first motion parameters to rotate the placement rod and adjust it to the target angle.

[0245] The second motor controller, controlled by the second motion parameters, makes the moving detection head move along the path trajectory.

[0246] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, performs the following steps:

[0247] The angle data of the moving detection head is obtained;

[0248] The placement rod was identified as a rotation node.

[0249] Calculate the initial connection weights corresponding to the rotated nodes;

[0250] The first matching parameter is calculated based on the angle data;

[0251] The distance parameter between the sample input vector and the output layer node is calculated using the initial connection weights;

[0252] The neighborhood range of the output layer is determined based on the distance parameter.

[0253] Using the first matching parameter as the adjustment amount, calculate the weights of the output layer nodes and the neighborhood range that meet the minimum distance adjustment. If the number of iterations meets the preset conditions, output the clustering result data of the rotated nodes.

[0254] The clustering results data are input to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head. The first motor controller and the second motor controller control the placement rod and the moving detection head to perform the detection operation of the seal.

[0255] Preferably, the moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the method includes:

[0256] The roll angle, pitch angle, and yaw angle of the moving detection head are calculated using the accelerometer, magnetometer, and gyroscope.

[0257] Calculate the swing characteristic parameters of the placement rod;

[0258] The Kalman filter is modified by the swing characteristic parameters to obtain the modified Kalman filter;

[0259] The roll angle, pitch angle, and yaw angle of the motor of the moving detection head are filtered by a modified Kalman filter to obtain optimized roll angle, pitch angle, and yaw angle.

[0260] The optimized roll angle, pitch angle, and yaw angle are transmitted to the second motor controller of the moving detection head for data comparison, and the operating angle of the second motor controller of the moving detection head is adjusted.

[0261] Preferably, determining the neighborhood range of the output layer based on the distance parameter includes:

[0262] Generate multiple candidate solutions corresponding to the distance parameters;

[0263] Obtain the preset destruction operator and preset repair operator for the path of the rotated node;

[0264] Calculate the parameter value of each operator in the search cycle, and select the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value;

[0265] The optimal solution is determined as the neighborhood range of the output layer.

[0266] Preferably, the step of calculating the first matching parameter based on the angle data includes:

[0267] Obtain the angle data sequence of the moving detection head;

[0268] A smoothed angle sequence is obtained by performing a smoothed window filter on the angle data sequence.

[0269] Calculate the rate of change of the angles in the smoothed angle sequence, and obtain the smoothness index vector from the rate of change of the angles.

[0270] The similarity between the smoothness index vector and the preset vectors in the preset detection pattern library;

[0271] Extract the weight parameters of the preset vector with the highest similarity;

[0272] The first matching parameter is calculated using the weight parameters and similarity.

[0273] Preferably, the step of calculating and adjusting the weights of the output layer nodes and their neighborhood ranges that satisfy the minimum distance using the first matching parameter as an adjustment amount, and outputting the clustering result data of the rotated nodes if the number of iterations meets a preset condition, includes:

[0274] The output layer node that meets the minimum distance requirement is determined as the best matching unit;

[0275] The neighborhood function is determined based on the distance parameter between the best matching unit and its neighboring nodes; this neighborhood function sets the update magnitude of each node within the neighborhood relative to the best matching unit. The neighborhood range is the topological neighborhood centered on the best matching unit.

[0276] For the best matching unit and its neighboring nodes, set the first matching parameter as the adjustment amount, adjust the weight vector of the BMU and its neighboring nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

[0277] Preferably, the step of inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the seal detection operation through the first motor controller and the second motor controller, includes:

[0278] The clustering results are mapped to the first motion parameters and the second running parameters;

[0279] The first motor controller is controlled by the first motion parameters to rotate the placement rod and adjust it to the target angle.

[0280] The second motor controller, controlled by the second motion parameters, makes the moving detection head move along the path trajectory.

[0281] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0282] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0283] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, embodiments of the present invention can take the form of entirely hardware embodiments, entirely software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0284] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of apparatus, terminal devices (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0285] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction methods implemented in a process. Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0286] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0287] Although preferred embodiments of the present invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the embodiments of the present invention.

[0288] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, apparatus, article, or terminal device that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, apparatus, article, or terminal device. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, apparatus, article, or terminal device that includes said element.

[0289] The foregoing has provided a detailed description of a precision testing method and system for sealing components, a computer device, and a storage medium provided by the present invention. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A precision inspection method for a sealing component, characterized in that, This invention relates to a fixed clamp and a movable detection head. The fixed clamp includes multiple placement rods arranged in a distributed manner for placing a sealing element. The movable detection head is used to detect the sealing element. The angle data of the moving detection head is obtained; The placement rod was identified as a rotation node. Calculate the initial connection weights corresponding to the rotated nodes; The first matching parameter is calculated based on the angle data; The distance parameter between the sample input vector and the output layer node is calculated using the initial connection weights; The neighborhood range of the output layer is determined based on the distance parameter. Using the first matching parameter as the adjustment amount, calculate the weights of the output layer nodes and the neighborhood range that satisfy the minimum distance. If the number of iterations meets the preset conditions, output the clustering result data of the rotated nodes. The clustering results data are input to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head. The first motor controller and the second motor controller control the placement rod and the moving detection head to perform the detection operation of the seal.

2. The method according to claim 1, characterized in that, The moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the method includes: The roll angle, pitch angle, and yaw angle of the moving detection head are calculated using the accelerometer, magnetometer, and gyroscope. Calculate the swing characteristic parameters of the placement rod; The corrected Kalman filter is obtained by modifying the swing characteristic parameters. The roll angle, pitch angle, and yaw angle of the motor of the moving detection head are filtered by a modified Kalman filter to obtain optimized roll angle, pitch angle, and yaw angle. The optimized roll angle, pitch angle, and yaw angle are transmitted to the second motor controller of the moving detection head for data comparison, and the operating angle of the second motor controller of the moving detection head is adjusted.

3. The method according to claim 1, characterized in that, Determining the neighborhood range of the output layer based on the distance parameter includes: Generate multiple candidate solutions corresponding to the distance parameters; Obtain the preset destruction operator and preset repair operator for the path of the rotated node; Calculate the parameter value of each operator in the search cycle, and select the optimal solution from multiple candidate solutions corresponding to the path of the rotating node based on the parameter value; The optimal solution is determined as the neighborhood range of the output layer.

4. The method according to claim 2, characterized in that, The calculation of the first matching parameter based on the angle data includes: Obtain the angle data sequence of the moving detection head; A smoothed angle sequence is obtained by performing a smoothed window filter on the angle data sequence. Calculate the rate of change of the angles in the smoothed angle sequence, and obtain the smoothness index vector from the rate of change of the angles. The similarity between the smoothness index vector and the preset vector in the preset detection pattern library; Extract the weight parameters of the preset vector with the highest similarity; The first matching parameter is calculated using the weight parameter and the similarity.

5. The method according to claim 1, characterized in that, The step involves using the first matching parameter as an adjustment amount to calculate and adjust the weights of the output layer nodes and their neighborhood ranges that satisfy the minimum distance. If the number of iterations meets a preset condition, the clustering result data of the rotated nodes is output, including: The output layer node that meets the minimum distance requirement is determined as the best matching unit; The neighborhood function is determined based on the distance parameter between the best matching unit and its neighboring nodes; this neighborhood function sets the update magnitude of each node within the neighborhood relative to the best matching unit. The neighborhood range is the topological neighborhood centered on the best matching unit. For the best matching unit and its neighborhood nodes, set the first matching parameter as the adjustment amount, adjust the weight vector of the BMU and its neighborhood nodes until the number of iterations meets the convergence condition, and output the clustering result data of the rotated nodes.

6. The method according to claim 5, characterized in that, The step of inputting the clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and controlling the placement rod and the moving detection head to perform the detection operation of the seal through the first motor controller and the second motor controller, includes: The clustering results are mapped to the first motion parameters and the second running parameters; The first motor controller is controlled by the first motion parameters to rotate the placement rod and adjust it to the target angle. The second motor controller, controlled by the second motion parameters, makes the moving detection head move along the path trajectory.

7. A precision inspection system for sealing components, characterized in that, This invention relates to a fixed clamp and a movable detection head. The fixed clamp includes multiple placement rods arranged in a distributed manner for placing a sealing element. The movable detection head is used to detect the sealing element. The first acquisition module is used to acquire the angle data of the moving detection head; The second acquisition module is used to acquire the placement rod as a rotation node; The first calculation module is used to calculate the initial connection weights corresponding to the rotated nodes; The second calculation module is used to calculate the first matching parameter based on the angle data; The third calculation module is used to calculate the distance parameters between the sample input vector and the output layer nodes using the initial connection weights; The determination module is used to determine the neighborhood range of the output layer based on the distance parameter; The node weight module is used to calculate and adjust the weights of the output layer nodes and their neighborhood ranges that meet the minimum distance using the first matching parameter as the adjustment amount. If the number of iterations meets the preset conditions, the clustering result data of the rotated nodes is output. The detection operation module is used to input clustering result data to the first motor controller corresponding to the placement rod and the second motor controller of the moving detection head, and to control the placement rod and the moving detection head to perform the detection operation of the seal through the first motor controller and the second motor controller.

8. The system according to claim 7, characterized in that, The moving detection head is equipped with an accelerometer, a magnetometer, and a gyroscope; the system includes: The fourth calculation module is used to calculate the roll angle, pitch angle and yaw angle of the moving detection head using the accelerometer, magnetometer and gyroscope; The fifth calculation module is used to calculate the swing characteristic parameters of the placement rod; The correction module is used to correct the Kalman filter using the swing characteristic parameters to obtain the corrected Kalman filter; The filtering operation module is used to filter the roll angle, pitch angle and yaw angle of the motor of the moving detection head through a modified Kalman filter to obtain optimized roll angle, pitch angle and yaw angle; The adjustment module is used to transmit the optimized roll angle, pitch angle and yaw angle to the second motor controller of the moving detection head for data comparison and to adjust the operating angle of the second motor controller of the moving detection head.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the precision detection method for the seal as described in any one of claims 1 to 6.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the precision detection method for the seal as described in any one of claims 1 to 6.

Images