A child electric scooter frame weld consistency detection method

Abstract

This invention relates to the field of weld inspection technology, and discloses a method for inspecting the consistency of welds on a children's electric scooter frame. The method includes: acquiring basic data of the children's electric scooter frame to obtain frame structural reference information; collecting multi-source inspection data of the frame to be inspected to obtain original weld acquisition information; performing attitude correction on the original weld acquisition information to obtain frame coordinate inspection information; extracting weld consistency features from the frame coordinate inspection information to obtain weld geometric consistency features; and marking the load-sensitive area of ​​the frame structural reference information to obtain weld load-sensitive area information. This invention employs attitude correction, coordinate transformation, and deviation compensation techniques based on the frame structural reference, achieving the technical effect of distinguishing between clamping attitude deviations, post-weld micro-deformation, and actual weld consistency anomalies. This enables accurate detection of the weld relative to the frame's stress reference, overcoming the shortcomings of fixed template inspection which is prone to misjudgment and missed detection.

Description

Technical Field

[0001] The present invention relates to the technical field of weld detection, and in particular to a method for detecting the consistency of welds of a children's electric scooter frame, an electronic device, and a storage medium. Background Art

[0002] In the scenario of online detection of key bearing welds of the folding riser seat, the front pedal connection part, the rear fork connection part, and the stem support connection part after the welding of the children's electric scooter frame and before surface spraying or vehicle assembly, the frame has problems such as thin-walled pipe fittings, small-sized welds, variable special-shaped connection angles, batch model changes, and slight post-weld deformation. It is necessary to solve the problem of judging the consistency deviation of the weld center line, weld width, weld continuity, and the weld relative to the force-bearing reference of the frame when there are deviations in the clamping attitude of the frame and post-weld micro-deformation, and to associate this consistency deviation with the sensitive areas of low-step impact, lateral fall, and pedal torsion load during the use of children, and output the detection result of the consistency of the frame welds that can be used for repair positioning, process adjustment, and batch release judgment;

[0003] Currently, for the weld detection of a children's electric scooter frame, the frame is usually placed on a detection tooling, and the weld appearance images and weld profile data are collected by manual visual inspection, planar camera photography detection, or laser profile scanning. Then, the collected results are compared with a preset template, a standard weld image, or a size threshold to determine whether there are problems such as incomplete welding, missed welding, weld bead, porosity, undercut, width out-of-tolerance, and abnormal reinforcement height in the weld; in an automated detection system, an image recognition model is used to classify the types of weld defects and output a conclusion of qualified or unqualified for the frame welds;

[0004] In the prior art, it is not easy to distinguish the clamping attitude deviation, post-weld micro-deformation, and real weld consistency abnormality by fixed template detection. The children's electric scooter frame has a small size, and there are curved surfaces, inclined surfaces, and thin-walled pipe intersection structures at the folding riser seat, the front pedal, and the rear fork connection positions. The existing detection methods usually use a fixed shooting angle, a fixed detection window, or a fixed weld template as the judgment basis. When the frame has a posture deviation during clamping or local thermal deformation occurs after welding, the position and shape of the weld in the collected image or profile data will shift as a whole. The prior art is likely to misjudge the change in the clamping attitude as weld inconsistency and is likely to judge a product with an offset frame structure reference but an approximately qualified weld appearance as qualified, resulting in the detection result not being able to truly reflect the consistency of the weld relative to the force-bearing position of the frame. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a method for detecting the consistency of welds of a children's electric scooter frame, an electronic device, and a storage medium to at least partially solve the problems raised in the above background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides a method for detecting the consistency of weld seams on a children's electric scooter frame, comprising:

[0008] Obtain the basic data of the children's electric scooter frame to obtain the frame structure reference information;

[0009] Collect multi-source inspection data of the chassis to be inspected to obtain the original information of the weld.

[0010] The original weld seam data is subjected to attitude correction to obtain the vehicle frame coordinate detection information;

[0011] Weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features;

[0012] The load-sensitive area of ​​the frame structure reference information is marked to obtain the load-sensitive area information of the weld.

[0013] By jointly evaluating the geometric consistency characteristics of the weld and the information on the load-sensitive area of ​​the weld, the consistency risk information of the weld segment is obtained.

[0014] The consistency risk information of the weld segment is classified into levels, and the consistency judgment result of the frame weld is output.

[0015] Preferably, the original weld seam acquisition information is subjected to attitude correction to obtain the vehicle frame coordinate detection information, including:

[0016] The frame structure reference information is analyzed to determine the line connecting the pedal plane, the front stem axis, and the center of the rear fork mounting hole;

[0017] A local coordinate system for the frame is constructed using the pedal plane as the plane reference, the front stem axis as the height direction reference, and the line connecting the centers of the rear fork mounting holes as the lateral correction reference.

[0018] The clamping deviation is extracted from the original information collected from the weld to obtain the actual clamping posture of the frame to be inspected;

[0019] Using the local coordinate system of the vehicle frame and the actual clamping posture as constraints, coordinate transformation and deviation compensation are performed on the original weld seam acquisition information to obtain the vehicle frame coordinate detection information.

[0020] Preferably, weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features, including:

[0021] Using the frame structure reference information as a constraint, the weld position is matched with the frame coordinate detection information to obtain the weld segment detection area;

[0022] Weld boundary identification and cross-sectional contour identification are performed on the weld segment detection area to obtain weld boundary information and weld cross-sectional information;

[0023] The centerline, width, reinforcement height, weld toe transition, and continuity are calculated on the weld boundary information and weld section information to obtain weld component consistency information;

[0024] The weld seam component consistency information is collected to obtain the weld seam geometric consistency characteristics.

[0025] Preferably, the load-sensitive area is marked on the frame structure reference information to obtain weld load-sensitive area information, including:

[0026] The key load-bearing positions of the frame structure reference information are identified to obtain the positions of the frame load-bearing welds;

[0027] The load-bearing weld locations of the vehicle frame are matched with child-use load types to obtain impact-sensitive weld locations, torsion-sensitive weld locations, rollover-sensitive weld locations, and ordinary connection weld locations;

[0028] The locations of impact-sensitive welds, torsion-sensitive welds, rollover-sensitive welds, and ordinary connection welds are marked to obtain information on weld load-sensitive zones.

[0029] Preferably, the weld geometric consistency characteristics and weld load-sensitive area information are jointly evaluated to obtain weld segment consistency risk information, including:

[0030] The geometric consistency characteristics of the weld are segmented and normalized to obtain the geometric deviation information of the weld segment.

[0031] The load-sensitive area information of the weld is matched with the sensitive area coefficient to obtain the load coefficient information of the weld segment;

[0032] The geometric deviation information of the weld segment is weighted by feature weight allocation to obtain the weighted deviation information of the weld segment;

[0033] Using the load coefficient information of the weld segment as a constraint, risk calculation is performed on the weighted deviation information of the weld segment to obtain the consistency risk information of the weld segment.

[0034] Preferably, the consistency risk information of the weld segment is graded, and the consistency judgment result of the frame weld is output, including:

[0035] Risk threshold matching is performed on the weld segment consistency risk information to obtain weld segment level information;

[0036] The weld segment grade information is summarized and judged for the whole vehicle to obtain the frame weld consistency grade.

[0037] The abnormal location and abnormal type are extracted from the weld segment grade information to obtain weld abnormal location information;

[0038] The results of the consistency level of the frame welds and the location information of weld anomalies are integrated to output the consistency judgment result of the frame welds.

[0039] Preferably, the original weld seam acquisition information is subjected to coordinate transformation and deviation compensation to obtain the frame coordinate detection information, including:

[0040] The actual clamping posture is analyzed to obtain clamping posture offset information;

[0041] A mapping relationship is constructed between the local coordinate system of the vehicle frame and the clamping posture offset information to obtain the coordinate transformation relationship;

[0042] Based on the coordinate transformation relationship, coordinate transformation is performed on the weld appearance image and weld contour point cloud in the original weld acquisition information to obtain the initial frame coordinate information;

[0043] The initial frame coordinate information is checked for reference residuals to obtain the coordinate deviation compensation amount;

[0044] Based on the coordinate deviation compensation amount, the initial frame coordinate information is offset to obtain the frame coordinate detection information.

[0045] Preferably, the weld boundary information and the weld cross-section information are calculated for centerline, width, reinforcement height, weld toe transition, and continuity to obtain weld component consistency information, including:

[0046] The weld boundary information is mapped to boundary points to obtain weld boundary point pairs;

[0047] By performing midpoint connection and spacing calculation on the weld boundary point pairs, the weld centerline information and weld width information are obtained;

[0048] The weld cross-section information is used to calculate the height difference to obtain the weld reinforcement information;

[0049] The edge slope change is calculated based on the weld boundary information and weld section information to obtain weld toe transition information;

[0050] Discontinuous region identification and length statistics are performed on the weld boundary information and weld cross-section information to obtain weld continuity information;

[0051] The weld centerline information, weld width information, weld reinforcement information, weld toe transition information, and weld continuity information are collected to obtain the weld component consistency information.

[0052] Preferably, the load-bearing weld locations on the vehicle frame are matched for child-use load types to obtain impact-sensitive weld locations, torsion-sensitive weld locations, rollover-sensitive weld locations, and ordinary connection weld locations, including:

[0053] The structural parts of the load-bearing welds on the frame are identified to obtain the locations of the front foot pedal connection welds, the front wheel support connection welds, the rear fork connection welds, the two sides of the foot pedal connection welds, the stem support connection welds, the side welds of the folding stem base, and the auxiliary connection welds.

[0054] Impact load matching was performed on the front end connection weld of the pedal, the front wheel support connection weld and the rear fork connection weld to obtain the impact-sensitive weld location.

[0055] Torsional load matching was performed on the connection welds on both sides of the pedal and the connection weld of the stem support to obtain the torsion-sensitive weld locations.

[0056] The position of the side weld of the folded riser is matched with the overturning load to obtain the position of the overturning sensitive weld.

[0057] The auxiliary connection weld positions are matched with ordinary connections to obtain the ordinary connection weld positions.

[0058] Preferably, using the load coefficient information of the weld segment as a constraint, risk calculation is performed on the weighted deviation information of the weld segment to obtain the consistency risk information of the weld segment, including:

[0059] The weighted deviation information of the weld segment is read in segments to obtain the weighted deviation value of each weld segment;

[0060] The load coefficient information of the weld segment is read accordingly to obtain the load coefficient value of each weld segment;

[0061] Using the positional correspondence of each weld segment as a constraint, the weighted deviation value and the load coefficient value of each weld segment are matched to obtain the risk calculation object of the weld segment;

[0062] The risk value of the weld segment is obtained by multiplying the risk calculation objects of the weld segment.

[0063] By associating the risk value of the weld segment with the corresponding weld segment location, the consistency risk information of the weld segment is obtained.

[0064] In a second aspect, the present invention provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, characterized in that: when the processor executes the computer program, it implements the steps of the method in the first aspect.

[0065] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, characterized in that: when the computer program is executed by a processor, it implements the steps of the method in the first aspect.

[0066] The present invention has the following beneficial effects:

[0067] 1. This invention adopts a posture correction, coordinate transformation and deviation compensation technical solution based on the frame structure reference, which achieves the technical effect of distinguishing the clamping posture deviation, post-weld micro deformation and the abnormal consistency of the actual weld, realizes the accurate detection of the weld relative to the frame stress reference, and solves the shortcomings of fixed template detection which is prone to misjudgment and omission.

[0068] 2. The present invention adopts a technical solution of joint evaluation of weld geometric consistency characteristics and weld load-sensitive area information to achieve the technical effect of associating weld anomalies with the risk of loads used by children, realize risk classification and anomaly location of weld segments, and solve the problem that simple appearance defect identification is not easy to reflect the consistency risk of key welds. Attached Figure Description

[0069] Figure 1 This is a flowchart illustrating the vehicle frame weld consistency detection method provided in an embodiment of the present invention. Detailed Implementation

[0070] To enable those skilled in the art to understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort should fall within the scope of protection of the present invention.

[0071] The present invention will now be described in detail with reference to the accompanying drawings:

[0072] Please see the appendix Figure 1This invention provides a method for inspecting the consistency of weld seams in children's electric scooter frames, applicable to online inspection stations after welding, before painting, or before final assembly. The inspection objects include weld seams of the folding stem base, front footplate connection weld seams, rear fork connection weld seams, front wheel support connection weld seams, side footplate connection weld seams, stem support connection weld seams, and auxiliary connection weld seams. The inspection system includes a frame positioning fixture, a multi-view industrial camera, a line laser profile sensor, a fixture positioning sensor, a light source assembly, a motion module, a controller, and a computing unit. The multi-view industrial camera acquires images of weld surface texture, grayscale changes, boundary direction, and weld toe lines; the line laser profile sensor acquires weld section height, excess height, width, and base material transition profile; the fixture positioning sensor acquires the clamping state of the frame in the inspection fixture, the contact state of the positioning pins, the offset of the reference hole, and the clamping posture offset.

[0073] Obtain the basic data of the children's electric scooter frame to obtain the frame structure reference information.

[0074] First, the model number of the children's electric scooter frame to be tested is obtained through the production management system, barcode scanning system, or manual model selection. The calculation unit calls up the basic frame data corresponding to the model number. The sources of the basic frame data include the frame's 3D design model, 2D engineering drawings, welding process documents, inspection specifications, tooling positioning diagrams, finite element analysis documents, bench test records, historical after-sales failure records, and measurement data of the first qualified sample.

[0075] The basic frame data includes at least: key frame reference points, pedal plane definition points, front seat tube axis definition points, rear fork mounting hole center points, folding seat positioning hole center points, designed weld paths, allowable weld size ranges, key load-bearing area identifiers, and weld numbers. The designed weld path is represented as a three-dimensional parametric curve, a sequence of discrete sampling points, or a set of paths composed of multiple straight line segments and circular arc segments. For each weld, the designed weld path records the weld number, start coordinates, end coordinates, path direction, theoretical weld centerline, theoretical weld leg or theoretical width, allowable reinforcement height range, allowable centerline offset range, allowable discontinuous notch length, and the structural location to which it belongs.

[0076] Key frame reference points are obtained as follows: If the frame design model already includes a unified coordinate system, directly read the pedal plane features, front seat tube axis features, and rear fork mounting hole features from the design model; if the design model does not display these features, fit the pedal plane using at least three non-collinear points on the upper surface or pedal positioning surface of the pedal, fit the front seat tube axis using the center of the inner hole or outer circular section of the front seat tube, and determine the lateral correction reference by connecting the centers of the left and right rear fork mounting holes. For the trial production stage of new models, the allowable weld size range is reduced from no less than... The measured data of qualified vehicle frames are determined, such as calculating the mean and standard deviation of weld width, reinforcement height, centerline offset and weld toe transition index, and forming an initial allowable threshold in combination with the enterprise's welding process specifications; in the mass production stage, the thresholds in the verified enterprise inspection specifications or design specifications are given priority.

[0077] When generating frame structural reference information from basic frame data, the calculation unit uniformly converts data from different sources into the vehicle design coordinate system. Specifically, arc length sampling is performed on the weld paths in the 3D frame model to form a set of design weld path points; key reference points are numbered to form a reference point table; the allowable weld size range is field-processed to form a weld feature threshold table; and structural parts are mapped to key load-bearing areas to form a load-bearing area identification table. The frame structural reference information includes the vehicle model number, reference point coordinates, pedal plane equation, front seat tube axis vector, rear fork mounting hole center connection vector, design weld paths, weld segmentation rules, allowable thresholds for each weld feature, key load-bearing area identification, and data version number.

[0078] To ensure that subsequent detection does not rely on a fixed image template, this implementation constructs a local coordinate system for the frame based on the line connecting the pedal plane, the front seat tube axis, and the center of the rear fork mounting hole. Let the unit normal vector of the pedal plane be... The direction vector of the front riser axis is The direction vector of the line connecting the centers of the rear fork mounting holes is The three directions of the local coordinate system of the chassis are determined by the following formula:

[0079] ,

[0080] ,

[0081] ,

[0082] in, Indicates the height direction of the local coordinate system of the vehicle frame. Indicates the lateral correction direction. Indicates the vertical direction. This is used to ensure that the height direction and the positive direction of the front riser axis are consistent. The coordinate origin is selected as the projection point on the pedal plane near the center of the folding riser seat positioning hole, or the center point of the front end of the pedal, or the frame detection origin defined by the enterprise tooling.

[0083] Multi-source inspection data of the chassis to be inspected are collected to obtain the original information of the weld.

[0084] After the vehicle frame to be inspected enters the inspection station, it is first positioned by the tooling positioning pins, clamping mechanism, and support block. The tooling positioning sensor detects whether the positioning pins are fully inserted, whether the clamping mechanism is in place, and whether the pedal support surface is in stable contact, and records the displacement of each positioning point. If the positioning status does not meet the preset requirements, such as the positioning pins not being inserted, insufficient clamping force, or obvious warping of the reference support surface, the controller can directly output a re-clamping prompt to avoid misjudging the clamping abnormality as a weld abnormality.

[0085] After clamping, the multi-view industrial camera images areas such as the folding stem base, pedal front end, rear fork connection, and handlebar support connection, according to preset shooting positions. Before camera acquisition, exposure compensation, polarization suppression, high-brightness weld reflection suppression, and distortion correction are performed. To ensure that small welds on children's electric scooters can be identified, the actual size corresponding to a single pixel of the camera is preferably no larger than the minimum allowable width of the weld. For example, the minimum allowable width of the weld is At that time, the actual size corresponding to a single pixel is no larger than The line laser profile sensor scans along the weld path or along the area where the weld is located, with the scanning interval preferably not exceeding the minimum allowable width of the weld. For example, the minimum allowable width of the weld is At that time, the scanning interval is no greater than .

[0086] The raw data acquired for the weld includes weld appearance images, weld contour point clouds, and chassis clamping position data. The weld appearance images include the original image, distortion-corrected image, illumination-equalized image, camera intrinsic parameters, and camera extrinsic parameters. The weld contour point cloud includes laser cross-section points, 3D point coordinates, reflection intensity, scan number, and point cloud confidence level. The chassis clamping position data includes locating pin displacement, fixture clamping status, tooling reference point detection values, and detection timestamps. To ensure consistency among multi-source data, all images, point clouds, and sensor data are accompanied by a unified timestamp and workstation number. Data association is established through calibration relationships between the camera, laser, and tooling.

[0087] After data acquisition, the computing unit performs a preliminary quality check on the original weld seam data. For image data, it checks the average brightness, overexposure ratio, underexposure ratio, edge sharpness, and occlusion area ratio; for point cloud data, it checks the point cloud density, missing point ratio, outlier ratio, and cross-sectional continuity; for clamping position data, it checks whether each positioning point is within the allowable range. If the quality of the original acquired information does not meet the requirements, it re-triggers the taking of a picture or re-scans the image, instead of directly proceeding to weld seam defect judgment.

[0088] The original weld seam data is subjected to attitude correction to obtain the chassis coordinate detection information.

[0089] The purpose of attitude correction is to unify the weld appearance image, weld contour point cloud and tooling positioning data into the local coordinate system of the vehicle frame, so that even if there are slight clamping deviations or micro deformations after welding in different vehicle frames, the consistency of the weld can be evaluated based on the same structural benchmark.

[0090] In the specific implementation, the calculation unit first parses the line connecting the pedal plane, the front seat tube axis, and the rear fork mounting hole center from the frame structure reference information, and calls the aforementioned frame local coordinate system. Then, based on the tooling positioning sensor data and the actual collected reference feature points, it extracts the actual clamping posture of the frame to be tested. The actual clamping posture includes translational offset, pitch deviation around the lateral axis, roll deviation around the longitudinal axis, and yaw deviation around the height axis.

[0091] Let the number of sensors detected in the sensor coordinate system be... The benchmark point is The corresponding design reference point in the local coordinate system of the chassis is The actual direction of the front riser is The former management theory direction is The rotation matrix is ​​obtained through weighted rigid body registration. Translation vector :

[0092] ,

[0093] in, Represents the optimal rotation matrix. Represents the optimal rotation matrix. This indicates the number of reference points participating in the registration. Indicates the first The credibility weight of the benchmark point This indicates the constraint weight of the front riser axis direction. This represents a set of three-dimensional rotation matrices. The reliability weight of the reference point is determined based on the machining accuracy of the positioning hole, the stability of the sensor measurement, and whether the point is obstructed. For example, the weight of the center point of the positioning hole is greater than the weight of the edge point of the thin plate that is easily deformed after welding.

[0094] get and Subsequently, a three-dimensional coordinate transformation is directly performed on the weld contour point cloud. The weld boundary points in the weld appearance image are mapped to coordinates using camera calibration parameters, projection models, and corresponding three-dimensional surface points. For image edge points that cannot be directly reconstructed in three dimensions, their approximate three-dimensional positions are obtained by intersecting the adjacent laser cross-section, the known base material surface where the weld is located, and the camera rays. This forms the initial frame coordinate information.

[0095] After attitude correction, a baseline residual check is required to determine the reliability of the clamping or data acquisition. The residual is calculated using the following formula:

[0096] ,

[0097] in, Indicates the baseline residual. This represents the directional residual reduction factor. If If the residual is less than or equal to the allowable residual, the initial frame coordinate information is compensated for the deviation based on the residual direction to obtain the frame coordinate detection information; if If the residual exceeds the allowable error, a prompt to re-clamp or re-acquire data will be output. The allowable residual is determined jointly by the tooling repeatability error and the measurement system repeatability error. For example, if the tooling repeatability error is... The repeatability error of the measurement system is At that time, the allowable residual can be set to about.

[0098] Weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features.

[0099] After obtaining the frame coordinate detection information, the calculation unit uses the designed weld paths in the frame structure reference information as constraints to match the weld positions on the measured data. Specifically, a three-dimensional search window is established centered on each designed weld path, according to the allowable centerline offset range of the weld, the maximum weld width, and the scanning error. For straight welds, the search window is a rectangular tubular region extending along the weld direction; for curved welds, the search window is a curved tubular region that unfolds segment by segment along the tangent of the path. Image edge points and point cloud points located within the search window and conforming to the characteristics of weld texture, contour protrusion, or weld toe transition are assigned to the corresponding weld segment detection area.

[0100] After the weld section detection area is determined, weld boundary identification and cross-sectional contour identification are performed. Weld boundary identification is accomplished through grayscale gradient, texture direction, weld color difference, edge continuity, and abrupt changes in point cloud height. Weld cross-sectional contour identification is accomplished through base material plane fitting in the laser cross-section, extraction of weld protrusion areas, weld toe point identification, and cross-sectional peak point identification. For each cross-section, candidate base material points are selected on both sides of the weld, and weld beads, spatter points, and obvious outliers are removed before fitting the local base material surface. The left and right weld toe points are determined based on the abrupt change in cross-sectional height relative to the base material surface. The weld reinforcement height and cross-sectional shape are determined based on the point cloud height distribution between the left and right weld toe points.

[0101] Weld seam segments are made according to a fixed arc length. The default segment length for ordinary areas can be set to [value missing]. For sections with significant curvature changes or load-sensitive areas such as the folding riser base, pedal front end, and rear fork connection, the segment length can be [determined]. During the trial production phase, the shortest reliably identifiable abnormal length among the rework samples is selected, ensuring that the segment length does not exceed this shortest abnormal length. Multiple sampling points along the weld length are taken within each weld segment to reduce the impact of single-section noise on the test results.

[0102] Let the first The first weld There are a total of Cross-sectional sampling point, number The midpoint of the left and right weld toe points of the sampling points is The designed weld path is The width of the weld in this section is The weld reinforcement height of this section is The design allowance height of this weld or the reference allowance height of the first piece is... The changes in slope of the left and right weld toes are respectively and , No. The length of the discontinuous abnormal region is The consistency information for each weld component is represented as follows:

[0103] ,

[0104] ,

[0105] ,

[0106] ,

[0107] ,

[0108] in, Indicates the offset of the center line. Indicates the width fluctuation amount. Represents the residual height deviation, Indicates the amount of abrupt change in weld toe transition. This indicates the length of the continuity gap. The abrupt change in weld toe transition is obtained from the slope change, curvature change, or height gradient change between the weld edge and the base metal region; the continuity gap is determined by a combination of factors, including weld boundary interruption in the image, disappearance of weld protrusion in the point cloud, abnormal laser reflection intensity, or discontinuity in the cross-sectional profile.

[0109] To avoid false positives caused by localized splashes, oil stains, or reflections, outliers are removed from each cross-sectional feature during feature extraction. Outlier removal methods include median filtering, local window consistency checks, or weighted averaging based on point cloud confidence levels. When the number of effective cross-sections within a weld segment is lower than a preset proportion, the weld segment is not directly judged as qualified; instead, it is marked as having insufficient acquisition confidence, and a re-inspection prompt is output in subsequent grade determinations.

[0110] By marking the load-sensitive areas of the vehicle frame structure reference information, the load-sensitive area information of the weld is obtained.

[0111] The weld risks of children's electric scooter frames are related to geometric deviations and the type of stress the weld is subjected to during child use. Therefore, this embodiment marks the welds as load-sensitive areas based on the key load-bearing area identifiers in the frame structure reference information.

[0112] The sources of load-sensitive area markings fall into four categories: The first category includes critical load-bearing welds pre-marked in structural design drawings or 3D models; the second category includes weld areas with high stress under low-step impact, pedal torsion, and lateral fall conditions in finite element simulations; the third category includes weld areas showing high strain or crack initiation in bench impact tests, torsion tests, and drop tests; and the fourth category includes weld areas with high-frequency failures in historical after-sales fracture records or repair records. If initial test data is lacking, the folding riser seat, pedal front connection, and rear fork connection will be marked as high-sensitive areas by default, and will be corrected subsequently based on test results and after-sales data.

[0113] In the specific marking process, the structural components of the frame's load-bearing welds are first identified, resulting in the locations of the front pedal connection weld, front wheel support connection weld, rear fork connection weld, side pedal connection welds, stem support connection weld, side folding seat tube weld, and auxiliary connection welds. Then, the front pedal connection weld, front wheel support connection weld, and rear fork connection weld are matched as impact-sensitive weld locations; the side pedal connection welds and stem support connection welds are matched as torsion-sensitive weld locations; the side folding seat tube weld and the lower stem connection weld are matched as rollover-sensitive weld locations; and welds on non-primary load-bearing decorative parts or auxiliary support parts are matched as ordinary connection weld locations.

[0114] The information for the load-sensitive zone of a weld includes the weld number, weld segment number, structural component name, load-sensitive zone type, source of the sensitive zone, load factor, source confidence level, and tag version number. The default value for the load factor is: impact-sensitive zone. Reverse sensitive area Side roll sensitive area Normal connection area If the same weld segment belongs to multiple sensitive areas, such as a weld near a folding seat being simultaneously affected by lateral and torsional loads, a larger load factor should be used, or a composite load factor should be used based on the company's test data.

[0115] By jointly evaluating the geometric consistency characteristics of the weld and the information on the load-sensitive area of ​​the weld, the consistency risk information of the weld segment is obtained.

[0116] Before joint evaluation, it is necessary to normalize the geometric consistency characteristics of welds with different dimensions. Let the set of characteristic types be... These represent centerline offset, width fluctuation, excess height deviation, abrupt transition at the weld toe, and continuity gap, respectively. weld number The first weld segment The measured eigenvalues ​​of the class are The risk-free benchmark is The allowed threshold is The confidence level of the data collection is The confidence level correction factor is The stabilization constant is The normalized deviation value can be calculated using the following formula:

[0117] ,

[0118] in, This represents the normalized geometric deviation information.

[0119] When the measured feature value reaches near the allowable threshold near ;

[0120] When the measured feature value is significantly smaller than the allowable threshold Less than ;

[0121] When the measured feature value exceeds the allowable threshold Greater than .

[0122] Collection confidence level The value is determined based on image sharpness, point cloud density, effective cross-section ratio, and sensor synchronization status, and its range is [value range missing]. to When the confidence level is insufficient, the probability of the weld segment being monitored or re-inspected is increased by using correction terms.

[0123] Feature weights are determined based on the company's initial experience, trial production samples, or manual re-inspection samples. In the initial stage, the basic weights for centerline offset, width fluctuation, excess height deviation, weld toe transition abrupt change, and continuous gaps are set equal. After accumulating a certain number of rework and qualified samples, the correlation between various features and manual rework conclusions is calculated, and the weights are updated. Let the first feature be... The basic weights of class features are The rework correlation index is The weight adjustment strength is The updated feature weights are:

[0124] ,

[0125] in, Indicates the first The final weights of the class features, and the sum of all feature weights is... If no rework samples have been accumulated yet, then all At this point, the weights remain at the base weights. If statistical results indicate a high correlation between centerline offset, continuity gaps, and repair conclusions, the weights are increased accordingly. and The weight.

[0126] After obtaining the normalized geometric deviation information and feature weights, the load coefficients in the weld load-sensitive zone information are further read. Let the first... weld number The load factor of the weld segment is The mean of the normalized deviation is The coupling enhancement coefficient is The weld section consistency risk value is calculated using the following formula:

[0127] ,

[0128] in, This represents the weld segment consistency risk value. The first term reflects the weighted summation of various geometric deviations, while the second term reflects the enhanced effect of local anomalies that are particularly prominent in a certain type of feature. For example, if the length of a continuous notch is not large, but it is significantly higher than other features in a highly sensitive area, the risk value will be increased. This formula enables this implementation method to determine whether the weld size is out of tolerance and to distinguish the risk differences of the same geometric deviation in different load-sensitive areas.

[0129] The weld segment consistency risk information includes weld number, weld segment number, weld segment start and end coordinates, weld segment center coordinates, load-sensitive zone type, load coefficient, normalized deviation value of each component, weight of each component, comprehensive risk value, main anomaly characteristics, and data confidence level. Main anomaly characteristics are identified through comparison. The contribution size is determined. For example, when the contribution of the continuity gap is the largest, the anomaly type of the weld segment is marked as a continuity anomaly; when the contribution of the centerline offset is the largest, the anomaly type of the weld segment is marked as a weld centerline offset anomaly.

[0130] The consistency risk information of the weld section is classified into levels, and the consistency judgment result of the frame weld is output.

[0131] During the risk assessment, the calculation unit performs threshold matching on the consistency risk values ​​of each weld segment. Let the risk thresholds be as follows: , , , No. weld number The grade of the weld segment is determined by the following formula:

[0132] ,

[0133] in, This indicates the weld segment level. If a weld segment is located in an impact-sensitive, torsion-sensitive, or rollover-sensitive zone and reaches a high-risk abnormality level, an interception command will be output to prohibit its flow into the painting or assembly process. If a weld segment is at the attention level, a re-inspection prompt will be output; if it is at the abnormal level, a rework suggestion will be output; if all weld segments of the entire vehicle are at the qualified level, a consistency qualified conclusion will be output.

[0134] The overall vehicle weld consistency level is determined by using the most severe weld segment grade. Simultaneously, the number of abnormal weld segments, abnormal weld lengths, the type of sensitive area where the abnormality is located, and the main abnormal characteristics are statistically analyzed. If multiple consecutive weld segments within the same weld exhibit centerline offset anomalies, they are grouped into an abnormal region, and the starting, ending, and center coordinates of this abnormal region are output. If the anomaly is located in a highly sensitive area (i.e., the geometric deviation is slightly below the rework threshold for a normal area), the rework priority is increased.

[0135] The consistency assessment results for chassis welds are output via a workstation display interface, inspection report, QR code traceability data, or production line control signals. The inspection report includes the vehicle model number, chassis number, inspection time, inspection equipment number, overall vehicle consistency level, abnormal weld number, abnormal weld segment location, abnormal feature type, load-sensitive area type, risk value, rework priority, re-inspection recommendations, prohibition of transfer instructions, and corresponding image screenshots and point cloud cross-sectional diagrams. Rework personnel directly locate the specific weld area on the chassis based on the abnormal weld segment location. Process engineers determine whether adjustments to tooling positioning, welding speed, wire feed stability, welding torch angle, or heat input are necessary based on the abnormality type.

[0136] In a further implementation, the inspection results of individual vehicle frames are summarized into batch process feedback information. The calculation unit statistically analyzes the consistency risk information of weld segments across multiple vehicle frames based on fields such as vehicle model, tooling number, welding shift, welder number, robot program number, welding current, welding voltage, wire feed speed, and welding speed. If the centerline offset anomaly occurs consecutively at the same weld location under the same vehicle frame local coordinates, a tooling positioning verification prompt is output; if the width fluctuation anomaly occurs consecutively at the same weld location, a welding speed or wire feed stability verification prompt is output; if the weld toe transition abruptly occurs consecutively in a highly sensitive area, a welding torch angle, heat input, and shielding gas status verification prompt is output; if multiple vehicle frames in the same batch exhibit overall offset in the same direction, a fixture wear or positioning reference offset prompt is output. This batch feedback is not a simple quality statistic, but rather an attribution of causes based on the same weld location, the same vehicle frame local coordinates, and the same anomaly characteristic type, forming a closed loop from inspection results to welding process adjustments.

[0137] The frame structure reference information is generated jointly from the vehicle model number, 3D design model, welding process documents, tooling positioning diagram, measurement results of the first qualified sample, and records of critical load-bearing areas. The generation process includes vehicle model identification, design weld path analysis, critical reference point fitting, weld threshold fielding, and load-bearing area mapping. It includes reference point coordinates, pedal plane, front seat tube axis, rear fork mounting hole center line, design weld path, weld allowable threshold, and critical load-bearing area identification. Its purpose is to provide a unified structural foundation for attitude correction, weld position matching, feature threshold determination, and load-sensitive area marking. Subsequent steps use this information to establish a local coordinate system for the frame and correlate measured weld data with the design weld path segment by segment.

[0138] The raw weld seam information is acquired synchronously by a multi-view industrial camera, a line laser profile sensor, and a tooling positioning sensor after the chassis is clamped. The generation process includes positioning status confirmation, image capture, laser profile scanning, timestamp synchronization, sensor calibration parameter binding, and acquisition quality checks. It includes weld seam appearance images, weld seam profile point clouds, camera intrinsic and extrinsic parameters, laser scan sequence number, positioning pin displacement, clamping status, and acquisition confidence level. Its purpose is to provide the raw data source for weld seam boundaries, weld cross-sections, and the actual clamping posture. Subsequent steps transform this information into the chassis local coordinate system based on calibration relationships and clamping posture.

[0139] The frame coordinate detection information is obtained from the original weld acquisition information after being constrained by the frame local coordinate system, clamping posture analysis, coordinate transformation, and residual compensation. The generation process includes datum feature recognition, rigid body registration, 3D mapping of image edges, point cloud coordinate transformation, and datum residual verification. It includes weld image boundary points, weld point cloud points, weld section points, actual datum point positions, and acquisition residuals located in the frame local coordinate system. Its purpose is to eliminate tooling clamping deviations and sensor coordinate differences, enabling welds of different frames to be compared under the same structural datum. Subsequent steps use this information as input to extract centerline, width, reinforcement height, weld toe transition, and continuity features along the designed weld path.

[0140] The weld geometric consistency feature is generated jointly by the chassis coordinate detection information and the designed weld path. The generation process includes weld position matching, weld segment detection area establishment, weld boundary identification, cross-sectional profile identification, segmented sampling, and feature calculation. This includes centerline offset, width fluctuation, excess height deviation, weld toe transition abrupt change, continuity notch length, and corresponding data confidence levels. Its function is to convert the weld appearance and 3D profile into quantifiable, comparable, and traceable consistency indicators. Subsequent steps normalize and weight this feature, and calculate the risk value together with weld load-sensitive area information.

[0141] The weld load-sensitive zone information is generated from the critical load-bearing area identifiers, structural component identification results, finite element simulation results, bench test results, or historical failure records in the vehicle frame structural reference information. The generation process includes identifying the load-bearing weld location, classifying structural components, matching child-use load types, and assigning sensitive zone coefficients. It includes weld number, weld segment number, structural component, impact-sensitive zone marker, torsion-sensitive zone marker, rollover-sensitive zone marker, ordinary connection zone marker, load coefficient, and marker source. Its purpose is to incorporate the risks of low-step impacts, pedal torsion, and lateral falls during child use into the weld consistency evaluation. Subsequent steps read the load coefficients from this information according to the weld segment positional relationships, and perform risk amplification or general evaluation of geometric deviations.

[0142] The weld segment consistency risk information is generated jointly from the normalized results of weld geometric consistency features, feature weights, and load coefficients from the weld load-sensitive area information. The generation process includes segment reading, feature normalization, feature weight allocation, load coefficient matching, comprehensive risk calculation, and identification of major anomalies. It includes weld segment location, normalized deviation values ​​for each component, contribution of each component, load-sensitive area type, load coefficient, comprehensive risk value, and major anomaly types. Its function is to transform simple geometric deviations into weld segment evaluation results that reflect the actual load-bearing risk of children's electric scooters. Subsequent steps use this risk information to determine the risk level, locate anomalies, prioritize rework, and provide process feedback.

[0143] The frame weld consistency assessment results include the overall weld consistency level, location of abnormal weld segments, type of abnormal features, type of load-sensitive area, risk value of weld segment, rework priority, re-inspection prompts, prohibition of transfer instructions, and corresponding image and point cloud evidence. This result is derived step-by-step from frame structural reference information, frame coordinate detection information, weld geometric consistency features, weld load-sensitive area information, and weld segment consistency risk information, rather than directly from a single image template or single defect classification result. This result guides online inspection release, rework location, manual re-inspection, pre-painting interception, welding process parameter adjustment, and fixture maintenance. This result addresses the problems in existing technologies where fixed template inspection is difficult to distinguish between clamping posture deviations and actual weld anomalies, and where simple visual inspection cannot reflect the load risk associated with child use. Through full-chain local coordinate reference correction, three-dimensional consistency feature extraction, and joint evaluation of load-sensitive areas, this implementation achieves stable identification and actionable handling of consistency risks in key load-bearing welds of children's electric scooters.

[0144] Example 1: Used to verify whether the interference of clamping posture deviation and post-weld micro-deformation on the weld consistency detection results can be reduced, and whether the abnormal weld consistency can be distinguished from the overall posture change of the frame.

[0145] Children's electric scooter frames of the same model were selected as the test objects, all in a state after welding and before surface painting. Before testing, basic data of the children's electric scooter frames were acquired to obtain frame structural reference information, including the pedal plane, the front stem axis, the center line connecting the rear fork mounting holes, and the location of key load-bearing welds. Subsequently, multi-source test data of the frames under test were collected to obtain raw weld information, including weld appearance images, weld contour point clouds, and frame clamping position data.

[0146] To verify the attitude correction effect, the chassis samples were divided into four categories: Category 1: chassis with normal clamping attitude and normal weld consistency; Category 2: chassis with skewed clamping attitude but normal welds relative to the chassis structural reference; Category 3: chassis with slight overall deformation after welding but welds still within the allowable range relative to the chassis structural reference; Category 4: chassis with welds exhibiting centerline offset, width fluctuation, or continuity abnormalities relative to the chassis structural reference. Inspectors verified the actual weld positions and their relationship to the chassis structural reference to generate a comparison result.

[0147] During inspection, a fixed template inspection method without frame structural benchmark correction was first used for identification, followed by a frame weld consistency inspection method. In the frame weld consistency inspection method, the frame structural benchmark information was analyzed to determine the line connecting the pedal plane, the front seat tube axis, and the center of the rear fork mounting holes. A local coordinate system for the frame was constructed using the pedal plane as the plane benchmark, the front seat tube axis as the height reference, and the line connecting the centers of the rear fork mounting holes as the lateral correction reference. Clamping deviations were extracted from the original weld data to obtain the actual clamping posture of the frame under inspection. Using the local coordinate system and the actual clamping posture as constraints, coordinate transformation and deviation compensation were performed on the original weld data to obtain the frame coordinate inspection information.

[0148] After obtaining the frame coordinate detection information, weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features. For welds with overall positional offsets during fixed template detection, it is further determined whether the offset still exists within the frame's local coordinate system. If the weld has an overall offset in the image, but after coordinate transformation and deviation compensation, the positional relationship between the weld and the line connecting the center of the pedal plane, the front riser axis, and the rear fork mounting hole is restored to normal, this situation is judged as clamping posture deviation or post-weld overall micro-deformation, and weld consistency anomalies are not directly output. If, after posture correction, the weld centerline, width, excess height, weld toe transition, or continuity still shows anomalies relative to the frame structural reference, weld anomaly positioning information is output.

[0149] Table 1 shows the identification results of the two detection methods for different parts.

[0150] The clamping is normal and the weld is normal. The clamping is misaligned, but the weld is normal. The weld seam was slightly deformed after welding, but the overall shape remained normal. Abnormal weld consistency

[0151] As shown in Table 1, the fixed template detection method resulted in numerous anomaly detections in samples with clamping misalignment and post-weld micro-deformation. However, the verification results indicated that the welds of these samples did not exhibit any real anomalies relative to the frame structural reference. This suggests that the fixed template detection method is prone to misjudging changes in the overall frame posture as weld consistency anomalies. Furthermore, the number of genuine weld consistency anomalies identified by the fixed template detection method was lower than the number of anomalies verified, indicating that when the frame structural reference has shifted but the weld appearance remains close to the standard template, it is prone to missed detections.

[0152] This invention corrects the orientation of the original weld acquisition information by using the local coordinate system of the vehicle frame, the actual clamping posture, coordinate transformation, and deviation compensation. This allows weld inspection to be performed without relying on fixed image positions or fixed inspection windows, instead judging the geometric consistency of the weld under a unified vehicle frame structural reference. Inspection results show that this invention can significantly reduce false positives caused by clamping misalignment and post-weld micro-deformation, and improve the completeness of identifying true weld consistency anomalies.

[0153] Therefore, this embodiment can verify that the present invention adopts a posture correction, coordinate transformation and deviation compensation technical solution based on the frame structure reference, which can achieve the technical effect of distinguishing the clamping posture deviation, post-weld micro deformation and the abnormal consistency of the actual weld, realize the accurate detection of the weld relative to the frame stress reference, and solve the shortcomings of the fixed template detection which is prone to misjudgment and omission.

[0154] Example 2: This example is used to verify whether abnormal weld geometry consistency can be correlated with the risk of child use loads, and whether different risk levels and repair priorities can be assigned to weld segments with similar appearance abnormalities.

[0155] The same model of children's electric scooter frame was selected as the test object. First, the basic data of the children's electric scooter frame was obtained to obtain the frame structure reference information; then, multi-source detection data of the frame to be tested was collected to obtain the original weld acquisition information; then, the attitude correction of the original weld acquisition information was performed to obtain the frame coordinate detection information, and the weld consistency feature was extracted from the frame coordinate detection information to obtain the weld geometric consistency feature.

[0156] In this embodiment, the geometric consistency characteristics of the weld include centerline offset, width fluctuation, excess height deviation, weld toe transition variation, and continuity anomaly. For different weld segments, if only the appearance image is observed, the geometric deviation of some weld segments is relatively similar. Traditional appearance inspection methods can usually only output the same or similar defect conclusions, which is not easy to reflect the differences in the loads experienced by children on the frame part where the weld segment is located.

[0157] To this end, the present invention marks the load-sensitive areas of the frame structure reference information to obtain the load-sensitive area information of the welds. Specifically, the critical load-bearing positions of the frame structure reference information are identified to obtain the positions of the load-bearing welds of the frame; the load-bearing weld positions of the frame are matched with the load types of child use, and the positions of the front foot pedal connection weld, the front wheel support connection weld, and the rear fork connection weld are matched as impact-sensitive weld positions, the positions of the side foot pedal connection welds and the stem support connection welds are matched as torsion-sensitive weld positions, the position of the side weld of the folding seat tube is matched as a rollover-sensitive weld position, and the auxiliary connection weld positions are matched as ordinary connection weld positions.

[0158] Then, the geometric consistency characteristics of the weld are segmented and normalized to obtain the geometric deviation information of the weld segment; the sensitive area coefficients of the weld load-sensitive area are matched to obtain the load coefficient information of the weld segment; feature weights are assigned to the geometric deviation information of the weld segment to obtain the weighted deviation information of the weld segment; using the load coefficient information of the weld segment as a constraint, risk calculation is performed on the weighted deviation information of the weld segment to obtain the consistency risk information of the weld segment. Finally, the consistency risk information of the weld segment is graded, and the consistency judgment result of the frame weld is output.

[0159] Table 2 shows the risk assessment results of different weld segments with similar geometric deviations under the vehicle frame weld consistency inspection method.

[0160] Front end weld of the pedal Shock sensitive Consistency anomaly Stand support connection weld Torture Sensitivity Consistency anomaly Side weld of folded riser seat Side roll sensitivity Consistency anomaly Auxiliary connection weld Normal sensitivity Consistency Focus

[0161] As shown in Table 2, with the same weighted deviation value, different weld segments exhibit varying risk values ​​and assessment results due to their different locations on the frame and corresponding load-sensitive areas. The front footpeg connection weld is associated with low-step impacts during child use, resulting in a high risk value; the stem support connection weld is associated with footpeg torsion or stem stress, leading to a higher risk value than ordinary connection welds; the folding seat tube side weld is associated with lateral fall loads, also resulting in a higher risk value than ordinary connection welds; although the auxiliary connection welds have similar geometric deviations, they are classified as consistency concern because they are not located in primary load-bearing areas.

[0162] The results indicate that this invention does not rely on the presence of weld defects as the sole criterion for judgment, but rather combines this with a joint evaluation of the load-sensitive area for children using the frame where the weld is located. For abnormal weld segments that appear minor but are located in high-risk load-bearing areas, this invention can improve the risk assessment level; for similar abnormalities located in ordinary connection areas, this invention can avoid simply making a high-risk judgment, ensuring that the test results reflect the actual load-bearing risk of the children's electric scooter frame.

[0163] Therefore, this embodiment can verify that the technical solution of the present invention, which uses the joint evaluation of weld geometric consistency characteristics and weld load-sensitive area information, achieves the technical effect of associating weld anomalies with the risk of loads used by children, realizes risk classification and anomaly location of weld segments, and solves the problem that simple appearance defect identification is not easy to reflect the consistency risk of key welds.

[0164] Embodiments of the present invention have been presented and described. It will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

[0165] The embodiments of the present invention described above are subject to modification and change of method by those skilled in the art without departing from the embodiments and broader aspects of the present invention. The appended claims are intended to include all such modifications and changes of method that do not depart from the present invention.

[0166] Those skilled in the art should understand that the embodiments of the present invention can be implemented using a pure hardware architecture, a pure software architecture, or an integrated hardware and software architecture. The present invention can be prepared as a computer program product, which can be stored in various non-volatile computer-readable storage media, including but not limited to solid-state drives, flash memory chips, mobile storage devices, optical discs, cloud storage servers, and other standardized storage media, and is not limited to traditional storage media.

[0167] The embodiments of the present invention described above are subject to modification and change of method by those skilled in the art without departing from the embodiments and broader aspects of the present invention. The appended claims are intended to include all such modifications and changes of method that do not depart from the present invention.

[0168] Based on the foregoing description in conjunction with the accompanying drawings, those skilled in the art will understand that the embodiments of this application can also be implemented by software programs. Therefore, this application also provides a computer-readable storage medium. This computer-readable storage medium stores computer-readable instructions thereon, which, when executed by one or more processors, implement the method described above in conjunction with the accompanying drawings.

[0169] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0170] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0171] It should be understood that when the terms "first," "second," "third," and "fourth," etc., are used in the claims, specification, and drawings of this application, they are used only to distinguish different objects and not to describe a specific order. The terms "comprising" and "including" as used in the specification and claims of this application indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.

[0172] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this specification and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this specification and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0173] Although the embodiments of this application are described above, the content is merely an example adopted for the purpose of facilitating understanding of this application and is not intended to limit the scope and application scenarios of this application. Any person skilled in the art described in this application may make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in this application, but the scope of patent protection of this application shall still be determined by the scope defined in the appended claims.

Claims

1. A method for detecting the consistency of weld seams in a children's electric scooter frame, characterized in that, include: Obtain the basic data of the children's electric scooter frame to obtain the frame structure reference information; Collect multi-source inspection data of the chassis to be inspected to obtain the original information of the weld. The original weld seam data is subjected to attitude correction to obtain the vehicle frame coordinate detection information; Weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features; The load-sensitive area of ​​the frame structure reference information is marked to obtain the load-sensitive area information of the weld. By jointly evaluating the geometric consistency characteristics of the weld and the information on the load-sensitive area of ​​the weld, the consistency risk information of the weld segment is obtained. The consistency risk information of the weld segment is classified into levels, and the consistency judgment result of the frame weld is output.

2. The method for detecting the consistency of chassis welds according to claim 1, characterized in that, The original weld seam acquisition information is subjected to attitude correction to obtain the frame coordinate detection information, including: The frame structure reference information is analyzed to determine the line connecting the pedal plane, the front stem axis, and the center of the rear fork mounting hole; A local coordinate system for the frame is constructed using the pedal plane as the plane reference, the front stem axis as the height direction reference, and the line connecting the centers of the rear fork mounting holes as the lateral correction reference. The clamping deviation is extracted from the original information collected from the weld to obtain the actual clamping posture of the frame to be inspected; Using the local coordinate system of the vehicle frame and the actual clamping posture as constraints, coordinate transformation and deviation compensation are performed on the original weld seam acquisition information to obtain the vehicle frame coordinate detection information.

3. The method for detecting the consistency of chassis welds according to claim 1, characterized in that, Weld consistency features are extracted from the frame coordinate detection information to obtain weld geometric consistency features, including: Using the frame structure reference information as a constraint, the weld position is matched with the frame coordinate detection information to obtain the weld segment detection area; Weld boundary identification and cross-sectional contour identification are performed on the weld segment detection area to obtain weld boundary information and weld cross-sectional information; The centerline, width, reinforcement height, weld toe transition, and continuity are calculated on the weld boundary information and weld section information to obtain weld component consistency information; The weld seam component consistency information is collected to obtain the weld seam geometric consistency characteristics.

4. The method for detecting the consistency of chassis welds according to claim 1, characterized in that, The load-sensitive area of ​​the frame structure reference information is marked to obtain the weld load-sensitive area information, including: The key load-bearing positions of the frame structure reference information are identified to obtain the positions of the frame load-bearing welds; The load-bearing weld locations of the vehicle frame are matched with child-use load types to obtain impact-sensitive weld locations, torsion-sensitive weld locations, rollover-sensitive weld locations, and ordinary connection weld locations; The locations of impact-sensitive welds, torsion-sensitive welds, rollover-sensitive welds, and ordinary connection welds are marked to obtain information on weld load-sensitive zones.

5. The method for detecting the consistency of chassis welds according to claim 1, characterized in that, The weld segment consistency risk information is obtained by jointly evaluating the weld geometric consistency characteristics and the weld load-sensitive area information, including: The geometric consistency characteristics of the weld are segmented and normalized to obtain the geometric deviation information of the weld segment. The load-sensitive area information of the weld is matched with the sensitive area coefficient to obtain the load coefficient information of the weld segment; The geometric deviation information of the weld segment is weighted by feature weight allocation to obtain the weighted deviation information of the weld segment; Using the load coefficient information of the weld segment as a constraint, risk calculation is performed on the weighted deviation information of the weld segment to obtain the consistency risk information of the weld segment.

6. The method for detecting the consistency of chassis welds according to claim 1, characterized in that, The consistency risk information of the weld segments is assessed to determine the level, and the consistency assessment result of the chassis welds is output, including: Risk threshold matching is performed on the consistency risk information of the weld segment to obtain the weld segment level information; The weld segment grade information is summarized and judged for the whole vehicle to obtain the frame weld consistency grade. The abnormal location and abnormal type are extracted from the weld segment grade information to obtain weld abnormal location information; The results of the consistency level of the frame welds and the location information of weld anomalies are integrated to output the consistency judgment result of the frame welds.

7. The method for detecting the consistency of chassis welds according to claim 2, characterized in that, The original weld seam data is transformed and offset compensation is performed to obtain the chassis coordinate detection information, including: The actual clamping posture is analyzed to obtain clamping posture offset information; A mapping relationship is constructed between the local coordinate system of the vehicle frame and the clamping posture offset information to obtain the coordinate transformation relationship; Based on the coordinate transformation relationship, coordinate transformation is performed on the weld appearance image and weld contour point cloud in the original weld acquisition information to obtain the initial frame coordinate information; The initial frame coordinate information is checked for reference residuals to obtain the coordinate deviation compensation amount; Based on the coordinate deviation compensation amount, the initial frame coordinate information is offset to obtain the frame coordinate detection information.

8. The method for detecting the consistency of chassis welds according to claim 3, characterized in that, The centerline, width, reinforcement height, weld toe transition, and continuity are calculated based on the weld boundary information and the weld cross-section information to obtain weld component consistency information, including: The weld boundary information is mapped to boundary points to obtain weld boundary point pairs; By performing midpoint connection and spacing calculation on the weld boundary point pairs, the weld centerline information and weld width information are obtained; The weld cross-section information is used to calculate the height difference to obtain the weld reinforcement information; The edge slope change is calculated based on the weld boundary information and weld section information to obtain weld toe transition information; Discontinuous region identification and length statistics are performed on the weld boundary information and weld cross-section information to obtain weld continuity information; The weld centerline information, weld width information, weld reinforcement information, weld toe transition information, and weld continuity information are collected to obtain the weld component consistency information.

9. The method for detecting the consistency of chassis welds according to claim 4, characterized in that, The load-bearing weld locations on the vehicle frame are matched with child-use load types to obtain impact-sensitive weld locations, torsion-sensitive weld locations, rollover-sensitive weld locations, and ordinary connection weld locations, including: Structural component identification was performed on the load-bearing weld locations of the frame to obtain the locations of the front foot pedal connection weld, the front wheel support connection weld, the rear fork connection weld, the two sides of the foot pedal connection weld, the stem support connection weld, the side weld of the folding stem seat, and the auxiliary connection weld. Impact load matching was performed on the front end connection weld of the pedal, the front wheel support connection weld and the rear fork connection weld to obtain the impact-sensitive weld location. Torsional load matching was performed on the connection welds on both sides of the pedal and the connection weld of the stem support to obtain the torsion-sensitive weld locations. The position of the side weld of the folded riser is matched with the overturning load to obtain the position of the overturning sensitive weld. The auxiliary connection weld positions are matched with ordinary connections to obtain the ordinary connection weld positions.

10. The method for detecting the consistency of chassis welds according to claim 5, characterized in that, Using the load coefficient information of the weld segment as a constraint, risk calculation is performed on the weighted deviation information of the weld segment to obtain the consistency risk information of the weld segment, including: The weighted deviation information of the weld segment is read in segments to obtain the weighted deviation value of each weld segment; The load coefficient information of the weld segment is read accordingly to obtain the load coefficient value of each weld segment; Using the positional correspondence of each weld segment as a constraint, the weighted deviation value and the load coefficient value of each weld segment are matched to obtain the risk calculation object of the weld segment; The risk value of the weld segment is obtained by multiplying the risk calculation objects of the weld segment. By associating the risk value of the weld segment with the corresponding weld segment location, the consistency risk information of the weld segment is obtained.

Images