Water pump forceps clamping assisting method and system based on visual positioning

Abstract

The invention provides a water pump clamp clamping assisting method and system based on visual localization, and relates to the field of machine vision, and the method comprises the steps: building a pipe fitting clamping simulation space in a three-dimensional clamping simulation platform according to a pipe fitting three-dimensional model and a jaw three-dimensional model; and on the basis of the clamping parameter adjustment space and the pipe fitting clamping simulation space, with maximization of the pipe fitting clamping stability and the pressure distribution uniformity as the target, iterative optimization search is conducted on the clamping parameters of the water pump clamp, the optimal clamping parameters are output, and the water pump clamp is controlled to conduct clamping operation on the to-be-clamped pipe fitting according to the optimal clamping parameters. The technical problems that clamping of traditional water pump forceps depends on manual experience, and clamping precision, uniformity and stability are difficult to consider are solved. By means of the method, accurate matching of geometrical characteristics of the pipe fitting and the jaw and intelligent adjustment of clamping parameters can be achieved, the surface damage rate of the pipe fitting can be reduced, and clamping precision and reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of machine vision, and in particular to a visual positioning-based water pump clamping auxiliary method and system. Background Technology

[0002] Pump clamps are a commonly used pipe clamping tool, widely used in pipe fixing, cutting, welding and assembly operations. In the traditional clamping process, operators mainly rely on experience to determine the clamping position, clamping force and closing amount. This method is greatly affected by human operation differences and is prone to insufficient clamping accuracy, affecting the processing or assembly quality.

[0003] Due to the lack of scientific quantification and optimization mechanisms, traditional water pump clamps result in uneven pressure distribution between the jaws and the pipe fittings during clamping, easily causing scratches, indentations, or deformation on the pipe surface. This uneven pressure is particularly damaging to thin-walled, precision, or fragile pipe fittings. Furthermore, traditional methods struggle to balance clamping stability and surface protection, and are ill-suited for pipe fittings of different materials, sizes, and surface characteristics, thus limiting the reliability and repeatability of automated clamping. Summary of the Invention

[0004] The purpose of this invention is to provide a vision-based positioning-based water pump clamping assistance method and system to solve the technical problem that traditional water pump clamping relies on manual experience and makes it difficult to simultaneously achieve clamping accuracy, uniformity, and stability. The method includes: In a first aspect, the present invention provides a visual positioning-based water pump clamping assistance method, comprising: acquiring an image of the surface of a pipe to be clamped, and simulating and constructing a three-dimensional model of the pipe by combining the pipe size information and the pipe material information; acquiring an image of the surface of the jaws of the water pump clamp, and simulating and constructing a three-dimensional model of the jaws by combining the jaw tooth pattern size information and the jaw tooth pattern material; constructing a pipe clamping simulation space in a three-dimensional clamping simulation platform based on the three-dimensional model of the pipe and the three-dimensional model of the jaws; and iteratively optimizing and searching the clamping parameters of the water pump clamp based on the clamping parameter adjustment space and the pipe clamping simulation space, with the goal of maximizing the clamping stability and pressure distribution uniformity of the pipe, outputting the optimal clamping parameters, and controlling the water pump clamp to perform clamping operations on the pipe to be clamped according to the optimal clamping parameters.

[0005] Preferably, the visual positioning-based pump clamping assistance method further includes: acquiring multi-angle images of the pipe fitting surface to be clamped using an industrial camera; inputting the multi-angle images of the pipe fitting surface into a first surface feature recognition plugin and outputting the pipe fitting surface feature distribution, wherein the first surface feature recognition plugin is constructed based on a convolutional neural network; and simulating and constructing a three-dimensional model of the pipe fitting based on the pipe fitting surface feature distribution, pipe fitting size information, and pipe fitting material information.

[0006] Preferably, the visual positioning-based water pump clamping assistance method further includes: acquiring multi-angle jaw surface images of the water pump clamp using an industrial camera; inputting the multi-angle jaw surface images into a second surface feature recognition plugin and outputting the jaw surface feature distribution, wherein the second surface feature recognition plugin is constructed based on a convolutional neural network; and simulating and constructing a three-dimensional model of the jaw based on the jaw surface feature distribution, jaw tooth size information, and jaw tooth material.

[0007] Preferably, the visual positioning-based water pump clamping assistance method further includes: configuring a clamping parameter adjustment space, wherein the clamping parameter adjustment space includes a clamping point threshold, a clamping angle threshold, a clamping force threshold, and a jaw closure amount threshold.

[0008] Preferably, the visual positioning-based pump clamping assistance method further includes: randomly selecting any parameter from the clamping point threshold, clamping angle threshold, clamping force threshold, and jaw closure threshold to obtain a first initial clamping parameter, and iteratively selecting Q initial clamping parameters, where Q is an integer greater than 200; within the pipe clamping simulation space, performing pipe clamping simulations according to the Q initial clamping parameters respectively, and outputting Q simulated anti-slip margins and Q simulated pressure distributions; setting a minimum anti-slip margin index based on the clamping task, and filtering the Q simulated anti-slip margins to obtain K qualified clamping parameters; constructing a parameter quality evaluation function with the goal of maximizing pipe clamping stability and pressure distribution uniformity, and evaluating K parameter quality coefficients based on the K simulated anti-slip margins and K simulated pressure distributions of the K qualified clamping parameters; based on the clamping parameter adjustment space and the pipe clamping simulation space, iteratively optimizing and searching the clamping parameters of the pump clamp according to the K qualified clamping parameters and K parameter quality coefficients, and outputting the optimal clamping parameters.

[0009] Preferably, the visual positioning-based pump clamping assistance method further includes: assessing the surface damage resistance of the pipe fitting based on the pipe fitting material information, clamp jaw tooth size information, and clamp jaw tooth material, and outputting the surface damage resistance strength of the pipe fitting; setting the ratio of the surface damage resistance strength of the pipe fitting to the preset standard damage resistance strength as a first weight adjustment coefficient, compensating the first initial weight according to the first weight adjustment coefficient to obtain a first adaptation weight, wherein the first adaptation weight is the weight of pressure distribution uniformity; subtracting the first adaptation weight from 1 to obtain a second adaptation weight, wherein the second adaptation weight is the weight of pipe fitting clamping stability; and constructing a parameter quality evaluation function based on the first adaptation weight and the second adaptation weight, with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity.

[0010] Preferably, the visual positioning-based pump clamping assistance method further includes: setting qualified clamping parameters as initial solutions; arranging K initial solutions in descending order of parameter quality coefficients to generate an initial solution sequence; selecting the first solution in the initial solution sequence as the optimal solution and the remaining K-1 initial solutions as inferior solutions; adjusting the K-1 inferior solutions according to a preset optimization step size, using the optimal solution as the adjustment direction, to obtain K-1 updated inferior solutions, and calculating K-1 parameter quality coefficients; reordering the optimal solution and the K-1 updated inferior solutions based on the K-1 parameter quality coefficients to obtain an updated solution sequence, eliminating inferior solutions in the updated solution sequence with a preset proportion, and supplementing the parameters with equal values ​​through the clamping parameter adjustment space, wherein the preset proportion decreases as the number of optimization attempts increases; continuing iterative optimization based on the optimization mechanism of optimal solution selection - inferior solution update - reordering - inferior solution elimination - inferior solution supplementation until a preset convergence number is reached, and outputting the optimal solution of the current updated solution sequence as the optimal clamping parameters.

[0011] Secondly, the present invention also provides a vision-based positioning water pump clamping assistance system for performing a vision-based positioning water pump clamping assistance method as described in the first aspect, comprising: a pipe fitting 3D model construction module for acquiring surface images of the pipe fitting to be clamped, and simulating and constructing a 3D model of the pipe fitting by combining pipe fitting size information and pipe fitting material information; a clamp jaw 3D model construction module for acquiring surface images of the clamp jaws of the water pump clamp, and simulating and constructing a 3D model of the clamp jaws by combining clamp jaw tooth pattern size information and clamp jaw tooth pattern material; a clamping simulation space construction module for constructing a pipe fitting clamping simulation space within a 3D clamping simulation platform based on the pipe fitting 3D model and the clamp jaw 3D model; and a clamping parameter optimization search module for iteratively optimizing and searching the clamping parameters of the water pump clamp based on the clamping parameter adjustment space and the pipe fitting clamping simulation space, with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity, outputting optimal clamping parameters, and controlling the water pump clamp to perform clamping operations on the pipe fitting to be clamped according to the optimal clamping parameters.

[0012] The embodiments of the present invention have the following advantages: By acquiring surface images of the pipe fitting to be clamped and combining them with the pipe fitting's size and material information, a 3D model of the pipe fitting is simulated and constructed. Simultaneously, by acquiring surface images of the water pump clamp's jaws and combining them with the jaw tooth size and material information, a 3D model of the jaws is simulated and constructed. Then, within a 3D clamping simulation platform, a pipe fitting clamping simulation space is constructed based on the pipe fitting's 3D model and the jaw's 3D model. Next, based on the clamping parameter adjustment space and the pipe fitting clamping simulation space, with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity, the clamping parameters of the water pump clamp are iteratively optimized and searched to output the optimal clamping parameters. Finally, the water pump clamp is controlled to clamp the pipe fitting according to the optimal clamping parameters. In other words, through visual acquisition, 3D modeling, and multi-objective optimization of clamping parameters, precise matching of the geometric features of the pipe fitting and the jaws and intelligent adjustment of clamping parameters can be achieved, maximizing clamping stability and improving pressure distribution uniformity, thereby reducing the surface damage rate of the pipe fitting and improving clamping accuracy and reliability. Attached Figure Description

[0013] Figure 1 This is a flowchart of the steps of a visual positioning-based water pump clamping auxiliary method of the present invention; Figure 2 This is a schematic diagram of the structure of a water pump clamping auxiliary system based on vision positioning according to the present invention.

[0014] Explanation of reference numerals in the attached figures: Pipe fitting 3D model building module 11, jaw 3D model building module 12, clamping simulation space building module 13, clamping parameter optimization search module 14. Detailed Implementation

[0015] This invention provides a vision-based water pump clamping assistance method and system, solving the technical problem that traditional water pump clamping relies on manual experience and struggles to balance clamping accuracy, uniformity, and stability. Through visual acquisition, 3D modeling, and multi-objective optimization of clamping parameters, precise matching of the pipe fitting's geometric features with the clamp jaws and intelligent adjustment of clamping parameters can be achieved. This maximizes clamping stability, improves pressure distribution uniformity, reduces pipe surface damage, and enhances clamping accuracy and reliability.

[0016] The technical solutions of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. It should also be noted that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, not all of them.

[0017] Example 1, please refer to the appendix. Figure 1 This invention provides a vision-based positioning-based water pump clamping assistance method, applied to a vision-based positioning-based water pump clamping assistance system, specifically including the following steps: S100: Acquire images of the pipe surface to be clamped, and simulate and construct a three-dimensional model of the pipe by combining the pipe size information and pipe material information.

[0018] Furthermore, step S100 of the present invention further includes: S110: Acquire multi-angle images of the pipe fitting surface to be clamped using an industrial camera; S120: Input the multi-angle images of the pipe fitting surface into a first surface feature recognition plugin and output the surface feature distribution of the pipe fitting, wherein the first surface feature recognition plugin is constructed based on a convolutional neural network; S130: Simulate and construct a three-dimensional model of the pipe fitting based on the surface feature distribution, pipe fitting size information, and pipe fitting material information.

[0019] Specifically, firstly, an industrial camera is used to acquire multi-angle surface images of the clamped pipe. The shooting distance and focal length are adjusted according to the pipe size to ensure clear imaging of the pipe surface features and obtain multi-angle images of the pipe surface.

[0020] Next, a first surface feature recognition plugin is constructed based on a convolutional neural network. This plugin consists of multiple convolutional layers (for extracting local texture features), pooling layers (for dimensionality reduction and feature concentration), batch normalization layers (for accelerating training and stabilizing convergence), and activation function layers (such as ReLU, used to introduce nonlinearity). It can effectively identify the microscopic and macroscopic features of the pipe fitting surface and is trained to convergence using sample data. Then, the collected multi-angle pipe fitting surface image sequence is input into the first surface feature recognition plugin. The network performs feature extraction and classification analysis on the images to generate a pipe fitting surface feature distribution map. The identified surface features include surface texture morphology, scratch location and length, height and diameter of pits or protrusions, weld direction and width, surface roughness level, as well as the spatial distribution density and confidence value of the features. Each feature point is labeled with its specific location and geometric parameters, providing an accurate data foundation for subsequent 3D model simulation construction based on pipe fitting size and material, and clamping pressure distribution calculation.

[0021] Then, in the 3D modeling platform, a basic geometric shape model of the pipe fitting (such as a cylinder or the overall outline of the pipe fitting) is generated based on the pipe fitting size information. The surface feature distribution is further mapped onto the model surface to form a point cloud. The Poisson surface reconstruction method is used to convert the point cloud into a continuous and smooth 3D surface, while retaining the true geometric shape of local micro-features such as scratches, pits, protrusions and welds. Finally, material properties are assigned to the model surface for subsequent clamping mechanics simulation and pressure distribution analysis, thereby obtaining a complete 3D model of the pipe fitting that has both accurate geometric features and material physical properties, providing basic data for intelligent clamping parameter optimization.

[0022] S200: Collects images of the jaw surface of the water pump clamp, and constructs a three-dimensional model of the jaw by combining the jaw tooth size information and jaw tooth material simulation.

[0023] Furthermore, step S200 of the present invention further includes: S210: Acquire multi-angle jaw surface images of the water pump clamp using an industrial camera; S220: Input the multi-angle jaw surface images into a second surface feature recognition plugin and output the jaw surface feature distribution, wherein the second surface feature recognition plugin is constructed based on a convolutional neural network; S230: Simulate and construct a three-dimensional model of the jaw based on the jaw surface feature distribution, jaw tooth pattern size information, and jaw tooth pattern material.

[0024] Specifically, multi-angle images of the jaw surface of the water pump clamp are acquired using an industrial camera, and the shooting distance, focal length, and exposure parameters are adjusted according to the jaw size to ensure that the tooth details and surface texture are clearly visible; finally, a multi-angle jaw surface image sequence is output, providing a data foundation for subsequent jaw feature recognition and 3D modeling.

[0025] Next, a second surface feature recognition plugin was constructed based on a convolutional neural network to extract the geometric and tooth pattern features of the water pump pliers jaws. This plugin consists of multiple convolutional layers (for capturing local details of the tooth patterns), pooling layers (for dimensionality reduction and feature aggregation), batch normalization layers (for accelerating training and stabilizing convergence), and activation function layers (such as ReLU, used to introduce nonlinearity). It can simultaneously identify key parameters such as jaw tooth pattern shape, tooth pitch, tooth height, and tooth tip angle, and accurately analyze the jaw surface texture, wear, or defect locations. Subsequently, multi-angle jaw surface images were input into the second surface feature recognition plugin. The network extracted and classified features from the images, outputting a jaw surface feature distribution map. The output features include the geometric dimensions, spatial location, orientation, tooth pitch, tooth tip angle, and local surface wear information for each tooth pattern, along with annotations of feature confidence and distribution density, providing a precise data foundation for subsequent 3D jaw modeling and clamping parameter optimization.

[0026] Then, within the 3D simulation platform, a complete geometric contour model of the jaws is generated based on the jaw teeth size information (such as tooth height, tooth pitch, tooth width, and tooth tip angle). Next, the tooth features are mapped onto the model surface to form a point cloud. Then, a continuous and smooth 3D surface is generated through Poisson surface reconstruction, while preserving the microscopic geometric features and wear information of the teeth. Finally, the tooth material properties are assigned to the model surface for subsequent pipe clamping mechanical analysis and pressure distribution simulation, thus obtaining a complete 3D jaw model that contains both accurate geometric features and material physical properties, providing a data foundation for clamping parameter optimization.

[0027] S300: Within the three-dimensional clamping simulation platform, a pipe clamping simulation space is constructed based on the three-dimensional model of the pipe fitting and the three-dimensional model of the jaws.

[0028] Specifically, within the 3D clamping simulation platform, the 3D models of the pipe fitting and the jaws are first imported into a unified coordinate system, and initial positioning and attitude correction are performed on the models to ensure that the relative positions of the pipe fitting and the jaws are consistent with the actual clamping state. Next, based on the geometry, surface features, and material properties of the pipe fitting, as well as the jaw's tooth geometry and material characteristics, the contact surface, contact area, and friction conditions are defined, providing a physical basis for clamping mechanics calculations. The forces and deformations of the pipe fitting are then preliminarily simulated using finite element or rigid body dynamics methods. Subsequently, a pipe fitting clamping simulation space is generated within the simulation platform to dynamically simulate the clamping state of the pump clamp under different clamping parameter conditions, including variables such as clamping point position, clamping angle, closing amount, and clamping force. This simulation space can calculate in real time the contact pressure distribution, force uniformity, and possible slippage or tilting trends between the jaws and the pipe fitting, and visualize the clamping state and potential stress concentration areas, providing a quantitative basis for clamping parameter optimization and achieving accurate simulation analysis that balances clamping stability and workpiece surface protection.

[0029] S400: Based on the clamping parameter adjustment space and the pipe clamping simulation space, with the goal of maximizing the pipe clamping stability and pressure distribution uniformity, the clamping parameters of the water pump clamp are iteratively optimized and searched to output the optimal clamping parameters, and the water pump clamp is controlled to perform clamping operations on the pipe to be clamped according to the optimal clamping parameters.

[0030] Furthermore, step S400 of the present invention further includes: S410: Configure clamping parameter adjustment space, wherein the clamping parameter adjustment space includes clamping point threshold, clamping angle threshold, clamping force threshold and jaw closure amount threshold.

[0031] Specifically, a clamping parameter adjustment space is configured within the 3D clamping simulation platform to systematically define the adjustable parameter ranges when the pump clamp grips the pipe fitting. This clamping parameter adjustment space includes clamping point thresholds, clamping angle thresholds, clamping force thresholds, and jaw closure thresholds. The clamping point thresholds limit the selectable clamping positions of the jaws on the pipe fitting surface; the clamping angle thresholds constrain the range of tilt angles of the jaws relative to the central axis of the pipe fitting; the clamping force thresholds limit the range of clamping force applied by the jaws to the pipe fitting; and the jaw closure thresholds control the jaw opening and closing amplitude, ensuring a stable clamping while avoiding excessive compression of the pipe fitting. By defining these parameter thresholds, the clamping parameter adjustment space can support systematic and controllable iterative optimization searches, providing operable constraints for subsequent clamping parameter optimization.

[0032] Furthermore, step S400 of the present invention further includes: S420: Randomly select any parameter from the clamping point threshold, clamping angle threshold, clamping force threshold, and jaw closure threshold to obtain the first initial clamping parameter, and iteratively select Q initial clamping parameters, where Q is an integer greater than 200; S430: In the pipe clamping simulation space, perform pipe clamping simulations according to the Q initial clamping parameters respectively, and output Q simulated anti-slip margins and Q simulated pressure distributions; S440: Based on the clamping task, set a minimum anti-slip margin index, and filter the Q simulated anti-slip margins to obtain K qualified clamping parameters.

[0033] Specifically, firstly, within the allowable range of clamping point threshold, clamping angle threshold, clamping force threshold, and jaw closure threshold, specific values ​​of each parameter are randomly selected and combined to generate the first set of initial clamping parameters for simulation analysis of the pipe clamping state. Subsequently, this process is repeated iteratively to generate Q sets of initial clamping parameters, where Q is an integer greater than 200, to ensure sufficient coverage of the parameter space and increase the diversity of the search. This provides a rich set of candidate parameters for subsequent optimization algorithms based on clamping stability and pressure distribution uniformity, thereby improving the efficiency and accuracy of clamping parameter optimization.

[0034] Next, the initial clamping parameters of group Q are sequentially input into the pipe clamping simulation space. For each group of parameters, pipe clamping simulation is performed. During the simulation, the force and friction distribution in the contact area between the jaws and the pipe are calculated, and the anti-slip performance of the pipe during the clamping process is evaluated. The corresponding simulated anti-slip margin is output. At the same time, the pressure distribution on the surface of the pipe is calculated, including the pressure magnitude, uniformity, and potential stress concentration areas. Through the simulation analysis of the initial parameters of group Q, Q anti-slip margin values ​​and Q pressure distribution data can be obtained, providing a quantitative basis for subsequent clamping parameter optimization and multi-objective optimization.

[0035] Then, a minimum anti-slip margin index is set according to the requirements of the specific clamping task to ensure that the pipe fitting has sufficient stability and anti-slip capability during clamping. Subsequently, Q simulated anti-slip margin values ​​are compared and screened one by one with this minimum index, eliminating clamping parameter combinations that are lower than the minimum anti-slip margin requirement, and retaining clamping parameters that meet the conditions. After screening, K sets of qualified clamping parameters are obtained. These parameters, while ensuring the clamping stability of the pipe fitting, provide candidate solutions for subsequent multi-objective optimization based on pressure distribution uniformity and surface protection.

[0036] S450: A parametric quality evaluation function is constructed with the goal of maximizing the clamping stability and pressure distribution uniformity of the pipe fitting. The K parametric quality coefficients are obtained by evaluating the K simulated anti-slip margins and K simulated pressure distributions of the K qualified clamping parameters.

[0037] Furthermore, step S450 of the present invention further includes: S451: Evaluate the surface damage resistance of the pipe fitting based on the pipe fitting material information, jaw tooth size information, and jaw tooth material, and output the surface damage resistance strength of the pipe fitting; S452: Set the ratio of the surface damage resistance strength of the pipe fitting to the preset standard damage resistance strength as the first weight adjustment coefficient, and compensate the first initial weight according to the first weight adjustment coefficient to obtain the first adaptation weight, wherein the first adaptation weight is the weight of pressure distribution uniformity; S453: Subtract the first adaptation weight from 1 to obtain the second adaptation weight, wherein the second adaptation weight is the weight of pipe fitting clamping stability; S454: Based on the first adaptation weight and the second adaptation weight, construct a parameter quality evaluation function with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity.

[0038] Specifically, firstly, by combining the pipe material information (such as hardness, elastic modulus, yield strength, etc.), the geometric dimensions of the jaw teeth (such as tooth height, tooth pitch, tooth width, and tooth tip angle), and the material properties of the jaw teeth (such as hardness, elastic modulus, and coefficient of friction), the surface stress and possible deformation of the pipe during the clamping process are analyzed, and a surface damage resistance assessment model for the pipe is established. Then, by calculating the local stress, friction distribution, and deformation amplitude in the contact area between the jaws and the pipe, the surface damage resistance strength index of the pipe is output. This index can quantify the ability of the pipe surface to resist scratches, indentations, or micro-damage under clamping force, providing a reliable reference for clamping parameter optimization and pressure distribution adjustment.

[0039] Next, the ratio of the surface damage resistance strength of the pipe fitting to the preset standard damage resistance strength is set as the first weight adjustment coefficient. This first weight adjustment coefficient reflects the relative superiority or inferiority of the current clamping parameters in protecting the pipe fitting surface. Subsequently, the initially set pressure distribution uniformity weight is dynamically compensated according to the first weight adjustment coefficient to generate a first adaptation weight. This allows the evaluation of pressure distribution uniformity to automatically adjust its importance in multi-objective optimization based on the pipe fitting's damage resistance, thereby achieving a balance between ensuring pipe fitting surface protection and pressure distribution optimization during clamping. Furthermore, 1 is subtracted from the first adaptation weight to obtain a second adaptation weight, where the second adaptation weight is the weight of pipe fitting clamping stability. This method achieves a dynamic balance between two types of evaluation indicators: when the pressure distribution uniformity weight increases, the clamping stability weight decreases accordingly; conversely, when the clamping stability weight increases, the pressure distribution uniformity weight decreases accordingly. This automatically coordinates the weight allocation between pressure uniformity and clamping stability during multi-objective optimization, ensuring that the optimization result balances pipe fitting surface protection and clamping reliability.

[0040] Then, a parameter quality evaluation function is constructed based on the first adaptation weight (pressure distribution uniformity weight) and the second adaptation weight (fitting clamping stability weight). This function aims to maximize fitting clamping stability and pressure distribution uniformity, and quantifies both types of indicators through a weighted approach. This ensures that in multi-objective optimization, both clamping mechanical stability and uniform stress and protection of the fitting surface are considered. Finally, using the parameter quality evaluation function, K parameter quality coefficients are obtained based on K simulated anti-slip margins and K simulated pressure distributions of the K qualified clamping parameters. Each parameter quality coefficient reflects the comprehensive performance of that set of clamping parameters in balancing clamping stability and pressure uniformity, providing a quantitative basis for subsequent iterative optimization search and optimal clamping parameter output.

[0041] S460: Based on the clamping parameter adjustment space and the pipe clamping simulation space, the clamping parameters of the water pump clamp are iteratively optimized and searched according to the K qualified clamping parameters and the K parameter quality coefficients, and the optimal clamping parameters are output.

[0042] Furthermore, step S460 of the present invention further includes: S461: Set the qualified clamping parameters as the initial solution, and arrange the K initial solutions in descending order of parameter quality coefficients to generate an initial solution sequence; S462: Select the first solution in the initial solution sequence as the optimal solution, and the remaining K-1 initial solutions as inferior solutions; S463: Using the optimal solution as the adjustment direction, adjust the K-1 inferior solutions according to a preset optimization step size to obtain K-1 updated inferior solutions, and calculate the K-1 parameter quality coefficients; S464: Based on the K-1 parameter quality coefficients, reorder the optimal solution and the K-1 updated inferior solutions respectively to obtain an updated solution sequence, and eliminate inferior solutions in the updated solution sequence with a preset proportion, and supplement the parameters with equal value through the clamping parameter adjustment space, wherein the preset proportion decreases with the increase of the optimization number; S465: Continue iterative optimization based on the optimization mechanism of optimal solution selection - inferior solution update - reordering - inferior solution elimination - inferior solution supplementation until a preset convergence number is reached, and output the optimal solution of the current updated solution sequence as the optimal clamping parameter.

[0043] Specifically, firstly, K qualified clamping parameters are used as initial solutions. These initial solutions are then sorted from largest to smallest based on their quality coefficients, generating an initial solution sequence that reflects the relative merits of each clamping parameter set in balancing clamping stability and pressure distribution uniformity. Subsequently, the first solution with the highest parameter quality coefficient is selected from the initial solution sequence as the optimal solution, representing the current best combination of clamping parameters. The remaining K-1 initial solutions are marked as inferior solutions and used for subsequent iterative optimization or local search to gradually approach the globally optimal clamping parameters. Then, using the selected optimal solution as the adjustment direction, the parameters of the remaining K-1 inferior solutions in the initial solution sequence are updated according to the preset optimization step size. That is, based on the differences between the optimal solution and each inferior solution in the parameter space such as clamping point, clamping angle, clamping force, and jaw closure, the parameters of the inferior solutions are finely adjusted along the direction of the optimal solution to generate K-1 updated inferior solutions. Subsequently, each group of updated inferior solutions is input into the parameter quality evaluation function to recalculate its parameter quality coefficient, which is used to evaluate the comprehensive performance of the updated clamping parameters in balancing clamping stability and pressure distribution uniformity, providing a quantitative basis for the next round of iterative optimization, and obtaining K-1 parameter quality coefficients.

[0044] Further, based on the calculated parameter quality coefficients of the K-1 updated inferior solutions, the superior solutions and the updated inferior solutions are reordered to generate a new sequence of updated solutions. During the sorting process, superior solutions remain at the front of the sequence, while the parameters of inferior solutions ranked lower are marked as potential elimination targets. This is used to control the gradual convergence of the search space, thereby ensuring the overall improvement of the quality of the clamping parameters in each iteration. Next, inferior solutions with a predetermined proportion in the updated solution sequence are eliminated, and their parameters are replenished by equivalent values ​​through the clamping parameter adjustment space to reduce the interference of low-quality parameters on iterative optimization. The predetermined proportion decreases as the number of optimization attempts increases. That is, a higher proportion is used in the early stage of optimization to expand the search range and increase parameter diversity; the proportion is gradually reduced in the later stage of optimization to improve search accuracy, making parameter updates more concentrated in the high-quality solution region. At the same time, the clamping parameter adjustment space is used to replenish the eliminated inferior solutions with equivalent parameters to ensure that the number of solution sequences remains stable, providing sufficient candidate parameters for the next round of iterative optimization.

[0045] Then, based on the iterative optimization mechanism constructed in steps S462 to S464, the clamping parameters are continuously optimized according to the process of "selecting excellent solutions - updating inferior solutions - reordering - eliminating inferior solutions - supplementing inferior solutions". In each iteration, the excellent solution guides the inferior solutions to converge to the high-quality parameter region, updates the parameters of the inferior solutions and recalculates the parameter quality coefficients. Then, the solution sequence is sorted and inferior solutions are eliminated. New equivalent inferior solutions are supplemented by adjusting the clamping parameters to ensure the dynamic adjustment of the search space and the stability of the optimization direction. This iterative process continues until the preset number of convergences is reached or the convergence condition is met, ensuring that the optimization effect of the clamping parameters fully reflects stability and pressure uniformity. After the iteration ends, the excellent solution with the highest parameter quality coefficient in the current updated solution sequence is output as the final optimal clamping parameter, which is used to control the clamping operation of the water pump clamp on the pipe, realizing the comprehensive optimization of workpiece surface protection, pressure distribution uniformity and clamping stability during the clamping process.

[0046] Finally, based on the optimal clamping parameters, the clamping point position, clamping angle, clamping force, and jaw closure amount are set to drive the water pump clamp to move precisely and complete the clamping action. This ensures that the jaws contact the pipe fitting in the predetermined posture, achieving uniform pressure distribution, minimizing damage to the pipe fitting surface, and maximizing clamping stability during the clamping process. This results in high-precision and high-reliability automated pipe fitting clamping operations.

[0047] In summary, the visual positioning-based water pump clamping assistance method provided by this invention has the following technical effects: By acquiring surface images of the pipe fitting to be clamped and combining them with the pipe fitting's size and material information, a 3D model of the pipe fitting is simulated and constructed. Simultaneously, by acquiring surface images of the water pump clamp's jaws and combining them with the jaw tooth size and material information, a 3D model of the jaws is simulated and constructed. Then, within a 3D clamping simulation platform, a pipe fitting clamping simulation space is constructed based on the pipe fitting's 3D model and the jaw's 3D model. Next, based on the clamping parameter adjustment space and the pipe fitting clamping simulation space, with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity, the clamping parameters of the water pump clamp are iteratively optimized and searched to output the optimal clamping parameters. Finally, the water pump clamp is controlled to clamp the pipe fitting according to the optimal clamping parameters. In other words, through visual acquisition, 3D modeling, and multi-objective optimization of clamping parameters, precise matching of the geometric features of the pipe fitting and the jaws and intelligent adjustment of clamping parameters can be achieved, maximizing clamping stability and improving pressure distribution uniformity, thereby reducing the surface damage rate of the pipe fitting and improving clamping accuracy and reliability.

[0048] Example 2: Based on the same inventive concept as the vision-based water pump clamping assistance method in the foregoing examples, this invention also provides a vision-based water pump clamping assistance system. Please refer to the appendix. Figure 2 The system includes: a pipe fitting 3D model construction module 11, used to acquire surface images of the pipe fitting to be clamped, and simulate and construct a 3D model of the pipe fitting by combining the pipe fitting size information and material information; a jaw 3D model construction module 12, used to acquire surface images of the jaws of the pump clamp, and simulate and construct a 3D model of the jaws by combining the jaw tooth pattern size information and jaw tooth pattern material; a clamping simulation space construction module 13, used to construct a pipe fitting clamping simulation space within the 3D clamping simulation platform based on the pipe fitting 3D model and the jaw 3D model; and a clamping parameter optimization search module 14, used to iteratively optimize and search the clamping parameters of the pump clamp based on the clamping parameter adjustment space and the pipe fitting clamping simulation space, with the goal of maximizing the pipe fitting clamping stability and pressure distribution uniformity, output the optimal clamping parameters, and control the pump clamp to perform clamping operations on the pipe fitting to be clamped according to the optimal clamping parameters.

[0049] Furthermore, the visual positioning-based water pump clamping auxiliary system is also used for: acquiring multi-angle images of the pipe fitting surface to be clamped using an industrial camera; inputting the multi-angle images of the pipe fitting surface into a first surface feature recognition plugin, and outputting the pipe fitting surface feature distribution, wherein the first surface feature recognition plugin is constructed based on a convolutional neural network; and simulating and constructing a three-dimensional model of the pipe fitting based on the pipe fitting surface feature distribution, pipe fitting size information, and pipe fitting material information.

[0050] Furthermore, the visual positioning-based water pump clamping auxiliary system is also used for: acquiring multi-angle jaw surface images of the water pump clamp using an industrial camera; inputting the multi-angle jaw surface images into a second surface feature recognition plugin and outputting the jaw surface feature distribution, wherein the second surface feature recognition plugin is constructed based on a convolutional neural network; and simulating and constructing a three-dimensional model of the jaw based on the jaw surface feature distribution, jaw tooth size information, and jaw tooth material.

[0051] Furthermore, the visual positioning-based water pump clamping auxiliary system is also used to: configure a clamping parameter adjustment space, wherein the clamping parameter adjustment space includes a clamping point threshold, a clamping angle threshold, a clamping force threshold, and a jaw closure amount threshold.

[0052] Furthermore, the vision-based water pump clamping auxiliary system is also used for: randomly selecting any parameter from the clamping point threshold, clamping angle threshold, clamping force threshold, and jaw closure threshold to obtain a first initial clamping parameter, and iteratively selecting Q initial clamping parameters, where Q is an integer greater than 200; within the pipe clamping simulation space, performing pipe clamping simulations according to the Q initial clamping parameters respectively, and outputting Q simulated anti-slip margins and Q simulated pressure distributions; setting a minimum based on the clamping task. The anti-slip margin index is used to screen the Q simulated anti-slip margins to obtain K qualified clamping parameters. A parameter quality evaluation function is constructed with the goal of maximizing pipe clamping stability and pressure distribution uniformity. K parameter quality coefficients are obtained based on the K simulated anti-slip margins and K simulated pressure distributions of the K qualified clamping parameters. Based on the clamping parameter adjustment space and the pipe clamping simulation space, the clamping parameters of the water pump clamp are iteratively optimized and searched according to the K qualified clamping parameters and the K parameter quality coefficients to output the optimal clamping parameters.

[0053] Furthermore, the visual positioning-based pump clamping auxiliary system is also used for: assessing the surface damage resistance of the pipe fitting based on the pipe fitting material information, clamp tooth pattern size information, and clamp tooth pattern material, and outputting the surface damage resistance strength of the pipe fitting; setting the ratio of the surface damage resistance strength of the pipe fitting to the preset standard damage resistance strength as a first weight adjustment coefficient, compensating the first initial weight according to the first weight adjustment coefficient to obtain a first adaptation weight, wherein the first adaptation weight is the weight of pressure distribution uniformity; subtracting the first adaptation weight from 1 to obtain a second adaptation weight, wherein the second adaptation weight is the weight of pipe fitting clamping stability; and constructing a parameter quality evaluation function based on the first adaptation weight and the second adaptation weight, with the goal of maximizing pipe fitting clamping stability and pressure distribution uniformity.

[0054] Furthermore, the visual positioning-based pump clamping auxiliary system is also used for: setting qualified clamping parameters as initial solutions; arranging K initial solutions in descending order of parameter quality coefficients to generate an initial solution sequence; selecting the first solution in the initial solution sequence as the optimal solution and the remaining K-1 initial solutions as inferior solutions; adjusting the K-1 inferior solutions according to a preset optimization step size with the optimal solution as the adjustment direction to obtain K-1 updated inferior solutions, and calculating K-1 parameter quality coefficients; reordering the optimal solution and the K-1 updated inferior solutions based on the K-1 parameter quality coefficients to obtain an updated solution sequence, eliminating inferior solutions in the updated solution sequence with a preset proportion, and supplementing the parameters with equal values ​​through the clamping parameter adjustment space, wherein the preset proportion decreases with the increase of the optimization number; continuing iterative optimization based on the optimization mechanism of optimal solution selection-inferior solution update-reordering-inferior solution elimination-inferior solution supplementation until a preset convergence number is reached, and outputting the optimal solution of the current updated solution sequence as the optimal clamping parameters.

[0055] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. The vision-based water pump clamping assistance method and specific examples in Embodiment 1 are also applicable to the vision-based water pump clamping assistance system of this embodiment. Through the foregoing detailed description of the vision-based water pump clamping assistance method, those skilled in the art can clearly understand the vision-based water pump clamping assistance system of this embodiment. Therefore, for the sake of brevity, it will not be described in detail here. As for the system disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple, and relevant parts can be referred to in the method section.

[0056] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0057] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of this invention and its equivalents, this invention also intends to include these modifications and variations.

Claims

1. A visual positioning-based water pump clamping assistance method, characterized in that the method... include: Acquire images of the pipe surface to be clamped, and simulate and construct a three-dimensional model of the pipe by combining the pipe size information and pipe material information; Collect images of the jaw surface of the water pump clamp, and construct a three-dimensional model of the jaw by combining the jaw tooth size information and jaw tooth material simulation. Within the three-dimensional clamping simulation platform, a pipe clamping simulation space is constructed based on the three-dimensional model of the pipe fitting and the three-dimensional model of the jaws. Based on the clamping parameter adjustment space and the pipe clamping simulation space, with the goal of maximizing the pipe clamping stability and pressure distribution uniformity, the clamping parameters of the water pump clamp are iteratively optimized and searched to output the optimal clamping parameters, and the water pump clamp is controlled to perform clamping operations on the pipe to be clamped according to the optimal clamping parameters.

2. The water pump clamping assistance method based on vision positioning according to claim 1, characterized in that, Acquire images of the pipe surface to be clamped, and combine these images with the pipe's dimensions and material information to simulate and construct a 3D model of the pipe, including: Multi-angle images of the pipe fitting surface to be clamped are captured using an industrial camera. The multi-angle pipe fitting surface image is input into the first surface feature recognition plugin, and the pipe fitting surface feature distribution is output. The first surface feature recognition plugin is constructed based on a convolutional neural network. A three-dimensional model of the pipe fitting is constructed based on the surface feature distribution, size information, and material information of the pipe fitting.

3. The water pump clamping assistance method based on vision positioning according to claim 1, characterized in that, Images of the jaws of the water pump clamp were collected, and a 3D model of the jaws was constructed by combining the jaw tooth size information and jaw tooth material simulation, including: Multi-angle images of the jaw surface of the water pump clamp were captured using an industrial camera. The multi-angle jaw surface image is input into the second surface feature recognition plugin, which outputs the jaw surface feature distribution. The second surface feature recognition plugin is constructed based on a convolutional neural network. A three-dimensional model of the jaws is constructed based on the surface feature distribution of the jaws, the size information of the jaw teeth, and the material of the jaw teeth.

4. The water pump clamping assistance method based on vision positioning according to claim 1, characterized in that, Configure clamping parameter adjustment space, wherein the clamping parameter adjustment space includes clamping point threshold, clamping angle threshold, clamping force threshold and jaw closure amount threshold.

5. The water pump clamping assistance method based on vision positioning according to claim 4, characterized in that, Based on the clamping parameter adjustment space and the pipe clamping simulation space, with the goal of maximizing pipe clamping stability and pressure distribution uniformity, the clamping parameters of the water pump clamp are iteratively optimized and searched to output the optimal clamping parameters, including: Randomly select any parameter from the clamping point threshold, clamping angle threshold, clamping force threshold, and jaw closure threshold to obtain the first initial clamping parameter, and iteratively select Q initial clamping parameters, where Q is an integer greater than 200; Within the pipe clamping simulation space, pipe clamping simulations are performed according to the Q initial clamping parameters, and Q simulated anti-slip margins and Q simulated pressure distributions are output. Based on the minimum anti-slip margin index set for the clamping task, the Q simulated anti-slip margins are screened to obtain K qualified clamping parameters; A parametric quality evaluation function is constructed with the goal of maximizing the clamping stability and pressure distribution uniformity of the pipe fittings. The K parameter quality coefficients are obtained by evaluating the K simulated anti-slip margins and K simulated pressure distributions of the K qualified clamping parameters. Based on the clamping parameter adjustment space and the pipe clamping simulation space, the clamping parameters of the water pump clamp are iteratively optimized and searched according to the K qualified clamping parameters and K parameter quality coefficients, and the optimal clamping parameters are output.

6. The visual positioning-based water pump clamping assistance method according to claim 5, characterized in that, A parametric quality evaluation function is constructed with the goal of maximizing pipe clamping stability and pressure distribution uniformity, including: Based on the pipe fitting material information, jaw tooth pattern size information, and jaw tooth pattern material, the surface damage resistance of the pipe fitting is assessed, and the surface damage resistance strength of the pipe fitting is output. The ratio of the surface damage resistance strength of the pipe fitting to the preset standard damage resistance strength is set as the first weight adjustment coefficient. The first initial weight is compensated according to the first weight adjustment coefficient to obtain the first adaptation weight, wherein the first adaptation weight is the weight of pressure distribution uniformity. The second adaptation weight is obtained by subtracting the first adaptation weight from 1, wherein the second adaptation weight is the weight of the pipe clamping stability; Based on the first and second adaptation weights, a parametric quality evaluation function is constructed with the goal of maximizing the clamping stability and pressure distribution uniformity of the pipe fitting.

7. The water pump clamping assistance method based on vision positioning according to claim 5, characterized in that, Based on the clamping parameter adjustment space and the pipe clamping simulation space, the clamping parameters of the water pump clamp are iteratively optimized and searched according to the K qualified clamping parameters and K parameter quality coefficients to output the optimal clamping parameters, including: Set the qualified clamping parameters as the initial solution, and arrange the K initial solutions in descending order of parameter quality coefficient to generate an initial solution sequence; The first solution in the initial solution sequence is selected as the optimal solution, and the remaining K-1 initial solutions are selected as the inferior solutions. Using the optimal solution as the adjustment direction, K-1 inferior solutions are adjusted according to the preset optimization step size to obtain K-1 updated inferior solutions, and K-1 parameter quality coefficients are calculated. Based on the K-1 quality coefficients, the optimal solution and the K-1 updated inferior solutions are reordered to obtain an updated solution sequence. Inferior solutions with a predetermined proportion in the updated solution sequence are eliminated, and the parameters are supplemented by equal value through the clamping parameter adjustment space. The predetermined proportion decreases as the number of optimization attempts increases. The optimization mechanism of selecting excellent solutions, updating inferior solutions, reordering, eliminating inferior solutions, and supplementing inferior solutions continues to iterate until the preset number of convergences is reached. The excellent solution of the current updated solution sequence is then set as the optimal clamping parameter.

8. A vision-based water pump clamping auxiliary system, characterized in that, The steps for implementing a vision-based positioning-based water pump clamping assistance method according to any one of claims 1 to 7 include: The pipe fitting 3D model construction module is used to acquire images of the pipe fitting surface to be clamped, and combine the pipe fitting size information and pipe fitting material information to simulate and construct a 3D model of the pipe fitting; The jaw 3D model construction module is used to acquire images of the jaw surface of the water pump clamp, and combine the jaw tooth size information and jaw tooth material simulation to construct a 3D model of the jaw. The clamping simulation space construction module is used to construct a pipe clamping simulation space within a three-dimensional clamping simulation platform based on the three-dimensional model of the pipe fitting and the three-dimensional model of the jaws. The clamping parameter optimization search module is used to iteratively optimize the clamping parameters of the water pump clamp based on the clamping parameter adjustment space and the pipe clamping simulation space, with the goal of maximizing the clamping stability and pressure distribution uniformity of the pipe fitting. It outputs the optimal clamping parameters and controls the water pump clamp to perform clamping operations on the pipe fitting to be clamped according to the optimal clamping parameters.

Images