A welding workpiece quality detection method and device, a storage medium and a vehicle

Abstract

The application discloses a welding workpiece quality detection method and device, a storage medium and a vehicle, relates to the technical field of machine vision detection, and comprises the following steps: determining all welding point regions of a target welding workpiece based on a two-dimensional image of the target welding workpiece, mapping the welding point regions to a three-dimensional point cloud space of the target welding workpiece, and determining welding point three-dimensional point cloud data corresponding to the welding point regions; in the three-dimensional point cloud space, determining workpiece three-dimensional point cloud data corresponding to a key detection region of the target welding workpiece, performing plane fitting on the workpiece three-dimensional point cloud data, detecting the height size of a to-be-detected region, and determining a first key detection region with a size out of tolerance; calculating actual welding point center positions of the welding point regions according to the welding point three-dimensional point cloud data and the two-dimensional image, and determining a first welding point with a position offset; and detecting the quality of the target welding workpiece according to the first key detection region and / or the first welding point, so that the comprehensive detection of the size and the welding point is completed in a single station.

Description

Technical Field

[0001] This application relates to the field of machine vision inspection technology, and in particular to a method for inspecting the quality of welded workpieces, a device for inspecting the quality of welded workpieces, a storage medium, and a vehicle. Background Technology

[0002] In industrial welding processes, the number of weld points, positional accuracy, and dimensional deviations in critical areas of the workpiece are core indicators that determine product quality, directly affecting product assembly accuracy and reliability. In related technologies, the detection of these indicators is mainly divided into three categories: manual inspection, contact measurement, and single machine vision inspection.

[0003] Among the aforementioned inspection methods, manual inspection is extremely inefficient, and the results are highly subjective due to the operator's experience and fatigue, making it difficult to guarantee accuracy. Contact measurement can cause potential scratches and pressure damage to the workpiece surface and cannot achieve high-speed automated inspection, making it completely unsuitable for the pace requirements of modern industrial production lines. Single machine vision inspection has low efficiency in integrating inspection results, cannot quickly generate unified quality assessment reports, and is also difficult to effectively feed the inspection results back to the production line control system.

[0004] Moreover, the presence of textures and welding spatter on the surface of the welded workpieces necessitates high precision in feature extraction, which impacts the overall production efficiency of the production line. Summary of the Invention

[0005] The purpose of this invention is to provide a method, device, electronic equipment, storage medium, and vehicle for inspecting the quality of welded workpieces, at least addressing the technical problem of how to improve the accuracy of feature extraction and how to accurately detect weld points, thereby improving the overall production efficiency of the production line.

[0006] This invention provides the following solution:

[0007] According to one aspect of the present invention, a method for inspecting the quality of welded workpieces is provided, comprising:

[0008] Based on the two-dimensional image of the target welding workpiece, all welding point regions of the target welding workpiece are determined, and the welding point regions are mapped into the three-dimensional point cloud space of the target welding workpiece to determine the three-dimensional point cloud data of the welding point corresponding to the welding point regions.

[0009] In the three-dimensional point cloud space, the workpiece three-dimensional point cloud data corresponding to the key detection area of ​​the target welded workpiece is determined, the workpiece three-dimensional point cloud data is fitted to a plane, the height dimension of the area to be detected is detected, and the first key detection area with out-of-tolerance dimensions is determined.

[0010] Based on the three-dimensional point cloud data of the solder joint and the two-dimensional image, calculate the actual center position of the solder joint in the solder joint area and determine the first solder joint with positional offset.

[0011] The quality of the target welded workpiece is inspected based on the first key detection area and / or the first weld point.

[0012] Preferably, the step of performing planar fitting on the three-dimensional point cloud data of the workpiece, detecting the height dimension of the area to be detected, and determining the first key detection area with out-of-tolerance dimensions includes:

[0013] Remove noise and invalid data from the workpiece's 3D point cloud data to obtain the processed workpiece 3D point cloud data.

[0014] In the processed 3D point cloud data of the workpiece, a plane fitting is performed on the 3D point cloud data of the workpiece to determine the plane normal vector of the key detection area;

[0015] Based on the plane normal vector, the height dimension deviation of each of the key detection regions is detected.

[0016] Preferably, detecting the height dimension deviation of each of the key detection regions based on the plane normal vector includes:

[0017] For each of the key detection areas, in response to the angular offset between the plane normal vector and the template plane normal vector, the number of measured points within the key detection area and the height dimension of each measured point are calculated.

[0018] Determine the rejection percentage, and based on the rejection percentage and the number of measured points, determine the number of height dimensions to be rejected;

[0019] Based on the number of height dimensions to be removed, the height dimensions are removed in descending order, and the average height dimension is calculated based on the removed height dimensions and the number of measured points.

[0020] The height dimension average and the template height dimension are compared to determine the height dimension deviation of the key detection area, until the height dimension deviation of all key detection areas is determined.

[0021] Preferably, the first key detection area for determining dimensional deviations includes:

[0022] Each height dimension deviation is compared with a preset tolerance threshold, and height dimension deviations that are greater than the tolerance threshold are determined to be out of tolerance.

[0023] The critical detection area corresponding to the dimensional deviation is determined as the first critical detection area.

[0024] Preferably, the step of calculating the actual center position of the solder joint in the solder joint region based on the three-dimensional point cloud data of the solder joint and the two-dimensional image, and determining the first solder joint with a positional offset, includes:

[0025] The two-dimensional image is preprocessed and its size is normalized to obtain a standard two-dimensional image;

[0026] Extract the solder joint features from the standard two-dimensional image, generate a solder joint detection box based on the solder joint features, and determine the center position of the solder joint detection box;

[0027] The centroid of the weld point region is calculated based on the three-dimensional point cloud data of the weld point, and the actual center position of the weld point is determined based on the centroid and the center position.

[0028] Obtain the center point position of the template weld point, and calculate the Euclidean distance between the actual weld point center position and the template weld point center position;

[0029] Obtain a preset distance threshold, compare the distance threshold with the Euclidean distance, and determine the first solder joint with a positional offset.

[0030] Preferably, the step of inspecting the quality of the target welded workpiece based on the first key detection area and / or the first weld point includes:

[0031] A confidence analysis is performed on the first key detection area and / or the first solder joint. The first key detection area with a confidence level not lower than the confidence level threshold is identified as a defect area, and the first solder joint with a confidence level not lower than the confidence level threshold is identified as a defective solder joint.

[0032] Based on the established quality standards, the distribution of the defective areas and the number of defective weld points are compared to detect the quality of the target welded workpiece.

[0033] Preferably, after comparing the distribution of the defective areas and the number of defective weld points based on the set quality acceptance criteria to detect the quality of the target welded workpiece, the method further includes:

[0034] The workpiece number of the target welded workpiece, the total number of weld points, the number of the first weld points, the height dimensional deviation, and the qualification status of the target welded workpiece are determined, and a quality inspection and evaluation result is generated.

[0035] The quality inspection and evaluation results are stored in a designated database.

[0036] According to a second aspect of the present invention, a welding workpiece quality inspection device is provided, comprising:

[0037] The data determination module is used to determine all the weld point regions of the target welding workpiece based on a two-dimensional image of the target welding workpiece, map the weld point regions to the three-dimensional point cloud space of the target welding workpiece, and determine the three-dimensional point cloud data of the weld points corresponding to the weld point regions.

[0038] The dimension detection module is used to determine the workpiece three-dimensional point cloud data corresponding to the key detection area of ​​the target welded workpiece in the three-dimensional point cloud space, perform plane fitting on the workpiece three-dimensional point cloud data, detect the height dimension of the area to be detected, and determine the first key detection area with out-of-tolerance dimensions.

[0039] The position detection module is used to calculate the actual center position of the solder joint in the solder joint area based on the three-dimensional point cloud data of the solder joint and the two-dimensional image, and to determine the first solder joint with positional offset.

[0040] The quality inspection module is used to inspect the quality of the target welded workpiece based on the first key inspection area and / or the first weld point.

[0041] According to three aspects of the present invention, an electronic device is provided, comprising: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0042] The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of the welding workpiece quality inspection method.

[0043] According to four aspects of the present invention, a computer-readable storage medium is provided, comprising: storing a computer program executable by an electronic device, wherein when the computer program is run on the electronic device, the electronic device causes the electronic device to perform the steps of a method for inspecting the quality of welded workpieces.

[0044] According to five aspects of the present invention, a vehicle is provided, comprising:

[0045] Electronic equipment used to implement the steps of a method for inspecting the quality of welded workpieces;

[0046] The processor runs a program, and when the program runs, it executes the steps of a welded workpiece quality inspection method based on data output from electronic devices.

[0047] Storage medium for storing programs that, when run, execute steps of a welding workpiece quality inspection method based on data output from an electronic device.

[0048] The above solution achieves the following beneficial technical effects:

[0049] According to the present application, all welding spot regions of the target welded workpiece are determined through the two-dimensional image of the target welded workpiece, the welding spot regions are mapped to the three-dimensional point cloud space of the target welded workpiece to determine the three-dimensional point cloud data corresponding to the welding spot regions, so as to avoid blind search in the full-image point cloud. The two-dimensional image and the three-dimensional point cloud data of the target welded workpiece are collected by the image acquisition device triggered by the logic controller, which can realize the linkage between detection and production line production, avoid the delay and error caused by manual triggering, and lay a foundation for subsequent closed-loop control.

[0050] According to the present application, plane fitting is performed on the three-dimensional point cloud data, the height dimension of all key detection regions is detected, and the first welding spot with out-of-tolerance dimension is determined, which breaks through the limitations that the traditional two-dimensional vision can only detect plane dimensions and the laser ranging can only perform single-point detection, realizes the global planar analysis of the workpiece height dimension deviation, accurately captures the height deviation caused by welding deformation and assembly error, and is fused with the welding spot data to form a comprehensive detection dimension.

[0051] According to the present application, the actual welding spot center position of the welding spot region is calculated based on the three-dimensional point cloud data and the two-dimensional image, and the first key detection region with position offset is determined, which can effectively avoid the interference of welding spatter and workpiece surface texture, realize high-precision positioning and quantity statistics of welding spots, and solve the problem of low detection accuracy of traditional template matching.

[0052] According to the present application, the quality of the target welded workpiece is detected based on the height dimension and the position offset, which can realize the comprehensive detection of dimension and welding spots in a single station for the industrial fixed-station detection scenario, replace the traditional mode of separate detection by multiple devices or manual detection, and reduce equipment investment and labor cost; the detection standards are uniformly set by the system, which avoids the subjectivity of manual detection, and effectively improves the consistency of product quality and process compliance. Brief Description of the Drawings

[0053] Figure 1 is a flowchart of a method for detecting the quality of a welded workpiece provided by one or more embodiments of the present invention.

[0054] Figure 2 is a schematic diagram of a quality detection scenario for a welded workpiece provided by a specific embodiment of the present invention.

[0055] Figure 3 is a schematic diagram of quality detection for a welded workpiece provided by a specific embodiment of the present invention.

[0056] Figure 4 is a schematic diagram of a quality detection evaluation result provided by a specific embodiment of the present invention.

[0057] Figure 5 is a flowchart of quality detection for a target vehicle model provided by a specific embodiment of the present invention.

[0058] Figure 6 This is a structural diagram of a welding workpiece quality inspection device provided in one or more embodiments of the present invention.

[0059] Figure 7 This is a structural block diagram of an electronic device for a method of inspecting the quality of welded workpieces provided in one or more embodiments of the present invention. Detailed Implementation

[0060] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] For the quality inspection of welded workpieces, traditional inspection methods mostly involve manual visual inspection of weld points and measurement of dimensions using contact measuring tools. Manual inspection is extremely inefficient, and the results are greatly affected by the operator's experience and fatigue, making it highly subjective and difficult to guarantee accuracy. Contact measurement can cause potential scratches and pressure damage to the workpiece surface and cannot achieve high-speed automated inspection, making it completely unsuitable for the pace requirements of modern industrial production lines.

[0062] Existing machine vision inspection technologies mostly use two-dimensional vision to inspect weld points or three-dimensional vision to inspect dimensions, which cannot simultaneously achieve a comprehensive judgment of the weld point position / number and the workpiece's Z-axis (height direction) dimensional deviation. Furthermore, different inspection modules are independent of each other, resulting in poor data interoperability, low efficiency in integrating inspection results, inability to quickly generate a unified quality assessment report, and difficulty in effectively feeding the inspection results back to the production line control system.

[0063] Some weld point detection solutions use simple template matching, which is easily affected by welding spatter and workpiece surface texture. The feature extraction accuracy is insufficient, resulting in large weld point positioning deviations. The dimensional deviation calculation does not combine the weld point position matching results, and the two detection data are disconnected, resulting in poor accuracy in the comprehensive defect judgment of welded workpieces, which cannot meet the quality control requirements of high-precision industrial manufacturing.

[0064] Existing vision inspection systems can only complete inspection operations independently. The inspection results need to be recorded manually and transmitted to the production line PLC control system. Then, the PLC is manually operated to perform subsequent actions such as sorting and alarms. The inspection and production processes are disconnected, and closed-loop automated control cannot be achieved, which affects the overall production efficiency of the production line.

[0065] Traditional testing systems often contain scattered records of test data without a unified database management system. This makes it impossible to achieve real-time storage, classification and statistics, and multi-condition querying of test data. When quality problems occur on the production line, it is difficult to quickly trace the cause of the problem and cannot provide effective data support for process optimization.

[0066] Inspection systems developed in a single language are prone to contradictions such as "poor interface adaptation to industrial scenarios" or "high difficulty in implementing visual algorithms." The stability and response speed of the user interface cannot be balanced with the efficiency of algorithm development and iteration. The maintainability and scalability of the system are insufficient, making it difficult to adapt to the inspection needs of different welded workpieces.

[0067] In summary, existing technologies cannot solve the comprehensive problems of low efficiency, insufficient accuracy, single dimension, poor production line linkage, difficult data traceability, and weak system scalability in welded workpiece inspection. There is an urgent need for an integrated, automated, high-precision welded workpiece inspection solution that can be deeply integrated with industrial production lines.

[0068] Based on this, this application proposes a method, device, storage medium, and vehicle for inspecting the quality of welded workpieces, forming a complete technical chain from data acquisition, processing, analysis to result output, production line linkage, and data storage. It addresses the problem of single-dimensional inspection in existing systems, achieving integrated synchronous inspection of weld point quantity / positional offset and workpiece Z-axis dimensional deviation. It solves the problems of weld point inspection being susceptible to interference and low dimensional inspection accuracy, improving the accuracy of comprehensive defect judgment for welded workpieces. It addresses the problem of poor linkage between the inspection system and the production line's programmable logic controller (PLC), achieving fully closed-loop automated control of triggering, acquisition, detection, feedback, and execution without manual intervention. It solves the problems of scattered inspection data and difficulty in traceability by constructing a unified database management system, enabling real-time storage and multi-dimensional traceability of inspection data. It addresses the problem of limited system development architecture by achieving decoupling and collaboration between the user interface and visual algorithms, improving system maintainability and scalability. It solves the problem of requiring multiple devices / manual cooperation for fixed-station inspection, achieving automated judgment of comprehensive quality of welded workpiece dimensions and weld points at a single station, reducing production line equipment and labor costs.

[0069] The following embodiments will be described in detail with reference to the accompanying drawings.

[0070] Figure 1 This is a flowchart of a method for inspecting the quality of welded workpieces provided in one or more embodiments of the present invention.

[0071] like Figure 1 The methods for inspecting the quality of welded workpieces shown include:

[0072] Step S1: Based on the two-dimensional image of the target welding workpiece, determine all the welding point areas of the target welding workpiece, map the welding point areas to the three-dimensional point cloud space of the target welding workpiece, and determine the three-dimensional point cloud data of the welding points corresponding to the welding point areas.

[0073] This embodiment is applied to the weld quality and dimensional deviation detection of welded workpieces in industrial manufacturing. It is designed for industrial fixed-station detection scenarios and enables a single station to complete the comprehensive detection of dimensions and welds. Figure 2 This is a schematic diagram of a welding workpiece quality inspection scenario provided in a specific embodiment of the present invention. For example... Figure 2 As shown in the figure, reference numeral 1 represents the aluminum profile frame, reference numeral 2 represents the clamp, reference numeral 3 represents the light source, reference numeral 4 represents the laser galvanometer stereo camera, reference numeral 5 represents the camera protective cover, reference numeral 6 represents the gripper mechanism, reference numeral 7 represents the detection platform, reference numeral 8 represents the industrial control cabinet, and reference numeral 9 represents the base support.

[0074] Two-dimensional images of the target welded workpiece and three-dimensional point cloud data of the weld points can be acquired using a set image acquisition device.

[0075] Among them, the PLC can trigger the photo-taking signal to trigger the image acquisition device in the designated area of ​​the industrial fixed station inspection scenario to acquire the initial image set of the target welding workpiece.

[0076] The image acquisition device can be a 3D / 2D area scan camera, deployed at a fixed workstation on an industrial production line. Responding to a photo-taking signal triggered by a PLC, it can simultaneously acquire an initial image set of the target welded workpiece. The initial image set includes 3D point cloud data and 2D image data of the target welded workpiece. The 3D point cloud data can identify the weld point's 3D point cloud data and the workpiece's 3D point cloud data for detecting key inspection areas. Simultaneous acquisition of 3D / 2D data at the same workstation enables integrated acquisition of inspection data, avoiding spatial discrepancies from multiple device acquisitions and improving data correlation. The camera is adaptable to fixed industrial workstation scenarios, and the acquisition range and accuracy can be adjusted according to the workpiece.

[0077] In this embodiment, the PLC photo-taking signal triggering module can use various industrial series PLCs as the core of the linkage between the production line and the detection system. It is compatible with mainstream industrial PLC models, has high communication stability, and ensures the real-time performance and accuracy of triggering and executing actions.

[0078] This allows for weld point detection on the acquired two-dimensional images, identifying all weld point areas of the target welded workpiece, and mapping these areas to the three-dimensional point cloud space of the target welded workpiece. In the acquired three-dimensional point cloud data, the corresponding three-dimensional point cloud data of each weld point area is determined.

[0079] Step S2: In the three-dimensional point cloud space, determine the three-dimensional point cloud data of the workpiece corresponding to the key detection area of ​​the target welded workpiece, perform plane fitting on the three-dimensional point cloud data of the workpiece, detect the height dimension of the area to be detected, and determine the first key detection area with out-of-tolerance dimensions.

[0080] The Open3D library (or PCL point cloud library) is used to process the 3D point cloud data of the workpiece, thereby detecting the processed workpiece point cloud data and determining the corresponding 3D point cloud data of all key detection areas of the target welding workpiece. The processing of the workpiece 3D point cloud data may include registration, filtering, etc.

[0081] Further, based on the 3D point cloud data of the workpiece corresponding to the key detection area, plane fitting is performed to calculate the height dimension of each measured point in the key detection area. By using the tolerance threshold, the first key detection area corresponding to the dimensional deviation is determined.

[0082] Step S3: Based on the 3D point cloud data and 2D image of the solder joint, calculate the actual center position of the solder joint in the solder joint area and determine the first solder joint with positional offset.

[0083] The acquired 2D image data is preprocessed and sized using a real-time object detection (YOLO) model to obtain a standard input image.

[0084] Furthermore, the original 2D image is read and converted into the channel order and data format required by the model. Then, size normalization is performed. According to the model input requirements, the image is adjusted to a specified square size by scaling proportionally and filling the edges to prevent the target object from being stretched or deformed. Next, the image pixel values ​​are normalized, scaling the pixel values ​​from 0-255 to the numerical range of 0-1 or -1 to 1. Then, standardization operations such as channel dimension conversion and color space adjustment are completed to finally obtain a preprocessed image that conforms to the YOLO model input format, has a uniform size, and standardized values.

[0085] This process involves detecting all solder joint detection boxes based on the standard input image, and then determining the actual center position of the solder joint based on the detection boxes and the corresponding 3D point cloud data of the solder joint region. The offset between the center position of each actual solder joint and the center of the template solder joint is calculated to identify the first solder joint with a positional offset.

[0086] Step S4: Detect the quality of the target welded workpiece based on the first key detection area and / or the first weld point.

[0087] The confidence level of the first critical inspection area and / or the first weld point that may have defects is calculated to determine the defect distribution of the first critical inspection area and the dimensional deviation of the first weld point, thereby inspecting the quality of the target welded workpiece.

[0088] This embodiment integrates 3D point cloud processing and 2D deep learning target detection technologies. It acquires 3D / 2D data synchronously through the same vision acquisition system, and can determine three core indicators—the number of solder joints, positional offset, and Z-axis dimensional deviation—in a single inspection. This solves the problem of separate implementation of traditional inspection methods and significantly improves inspection efficiency.

[0089] Furthermore, the following is an implementation method for detecting the height dimensional deviation of all key detection areas of the target welded workpiece based on the workpiece's 3D point cloud data.

[0090] Registration is performed on the 3D point cloud data of the workpiece. Specifically, coarse registration is completed based on local feature matching of the 3D point cloud data from key detection, sampling consistency, or normal distribution transformation to obtain the approximate spatial pose transformation relationship and eliminate large initial deviations. Then, fine registration is performed using the iterative nearest point algorithm and its improved forms such as point-to-surface and generalized ICP. Finally, by iteratively optimizing the distance error between corresponding points, high-precision alignment of the 3D point cloud data of the target welding workpiece is achieved. Furthermore, noise and invalid data are removed from the aligned 3D point cloud data, resulting in processed 3D point cloud data of the workpiece.

[0091] In the processed 3D point cloud data for the key inspection area, plane fitting is performed on the workpiece 3D point cloud data corresponding to each key inspection area to determine the plane normal vector of the key inspection area. The plane normal vector is compared with the template plane normal vector. An angular offset between the plane normal vector and the template plane normal vector is determined, and the number of measured points within the key inspection area and the height dimension of each measured point are calculated. The rejection percentage is determined, and based on the rejection percentage and the number of measured points, the number of height dimensions to be rejected is determined. For example, the rejection percentage is set to 5%.

[0092] Therefore, based on the number of height dimensions to be eliminated, height dimensions are eliminated in descending order. The average height dimension is then calculated based on the eliminated height dimensions and the number of measured points. Specifically, the Z-axis average is calculated after eliminating 5% of the Z-axis high points. The average height dimension is compared with the template height dimension to determine the height dimension deviation of key detection areas, until the height dimension deviation of all key detection areas is determined.

[0093] Each height dimensional deviation is compared with a preset tolerance threshold. Height dimensional deviations exceeding the tolerance threshold are identified as dimensional out-of-tolerance. The first measured point corresponding to the dimensional out-of-tolerance is determined and designated as the first critical inspection area. Here, tolerance indicates that the Z-axis dimensional deviation at the inspection location has no positional deviation. The Z-axis dimensional deviation is determined by interpolation calculation using point cloud data collected by a 3D camera and a template based on a preset standard part mold. For example, if the template value is 0, and the Z-axis dimensional deformation of the target welded workpiece is 0.1mm, the inspection result is +0.1mm, which is the Z-axis deviation value, i.e., the height dimensional deviation of that weld point.

[0094] This embodiment uses quantitative data to determine the first key detection area, replacing subjective human judgment. The judgment standard is unified, the results are more objective and accurate, and it is convenient for subsequent quality assessment and problem tracing.

[0095] In another embodiment of this example, based on the three-dimensional point cloud data of the solder joint and the two-dimensional image of the solder joint area, the actual center position of the solder joint in the solder joint area is calculated, and the first solder joint with positional offset is determined as follows.

[0096] After preprocessing and size normalization of the 2D image to obtain a standard 2D image, solder joint features are extracted from the standard 2D image using a YOLO (e.g., YOLOv8, YOLOv9, YOLOX) convolutional neural network. Solder joint detection boxes are generated based on these features, and their center points are determined. The centroid of the solder joint region is calculated based on the 3D point cloud data, and the 3D data of the actual solder joint center is calculated based on the centroid. Therefore, the actual solder joint center position is determined based on the 3D data of the actual solder joint center and the center point of the solder joint detection box. The target solder joint center position is obtained, and the Euclidean distance between the actual solder joint center position and the template solder joint center position is calculated. A preset distance threshold is obtained, and the distance threshold and the Euclidean distance are compared to determine the first solder joint with a positional offset.

[0097] The YOLO deep learning model is used to extract solder joint features, overcoming the interference problem of template matching. Combined with Euclidean distance calculation, high-precision positioning and offset determination of solder joints are achieved, enabling quantitative determination of solder joint position deviation. Moreover, the distance threshold can be flexibly adjusted to adapt to the process requirements of different workpieces, realizing automated and standardized statistics of solder joint offset.

[0098] In this embodiment, the distance threshold for matching weld point positions and the tolerance threshold for Z-axis dimensional deviation can be flexibly adjusted according to the process requirements of different welded workpieces, supporting the inspection needs of different vehicle models and different welding processes, without the need to rebuild the inspection system, thus reducing the production line modification cost.

[0099] Figure 3This is a schematic diagram of welded workpiece quality inspection provided in a specific embodiment of the present invention. For example... Figure 3 As shown, the first key detection area and / or the first weld point are determined by combining 3D point cloud data and 2D image data. Confidence analysis is performed on each key detection area and each weld point within the first key detection area and / or the first weld point. Key detection areas with a confidence level not lower than a confidence threshold are identified as defect areas, and weld points with a confidence level not lower than the confidence threshold are identified as defective weld points. Based on the set quality acceptance standards, the distribution of defective areas and the number of defective weld points are compared to detect the quality of the target welded workpiece. For example, if a key detection area has a height dimensional deviation ≥1mm, the target welded workpiece is determined to be unqualified; similarly, if a weld point has incomplete welding, the target welded workpiece is also determined to be unqualified.

[0100] This embodiment combines a comprehensive evaluation of defect confidence level and acceptance criteria to avoid the one-sidedness of judging by a single indicator, improve the reliability of quality judgment, and meet the quality control needs of high-precision industrial manufacturing.

[0101] This embodiment is implemented using Python and integrates a 3D dimension analysis unit and a solder joint detection unit. As mentioned above, the 3D dimension analysis unit uses the Open3D library to complete point cloud registration, filtering, and Z-axis dimension deviation calculation; the solder joint detection unit uses the YOLOv8 deep learning model to complete 2D image feature extraction, solder joint detection, position matching, and dimensional deviation determination.

[0102] Figure 4 This is a schematic diagram illustrating the quality inspection and evaluation results provided in a specific embodiment of the present invention. For example... Figure 4 As shown, after inspecting the target welded workpiece, a quality inspection evaluation result can be output. This result includes the workpiece number, total number of weld points, number of first weld points, height dimensional deviation (i.e., Z-axis deviation in the figure), and the workpiece's pass / fail status. The generated quality inspection evaluation result is stored in a designated database. For example, Python can be used to write the result to a Microsoft SQL Server database or other relational databases such as MySQL or Oracle, and a pass / fail signal can be output to the PLC via industrial Ethernet and communication protocols, allowing the PLC to execute subsequent production line actions. A closed-loop linkage control mechanism between vision inspection and the PLC is designed, enabling deep interaction between the inspection system and the PLC via industrial Ethernet and communication protocols. The inspection result is directly transmitted back to the PLC, triggering subsequent sorting and alarm actions. The inspection process requires no manual intervention, has a fast response speed, and is fully adapted to the high-speed production needs of modern production lines.

[0103] The test results are written to the database in real time, achieving standardized data storage. A Microsoft SQL Server database is used to store all quality assessment results in real time. Database fields include at least the workpiece number, test time, total number of weld points, weld point offset, Z-axis deviation, and pass / fail status. It supports fast data reading and writing and historical traceability. Database fields correspond one-to-one with test indicators, achieving standardized and structured storage of test data. This enables full lifecycle management of test data, providing reliable data support for tracing production line quality issues and optimizing process parameters. It supports high-speed reading and writing and multi-condition queries, solving the problems of fragmented data records and difficult traceability in traditional systems, providing data support for production line process optimization. Results are directly transmitted back to the PLC and trigger automated actions, achieving seamless integration between testing and the production line, completing a fully closed-loop control without manual intervention.

[0104] The Python language leverages the technological advantages of Open3D and YOLOv8 to achieve efficient development and iteration of visual algorithms; the two units work together to complete the fusion analysis of multi-dimensional detection data, providing core algorithm support for comprehensive quality judgment.

[0105] This embodiment also includes a human-machine interface (HMI) that can be developed using C# (.NET Framework). The HMI integrates detection control, real-time result display, and historical data query functions. It allows adjustment of parameters such as detection distance thresholds and tolerance thresholds based on workpiece type, and can also display defect severity distribution and quality assessment results. This ensures the stability, response speed, and PLC communication compatibility of the HMI in industrial settings. The visual interface enables real-time monitoring of the detection process, flexible parameter adjustment, and rapid data querying, reducing the operational threshold for operators and improving system usability.

[0106] This embodiment employs a cross-language modular architecture combining a C# (.NET Framework) front-end interface with Python vision algorithms. This ensures the stability, responsiveness, and PLC communication compatibility of the operating interface in industrial scenarios, achieving decoupling and efficient collaboration between "operation control" and "algorithm analysis." The C# interface handles communication with the PLC and database, as well as human-machine interaction, while the Python module is responsible for the implementation and iteration of the vision algorithm. The two modules achieve high-speed data exchange through an internal communication protocol. Simultaneously, C# directly connects to the Microsoft SQL Server database, enabling efficient reading and writing of algorithm results, operation instructions, and PLC feedback data. This resolves the contradictions inherent in traditional single-language development, such as poor interface adaptation to industrial scenarios or high algorithm implementation difficulty. Each module can be independently developed, iterated, and maintained, significantly improving system maintainability and scalability. Cross-language collaboration ensures operational stability in industrial scenarios while leveraging Python's technical advantages in vision algorithm development. Python utilizes Open3D and YOLOv8 for efficient algorithm development and iteration, and the decoupling significantly enhances maintainability and scalability. The visual human-machine interface integrates detection control, result display, and parameter adjustment functions, making it easy for operators to learn and lowering the operational threshold.

[0107] For example, this embodiment is deployed at a fixed workstation on an automotive manufacturing welding production line, targeting the detection of weld points and dimensional deviations in automotive welded workpieces. The system hardware and software configuration is as follows:

[0108] Hardware configuration: The PLC trigger module adopts Siemens S7-1500 PLC; the 3D / 2D image acquisition device adopts a high-precision industrial 3D camera (resolution ≥1280×1024) and a 2D area scan camera (frame rate ≥30fps); the vision algorithm processing module is deployed on an industrial control computer (CPU ≥ i7-12700K, memory ≥32G, graphics card ≥ RTX3060); all hardware are connected via industrial Ethernet (gigabit), and the communication protocol is S7.

[0109] Software configuration: The visual algorithm processing module is developed based on Python 3.8 and integrates the Open3D 0.17.0 and YOLOv8.0 deep learning frameworks; the database storage module uses Microsoft SQL Server 2019; the human-computer interaction module is developed based on C#.NET Framework 4.7; the operating system is Windows 10 Industrial Edition.

[0110] The following embodiment takes the quality inspection of welded workpieces of the C100 model in the automotive manufacturing field as an example. Based on the above software and hardware configuration, the quality inspection of the weld center position and dimensional deviation of the target inspection object is the welded workpiece of the C100 model.

[0111] Figure 5 This is a flowchart of a target vehicle quality inspection provided in a specific embodiment of the present invention. For example... Figure 5 As shown, the parameters include: for welding workpieces of C100 model, the distance threshold for matching the welding point position (i.e., the center point position of the template welding point) is set to 5000px in the human-machine interface. The Z-direction dimensional deviation tolerance threshold is set according to the process requirements (Top direction ≤ ±2mm, Down direction ≤ ±2mm). If the distance deviation exceeds the threshold, it is judged as welding point position offset. If the Z-direction dimensional deviation value exceeds the threshold, it is judged as the target welding workpiece size is out of tolerance. If any defect occurs, the workpiece is judged as qualified according to the severity.

[0112] Image acquisition: The PLC trigger module sends a photo-taking trigger signal, and the 3D / 2D image acquisition module simultaneously acquires 3D point cloud data and 2D image data of the welding workpiece of C100 model, forming an initial image set and transmitting it to the vision algorithm processing module.

[0113] Image preprocessing: The visual algorithm processing module uses the Open3D library to register and filter 3D point cloud data to remove point cloud noise caused by welding spatter and impurities on the workpiece surface; it performs size normalization (uniformly adjusted to 640×640 pixels) and YOLO model preprocessing on 2D images to obtain standard input images.

[0114] Solder joint target detection: The YOLOv8 model is used to extract features and detect targets in the standard input image, generate solder joint detection boxes, extract the center point coordinates of each detection box, and count the total number of detected solder joints.

[0115] Weld point position matching: Calculate the Euclidean distance between the center point coordinates and the preset weld point reference position of the C100 model, determine whether each weld point is offset based on the 5000px distance threshold, and count the number of offset weld points.

[0116] Z-axis dimensional deviation calculation: Based on the processed point cloud, the Z-axis dimensional deviation of the key detection area of ​​the workpiece is automatically calculated to obtain the specific deviation values ​​(i.e., height dimensional deviation) of the Top (upper part) and Down (lower part).

[0117] Defect and quality assessment: Calculate the distance deviation and Z-axis dimensional deviation of the weld point position, compare them with the preset threshold, and determine whether there are defects such as dimensional deviations or positional offsets in the key inspection area of ​​the target welded workpiece; perform confidence analysis on the defect results (confidence ≥ 95% is considered a valid defect), and combine the number of qualified weld points and the number of offsets to comprehensively determine whether the workpiece is qualified.

[0118] Results Output and Storage: Generate a quality assessment report, including vehicle model (C100), workpiece number, inspection time, total number of weld points, number of qualified welds, number of offsets, Z-axis deviation (Top / Down), and qualification status. Write the report to the Microsoft SQL Server 2019 database in real time and output qualified / unqualified signals to the PLC via the S7 communication protocol (or other industrial Ethernet communication protocols such as Profinet and Modbus TCP). The PLC automatically performs sorting actions based on the signals (qualified workpieces are transferred to the next process, and unqualified workpieces trigger an alarm and are sorted to the rework area).

[0119] The quality inspection and evaluation results of two batch inspections of the welded parts for the C100 model are as follows:

[0120] Inspection 1: Total number of standard weld points: 24; Number of weld points detected: 31; Number of qualified weld points: 31; Number of offsets: 0; Z-axis deviation: Top: 1.28mm, Down: 0.77mm; No dimensional deviation or positional offset defects; Workpiece is deemed qualified; PLC executes normal operation.

[0121] Inspection 2: Total number of standard solder joints: 21; Number of solder joints detected: 13; Number of qualified solder joints: 13; Number of offsets: 0; Z-axis deviation: Top: -1.06mm, Down: -0.97mm; No dimensional deviation or positional offset defects; All detected solder joints are deemed qualified; PLC executes corresponding actions according to production line process requirements.

[0122] Figure 6 This is a structural diagram of a welding workpiece quality inspection device provided in one or more embodiments of the present invention.

[0123] like Figure 6 The welding workpiece quality inspection device shown includes: a data acquisition module, a dimension detection module, a position detection module, and a quality inspection module.

[0124] The data determination module is used to determine all the weld point areas of the target welded workpiece based on the two-dimensional image of the target welded workpiece, map the weld point areas to the three-dimensional point cloud space of the target welded workpiece, and determine the three-dimensional point cloud data of the weld point corresponding to the weld point area.

[0125] The dimension detection module is used to determine the workpiece 3D point cloud data corresponding to the key detection area of ​​the target welded workpiece in 3D point cloud space, perform plane fitting on the workpiece 3D point cloud data, detect the height dimension of the area to be detected, and determine the first key detection area with out-of-tolerance dimensions.

[0126] The position detection module is used to calculate the actual center position of the solder joint in the solder joint area based on the 3D point cloud data and 2D image of the solder joint, and to determine the first solder joint with positional offset.

[0127] The quality inspection module is used to inspect the quality of the target welded workpiece based on the first critical inspection area and / or the first weld point.

[0128] The dimension detection module removes noise and invalid data from the workpiece's 3D point cloud data, resulting in processed workpiece 3D point cloud data. In this processed data, plane fitting is performed to determine the plane normal vectors of key detection areas. Based on these normal vectors, the height dimension deviation of each key detection area is detected. For each key detection area, in response to an angular offset between the plane normal vector and the template plane normal vector, the number of measured points within the key detection area and the height dimension of each measured point are calculated. A rejection percentage is determined, and based on this percentage and the number of measured points, the number of height dimensions to be rejected is determined. Based on this rejection number, height dimensions are rejected in descending order, and the average height dimension is calculated based on the rejected height dimensions and the number of measured points. The average height dimension is compared with the template height dimension to determine the height dimension deviation of the key detection areas, until the height dimension deviation of all key detection areas is determined. Each height dimension deviation is compared with a preset tolerance threshold; deviations exceeding this threshold are identified as out-of-tolerance. The key detection area corresponding to this out-of-tolerance deviation is designated as the first key detection area.

[0129] The position detection module is used to preprocess and normalize the size of the two-dimensional image to obtain a standard two-dimensional image; extract the solder joint features of the standard two-dimensional image, and generate a solder joint detection box based on the solder joint features to determine the center position of the solder joint detection box; calculate the centroid of the solder joint region based on the three-dimensional point cloud data of the solder joint, determine the actual center position of the solder joint based on the centroid and the center position, obtain the center position of the template solder joint, calculate the Euclidean distance between the actual center position of the solder joint and the center position of the template solder joint; obtain a preset distance threshold, compare the distance threshold and the Euclidean distance, and determine the first solder joint with positional offset.

[0130] The quality inspection module is used to perform confidence analysis on the first key inspection area and / or the first weld point, identify the first key inspection area with a confidence level not lower than the confidence level threshold as the defect area, and identify the first weld point with a confidence level not lower than the confidence level threshold as the defective weld point; based on the set quality acceptance standards, the distribution of the defect area and the number of defective weld points are compared to inspect the quality of the target welded workpiece.

[0131] The quality inspection module is also used to determine the workpiece number of the target welded workpiece, the total number of weld points, the number of first weld points, the height dimensional deviation, and the qualification status of the target welded workpiece, generate quality inspection evaluation results, and store the quality inspection evaluation results in a designated database.

[0132] Figure 7 This is a structural block diagram of an electronic device for a method of inspecting the quality of welded workpieces provided in one or more embodiments of the present invention.

[0133] like Figure 7 As shown, this application provides an electronic device, including: a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0134] The memory stores a computer program, which, when executed by the processor, causes the processor to perform the steps of a method for inspecting the quality of welded workpieces.

[0135] This application also provides a computer-readable storage medium storing a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of a method for inspecting the quality of welded workpieces.

[0136] This application also provides a vehicle, including:

[0137] Electronic equipment for implementing steps in a method for inspecting the quality of welded workpieces;

[0138] The processor runs a program, and when the program runs, it executes the steps of a welded workpiece quality inspection method based on data output from electronic devices.

[0139] Storage medium for storing programs that, when run, execute steps of a welding workpiece quality inspection method based on data output from an electronic device.

[0140] The communication bus mentioned in the above electronic devices can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the diagram, but this does not mean that there is only one bus or one type of bus.

[0141] The electronic device comprises a hardware layer, an operating system layer running on top of the hardware layer, and an application layer running on the operating system. The hardware layer includes hardware such as a central processing unit (CPU), a memory management unit (MMU), and memory. The operating system can be any one or more computer operating systems that control the electronic device through processes, such as Linux, Unix, Android, iOS, or Windows. Furthermore, in this embodiment of the invention, the electronic device can be a smartphone, tablet computer, or other handheld device, or a desktop computer, portable computer, or other electronic device; there is no particular limitation in this embodiment.

[0142] In this embodiment of the invention, the executing entity for electronic device control can be an electronic device itself, or a functional module within an electronic device capable of calling and executing a program. The electronic device can obtain the firmware corresponding to the storage medium. This firmware is provided by the supplier, and different storage media may have the same or different firmware; no limitation is made here. After obtaining the firmware corresponding to the storage medium, the electronic device can write this firmware into the storage medium; specifically, it burns the firmware corresponding to the storage medium into the storage medium. The process of burning the firmware into the storage medium can be implemented using existing technology, and will not be elaborated upon in this embodiment of the invention.

[0143] Electronic devices can also obtain reset commands corresponding to the storage media. The reset commands corresponding to the storage media are provided by the supplier. The reset commands corresponding to different storage media can be the same or different, and no restrictions are imposed here.

[0144] At this time, the storage medium of the electronic device is a storage medium on which the corresponding firmware has been written. The electronic device can respond to the reset command corresponding to the storage medium on which the corresponding firmware has been written, thereby resetting the storage medium on which the corresponding firmware has been written according to the reset command. The process of resetting the storage medium according to the reset command can be implemented by existing technology and will not be described in detail in this embodiment of the invention.

[0145] For ease of description, the above devices are described separately by function as various units and modules. Of course, in implementing this application, the functions of each unit and module can be implemented in one or more software and / or hardware.

[0146] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the meaning consistent with their meaning in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined.

[0147] For the sake of simplicity, the method embodiments are described as a series of actions. However, those skilled in the art should understand that the embodiments of the present invention are not limited to the described order of actions, because according to the embodiments of the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to the embodiments of the present invention.

[0148] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solution of this application, 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 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 various embodiments or some parts of the embodiments of this application.

[0149] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for inspecting the quality of welded workpieces, characterized in that, The method for inspecting the quality of welded workpieces includes: Based on the two-dimensional image of the target welding workpiece, all welding point regions of the target welding workpiece are determined, and the welding point regions are mapped into the three-dimensional point cloud space of the target welding workpiece to determine the three-dimensional point cloud data of the welding point corresponding to the welding point regions. In the three-dimensional point cloud space, the three-dimensional point cloud data of the workpiece corresponding to the key detection area of ​​the target welding workpiece is determined. Plane fitting is performed on the three-dimensional point cloud data of the workpiece to detect the height dimension of each measured point in the key detection area and calculate the average height dimension. The average height dimension is compared with the template height dimension to determine the height dimension deviation of the key detection area. The height dimension deviation is compared with a preset tolerance threshold. The height dimension deviation that is greater than the tolerance threshold is determined as a dimension out-of-tolerance. The first measured point corresponding to the dimension out-of-tolerance is determined, and the first measured point is determined as the first key detection area of ​​the dimension out-of-tolerance. Based on the 3D point cloud data of the solder joint, the centroid of the solder joint region is calculated. Based on the centroid, the 3D data of the actual solder joint center is calculated, and the solder joint features of the standard 2D image are extracted. Based on the solder joint features, a solder joint detection box is generated, and the center point of the solder joint detection box is determined. Based on the 3D data of the actual solder joint center and the center point of the solder joint detection box, the actual solder joint center position is determined. The Euclidean distance between the actual solder joint center position and the center point position of the template solder joint is calculated. A preset distance threshold is obtained. The distance threshold and the Euclidean distance are compared to determine the first solder joint with a positional offset. The quality of the target welded workpiece is inspected based on the first key detection area and / or the first weld point.

2. The method for inspecting the quality of welded workpieces according to claim 1, characterized in that, The step of performing planar fitting on the three-dimensional point cloud data of the workpiece, detecting the height dimension of the key detection area, and determining the first key detection area with out-of-tolerance dimensions includes: Remove noise and invalid data from the workpiece's 3D point cloud data to obtain the processed workpiece 3D point cloud data. In the processed 3D point cloud data of the workpiece, a plane fitting is performed on the 3D point cloud data of the workpiece to determine the plane normal vector of the key detection area; Based on the plane normal vector, the height dimension deviation of each of the key detection regions is detected.

3. The method for inspecting the quality of welded workpieces according to claim 2, characterized in that, The step of detecting the height dimension deviation of each key detection region based on the plane normal vector includes: For each of the key detection areas, in response to the angular offset between the plane normal vector and the template plane normal vector, the number of measured points within the key detection area and the height dimension of each measured point are calculated. Determine the rejection percentage, and based on the rejection percentage and the number of measured points, determine the number of height dimensions to be rejected; Based on the number of height dimensions to be removed, the height dimensions are removed in descending order, and the average height dimension is calculated based on the removed height dimensions and the number of measured points. The height dimension average and the template height dimension are compared to determine the height dimension deviation of the key detection area, until the height dimension deviation of all key detection areas is determined.

4. The method for inspecting the quality of welded workpieces according to claim 1, characterized in that, Before extracting solder joint features from the standard two-dimensional image, the process includes: The two-dimensional image is preprocessed and its size is normalized to obtain a standard two-dimensional image.

5. The method for inspecting the quality of welded workpieces according to claim 1, characterized in that, The step of inspecting the quality of the target welded workpiece based on the first key detection area and / or the first weld point includes: A confidence analysis is performed on the first key detection area and / or the first solder joint. The first key detection area with a confidence level not lower than the confidence level threshold is identified as a defect area, and the first solder joint with a confidence level not lower than the confidence level threshold is identified as a defective solder joint. Based on the established quality standards, the distribution of the defective areas and the number of defective weld points are compared to detect the quality of the target welded workpiece.

6. The method for inspecting the quality of welded workpieces according to claim 5, characterized in that, After comparing the distribution of the defective areas and the number of defective weld points based on the set quality acceptance standards to detect the quality of the target welded workpiece, the method further includes: The workpiece number of the target welded workpiece, the total number of weld points, the number of the first weld points, the height dimensional deviation, and the qualification status of the target welded workpiece are determined, and a quality inspection and evaluation result is generated. The quality inspection and evaluation results are stored in a designated database.

7. A welding workpiece quality inspection device, characterized in that, The welded workpiece quality inspection device includes: The data determination module is used to determine all the weld point regions of the target welding workpiece based on a two-dimensional image of the target welding workpiece, map the weld point regions to the three-dimensional point cloud space of the target welding workpiece, and determine the three-dimensional point cloud data of the weld points corresponding to the weld point regions. The dimension detection module determines the workpiece's 3D point cloud data corresponding to the key detection area of ​​the target welding workpiece in the 3D point cloud space, performs plane fitting on the workpiece's 3D point cloud data, detects the height dimension of each measured point in the key detection area, calculates the average height dimension, compares the average height dimension with the template height dimension, determines the height dimension deviation of the key detection area, compares the height dimension deviation with a preset tolerance threshold, and determines height dimension deviations greater than the tolerance threshold as dimensional out-of-tolerance; determines the first measured point corresponding to the dimensional out-of-tolerance, and identifies the first measured point as the first key detection area of ​​the dimensional out-of-tolerance. The position detection module calculates the centroid of the solder joint region based on the 3D point cloud data of the solder joint, calculates the 3D data of the actual solder joint center based on the centroid, extracts solder joint features from the standard 2D image, generates a solder joint detection box based on the solder joint features, and determines the center point of the solder joint detection box; it determines the actual solder joint center position based on the 3D data of the actual solder joint center and the center point of the solder joint detection box, calculates the Euclidean distance between the actual solder joint center position and the center point position of the template solder joint, obtains a preset distance threshold, compares the distance threshold and the Euclidean distance, and determines the first solder joint with positional offset. The quality inspection module is used to inspect the quality of the target welded workpiece based on the first key inspection area and / or the first weld point.

8. A computer-readable storage medium, characterized in that, The device stores a computer program executable by an electronic device, which, when run on the electronic device, causes the electronic device to perform the steps of the welded workpiece quality inspection method as described in any one of claims 1 to 6.

9. A vehicle, characterized in that, include: An electronic device for implementing the steps of the welded workpiece quality inspection method as described in any one of claims 1 to 6; A processor that runs a program, which, when running, executes the steps of the welded workpiece quality inspection method as described in any one of claims 1 to 6 from data output by an electronic device. A storage medium for storing a program that, when run, performs the steps of the welded workpiece quality inspection method as described in any one of claims 1 to 6 on data output from an electronic device.

Images