Robot adaptive impeller welding control method and system fusing multi-modal data

Abstract

The application belongs to the field of robot control, and particularly relates to a robot adaptive impeller welding control method and system fusing multi-modal data, the method obtains impeller workpiece images and three-dimensional point cloud data through an industrial vision subsystem, and extracts a three-dimensional weld position point set representing a weld track; based on weld geometric feature analysis, a candidate weld path set is generated; combining weld appearance and process requirements, optimal welding process parameters are predicted through a regression model, and weld gun attitude angles are calculated using quaternion conversion according to surface normal vectors; during the welding process, the relative position deviation of the robot and the workpiece and the parameter deviation of the welding current / voltage and the planned value are monitored in real time, when the deviation exceeds the preset tolerance interval, a PID control algorithm is used to dynamically adjust the robot motion parameters or the welding process parameters, and closed-loop control of the welding process is realized; the application realizes accurate welding path planning and process adaptive regulation and control, and significantly improves welding quality and forming consistency.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of robot control, and particularly relates to a robot adaptive impeller welding control method and system fusing multi-modal data. BACKGROUND

[0002] In the field of energy and aerospace, the welds of complex curved impellers are mostly located between narrow and variable-curvature blades, which poses a severe challenge to the precision and adaptability of robot welding. The currently widely used traditional technologies mainly include prior offline programming relying on three-dimensional CAD models, path reproduction based on manual teaching, and weld tracking systems using two-dimensional laser vision. However, these existing technologies all have obvious limitations: offline programming cannot compensate for the pose deviation in actual workpiece processing and assembly, leading to the possibility of collision or deviation of the welding torch from the groove; the welding parameters and torch posture set by the teaching reproduction method are fixed values, which cannot be dynamically adjusted according to the surface curvature and groove shape, resulting in unstable welding quality; the two-dimensional vision system can achieve a certain degree of horizontal tracking, but lacks three-dimensional shape perception ability, and the process parameters still rely heavily on manual experience configuration, which has a long changeover and debugging period and insufficient production line flexibility; more importantly, most existing systems are in an open-loop execution state, lacking real-time monitoring and compensation mechanisms for dynamic disturbances such as welding thermal deformation and robot positioning drift, making it difficult to guarantee process consistency and product qualification rate in mass production.

[0003] The above existing technologies have the following problems: existing welding technology solutions often rely on preset trajectories and fixed parameters, and are limited to offline programming of ideal workpiece models. However, for impeller parts with complex curved surfaces and spatial welds, it is difficult to achieve high-quality automated welding by relying only on geometric models and ignoring the dynamic changes of actual working conditions. Especially in actual production, impellers have various disturbances such as pose errors, processing errors, and welding thermal deformation. Therefore, the present application provides an intelligent welding system and method based on machine vision and adaptive control. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application proposes a robot adaptive impeller welding control method and system fusing multi-modal data, which obtains impeller workpiece image and three-dimensional point cloud data through an industrial vision subsystem, and extracts a three-dimensional weld position point set representing a weld trajectory; based on weld geometric features, the weld direction, groove angle and accessibility are analyzed to generate a candidate weld path set; combining weld topography and process requirements, optimal welding process parameters are predicted through a regression model, and the weld gun attitude angle is calculated using quaternion conversion based on the surface normal vector; during the welding process, the relative position deviation of the robot and the workpiece and the parameter deviation of the welding current / voltage and the planned value are monitored in real time, and when the deviation exceeds the preset tolerance interval, the PID control algorithm is used to dynamically adjust the robot motion parameters or the welding process parameters to realize closed-loop control of the welding process. The present application solves the problems of inaccurate trajectory positioning, process parameter adaptive matching and real-time dynamic compensation in impeller welding, realizes precise planning of the welding path and adaptive process control, and significantly improves the welding quality and forming consistency.

[0005] To achieve the above object, the present application provides the following technical scheme:

[0006] The robot adaptive impeller welding control method fusing multi-modal data comprises:

[0007] The image and three-dimensional point cloud data of the impeller workpiece surface are obtained using the configured industrial vision subsystem, and a three-dimensional weld position point set representing the impeller weld trajectory is extracted;

[0008] Based on the three-dimensional weld position point set, the weld direction, groove angle and weld accessibility are analyzed and distinguished by combining surface geometric features analysis to generate a candidate weld path set;

[0009] Based on the weld topography features corresponding to the candidate weld path set and the welding process requirements, the optimal welding process parameter set is obtained through parameter prediction based on the regression model, and the weld gun attitude angle is obtained based on the surface normal vector corresponding to the candidate weld path set and combined with quaternion conversion;

[0010] The preset welding instruction containing the optimal welding process parameters and the weld gun attitude angle is responded, and the relative position deviation of the six-axis industrial robot end and the impeller workpiece weld trajectory, as well as the parameter deviation of the actual welding current or voltage, welding speed and process planning value, are monitored in real time during the welding process;

[0011] When the relative position deviation or the parameter deviation does not meet the corresponding preset process tolerance interval, the PID control algorithm is used to dynamically adjust the six-axis motion parameters or the welding process parameters according to the deviation type and size, so that the welding process is always within the preset process tolerance interval.

[0012] Specifically, the process of extracting the three-dimensional weld position point set representing the impeller weld trajectory comprises:

[0013] S105, acquire a left-right view image pair sequence of the impeller workpiece to be welded at the same time and three-dimensional point cloud data with the same timestamp, perform noise and contrast preprocessing by combining a Gaussian convolution and a limited contrast adaptive histogram equalization algorithm, and perform timestamp alignment to obtain an enhanced left-right view impeller workpiece image pair sequence;

[0014] S106, based on the enhanced left-right view impeller workpiece image pair sequence, calculate gradient components of images in horizontal and vertical directions respectively by combining a Sobel operator, and obtain gradient amplitude and gradient direction of each pixel point in each left-right view impeller workpiece image pair;

[0015] S107, based on the gradient amplitude and gradient direction, traverse each pixel point in the image, compare the gradient amplitudes of the pixel point and its adjacent pixel points along the gradient direction of the pixel point, retain the pixel points that are local maximum values in the gradient direction, and obtain an edge pixel candidate map;

[0016] S108, based on the edge pixel candidate map, discriminate and screen according to a preset screening rule, set all pixel points marked as final edges to foreground color, set all non-edge pixel points to background color, and generate binary contour maps respectively representing impeller weld seam edges under left-right views.

[0017] Specifically, the process of extracting the three-dimensional weld seam position point set representing the impeller weld seam trajectory further includes:

[0018] S109, based on the binary contour map, slide an evaluation window of a preset size on the image, calculate the gray scale change caused by moving the evaluation window in horizontal and vertical directions by a preset length at each pixel position on the contour map, and calculate the corner point response function value of each pixel based on the gray scale change;

[0019] S110, compare the calculated corner point response function value with a preset screening threshold, and screen out all pixel points whose corner point response function value exceeds the threshold as a candidate corner point set under left-right views;

[0020] S111, based on the candidate corner point set under left-right views, obtain a left image feature point set and a right image feature point set under left-right views by combining a non-maximum suppression process;

[0021] S112, based on the left image feature point set and the right image feature point set, process by combining a stereo vision matching algorithm to obtain a feature point matching pair set under left-right views;

[0022] S113, based on the feature point matching pair set under the left-right view angle, combining the internal parameter matrix corresponding to the left-right view angle, the lens distortion coefficient set and the relative spatial pose relationship between the cameras under the left-right view angle, the three-dimensional space coordinates of each feature point in the industrial vision subsystem coordinate system are obtained through the principle of triangulation.

[0023] Specifically, the process of extracting the three-dimensional weld position point set representing the impeller weld seam trajectory further includes:

[0024] S114, based on the three-dimensional space coordinates of each feature point in the industrial vision subsystem coordinate system, the first three-dimensional coordinates of each feature point in the robot base coordinate system are obtained through coordinate transformation operation, as the image three-dimensional weld position point set representing the impeller weld seam trajectory;

[0025] S115, based on the timestamp aligned three-dimensional point cloud data, the second three-dimensional coordinates of each feature point in the robot base coordinate system in the three-dimensional point cloud data are calculated as the point cloud three-dimensional weld position point set representing the impeller weld seam trajectory;

[0026] S116, based on the point cloud three-dimensional weld position point set and the image three-dimensional weld position point set, the position point deviation calculation is performed to obtain the image-point cloud feature position point deviation sequence;

[0027] S117, when at least one image-point cloud feature position point deviation is greater than a preset deviation threshold, the mapped image three-dimensional weld position point set and the point cloud three-dimensional weld position point set are adjusted in real time, so that the image-point cloud feature position point deviation is less than or equal to the preset deviation threshold in real time, and the three-dimensional weld position point set after secondary optimization is obtained.

[0028] Specifically, the process of generating the candidate weld path set includes:

[0029] S201, based on the three-dimensional weld position point set after secondary optimization, a local neighborhood point cloud is established for each feature point using K-nearest neighbor algorithm, and the surface normal vector and curvature feature of each local neighborhood are calculated by principal component analysis method;

[0030] S202, based on the surface normal vector, the curvature feature, and the weld seam direction and the bevel angle, the local surface is fitted by moving least squares method, and the Gaussian curvature and the average curvature at each feature point are calculated, and the absolute value of the average curvature is taken as the curvature feature value of the point;

[0031] S203, the base material and welding material type, heat input sensitive coefficient of the current impeller workpiece are obtained from the preset impeller base material and welding material impeller base material and welding material process database, and combined with the preset material-curvature-heat input correlation model and the bevel angle, the welding curvature threshold range under different material combinations and bevel shapes is determined;

[0032] S204, compare the curvature feature value with the welding curvature threshold range, and screen out feature points with a curvature feature value not exceeding a preset maximum allowed welding curvature threshold as preliminary path candidate points.

[0033] Specifically, the process of generating the candidate welding path set further includes:

[0034] S205, based on the obtained three-dimensional coordinates of the preliminary path candidate points, spatial clustering analysis is performed in combination with a DBSCAN clustering algorithm, noise points not satisfying a preset minimum number of neighborhood points are removed, and the distribution density of the preliminary path candidate points in the corresponding local surface area satisfies a preset path point distribution density threshold;

[0035] S206, based on the curvature feature value, the material heat input sensitivity coefficient, the groove angle, and the path point distribution density of each local surface area obtained in S205, a comprehensive welding accessibility score of each preliminary path candidate point is calculated through a linear weighting algorithm;

[0036] S207, based on the comprehensive welding accessibility score of each preliminary path candidate point obtained in S206, a sorting and screening algorithm is used to select the top N highest scoring feature points to form the final candidate welding path set.

[0037] Specifically, the process of obtaining the welding gun attitude angle includes:

[0038] S301, based on the candidate welding path set, the surface normal vector calculated in S201 for each candidate welding path point is obtained, and a target attitude of the welding gun is calculated in combination with a preset welding gun working angle and walking angle rule;

[0039] S302, the target attitude is converted into X-Y-Z axis rotation angles in the robot base coordinate system to obtain the welding gun attitude angle.

[0040] Specifically, the process of obtaining the optimal welding process parameter set includes:

[0041] S304, based on the candidate welding path set, the curvature feature value calculated in S201 for each candidate welding path point is obtained;

[0042] S305, based on a preset welding heat input base number calculation formula, in combination with the base material weldability parameters obtained through the impeller base material and welding material process database, a reference heat input value is calculated;

[0043] S306, based on the curvature feature value, a heat input coefficient of each candidate welding path point is determined through a preset curvature-heat input coefficient mapping relationship;

[0044] S307, multiply the reference heat input value by the heat input coefficient to obtain an optimal heat input value for each candidate welding path point;

[0045] S308, based on the optimal heat input value, calculate the corresponding welding current, voltage and welding speed parameters through a preset welding parameter mapping model to obtain an optimal welding process parameter set.

[0046] Specifically, in response to a preset welding instruction containing optimal welding process parameters and a welding gun attitude angle, the method comprises:

[0047] S401, based on the optimal welding process parameter set output by S308 and the welding gun attitude angle output by S302, generate a robot welding action instruction sequence containing target space coordinates corresponding to the candidate welding path, the welding gun attitude angle and the optimal welding process parameters;

[0048] S402, uniformly convert the target space coordinates and attitude parameters in the action instruction generated by S401 to executable instructions in the robot base coordinate system;

[0049] S403, based on the converted executable instructions in the robot base coordinate system, calculate the target angles of each joint axis of the robot through a robot inverse kinematics algorithm, and generate a smooth motion trajectory;

[0050] S404, during the welding process, the following monitoring tasks are performed synchronously:

[0051] S404-1, based on the industrial vision subsystem, real-time acquisition of the relative position between the robot end and the impeller workpiece at a preset sampling frequency, and calculation of the position deviation;

[0052] S404-2, based on the welding power supply system, real-time acquisition of the actual welding current and voltage at a preset sampling rate, and calculation of the parameter deviation from the optimal welding process parameters obtained by S308;

[0053] S405, when the monitored position deviation exceeds the preset position deviation threshold or the parameter deviation exceeds the preset parameter deviation threshold, trigger the corresponding compensation mechanism, which comprises:

[0054] S405-1, for the position deviation, calculate the motion correction amount of each axis of the robot based on the vision positioning deviation compensation algorithm, and output to the robot controller to adjust the motion trajectory;

[0055] S405-2, for the parameter deviation, calculate the welding parameter adjustment amount based on the process parameter compensation algorithm, and adjust the welding current and voltage through the welding power supply system;

[0056] S406, after executing the compensation, re-execute the monitoring task of S404 to verify whether the position deviation and the parameter deviation have returned to the preset deviation range.

[0057] S407. When both the position deviation and parameter deviation are stable within the preset deviation range, and the current welding path segment is completed, the current welding stage is determined to be successfully executed, and preparations are made to execute the welding command for the next welding path segment.

[0058] The robot adaptive impeller welding control system, which integrates multimodal data, includes: a vision module, an analysis module, an attitude parameter module, a response module, and a discrimination and adjustment module.

[0059] The vision module uses the configured industrial vision subsystem to acquire images and three-dimensional point cloud data of the impeller workpiece surface, and extracts a set of three-dimensional weld position points that characterize the impeller weld trajectory.

[0060] The analysis module, based on a three-dimensional weld location point set and combined with surface geometric feature analysis, analyzes and judges the weld direction, bevel angle and welding accessibility, and generates a set of candidate welding paths.

[0061] The attitude parameter module obtains the optimal welding process parameter set by parameter prediction based on regression model, based on the weld morphology features and welding process requirements corresponding to the candidate welding path set. At the same time, it obtains the welding gun attitude angle by combining the surface normal vector corresponding to the candidate welding path set with quaternion transformation.

[0062] The response module responds to preset welding commands containing optimal welding process parameters and welding torch attitude angles, and monitors in real time the relative positional deviation between the end effector of the six-axis industrial robot and the weld seam trajectory of the impeller workpiece, as well as the parameter deviations between the actual welding current or voltage, welding speed and process planning values.

[0063] When the relative position deviation or parameter deviation does not meet the corresponding preset process tolerance range, the discrimination and adjustment module dynamically adjusts the six-axis motion parameters or welding process parameters according to the type and magnitude of the deviation, so that the welding process is always within the preset process tolerance range.

[0064] Compared with the prior art, the beneficial effects of the present invention are:

[0065] The present application aims at the deficiencies of the prior art, by configuring a dual-camera vision subsystem, using multi-source data fusion technology to accurately obtain the three-dimensional weld position point set of the impeller weld, realizing sub-millimeter level accurate positioning; based on the analysis of the geometric characteristics of the curved surface, the optimal welding path is automatically generated, the welding process parameters are predicted by the regression model combined with the material properties and curvature characteristics, and the welding gun attitude angle is planned according to the normal vector of the curved surface, realizing the adaptive matching of the welding parameters; during the welding execution process, the relative position deviation and parameter deviation of the robot and the workpiece are monitored in real time, the PID control algorithm is used to dynamically adjust the motion parameters and process parameters, forming a closed loop control, and effectively compensating the interference such as thermal deformation. The present application significantly improves the welding precision and quality consistency, greatly reduces the debugging time, enhances the production line flexibility, realizes the full-process automation and intelligentization of complex curved surface impeller welding. BRIEF DESCRIPTION OF DRAWINGS

[0066] Figure 1 The present application provides a robot adaptive impeller welding control method based on multi-modal data flow chart;

[0067] Figure 2 The present application provides a robot adaptive impeller welding control system module diagram based on multi-modal data. DETAILED DESCRIPTION

[0068] Embodiment 1

[0069] Please refer to Figure 1 The present application provides an embodiment: a robot adaptive impeller welding control method based on multi-modal data, the steps comprising:

[0070] S1, using the configured industrial vision subsystem to obtain the image and three-dimensional point cloud data of the impeller workpiece surface, and extracting the three-dimensional weld position point set representing the impeller weld trajectory;

[0071] S2, based on the three-dimensional weld position point set, combining the curved surface geometric feature analysis, analyzing and judging the weld direction, groove angle and weld accessibility, and generating a candidate weld path set;

[0072] S3, based on the weld appearance characteristics corresponding to the candidate set of welding paths and the welding process requirements, obtaining the optimal set of welding process parameters through parameter prediction based on the regression model, and simultaneously obtaining the welding torch posture angle based on the normal vector of the surface corresponding to the candidate set of welding paths and combining the quaternion conversion; the weld appearance characteristics in the embodiment mainly include: the spatial three-dimensional coordinates of the weld trajectory, the tangent direction of the weld direction, the groove angle, the groove depth, the local surface curvature characteristic value, the surface normal vector direction, the weld cross-sectional area, and the welding accessibility between the blades; the welding process requirements in the embodiment mainly include: the type and heat input sensitive coefficient of the base material and welding material, the recommended unit length heat input range, the deposited metal amount requirement, the welding material melting efficiency, the heat input adjustment coefficient, the matching range of welding current and voltage, the welding speed constraint, the interlayer temperature control requirement, and the quality standard of weld formation.

[0073] S4, in response to the preset welding instruction containing the optimal welding process parameters and the welding torch posture angle, and real-time monitoring the relative position deviation between the six-axis industrial robot end and the impeller workpiece weld trajectory in the welding process, as well as the parameter deviation between the actual welding current or voltage, welding speed and process planning value;

[0074] S5, when the relative position deviation or the parameter deviation does not meet the corresponding preset process tolerance interval, according to the deviation type and size, the PID control algorithm is used to dynamically adjust the six-axis motion parameters or the welding process parameters, so that the welding process is always within the preset process tolerance interval.

[0075] It needs to be further explained that the industrial vision subsystem in the embodiment includes two 12 million pixel industrial cameras, which are installed on both sides of the workstation rack, with a shooting angle of 60°, covering the workpiece operation area; it needs to be further explained that the process of extracting the three-dimensional weld position point set representing the impeller weld trajectory in the embodiment includes:

[0076] S101, obtaining an image sequence of a calibration board containing a known two-dimensional planar pattern at multiple different spatial poses, based on the correspondence between the two-dimensional image points and the three-dimensional spatial points extracted from the image sequence, and through a nonlinear optimization algorithm, the internal parameter matrix and lens distortion coefficient set of each camera are solved; the nonlinear optimization algorithm is preferably Levenberg-Marquardt algorithm;

[0077] S102, based on the camera internal parameters obtained in S101, synchronously collecting a plurality of stereoscopic image pairs of the same calibration board in different spatial poses by an industrial vision subsystem, calculating a relative spatial pose relationship between the two cameras based on the matching feature points extracted from each stereoscopic image pair, and establishing a complete spatial transformation model between the left and right camera coordinate systems; wherein the relative spatial pose relationship includes a three-dimensional space rotation matrix, a three-dimensional translation vector, and a physical baseline distance between the optical centers of the left and right cameras; the physical baseline distance between the optical centers of the left and right cameras is calculated based on the three-dimensional translation vector combined with the Euclidean norm formula;

[0078] S103, controlling the robot end effector to fixedly carry the calibration board, and performing at least three non-coplanar spatial pose transformations in the robot workspace, and synchronously completing the following operations at each pose:

[0079] collecting a complete image of the calibration board by the industrial vision subsystem;

[0080] reading and recording the accurate pose of the end effector in the robot base coordinate system at this time from the robot controller;

[0081] S104, based on the plurality of sets of synchronous data obtained in S103, i.e., each set containing the pose of the end effector in the robot base coordinate system, and the pose of the calibration board in the industrial vision subsystem coordinate system solved by the image of the calibration board collected at the corresponding time, establishing a hand-eye calibration equation, and solving a fixed homogeneous transformation matrix from the industrial vision subsystem coordinate system to the robot base coordinate system by a least squares optimization algorithm; it needs to be further explained that the construction of the industrial vision subsystem coordinate system in the embodiment takes the optical center of the left camera as the coordinate origin, the Z axis is along the optical axis direction of the left camera, the X axis is parallel to the horizontal direction of the image plane, and the Y axis is perpendicular to the X-Z plane to form a right-hand coordinate system; the construction of the robot base coordinate system takes the center of the robot mounting base as the coordinate origin, the Z axis is vertically upward, the X axis is along the horizontal direction of the front of the robot, and the Y axis is determined according to the right-hand rule; the two coordinate systems establish an accurate spatial mapping relationship through the hand-eye calibration equation obtained in S104, which is obtained by the least squares optimization algorithm in S104 based on the non-coplanar pose transformation of the robot end effector carrying the calibration board in S103 and the binocular vision system model established in S102.

[0082] S105, collecting a left-right view image pair sequence of a to-be-welded impeller workpiece at the same time and three-dimensional point cloud data with the same timestamp, combining a Gaussian convolution and a limited contrast self-adaptive histogram equalization algorithm to perform noise and contrast preprocessing and timestamp alignment, and obtaining an enhanced left-right view impeller workpiece image pair sequence and timestamp-aligned three-dimensional point cloud data;

[0083] S106, based on the enhanced left-right view vane workpiece image pair sequence, gradient components of the images in horizontal and vertical directions are calculated respectively by using a Sobel operator, to obtain gradient amplitude and gradient direction of each pixel point in each left-right view vane workpiece image pair;

[0084] S107, based on the gradient amplitude and gradient direction, each pixel point in the image is traversed, gradient amplitudes of the pixel point and its adjacent pixel points are compared along the gradient direction of the pixel point, and pixel points with local maximum gradient direction are retained to obtain an edge pixel candidate map;

[0085] S108, based on the edge pixel candidate map and a preset screening rule, discrimination and screening are performed, all pixel points marked as final edges are set as foreground color, and all non-edge pixel points are set as background color, to generate a binary contour map respectively representing vane weld edge under left-right view angles; wherein the screening rule includes:

[0086] pixel points with gradient amplitude higher than a preset high threshold value are determined as strong edge pixels and directly marked as final edges;

[0087] pixel points with gradient amplitude lower than a preset low threshold value are determined as non-edge pixels and removed;

[0088] pixel points with gradient amplitude between the high threshold value and the low threshold value are determined as weak edge pixels, and discrimination is performed based on their connectivity: if the weak edge pixel is connected with any strong edge pixel, the weak edge pixel is marked as a final edge, otherwise, the weak edge pixel is removed; in this embodiment, the high threshold value and the low threshold value are set based on the gradient amplitude statistical distribution of the edge pixel candidate map; wherein the high threshold value is determined based on a preset high quantile of the gradient amplitude, and the low threshold value is determined based on the product of the high threshold value and a preset proportion factor; the high quantile and the proportion factor are configured according to image noise level and edge detection sensitivity requirement.

[0089] S109, based on the binary contour map, an evaluation window with a preset size is slid on the image, the gray scale variation caused by the movement of the evaluation window by a preset length in horizontal and vertical directions at each pixel position on the contour map is calculated, and the corner point response function value of each pixel is calculated based on the gray scale variation; it needs to be further explained that, in this embodiment, the corner point response function value is calculated to detect points with sharp gray scale variation in the image, and these points correspond to points with sudden curvature change, such as corner edges, corner points or edges of positioning reference holes, and corner points of process grooves.

[0090] S110, the calculated corner point response function value is compared with a preset screening threshold value, and all pixel points with corner point response function value exceeding the threshold value are selected as a candidate corner point set under left-right view angles;

[0091] S111, based on the candidate corner set under left-right view angle, the left image feature point set and the right image feature point set under left-right view angle are obtained by combining the non-maximum suppression process; in this embodiment, the candidate corner set is extracted by the non-maximum suppression process, and the left image feature point set and the right image feature point set are output, wherein the left image feature point set and the right image feature point set directly contain the pixel coordinates of the key geometric features such as the leading edge and the trailing edge of the impeller blade, the intersection of the blade and the hub, and the start and end points of the weld. Each feature point in the left image feature point set and the right image feature point set corresponds to a specific key geometric feature on the surface of the impeller workpiece, including: the curvature mutation point of the leading edge and the trailing edge of the blade: represented by the local maximum value of the corner response function value; the positioning reference hole on the impeller: represented by the corner point or the center feature point of the hole edge; and the process heat dissipation groove on the blade surface: represented by the corner point or the profile mutation point of the groove edge.

[0092] It should be further pointed out that the specific steps of the non-maximum suppression process in this embodiment include:

[0093] c1. For each candidate corner in the candidate corner set, a rectangular neighborhood of a predetermined size is defined with it as the center;

[0094] c2. Traverse each candidate corner, and compare the corner response function value of the candidate corner with the corner response function values of all other candidate corners in its defined neighborhood;

[0095] c3. If the corner response function value of the current candidate corner is the maximum value in its defined neighborhood, it is retained; otherwise, it is suppressed and removed from the candidate corner set;

[0096] c4. After traversing all candidate corners and performing steps c2 and c3, the finally retained candidate corner set is output as the refined corner set;

[0097] S112, based on the left image feature point set and the right image feature point set, the feature point matching pair set under left-right view angle is obtained by combining the stereo vision matching algorithm; wherein the specific process of combining the stereo vision matching algorithm for processing includes:

[0098] S1121, based on the spatial transformation relationship between the left and right camera image coordinate systems established in S102, the original image containing the feature points is rectified, so that the corresponding epipolar lines of the left and right images are horizontally aligned;

[0099] S1122, for each feature point in the left image feature point set and the right image feature point set, a feature descriptor representing local texture features is calculated based on the gray scale distribution of its neighborhood pixels; the construction process of the feature descriptor includes:

[0100] S1122a, based on each feature point in the left / right image feature point set obtained in S111, extract a neighborhood window of a preset size centered on the feature point coordinates;

[0101] S1122b, based on the gradient amplitude and gradient direction obtained in S106, calculate a gradient direction histogram in the neighborhood window, divide 360 degrees of direction into a preset number of direction intervals;

[0102] S1122c, based on the edge distribution features in the binary contour map, calculate three texture feature parameters of edge density, edge direction consistency and local contrast in the neighborhood;

[0103] S1122d, combine the corner response function value obtained in S109 to encode feature point type information into the descriptor;

[0104] S1122e, normalize the extracted gradient direction histogram and texture feature parameters to eliminate the influence of light changes;

[0105] S1122f, convert the normalized feature vector into a compact binary descriptor format;

[0106] S1122g, based on the neighborhood size of the feature point retained in the non-maximum suppression process in S111, adaptively adjust the scale parameter of the descriptor;

[0107] S1123, for each feature point in the left image feature point set, in a preset search range on the same scan line of the right image after correction, based on the similarity measure of the feature descriptor, traverse to obtain the candidate matching feature point with the highest similarity;

[0108] S1124, perform bidirectional consistency verification on the preliminary matched feature point pairs, eliminate the false matching pairs that do not satisfy the uniqueness constraint, and obtain a feature point matching pair set;

[0109] S113, based on the feature point matching pair set under the left-right view angles, combined with the internal parameter matrix corresponding to the left-right view angles, the lens distortion coefficient set and the relative spatial pose relationship between the cameras under the left-right view angles, calculate the three-dimensional space coordinates of each feature point in the industrial vision subsystem coordinate system through the triangulation principle, the specific process includes:

[0110] S1131, for each feature point matching pair, calculate the horizontal coordinate difference value in the image pair after correction in S1121 to obtain the disparity value of the feature point;

[0111] S1132, based on the obtained disparity value, the camera intrinsic parameter matrix obtained in S101 and the physical baseline distance obtained in S102, calculate the three-dimensional space coordinates of the feature point in the industrial vision subsystem coordinate system through the triangulation principle;

[0112] S1133, based on the lens distortion coefficient obtained in S101, performing distortion correction on the three-dimensional coordinates calculated in S1132 to obtain three-dimensional space coordinates of the feature points in the industrial vision subsystem coordinate system;

[0113] S114, based on the three-dimensional space coordinates of the feature points in the industrial vision subsystem coordinate system, combining the fixed homogeneous transformation matrix from the industrial vision subsystem coordinate system to the robot base coordinate system, and through coordinate transformation operation, obtaining the first three-dimensional coordinates of the feature points in the robot base coordinate system as the image three-dimensional weld position point set representing the impeller weld seam trajectory;

[0114] The specific process of the coordinate transformation operation includes:

[0115] S1141, applying the fixed homogeneous transformation matrix obtained in S104 to perform coordinate transformation operation on the three-dimensional space coordinates of each feature point in the industrial vision subsystem coordinate system to calculate the preliminary three-dimensional coordinates of the feature points in the robot base coordinate system;

[0116] S1142, performing validity verification on all the preliminary three-dimensional coordinates of the feature points obtained through the coordinate transformation operation, eliminating abnormal coordinate points beyond the robot working space, and integrating all the three-dimensional coordinates of the feature points that pass the verification into a feature point three-dimensional coordinate set in the robot base coordinate system;

[0117] S1143, performing spatial distribution optimization processing on the feature point three-dimensional coordinate set in the robot base coordinate system to remove redundant feature points and ensure the spatial distribution uniformity of key position points, and obtaining the optimized three-dimensional weld position point set representing the impeller weld seam trajectory.

[0118] S115, based on the timestamp-aligned three-dimensional point cloud data, combining the fixed homogeneous transformation matrix obtained in S104 to perform coordinate transformation operation to calculate the second three-dimensional coordinates of the feature points in the robot base coordinate system in the three-dimensional point cloud data as the point cloud three-dimensional weld position point set representing the impeller weld seam trajectory;

[0119] S116, based on the point cloud three-dimensional weld position point set and the image three-dimensional weld position point set, performing position point deviation calculation to obtain an image-point cloud feature position point deviation sequence;

[0120] S117, when at least one image-point cloud feature position point deviation is greater than a preset deviation threshold, performing real-time adjustment on the mapped image three-dimensional weld position point set and the point cloud three-dimensional weld position point set so that the image-point cloud feature position point deviation is less than or equal to the preset deviation threshold in real time, and obtaining a twice-optimized three-dimensional weld position point set.

[0121] The three-dimensional weld seam trajectory point set acquisition method realizes sub-millimeter positioning of the complex curved surface weld of the impeller by constructing a high-precision visual measurement model and a multi-source information fusion mechanism. In particular, firstly, high-precision calculation is performed on camera internal participation and non-coplanar hand-eye calibration through the Levenberg-Marquardt algorithm, laying the foundation for accurate measurement of the entire system. Secondly, adaptive double-threshold edge detection based on gradient amplitude statistics and non-maximum suppression corner extraction are used, which can sensitively capture the curvature mutation characteristics of the blade edge, weld groove and effectively suppress noise interference. In addition, in terms of three-dimensional reconstruction and fusion, on the one hand, the robustness of stereo matching is ensured through SIFT-like binary descriptors and bidirectional consistency testing, and high-confidence image three-dimensional point sets are obtained through triangulation and lens distortion correction. On the other hand, two-dimensional image features and three-dimensional point cloud data are mapped and deviation analyzed in the robot base coordinate system through timestamp alignment, and the absolute spatial accuracy of the point cloud is used to cross-verify and compensate the image reconstruction results. This fusion strategy effectively overcomes the limitations of single visual technology being easily affected by workpiece surface reflection and single texture. Finally, through spatial distribution optimization and multi-source data closed-loop adjustment, the secondary optimized three-dimensional weld seam position point set not only has better single-point accuracy, but also more completely and uniformly covers the spatial trajectory of the impeller weld, providing reliable spatial topological information and quality assurance for subsequent adaptive planning of welding paths and process parameters, significantly improving the success rate of complex curved surface welding and the quality of weld formation.

[0122] It should be further explained that the process of generating the candidate weld path set in the embodiment includes:

[0123] S201, based on the secondary optimized three-dimensional weld seam position point set, a K-nearest neighbor algorithm is used to establish a local neighborhood point cloud for each feature point, and a principal component analysis method is used to calculate the surface normal vector and curvature feature of each local neighborhood, wherein the specific acquisition process of the surface normal vector and curvature feature of each local neighborhood in the embodiment includes:

[0124] S2011, based on the secondary optimized three-dimensional weld seam position point set in the robot base coordinate system obtained in S116, a K-nearest neighbor algorithm is used to search for a preset number of nearest neighbor feature points for each feature point, and a local neighborhood point cloud is constructed;

[0125] S2012, based on the three-dimensional coordinates of the local neighborhood point cloud, the covariance matrix of each local neighborhood is calculated;

[0126] S2013, eigenvalue decomposition is performed on the covariance matrix to obtain three eigenvalues and their corresponding eigenvectors; wherein: the maximum eigenvalue represents the distribution discrete degree of the point cloud in the main direction of the tangent plane, the intermediate eigenvalue represents the distribution discrete degree in the secondary direction of the tangent plane, and the minimum eigenvalue and its corresponding eigenvector jointly determine the surface geometric characteristics, the eigenvector defines the surface normal vector direction, and the ratio of the minimum eigenvalue to the sum of the eigenvalues quantifies the local curvature feature quantity of the point;

[0127] S2014, the eigenvector corresponding to the minimum eigenvalue is taken as the surface normal vector of the feature point, and the ratio of the minimum eigenvalue to the sum of the three eigenvalues is taken as the local curvature feature quantity of the point.

[0128] In this embodiment, based on the surface normal vector and the local neighborhood point cloud obtained in S201, the geometric characteristics of the weld trace and the groove angle are analyzed, and the specific process includes:

[0129] S201A1, based on the optimized three-dimensional weld position point set, a center line of the weld is fitted by using a principal curve extraction algorithm, and a tangent direction at each feature point is calculated along the center line, which represents the local trace of the weld at the point.

[0130] S201A2, in the normal plane of the weld center line, the local neighborhood point cloud constructed in S2011 is analyzed, and the upper and lower edge points of the groove are located by identifying the gradient mutation points existing in the cross-section point cloud.

[0131] S201A3, based on the identified upper and lower edge points of the groove, two fitting straight lines representing the bevels on both sides of the groove are constructed in the normal plane, and the included angle between the two fitting straight lines representing the bevels on both sides of the groove is calculated to obtain the groove angle value at the feature point;

[0132] S202, based on the surface normal vector, the curvature feature, and the weld trace and the groove angle analyzed in S201A, a local surface is fitted by using the moving least squares method, and the Gaussian curvature and the average curvature at each feature point are calculated, and the absolute value of the average curvature is taken as the curvature feature value of the point; it needs to be further explained that the process of obtaining the curvature feature value in this embodiment includes:

[0133] S2021, based on the surface normal vector and the local neighborhood point cloud obtained in S201, and combined with the weld trace and the groove angle feature, a weighted covariance matrix is constructed by using the moving least squares method to fit a local quadratic surface equation; wherein, the weld trace is used to determine the main direction of the surface fitting, and the groove angle feature is used to constrain the boundary condition of the fitted surface;

[0134] S2022, based on the local quadric surface equation, the first fundamental form coefficient and the second fundamental form coefficient of the surface are calculated; in the calculation process, the weld extension direction is taken as one of the principal directions of the surface, so as to improve the curvature calculation accuracy in the weld extension direction;

[0135] S2023, based on the first fundamental form coefficient and the second fundamental form coefficient, two principal curvatures are calculated by eigenvalue solution, and the Gaussian curvature and the mean curvature are calculated according to the principal curvatures; wherein the direction perpendicular to the weld extension direction is defined as the priority direction of the principal curvature calculation, so as to ensure that the curvature eigenvalue can accurately reflect the geometric change in the cross-sectional direction of the weld; it needs to be further explained that the first fundamental form coefficient and the second fundamental form coefficient in the embodiment are the differential geometric quantities calculated by the first order partial derivative and the second order partial derivative of the surface based on the local quadric surface fitted by the moving least square method; wherein the first fundamental form coefficient represents the local intrinsic metric property of the surface, including length and area elements, and the second fundamental form coefficient represents the external bending degree of the surface in the three-dimensional space, which together constitute the mathematical basis for calculating the principal curvature, the Gaussian curvature and the mean curvature; the calculation of the curvature eigenvalue particularly pays attention to the curvature change in the cross-sectional direction of the weld, which is perpendicular to the weld extension direction and can effectively represent the geometric characteristics of the groove area, thereby providing accurate geometric basis for the determination of the subsequent welding process parameters.

[0136] S203, the base material and welding material types, heat input sensitivity coefficients of the current impeller workpiece are obtained from the preset impeller base material and welding material impeller base material and welding material process database, and the welding curvature threshold range under different material combinations and groove shapes is determined in combination with the preset material-curvature-heat input correlation model and the groove angle;

[0137] It needs to be further explained that the process of determining the welding curvature threshold range under different material combinations and groove shapes in the embodiment includes:

[0138] S2031, the base material type, welding material type and corresponding heat input sensitivity coefficient associated with the current impeller workpiece identification code are obtained by querying the data record from the preset impeller base material and welding material impeller base material and welding material process database; the heat input sensitivity coefficient is divided into three levels: zero point eight for low sensitivity, one point zero for medium sensitivity, and one point two for high sensitivity; at the same time, the calculated groove angle value at the current feature point is obtained;

[0139] S2032, based on the obtained base material type and welding material type, the reference welding curvature threshold mapping table in the impeller base material and welding material process database is queried to obtain the corresponding reference curvature threshold; the reference curvature threshold is set in the range of zero point zero five per millimeter to zero point one five per millimeter according to the different material combinations;

[0140] S2033, when the heat input sensitivity coefficient is low sensitivity, the reference curvature threshold is kept unchanged; when the heat input sensitivity coefficient is medium sensitivity, the reference curvature threshold is reduced by 10%; when the heat input sensitivity coefficient is high sensitivity, the reference curvature threshold is reduced by 20%;

[0141] S2034, taking sixty degrees as the reference bevel angle, when the actual bevel angle is less than sixty degrees, the curvature threshold is reduced by 5% for every ten degrees of decrease; when the actual bevel angle is greater than sixty degrees, the curvature threshold is increased by 5% for every ten degrees of increase; the correction range of the bevel angle is limited between forty-five degrees and seventy-five degrees;

[0142] S2035, the final curvature threshold obtained after two corrections is set as the recommended upper limit of the welding curvature threshold at the current feature point; at the same time, the lower limit of the threshold is fixed as 0, so as to form the welding curvature threshold range of the point. It needs to be further explained that the specific value of the reference curvature threshold in the parent material and welding material impeller parent material and welding material process database is set by the person skilled in the art according to the corresponding material type; all the correction coefficients and the reference parameters are stored in the system configuration file in the form of fixed values.

[0143] S204, compare the curvature feature value with the welding curvature threshold range, and select the feature point with the curvature feature value not exceeding the preset maximum allowable welding curvature threshold as the preliminary path candidate point; it needs to be further explained that the preset maximum allowable welding curvature threshold in the embodiment is the final curvature threshold upper limit determined by querying the material process database to obtain the reference value, and then correcting twice according to the heat input sensitivity coefficient and the bevel angle of the specific workpiece.

[0144] S205, based on the three-dimensional coordinates of the obtained preliminary path candidate points, the DBSCAN clustering algorithm is combined to perform spatial clustering analysis, and the noise points not meeting the preset minimum neighborhood point number are removed, so that the distribution density of the preliminary path candidate points in the corresponding local surface area meets the preset path point distribution density threshold;

[0145] It needs to be further explained that the specific process of combining the DBSCAN clustering algorithm to perform spatial clustering analysis in the embodiment includes:

[0146] S2051, based on the three-dimensional coordinates of the preliminary path candidate points obtained in S204, set the preset neighborhood search radius and the minimum neighborhood point number as the clustering parameters;

[0147] S2052, traverse all the preliminary path candidate points, calculate the number of adjacent points of each point within its neighborhood radius, and mark the points meeting the minimum neighborhood point number requirement as core points;

[0148] S2053, recursively expand the cluster area from the core point as the starting point through the density accessibility principle, and all points meeting the density connectivity are included in the same cluster;

[0149] S2054, mark the points not included in any cluster area as noise points and exclude them from the path candidate point set;

[0150] S2055, calculate the number of path candidate points per unit surface area in each cluster area, and ensure that the distribution density of all areas is not lower than the preset path point distribution density threshold.

[0151] S206, based on the curvature characteristic value, the material heat input sensitivity coefficient, the bevel angle and the path point distribution density of each local surface area obtained in S205, the comprehensive welding accessibility score of each preliminary path candidate point is calculated through linear weighting algorithm, wherein the weight coefficient is preset according to the welding quality and accessibility requirement;

[0152] S207, based on the comprehensive welding accessibility score of each preliminary path candidate point obtained in S206, the top N highest scoring feature points are selected to form the final candidate welding path set by using sorting and screening algorithm, wherein the value of N is dynamically calculated according to the product of the diagonal line length of the three-dimensional bounding box of the impeller workpiece and the preset path point density parameter.

[0153] Further, the process of obtaining the welding gun posture angle in the embodiment includes:

[0154] S301, based on the candidate welding path set, the surface normal vector of each candidate welding path point calculated in S201 is obtained, and the target posture of the welding gun is calculated combined with the preset welding gun working angle and walking angle rule;

[0155] S302, convert the target posture into X-Y-Z axis rotation angle in the robot base coordinate system to obtain the welding gun posture angle; it needs to be further explained that the process of obtaining the welding gun posture angle in the embodiment includes:

[0156] S3021, based on the complete target posture of the robot end effector in the robot base coordinate system obtained in S3024, the 3×3 rotation matrix component is extracted;

[0157] S3022, convert the rotation matrix component into Euler angle representation form of X-Y-Z fixed axis rotation sequence in the robot base coordinate system;

[0158] S3023, adjust the converted Euler angle to the angle range specified by the robot control system to ensure that the rotation angles of each axis meet the kinematic constraints of the robot;

[0159] S3024, verify whether the converted X-Y-Z axis rotation angle precision reaches a preset angle deviation threshold value;

[0160] S3025, output the normalized X-axis rotation angle, Y-axis rotation angle and Z-axis rotation angle as the final welding gun posture angle to the robot control system;

[0161] S303, store the optimal welding process parameters of each candidate welding path point in association with the welding gun posture angle, and output a welding parameter set containing the candidate welding path point position, the optimal welding process parameters and the welding gun posture angle.

[0162] Further, the process of obtaining the optimal welding process parameter set in the embodiment includes:

[0163] S304, based on the candidate welding path set obtained in S207, obtain the curvature characteristic value calculated in S201 for each candidate welding path point;

[0164] S305, based on the preset welding heat input base formula, combined with the base material weldability parameters obtained through the impeller base material and welding material process database, calculate the reference heat input value; It needs to be further explained that the process of obtaining the reference heat input value in the embodiment includes:

[0165] S3051, obtain the recommended unit length heat input range of the current impeller base material and welding material combination from the preset welding impeller base material and welding material process database;

[0166] S3052, based on the weld cross-sectional feature, determine the deposited metal amount per unit length of the weld by weld cross-sectional area calculation;

[0167] S3053, according to the deposited metal amount and the welding material melting efficiency, calculate the basic heat input value that meets the welding forming requirement;

[0168] S3054, based on the base material heat sensitivity and the welding position requirement, obtain the corresponding heat input adjustment coefficient from the process parameter table;

[0169] S3055, multiply the basic heat input value by the heat input adjustment coefficient to obtain the final reference heat input value;

[0170] S306, determining a heat input coefficient of each candidate welding path point based on the curvature characteristic value through a preset curvature-heat input coefficient mapping relationship; wherein the greater the curvature characteristic value, the smaller the corresponding heat input coefficient, and the value range of the heat input coefficient is 0.8-1.2; the acquisition method of the curvature-heat input coefficient mapping relationship is: acquiring the weld forming quality data under different curvature characteristic values through welding process test, establishing a corresponding relationship model of the curvature characteristic value and the heat input coefficient, and storing the relationship model in the form of a data mapping table in the control system, wherein the greater the curvature characteristic value, the smaller the corresponding heat input coefficient;

[0171] S307, multiplying the reference heat input value and the heat input coefficient to obtain an optimal heat input value of each candidate welding path point;

[0172] S308, calculating corresponding welding current, voltage and welding speed parameters through a preset welding parameter mapping model based on the optimal heat input value to obtain an optimal welding process parameter set. It needs to be further explained that the construction method of the welding parameter mapping model in the embodiment is: based on welding process test data, collecting process parameters and weld quality data under various working conditions of different material combinations, and establishing a multiple regression model taking heat input value as input variable and welding current, voltage and welding speed as output variable; first, the experimental data is standardized and abnormal values are removed, then the quantitative relationship between parameters is obtained by least square method fitting, finally the determined model parameters are stored in the impeller base material and welding material process database for real-time calling, realizing the function of accurately backstepping the execution layer process parameters according to the optimal heat input value.

[0173] It needs to be further explained that the preset welding instruction in the embodiment in response to the optimal welding process parameters and the welding gun attitude angle includes:

[0174] S401, based on the optimal welding process parameter set output in S308 and the welding torch posture angle output in S3025, a robot welding action instruction sequence containing target space coordinates corresponding to the candidate welding path, the welding torch posture angle and the optimal welding process parameters is generated; the welding action instruction in the embodiment includes: a target point position instruction for determining the space coordinates of the welding torch in the robot base coordinate system; a welding torch posture angle instruction for controlling the spatial orientation of the welding torch; a welding process parameter instruction for adjusting the welding current, voltage and welding speed; a motion trajectory instruction for planning the continuous motion path of each joint axis; a real-time monitoring instruction for collecting the actual values of the position and welding parameters; a dynamic compensation instruction for generating parameter correction amounts; a state feedback instruction for verifying the system state; a welding completion instruction for confirming the task termination; these instructions establish a cooperative working mechanism through the transformation relationship between the industrial vision subsystem coordinate system and the robot base coordinate system, forming a complete welding control closed loop. The instructions in the embodiment are generated by combining the corresponding acquired control parameter characteristics with the PID control algorithm. For example, the generation method of the welding current control instruction in the welding process parameter instruction in the embodiment is: based on the real-time collected welding current deviation and the curvature characteristic value of the candidate welding path point, the current deviation is processed through the PID control algorithm, the PID parameter adjustment coefficient is output combined with the preset control rule, the current adjustment amount is calculated through the PID control algorithm, and finally the welding current control instruction is generated.

[0175] S402, the target space coordinates and posture parameters in the action instruction generated in S401 are uniformly converted to executable instructions in the robot base coordinate system through the fixed homogeneous transformation matrix of the industrial vision subsystem coordinate system to the robot base coordinate system obtained in S104;

[0176] S403, based on the converted executable instructions in the robot base coordinate system, the target angles of each joint axis of the robot are calculated through the robot inverse kinematics algorithm, and a smooth motion trajectory is generated;

[0177] It should be further explained that the process of calculating the target angles of each joint axis of the robot and generating a smooth motion trajectory through the robot inverse kinematics algorithm in the embodiment includes:

[0178] S4031, based on the link parameters and joint constraint conditions of the six-axis industrial robot, a robot kinematics model containing DH parameters is established;

[0179] S4032, based on the complete target posture of the robot end effector in the robot base coordinate system obtained in S3024, the position vector and rotation matrix are extracted;

[0180] S4033, based on the position vector and the rotation matrix, solving the robot inverse kinematics equation by analytical method or numerical method to calculate all possible joint angle combinations that satisfy the target pose;

[0181] S4034, based on the joint position constraint and the motion smoothness requirement, screening the optimal joint angle combination from all possible joint angle combinations;

[0182] S4035, based on the current joint angle and the optimal joint angle obtained in S4034, generating a smooth motion trajectory for each joint axis using a quintic polynomial interpolation algorithm;

[0183] S4036, performing singularity point detection and interference checking on the generated joint trajectory, and re-planning the trajectory when detecting that the trajectory passes through a singular configuration or interferes with the impeller blade;

[0184] S4037, outputting the final determined target angle sequence of each joint axis to the robot control system.

[0185] S404, during the welding process, the following monitoring tasks are performed synchronously:

[0186] S404-1, based on the industrial vision subsystem, real-time acquisition of the relative position between the robot end and the impeller workpiece at a preset sampling frequency, and calculation of the position deviation;

[0187] S404-2, based on the welding power supply system, real-time acquisition of the actual welding current and voltage at a preset sampling rate, and calculation of the parameter deviation from the optimal welding process parameters obtained in S308;

[0188] S405, when the monitored position deviation exceeds the preset position deviation threshold or the parameter deviation exceeds the preset parameter deviation threshold, triggering the corresponding compensation mechanism, including:

[0189] S405-1, for the position deviation, based on the visual positioning deviation compensation algorithm to calculate the motion correction amount of each axis of the robot, and output to the robot controller to adjust the motion trajectory; further, the process of calculating the motion correction amount of each axis of the robot based on the visual positioning deviation compensation algorithm in this embodiment includes:

[0190] S405-1a, based on the real-time position deviation between the robot end and the impeller workpiece obtained in S404-1, extracting a three-dimensional position deviation vector in the industrial vision subsystem coordinate system;

[0191] S405-1b, based on the fixed homogeneous transformation matrix from the industrial vision subsystem coordinate system to the robot base coordinate system obtained in S104, converting the three-dimensional position deviation vector to the robot base coordinate system;

[0192] S405-1c, based on the current joint angles of the robot, calculate the Jacobian matrix of the robot in the current configuration;

[0193] S405-1d, based on the converted position deviation vector and the Jacobian matrix, solve the angle correction amount of each joint axis by inverse Jacobian algorithm or damped least squares method;

[0194] S405-1e, perform kinematic constraint check on the solved joint axis angle correction amount to ensure that the correction amount is within the joint speed and acceleration limit range;

[0195] S405-1f, convert the angle correction amount after constraint processing into robot axis motion parameter correction instruction and output to the robot controller.

[0196] S405-2, for parameter deviation, calculate the welding parameter adjustment amount based on the process parameter compensation algorithm, and adjust the welding current and voltage through the welding power supply system; further, the specific process of calculating the welding parameter adjustment amount based on the process parameter compensation algorithm in this embodiment includes:

[0197] S405-2a, based on the real-time parameter deviation value of the actual welding current / voltage obtained in S404-2 and the optimal welding process parameter in S308, obtain the parameter deviation value;

[0198] S405-2b, based on the parameter deviation value, calculate the initial welding parameter adjustment amount by PID control algorithm;

[0199] S405-2c, based on the current welding path point curvature characteristic value calculated in S202, perform adaptive gain adjustment on the initial welding parameter adjustment amount;

[0200] S405-2d, convert the adjusted welding parameter adjustment amount into welding power supply control signal, which contains the accurate correspondence relationship of voltage-current;

[0201] S405-2e, limit the amplitude and change rate of the welding power supply control signal to ensure that the welding parameter adjustment is within the process allowed range;

[0202] S405-2f, send the constrained welding power supply control signal to the welding power supply system in real time;

[0203] S405-2g, based on the subsequent sampling data of the welding monitoring system, verify whether the actual welding parameter returns to the preset deviation range of the optimal welding process parameter in S308.

[0204] S406. After compensation, the monitoring task of S404 is re-executed to verify whether the position deviation and parameter deviation have returned to the preset deviation range. In this embodiment, the preset deviation range is set based on the welding process quality standard and the robot system positioning accuracy. The position deviation range is set by those skilled in the art based on the weld seam tracking accuracy requirements and the robot repeatability positioning accuracy. In this embodiment, the parameter deviation range is set according to the welding quality specification.

[0205] S407. When both the position deviation and parameter deviation are stable within the preset deviation range, and the current welding path segment is completed, the current welding stage is determined to be successfully executed, and preparations are made to execute the welding command for the next welding path segment.

[0206] This application first establishes a high-precision measurement foundation by configuring a dual-industry-camera vision subsystem and employing a multi-stage calibration method. First, nonlinear optimization algorithms are used to accurately calculate the camera's intrinsic and extrinsic parameters. Then, a robot carrying a calibration board completes non-coplanar hand-eye calibration, establishing a precise coordinate mapping relationship between the vision system and the robot. In the weld trajectory extraction stage, the system combines image and point cloud data, employing multi-level image processing techniques: including gradient-statistic-based adaptive edge detection, non-maximum suppression corner extraction, and a stereo matching algorithm that fuses texture and geometric features. This achieves sub-millimeter-level positioning of key features such as blade edges and weld bevels. Furthermore, it innovatively overcomes interference from workpiece surface reflection and monotonous textures through multi-source data fusion and cross-validation mechanisms. Second, based on a precise 3D point set, a local geometric model is constructed using the K-nearest neighbor algorithm. Principal component analysis is combined to calculate surface normal vectors and curvature features. Weld direction analysis and bevel angle detection are introduced, and the moving least squares method is used to accurately fit the local surface geometry. By establishing a multi-parameter correlation model of material-curvature-heat input, and comprehensively considering the coupling effects of material thermal sensitivity, bevel morphology, and surface curvature, the welding process threshold is dynamically determined. Then, through DBSCAN clustering optimization and comprehensive reachability scoring, an optimized welding path that satisfies both geometric constraints and process requirements is generated. Finally, based on the material database and weld cross-sectional features, the baseline heat input is calculated. Dynamic adaptive adjustment of parameters is achieved through the curvature-heat input mapping relationship, ensuring precise matching of heat input to surface geometric changes. The welding torch attitude angle is obtained based on the aforementioned normal vector through coordinate transformation and kinematic constraint detection, ensuring precise contact between the welding torch and the weld surface. During execution, the system monitors position and parameter deviations in real time, uses a visual compensation algorithm to calculate robot joint corrections, and dynamically adjusts the welding current and voltage in conjunction with a process parameter compensation algorithm, forming a fully closed-loop control. Finally, a smooth motion trajectory is generated through fifth-order polynomial interpolation, ensuring that the welding process is always in the optimal process state.

[0207] Example 2

[0208] Please see Figure 2Another embodiment provided by the application: a robot adaptive impeller welding control system fusing multi-modal data, comprising: a vision module, an analysis module, a posture parameter module, a response module, a discrimination adjustment module;

[0209] The vision module acquires images and three-dimensional point cloud data of the impeller workpiece surface by using a configured industrial vision subsystem, and extracts a three-dimensional weld position point set representing the impeller weld seam trajectory;

[0210] The analysis module analyzes and discriminates the weld seam direction, the groove angle and the weld accessibility based on the three-dimensional weld position point set and the surface geometry feature analysis, and generates a candidate weld path set;

[0211] The posture parameter module acquires an optimal weld process parameter set by parameter prediction based on a regression model based on the weld seam topographic features corresponding to the candidate weld path set and the weld process requirements, and obtains a welding gun posture angle based on the surface normal vectors corresponding to the candidate weld path set and a quaternion conversion;

[0212] The response module responds to a preset welding instruction containing the optimal weld process parameters and the welding gun posture angle, and monitors the relative position deviation between the end of the six-axis industrial robot and the impeller workpiece weld seam trajectory, and the parameter deviation between the actual welding current or voltage, welding speed and process planning value in the welding process;

[0213] The discrimination adjustment module dynamically adjusts the six-axis motion parameters or the weld process parameters by using a PID control algorithm according to the deviation type and size when the relative position deviation or the parameter deviation does not meet the corresponding preset process tolerance interval, so that the welding process is always within the preset process tolerance interval.

[0214] The embodiments of the application are described above in combination with the drawings, but the application is not limited to the above specific embodiments, and the above specific embodiments are only illustrative and not restrictive, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments without departing from the purpose and scope of the application, which are all within the protection of the application.

Claims

1. A robot adaptive impeller welding control method integrating multimodal data, characterized in that, include: The configured industrial vision subsystem is used to acquire images and 3D point cloud data of the impeller workpiece surface, and to extract a set of 3D weld position points that characterize the impeller weld trajectory. Based on the three-dimensional weld location point set and combined with surface geometric feature analysis, the weld direction, bevel angle and welding accessibility are analyzed and judged to generate a set of candidate welding paths. Based on the weld morphology features and welding process requirements corresponding to the candidate welding path set, the optimal welding process parameter set is obtained by parameter prediction based on the regression model. At the same time, based on the surface normal vector corresponding to the candidate welding path set, combined with quaternion transformation, the welding gun attitude angle is obtained. The system responds to preset welding commands containing optimal welding process parameters and welding torch attitude angles, and monitors in real time the relative positional deviation between the end effector of the six-axis industrial robot and the weld seam trajectory of the impeller workpiece, as well as the parameter deviations between the actual welding current or voltage, welding speed and process planning values. When the relative position deviation or parameter deviation does not meet the corresponding preset process tolerance range, the six-axis motion parameters or welding process parameters are dynamically adjusted according to the type and magnitude of the deviation, so that the welding process is always within the preset process tolerance range. The process of extracting the three-dimensional weld position point set representing the impeller weld trajectory includes: S105. Collect the left-right view image pair sequence of the impeller workpiece to be welded at the same time and the three-dimensional point cloud data with the same timestamp. Combine Gaussian convolution and contrast-limited adaptive histogram equalization algorithm to preprocess noise and contrast and perform timestamp alignment to obtain the enhanced left-right view impeller workpiece image pair sequence. S106. Based on the enhanced left-right view impeller workpiece image pair sequence, the Sobel operator is used to calculate the gradient components of the image in the horizontal and vertical directions respectively, so as to obtain the gradient magnitude and gradient direction of each pixel in each left-right view impeller workpiece image pair. S107. Based on the gradient magnitude and gradient direction, traverse each pixel in the image, compare the gradient magnitude of the pixel with its neighboring pixels along its gradient direction, retain the pixels with local maxima in the gradient direction, and obtain the edge pixel candidate image. S108. Based on the edge pixel candidate image and the preset filtering rules, the edge pixel candidate image is judged and filtered. All pixels marked as the final edge are set as the foreground color and all non-edge pixels are set as the background color. A binary contour image representing the edge of the impeller weld is generated from the left and right perspectives.

2. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 1, characterized in that, The process of extracting the three-dimensional weld position point set characterizing the impeller weld trajectory also includes: S109. Based on the binarized contour map, slide an evaluation window of a preset size on the image, calculate the amount of grayscale change caused by the evaluation window moving a preset length along the horizontal and vertical directions at each pixel position on the contour map, and calculate the corner response function value of each pixel based on the amount of grayscale change. S110. The calculated corner response function value is compared with a preset filtering threshold, and all pixels whose corner response function value exceeds the threshold are selected as candidate corner points in the left-right view. S111. Based on the candidate corner point set under the left-right perspective and the non-maximum suppression process, obtain the feature point set of the left image and the feature point set of the right image under the left-right perspective. S112. Based on the feature point set of the left image and the feature point set of the right image, the feature point matching algorithm is used to obtain a set of feature point matching pairs from the left and right perspectives. S113. Based on the feature point matching pair set under the left and right viewpoints, combined with the internal parameter matrix corresponding to the left and right viewpoints, the lens distortion coefficient set, and the relative spatial pose relationship between the cameras under the left and right viewpoints, the three-dimensional spatial coordinates of each feature point in the industrial vision subsystem coordinate system are obtained by calculation using the triangulation principle.

3. The robot adaptive impeller welding control method integrating multimodal data as described in claim 2, characterized in that, The process of extracting the three-dimensional weld position point set characterizing the impeller weld trajectory also includes: S114. Based on the three-dimensional spatial coordinates of each feature point in the industrial vision subsystem coordinate system, the first three-dimensional coordinates of each feature point in the robot base coordinate system are obtained through coordinate transformation calculation, which serves as the image three-dimensional weld position point set representing the impeller weld trajectory. S115. Based on the timestamp-aligned 3D point cloud data, calculate the second 3D coordinates of each feature point in the 3D point cloud data in the robot base coordinate system, and use it as the point cloud 3D weld position point set representing the impeller weld trajectory. S116. Based on the point cloud three-dimensional weld position point set and the image three-dimensional weld position point set, calculate the position point deviation to obtain the image-point cloud feature position point deviation sequence. S117. When at least one image-point cloud feature position point deviation is greater than a preset deviation threshold, the mapped image three-dimensional weld position point set and the point cloud three-dimensional weld position point set are adjusted in real time so that the image-point cloud feature position point deviation is less than or equal to the preset deviation threshold in real time, and a second-optimized three-dimensional weld position point set is obtained.

4. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 3, characterized in that, The process of generating the candidate welding path set includes: S201. Based on the three-dimensional weld location point set after secondary optimization, the K-nearest neighbor algorithm is used to establish a local neighborhood point cloud for each feature point. At the same time, the surface normal vector and curvature features of each local neighborhood are calculated by principal component analysis. S202. Based on the surface normal vector, curvature features, weld direction and bevel angle, the local surface is fitted by moving least squares method, and the Gaussian curvature and average curvature at each feature point are calculated. The absolute value of the average curvature is used as the curvature feature value of that point. S203. Obtain the type of base material and welding material, and the heat input sensitivity coefficient of the current impeller workpiece from the preset impeller base material and welding material process database, and determine the welding curvature threshold range under different material combinations and bevel shapes by combining the preset material-curvature-heat input correlation model and bevel angle. S204. Compare the curvature feature value with the welding curvature threshold range, and select feature points whose curvature feature value does not exceed the preset maximum allowable welding curvature threshold as preliminary path candidate points.

5. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 4, characterized in that, The process of generating the candidate welding path set also includes: S205. Based on the obtained three-dimensional coordinates of the preliminary path candidate points, spatial clustering analysis is performed using the DBSCAN clustering algorithm to remove noise points that do not meet the preset minimum number of neighborhood points, so that the distribution density of the preliminary path candidate points in the corresponding local surface region meets the preset path point distribution density threshold. S206. Based on the curvature characteristic value, material heat input sensitivity coefficient, bevel angle, and path point distribution density of each local curved surface region obtained in S205, calculate the comprehensive welding accessibility score of each preliminary path candidate point using a linear weighted algorithm. S207. Based on the comprehensive welding accessibility scores of each preliminary path candidate point obtained in S206, a sorting and filtering algorithm is used to select the top N feature points with the highest scores to form the final candidate welding path set.

6. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 5, characterized in that, The process of obtaining the welding torch attitude angle includes: S301. Based on the candidate welding path set, obtain the surface normal vector of each candidate welding path point calculated in S201, and calculate the target posture of the welding gun by combining the preset welding gun working angle and travel angle rules. S302. Convert the target posture into XYZ axis rotation angles in the robot base coordinate system to obtain the welding gun posture angle.

7. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 6, characterized in that, The process of obtaining the optimal welding process parameter set includes: S304. Based on the candidate welding path set, obtain the curvature feature value of each candidate welding path point calculated in S201. S305. Based on the preset formula for calculating the base welding heat input, and combined with the weldability parameters of the base material obtained from the impeller base material and welding material process database, calculate the base heat input value. S306. Based on the curvature feature value, determine the heat input coefficient of each candidate welding path point through a preset curvature-heat input coefficient mapping relationship; S307. Multiply the reference heat input value by the heat input coefficient to obtain the optimal heat input value for each candidate welding path point; S308. Based on the optimal heat input value, the corresponding welding current, voltage and welding speed parameters are calculated through a preset welding parameter mapping model to obtain the optimal welding process parameter set.

8. The robot adaptive impeller welding control method that integrates multimodal data as described in claim 7, characterized in that, The response includes preset welding commands for optimal welding process parameters and welding torch attitude angle, including: S401. Based on the optimal welding process parameter set output by S308 and the welding gun attitude angle output by S302, generate a robot welding action instruction sequence containing the target space coordinates, welding gun attitude angle and optimal welding process parameters corresponding to the candidate welding path. S402. Convert the target space coordinates and attitude parameters in the motion instructions generated in S401 into executable instructions in the robot base coordinate system. S403. Based on the executable instructions in the transformed robot base coordinate system, the target angles of each joint axis of the robot are calculated through the robot inverse kinematics algorithm, and a smooth motion trajectory is generated. S404. During the welding process, the following monitoring tasks shall be performed simultaneously: S404-1. The relative position between the robot end effector and the impeller workpiece is collected in real time at a preset sampling frequency based on the industrial vision subsystem, and the position deviation is calculated. S404-2. Based on the welding power supply system, the actual welding current and voltage are collected in real time at a preset sampling rate, and the parameter deviation from the optimal welding process parameters obtained in S308 is calculated. S405. When the detected position deviation exceeds a preset position deviation threshold or the parameter deviation exceeds a preset parameter deviation threshold, a corresponding compensation mechanism is triggered, including: S405-1. For positional deviation, calculate the motion correction amount of each axis of the robot based on the visual positioning deviation compensation algorithm, and output it to the robot controller to adjust the motion trajectory. S405-2. For parameter deviations, calculate the welding parameter adjustment amount based on the process parameter compensation algorithm, and adjust the welding current and voltage through the welding power supply system. S406. After performing compensation, re-execute the monitoring task of S404 to verify whether the position deviation and parameter deviation have returned to the preset deviation range. S407. When both the position deviation and parameter deviation are stable within the preset deviation range, and the current welding path segment is completed, the current welding stage is determined to be successfully executed, and preparations are made to execute the welding command for the next welding path segment.

9. A robot adaptive impeller welding control system integrating multimodal data, implemented based on the robot adaptive impeller welding control method integrating multimodal data as described in any one of claims 1-8, characterized in that, include: Vision module, analysis module, posture parameter module, response module, and discrimination adjustment module; The vision module uses the configured industrial vision subsystem to acquire images and three-dimensional point cloud data of the impeller workpiece surface, and extracts a set of three-dimensional weld position points that characterize the impeller weld trajectory. The analysis module, based on a three-dimensional weld location point set and combined with surface geometric feature analysis, analyzes and judges the weld direction, bevel angle and welding accessibility, and generates a set of candidate welding paths. The attitude parameter module obtains the optimal welding process parameter set by parameter prediction based on regression model, based on the weld morphology features and welding process requirements corresponding to the candidate welding path set. At the same time, it obtains the welding gun attitude angle by combining the surface normal vector corresponding to the candidate welding path set with quaternion transformation. The response module responds to preset welding commands containing optimal welding process parameters and welding torch attitude angles, and monitors in real time the relative positional deviation between the end effector of the six-axis industrial robot and the weld seam trajectory of the impeller workpiece, as well as the parameter deviations between the actual welding current or voltage, welding speed and process planning values. When the relative position deviation or parameter deviation does not meet the corresponding preset process tolerance range, the discrimination and adjustment module dynamically adjusts the six-axis motion parameters or welding process parameters according to the type and magnitude of the deviation, so that the welding process is always within the preset process tolerance range.

Images