Welding quality detection method for connector

Abstract

The invention discloses a welding quality detection method for a connector, and relates to the technical field of quality detection, and the method comprises the steps: collecting a to-be-detected connector through a 3D scanner, obtaining point cloud information, extracting depth features to obtain welding spot features, and recognizing defects based on a deep learning model of a training embedded template matching technology, a database containing defect type, shape and size information is constructed according to historical examples, three welding quality detection sub-cycles associated with a preset quality standard are set based on the database, and the three sub-cycles are connected to conduct welding quality detection. The technical problems that the detection result is inaccurate and unreliable and the product quality and reliability are affected due to the fact that the welding quality of the connector is not comprehensively and accurately detected in the prior art are solved, and the technical effects that the defective rate and the use risk are reduced, and the quality and reliability of the connector product are improved are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of quality inspection, and in particular to a method for inspecting the welding quality of a connector. Background Art

[0002] In the connector welding quality inspection scenario, the reliability and stability of connector welding quality are particularly prominent, and the contradiction between high-quality inspection technology and demand is relatively more prominent. Since the connector welding quality is directly related to the stability and reliability of the equipment, the development of efficient and accurate inspection methods to better meet the accurate identification of various welding defects has become a crucial link in solving the connector welding quality inspection. Traditional connector welding quality inspection methods are often relatively simple and localized, relying only on manual visual inspection or simple gage measurement, and lack of overall grasp of solder joint characteristics. There is a lack of in-depth analysis and utilization of the complex structure and potential defects of solder joints, making it difficult to comprehensively and accurately detect various welding defects. In terms of inspection results, there are inaccuracies, resulting in some subtle and critical defects being missed, while some non-critical issues are overly concerned. The formulation of inspection standards is relatively vague and inconsistent, and cannot well cope with the ever-increasing quality requirements of connectors.

[0003] At present, relevant technologies have the technical problem that the detection of connector welding quality is not comprehensive and accurate enough, resulting in inaccurate and unreliable detection results, which affects product quality and reliability. Summary of the invention

[0004] The present application provides a method for detecting the welding quality of connectors, which uses a 3D scanner to collect point cloud information of the connector to be inspected and extracts deep features to obtain solder joint features. Based on this, a deep learning model with embedded template matching technology is trained to identify defects. A database containing defect type, morphology, and size information is constructed based on historical examples. Based on this, three welding quality detection sub-loops associated with preset quality standards are respectively set, and the three sub-loops are connected to perform welding quality detection. This achieves accurate and efficient detection and evaluation of the connector welding quality, comprehensively considers various information such as defect type, morphology, and size, and achieves the technical effect of reducing the defective rate and usage risk, and improving the quality and reliability of connector products.

[0005] The present application provides a method for detecting the welding quality of a connector, comprising: Use a 3D scanner to collect the connector to be inspected, obtain point cloud information, and perform deep feature extraction to obtain solder joint feature information; based on the solder joint feature information, train a deep learning model to classify and identify defects, and the deep learning model is embedded with template matching technology, and the template matching technology is used to quickly identify known defect patterns; based on historical welding quality examples, statistical distribution laws are calculated and bound to the known defect patterns to build a defect database, which contains type information, morphological information, and size information of various welding defects; based on the type information of various welding defects in the defect database, set a first welding quality detection sub-cycle, wherein the termination condition corresponding to the first welding quality detection sub-cycle is associated with a first preset quality standard, and the first preset quality standard is based on industry specifications and connector product design requirements. Determined by constraints, including welding defect types that are not allowed to appear; based on the morphological information of various welding defects in the defect database, a second welding quality detection sub-cycle is set, and the termination condition corresponding to the second welding quality detection sub-cycle is associated with a second preset quality standard, and the second preset quality standard is determined by industry specifications and welding process conditions as constraints; based on the size information of various welding defects in the defect database, a third welding quality detection sub-cycle is set, and the termination condition corresponding to the third welding quality detection sub-cycle is associated with a third preset quality standard, and the third preset quality standard is determined by industry specifications and welding material characteristics as constraints; the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle are connected to perform welding quality inspection on the connector to be inspected.

[0006] A method for detecting the welding quality of connectors proposed in this application first uses a 3D scanner to collect point cloud information of the connector to be inspected and extracts deep features to obtain solder joint features. Based on this, a deep learning model embedded with template matching technology is trained to identify defects. A database containing defect type, morphology, and size information is constructed based on historical examples. Based on this, three welding quality detection sub-loops associated with preset quality standards are set respectively, and the three sub-loops are connected for welding quality detection. This achieves the technical effect of reducing the defective rate and usage risk, and improving the quality and reliability of connector products by realizing accurate and efficient detection and evaluation of the connector welding quality, and comprehensively considering various information such as defect type, morphology, and size. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the accompanying drawings of the embodiment of the present invention will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0008] Figure 1 A schematic diagram of a flow chart of a method for detecting the welding quality of a connector provided in an embodiment of the present application; Figure 2 A schematic diagram of a process for determining an optimal scanning viewing angle for a method for detecting welding quality of a connector provided in an embodiment of the present application. DETAILED DESCRIPTION

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0010] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0011] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0012] The present application embodiment provides a method for detecting the welding quality of a connector, such as Figure 1 As shown, the method includes: Step S100, use a 3D scanner to collect the connector to be inspected, obtain point cloud information, and perform deep feature extraction to obtain solder joint feature information. Specifically, when performing connector welding quality inspection, start a high-precision 3D scanner, place the connector to be inspected in the working area of ​​the scanner, and ensure that its position is accurate and stable. The 3D scanner emits dense light or laser beams to the connector to be inspected. The light or laser beam interacts with the surface of the connector and reflects from all angles. The scanner's sensor receives these reflected light or laser beams and converts them into a large number of data points. The collection of data points constitutes point cloud information. The point cloud information contains detailed information such as the three-dimensional geometric shape and position of the connector surface. Deep feature extraction is performed on the acquired point cloud information, and a Gaussian filter algorithm is used to remove noise. Gaussian filtering effectively smoothes the data by weighted averaging the point cloud data in the neighborhood, reducing the impact of random noise. The bilateral filtering technique is used to remove noise while maintaining The clarity of important features such as the edge of the solder joint, bilateral filtering not only considers the spatial distance, but also the difference in pixel values, so as to better retain the details while smoothing the noise, and use principal component analysis technology to reduce the data dimension and extract the main features. By calculating the covariance matrix of the data, the main direction of the data is found, and the high-dimensional data is projected into a low-dimensional space while retaining most of the information. Through data processing and analysis operations, possible noise and interference factors are effectively removed. After removing the noise, the feature information directly related to the solder joint is further extracted. The feature information includes key parameters such as the position, shape, size, and height of the solder joint, as well as related features such as the surface roughness and flatness around the solder joint. After deep feature extraction, the feature information that can accurately reflect the characteristics of the solder joint is finally obtained, which provides an important data basis for subsequent welding quality evaluation and defect detection.

[0013] In a possible implementation, a 3D scanner is used to collect the connector to be inspected, obtain point cloud information, and perform deep feature extraction to obtain solder joint feature information. Step S100 further includes step S110, configuring 3D scanner parameters, wherein the 3D scanner parameters are resolution and scanning speed. Specifically, the key parameters of the 3D scanner are configured, wherein the resolution determines the accuracy of the details that the scanner can capture when scanning the connector to be inspected. If the resolution is set to a higher level, the point cloud information obtained will be very fine and can clearly present the tiny features on the surface of the connector. However, high resolution also means that the amount of data generated will increase significantly, and the time and computing resources required for subsequent data processing and analysis will also increase accordingly. The scanning speed is a key parameter. A faster scanning speed can significantly improve the detection efficiency and complete the scanning of the connector in a short time. However, a too fast scanning speed causes the scanner to be inaccurate and incomplete when capturing information, thereby affecting the final detection quality. It is necessary to weigh the pros and cons and reasonably set the resolution and scanning speed according to the specific characteristics of the connector to be inspected, the requirements for detection accuracy, and the actual work efficiency requirements.

[0014] Step S120, based on the 3D scanner parameters, the connector to be inspected is scanned for full coverage of the solder joints, and the coverage position corresponding to the point cloud information includes M solder joints and M peripheral areas, and the peripheral area refers to the surrounding area that closely surrounds each solder joint and has slight differences in welding. Specifically, the connector to be inspected is scanned according to the previously set 3D scanner parameters. The goal of the scan is to achieve full coverage of the solder joints to ensure that no solder joint is missed. During the scanning process, the coverage position corresponding to the acquired point cloud information clearly covers the M solder joints, including the M peripheral areas that are closely connected to each solder joint. Although the peripheral area is not large in scope, it is crucial for accurately evaluating the welding quality due to the slight differences in welding. By fully scanning the M solder joints and M peripheral areas, rich and comprehensive information can be collected to provide sufficient data support for subsequent quality analysis.

[0015] Step S130, the M welding points and the M perimeter areas are combined, and the 3D scanner is synchronously adjusted during the scanning process. Specifically, during the scanning process, the M welding points and the M perimeter areas are analyzed as a whole, and the 3D scanner is synchronously adjusted according to the actual situation during the scanning process. The adjustment is mainly focused on the scanning angle and distance. In terms of scanning angle, the angle of the scanner is changed in real time according to the distribution and shape characteristics of the welding points and the perimeter area to ensure that the light or laser beam can irradiate the target area at the best angle, so as to obtain more accurate and complete point cloud information. In terms of scanning distance, the distance between the scanner and the connector is dynamically adjusted according to the size and shape of the connector to be inspected, as well as the position of the welding points and the perimeter area. If the distance is too far, it may lead to incomplete information collection or reduced accuracy. If the distance is too close, it may cause a scanning blind spot or cause unnecessary interference to the connector. Through the precise synchronous adjustment of the scanning angle and distance, the quality and efficiency of the scanning can be maximized, and high-quality point cloud information can be obtained, laying a solid foundation for subsequent welding quality inspection and analysis.

[0016] In a possible implementation, the M welding points and the M peripheral areas are combined, and the 3D scanner is synchronously adjusted during the scanning process. Step S130 further includes step S131, adding P position states according to the shape structure of the connector to be inspected, and the P position states correspond to the relative position of the connector to be inspected and the 3D scanner. Specifically, the shape structure of the connector to be inspected is analyzed. Since the shapes of the connectors are different, there are complex geometric shapes and structural features. In order to be able to comprehensively and accurately obtain its point cloud data, P different position states are set according to the specific shape and structure of the connector. Each position state clearly corresponds to a specific relative position relationship between the connector to be inspected and the 3D scanner. For example, if the connector is a curved tubular structure, multiple position states may be set, including horizontal placement, vertical placement, and tilted placement at different angles.

[0017] Step S132, based on the P position states, the position adjustment instruction is issued in comparison with the blocked area. Specifically, after determining the P position states, the connector is preliminarily scanned by a 3D scanner. During the scanning process, the information obtained by the actual scan is compared with the preset position state, with special attention paid to those areas that may be blocked due to improper position. Once the blocked area is found, the system will intelligently issue the corresponding position adjustment instruction according to the current position state and the situation of the blocked area. For example, if it is found that a certain part is blocked by other parts of the connector due to improper position of the connector, the system will issue an adjustment instruction to change the position of the connector.

[0018] Step S133, after the connector to be inspected receives the position adjustment instruction, it synchronously executes the rotation action, tilt action, and translation action corresponding to the position adjustment instruction. Specifically, when the connector to be inspected receives the position adjustment instruction issued by the system, it will respond immediately. The connector will synchronously execute the corresponding actions according to the requirements of the instruction. The actions include rotation, that is, rotation around a certain axis to change the angle; tilt action, so that the connector tilts in a plane; translation action, linear movement along a specific direction. Through precise action adjustment, it is ensured that the 3D scanner can obtain complete and accurate point cloud data for complex structural parts, thereby providing a reliable basis for subsequent quality inspection and analysis. Especially for parts with complex structures, only through such precise position adjustment can data loss be avoided and the accuracy and comprehensiveness of the inspection results be guaranteed.

[0019] In a possible implementation, according to the shape structure of the connector to be inspected, P position states are added, and the P position states correspond to the relative position of the connector to be inspected and the 3D scanner. Step S131 further includes step S1311, in the first position state, q scanning angles are added, and the first position state is any one of the P position states. Specifically, any one of the P position states is selected as the first position state, and on the basis of the first position state, q different scanning angles are added. The setting of the scanning angle is to scan the connector to be inspected from multiple angles to obtain more comprehensive and accurate information. For example, if the first position state is to place the connector horizontally, then the added q scanning angles may include different directions such as vertically downward from the top, horizontally to the right from the left, and horizontally to the left from the right.

[0020] Step S1312, based on the q scanning angles, the number of solder joints in each scanning angle is arranged in descending order to obtain a first scanning angle sequence. Specifically, after the q scanning angles are determined, the number of solder joints that can be covered by each scanning angle is calculated respectively, and the q scanning angles are sorted in descending order according to the number of solder joints, with the purpose of giving priority to scanning angles that can capture more solder joints, because the number of solder joints reflects the ability of the angle to obtain key information to a certain extent. For example, if a scanning angle can cover 20 solder joints, while another can only cover 10, then the former will be higher in the sorting.

[0021] Step S1313: Based on the q scanning angles, the coverage ratio of the peripheral area in each scanning angle is arranged in descending order to obtain a second scanning angle sequence. Specifically, based on the q scanning angles, the coverage ratio of each angle to the peripheral area is calculated. The peripheral area is also very important for evaluating the welding quality. The larger the coverage ratio, the more complete the peripheral area information obtained. The q scanning angles are sorted again in descending order according to the coverage ratio of the peripheral area to obtain a second scanning angle sequence.

[0022] Step S1314, according to the first scanning angle sequence and the second scanning angle sequence, q' optimal scanning angles in the first position state are determined. Specifically, the first scanning angle sequence (sorted based on the number of welds) and the second scanning angle sequence (sorted based on the percentage of perimeter area coverage) are comprehensively considered, and by analyzing and comparing the two sequences, q' scanning angles with relatively good performance in terms of weld point coverage and perimeter area coverage are screened out and determined as the optimal scanning angles in the first position state. The optimal scanning angle can provide the most valuable information for subsequent welding quality inspections, ensuring the accuracy and reliability of the inspections.

[0023] In one possible implementation, Figure 2 As shown, according to the first scanning angle sequence and the second scanning angle sequence, q′ optimal scanning angles in the first position state are determined, and step S1314 further includes step S13141, comparing the first scanning angle sequence and the second scanning angle sequence, and screening out N candidate scanning angles with the same serial number in the first scanning angle sequence and the second scanning angle sequence. Specifically, the first scanning angle sequence (sorted based on the number of solder joints) and the second scanning angle sequence (sorted based on the percentage of peripheral area coverage) obtained previously are compared and analyzed, and the serial number of each scanning angle in the two sequences is checked one by one. If the serial number of a scanning angle in the two sequences is the same, it means that its performance in terms of solder joint coverage and peripheral area coverage is relatively stable and excellent, and the scanning angle is screened out, and a total of N candidate scanning angles are obtained.

[0024] Step S13142, receiving the first verification signal, and verifying the solder joint coverage integrity of the N alternative scanning angles. If the verification passes, a second verification signal is issued. Specifically, after receiving the first verification signal, the system starts to verify the solder joint coverage integrity of the N alternative scanning angles, and checks whether these angles can completely cover all solder joints through calculation and analysis. If the verification passes, it indicates that these alternative scanning angles meet the requirements in terms of solder joint coverage, and the system will issue a second verification signal.

[0025] Step S13143, through the second verification signal, the N alternative scanning angles are verified for coverage integrity of the perimeter area. If passed, the q′ best scanning angles in the first position state are determined. Specifically, after receiving the second verification signal, the N alternative scanning angles are verified for coverage integrity of the perimeter area. Through calculation and analysis, it is determined whether the angle can completely cover the perimeter area. If the verification is passed, it means that the alternative scanning angle performs well in both solder joint coverage and perimeter area coverage. The q′ best scanning angles in the first position state can be determined from the N alternative scanning angles for subsequent detection work to ensure the accuracy and comprehensiveness of the detection results.

[0026] In a possible implementation, a first verification signal is received, and the N alternative scanning angles are verified for solder joint coverage integrity. If the verification passes, a second verification signal is issued, and step S13142 further includes step S131421, and if the verification fails, uncovered solder joints are determined based on the N alternative scanning angles. Specifically, if the solder joint coverage integrity verification fails for the N alternative scanning angles, it means that there are some solder joints that are not covered by these angles. Based on the N alternative scanning angles, the uncovered solder joints are accurately found through detailed comparison and analysis.

[0027] Step S131422, multiple scanning angles other than the N candidate scanning angles in the first scanning angle sequence are recorded as remaining scanning angles, and the number of the uncovered solder joints in the remaining scanning angles is arranged in descending order, so as to obtain a first scanning angle rearrangement sequence. Specifically, multiple scanning angles other than the N candidate scanning angles in the initial first scanning angle sequence are defined as remaining scanning angles, the number of uncovered solder joints in the remaining scanning angles is calculated, and the remaining scanning angles are rearranged in descending order, so as to obtain a first scanning angle rearrangement sequence.

[0028] Step S131423, based on the first scanning angle rearrangement sequence, the N candidate scanning angles are supplemented and verified until the solder joint coverage integrity verification is passed. Specifically, based on the obtained first scanning angle rearrangement sequence, the N candidate scanning angles that have not passed the verification are supplemented and verified, and the supplementary verification process is repeated, and the coverage integrity of the solder joints is continuously checked until the solder joint coverage integrity verification is finally passed. Through repeated supplementation and verification, it is ensured that a sufficiently comprehensive and accurate scanning angle can be found, providing a reliable basis for subsequent inspection work.

[0029] In a possible implementation, based on the first scanning angle reordering sequence, the N candidate scanning angles are supplementarily verified until the solder joint coverage integrity verification is passed, and step S131423 further includes step S1314231, determining U key solder joints and V conventional solder joints among the M solder joints according to the shape structure of the connector to be inspected. Specifically, the shape and structure characteristics of the connector to be inspected are analyzed, and based on the characteristics and the importance and function of the connector in practical applications, U key solder joints and V conventional solder joints are distinguished from a total of M solder joints. Key solder joints are usually the parts that have a greater impact on the performance and safety of the connector, while conventional solder joints are relatively less important.

[0030] Step S1314232, in reference to the U key solder joints, a solder joint coverage integrity verification strategy is set, and the solder joint coverage integrity verification strategy includes a minimum scanning number threshold for the U key solder joints. Specifically, for the determined U key solder joints, a special solder joint coverage integrity verification strategy is set, and the strategy clearly stipulates the minimum scanning number threshold for each key solder joint, and the threshold is determined comprehensively based on factors such as the importance of the solder joint, welding process requirements, and quality standards.

[0031] Step S1314233, when performing supplementary verification, a list of key solder joints is generated through the solder joint coverage integrity verification strategy, and the number of scans of each key solder joint is counted. Specifically, when performing the supplementary verification operation, a list of all key solder joints is generated according to the previously set solder joint coverage integrity verification strategy, and during the verification process, the number of scans that have been completed for each key solder joint is accurately counted so as to be compared with the set minimum scan number threshold.

[0032] Step S1314234, if the number of scans of the key solder joint does not reach the corresponding minimum scan number threshold, then continue to perform supplementary scanning from the first scan angle rearrangement sequence, and the scan angles in the first scan angle rearrangement sequence are allowed to perform supplementary scanning only once. Specifically, if it is found that the number of scans of a certain key solder joint does not reach the preset minimum scan number threshold, then it is necessary to continue to perform supplementary scanning, and select a suitable scan angle from the first scan angle rearrangement sequence for supplementation, but it should be noted that each scan angle is only allowed to perform supplementary scanning only once in this process, so as to ensure the efficiency and rationality of the scan, and avoid waste of resources caused by excessive scanning.

[0033] In a possible implementation, according to the shape structure of the connector to be inspected, U key solder joints and V regular solder joints among the M solder joints are determined, and step S1314231 further includes step S13142311, if the scanning angles in the first scanning angle rearrangement sequence are all supplemented and scanned once, in combination with the V regular solder joints, an auxiliary tool is used to perform supplementary scanning, and the auxiliary tool includes a reflector. Specifically, if the scanning angles in the first scanning angle rearrangement sequence are all supplemented and scanned once, when all the scanning angles in the first scanning angle rearrangement sequence have completed a supplementary scanning operation, in combination with the V regular solder joints, an auxiliary tool is used to perform supplementary scanning, and the auxiliary tool includes a reflector. In order to further improve the scanning effect, especially for the situation that there may be insufficient scanning before, the V regular solder joints are also taken into consideration, and supplementary scanning is performed using an auxiliary tool. The auxiliary tool mentioned includes a reflector, and the propagation path of the light is changed by the reflector, so that the information of the solder joints can be obtained from more angles, and the possible scanning blind spots or incomplete parts can be compensated to improve the comprehensiveness and accuracy of the entire scan.

[0034] Step S200, based on the solder joint feature information, a deep learning model is trained to classify and identify defects, and the deep learning model is embedded with template matching technology, and the template matching technology is used to quickly identify known defect patterns. Specifically, the acquired solder joint feature information is the basic data for subsequent training of the deep learning model. The feature information contains various attributes and feature descriptions of the solder joint. The deep learning model is trained using the solder joint feature information. During the training process, a large amount of annotated solder joint feature data is input into the model, so that the model learns how to judge whether there are defects in the solder joint and the type of defects based on these features. Template matching technology is embedded in this deep learning model, and template matching technology can quickly identify those known defect patterns. Known defect patterns are relatively broad and common categories of defects, such as common large categories of patterns such as weak welding and abnormal solder joint shapes.

[0035] Step S300, based on historical welding quality examples, statistical distribution rules are collected and bound with the known defect patterns to build a defect database, which contains type information, morphology information, and size information of various welding defects. Specifically, historical examples are analyzed and studied, and statistical methods are used to mine the distribution patterns therein. For example, the frequency of welding defects of different types and under different process conditions, or the distribution characteristics of defects in welding of different materials and structures, etc., are counted, and the mined distribution patterns are closely bound to known defect patterns. Known defect patterns are relatively broad categories, and when constructing a defect database, they will be further subdivided into specific sub-category type information, such as subdividing welding defect types into sub-categories such as cracks, lack of fusion, and lack of penetration. In addition to type information, the database will also include morphological information to describe the shape, direction, and edge characteristics of the defects; as well as size information, recording specific numerical data such as the length, width, and depth of the defects. By integrating and associating distribution patterns, known defect patterns (including subdivided sub-category type information), morphological information, and size information, a comprehensive and detailed defect database is constructed. In subsequent welding quality inspection and evaluation, comparison and analysis based on the database can be performed to more accurately determine whether the currently inspected weld has defects, what kind of defects exist, and the severity of the defects, etc., providing strong support and reference for improving welding quality and improving processes.

[0036] Step S400, based on the type information of various welding defects in the defect database, set a first welding quality detection sub-loop, wherein the termination condition corresponding to the first welding quality detection sub-loop is associated with a first preset quality standard, and the first preset quality standard is determined based on industry specifications and connector product design requirements as constraints, including welding defect types that are not allowed to occur. Specifically, the defect database that has been constructed is analyzed, with special attention paid to the type information of various welding defects therein. The type information includes specific types of welding defects such as cracks, lack of fusion, and lack of penetration. Based on the type information of welding defects, the first welding quality detection sub-cycle is set. The sub-cycle is an important link in the entire welding quality detection process. When setting the first welding quality detection sub-cycle, its corresponding termination condition is clarified. The termination condition is closely related to the first preset quality standard. The determination of the first preset quality standard is a rigorous process. It is based on industry specifications and standards and fully considers the specific design requirements of connector products. Under constraints, specific welding defect types that are not allowed to appear are clearly specified. For example, industry specifications and standards may stipulate that cracks are not allowed in certain key parts, and connector product design requirements may further point out that lack of fusion is not allowed in specific locations. During the operation of the first welding quality detection sub-cycle, the welding quality will continue to be tested and evaluated until the termination condition associated with the first preset quality standard is reached. The sub-cycle will end only when the standards and conditions are met, thereby ensuring the accuracy and reliability of the detection and ensuring that the welding quality meets the requirements of industry specifications and product design.

[0037] Step S500, based on the morphological information of various welding defects in the defect database, a second welding quality detection sub-cycle is set, and the termination condition corresponding to the second welding quality detection sub-cycle is associated with the second preset quality standard, and the second preset quality standard is determined by the industry specification standards and welding process conditions as constraints. Based on the size information of various welding defects in the defect database, a third welding quality detection sub-cycle is set, and the termination condition corresponding to the third welding quality detection sub-cycle is associated with the third preset quality standard, and the third preset quality standard is determined by the industry specification standards and welding material characteristics as constraints. Specifically, the morphological information of various welding defects in the defect database is studied in depth. The morphological information may include detailed descriptions of the shape of the defect (such as long strips, circles, irregular shapes, etc.), direction (straight lines, curves, bends, etc.), edge features (smooth, jagged, etc.), etc. Based on the morphological information, a second welding quality detection sub-cycle is set. When setting the second welding quality detection sub-cycle, its corresponding termination condition is clarified. The termination condition is closely related to the second preset quality standard. The determination of the second preset quality standard comprehensively considers the industry specifications and standards and the actual welding process conditions. The industry specifications and standards will have a certain allowable range and restriction requirements for the morphology of welding defects, and the welding process conditions (such as welding methods, welding parameters, etc.) will also affect the acceptance of the defect morphology. The various welding defects in the defect database are analyzed. The dimensional information of the defects, such as the length, width, depth, area and other specific values ​​of the defects, is collected. According to the dimensional information, the third welding quality detection sub-cycle is set. The termination condition of the third welding quality detection sub-cycle is associated with the third preset quality standard. The formulation of the third preset quality standard is based on industry specifications and standards, and fully considers the characteristics of welding materials. Different welding materials have different performances and the ability to tolerate defects. Therefore, the characteristics of welding materials are taken into consideration when determining the preset quality standards. During the operation of the second and third welding quality detection sub-cycles, the shape and size of welding defects are continuously detected and evaluated until the termination conditions specified by the corresponding preset quality standards are met. Through meticulous and targeted settings, the welding quality can be comprehensively and accurately evaluated to ensure that the welding products meet the relevant requirements and standards.

[0038] Step S600, connect the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle to perform welding quality detection on the connector to be inspected. Specifically, the functions and key points of the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle are clarified. The first welding quality detection sub-cycle mainly performs detection based on the type information of welding defects, the second welding quality detection sub-cycle focuses on the morphological information of welding defects, and the third welding quality detection sub-cycle focuses on the size information of welding defects. When connecting these three sub-cycles, the sequence and logical relationship between them are carefully designed, and the termination condition of the first welding quality detection sub-cycle is defined as the starting condition of the second welding quality detection sub-cycle, which means that the second welding quality detection sub-cycle will only be started when the first sub-cycle completes the detection of the welding defect type and meets the corresponding standards. The termination condition of the second welding quality detection sub-cycle is set as the starting condition of the third welding quality detection sub-cycle. The connection method forms a coherent and orderly detection. Process, when preparing to conduct welding quality inspection on the connector to be inspected, first start the first welding quality inspection sub-cycle, in the sub-cycle, according to the first preset quality standard, the type of welding defects of the connector is comprehensively and carefully inspected and evaluated, once the first welding quality inspection sub-cycle meets the termination condition, the second welding quality inspection sub-cycle is automatically triggered, according to the second preset quality standard, the form of the welding defect is deeply analyzed and judged, when the second welding quality inspection sub-cycle reaches the termination condition, the third welding quality inspection sub-cycle is immediately started, according to the third preset quality standard, the size of the welding defect is accurately measured and evaluated, by connecting and running the three sub-cycles in sequence, the welding quality of the connector to be inspected can be comprehensively and systematically inspected from multiple dimensions, to ensure the accuracy, completeness and reliability of the inspection results, and provide sufficient basis for judging whether the welding quality of the connector is qualified.

[0039] In one possible implementation, the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle are connected to perform welding quality detection on the connector to be inspected, and step S600 further includes step S610, defining the termination condition corresponding to the first welding quality detection sub-cycle as the starting condition corresponding to the second welding quality detection sub-cycle, defining the termination condition corresponding to the second welding quality detection sub-cycle as the starting condition corresponding to the third welding quality detection sub-cycle, and connecting the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle. Specifically, the termination condition of the first welding quality detection sub-cycle is determined. The termination condition is set based on the detection and evaluation of the welding defect type. When a specific detection standard is reached or the predetermined detection steps and requirements are completed, the first welding quality detection sub-cycle is judged to be terminated. The termination condition is directly defined as the starting condition of the second welding quality detection sub-cycle, which means that only when the first welding quality detection sub-cycle is successfully completed and meets its termination condition, the second welding quality detection sub-cycle will be triggered to start. The termination condition of the second welding quality detection sub-cycle is clarified. The condition is set based on the detection and judgment of the welding defect morphology. When the second welding quality detection sub-cycle completes the relevant detection and evaluation of the welding defect morphology and meets the set termination standard, the second welding quality detection sub-cycle is declared to be terminated, and the termination condition of the second welding quality detection sub-cycle is set as the start condition of the third welding quality detection sub-cycle. The starting condition of the loop is that the third welding quality detection sub-loop will only start to run after the second welding quality detection sub-loop is successfully completed. Through the above method, the first, second and third welding quality detection sub-loops are connected in sequence. In actual operation, when the first welding quality detection sub-loop completes the detection of welding defect types according to predetermined standards and requirements, its termination condition is met, and the second welding quality detection sub-loop is immediately triggered to start detecting the welding defect morphology. When the second welding quality detection sub-loop also completes the task and meets the termination condition, the third welding quality detection sub-loop is started to detect the size of the welding defect. The connection method forms a logically tight and orderly detection process, which ensures a comprehensive, in-depth and gradually refined detection of welding quality, thereby improving the accuracy and reliability of the detection, and being able to more comprehensively and accurately evaluate whether the welding quality meets the requirements.

[0040] The embodiment of the present application uses a 3D scanner to collect point cloud information of the connector to be inspected and extracts deep features to obtain solder joint features. Based on this, a deep learning model embedded with template matching technology is trained to identify defects. A database containing defect type, morphology, and size information is constructed based on historical examples. Based on this, three welding quality detection sub-loops associated with preset quality standards are set respectively, and the three sub-loops are connected to perform welding quality inspection. This achieves the technical effect of achieving accurate and efficient inspection and evaluation of connector welding quality, comprehensively considering various information such as defect type, morphology, and size, and reducing the defective rate and usage risk, and improving the quality and reliability of connector products.

[0041] The above specific implementation manner does not constitute a limitation to the protection scope of the present application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present application should be included in the protection scope of the present application. In some cases, the actions or steps recorded in the present application can be performed in an order different from that in the embodiment and can still achieve the desired results. In addition, the process depicted in the accompanying drawings does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A method for detecting the welding quality of a connector, characterized in that: The method comprises: Use a 3D scanner to collect the connector to be inspected, obtain point cloud information, and perform deep feature extraction to obtain solder joint feature information; Based on the solder joint feature information, a deep learning model is trained to classify and identify defects, wherein the deep learning model is embedded with a template matching technology, and the template matching technology is used to quickly identify known defect patterns; According to historical welding quality examples, statistical distribution rules are obtained and bound with the known defect patterns to construct a defect database, which contains type information, morphology information, and size information of various welding defects; Based on the type information of various welding defects in the defect database, a first welding quality detection sub-cycle is set, wherein the termination condition corresponding to the first welding quality detection sub-cycle is associated with a first preset quality standard, and the first preset quality standard is determined by industry specifications and connector product design requirements as constraints, including welding defect types that are not allowed to occur; Based on the morphological information of various welding defects in the defect database, a second welding quality detection sub-cycle is set, and the termination condition corresponding to the second welding quality detection sub-cycle is associated with a second preset quality standard, and the second preset quality standard is determined by industry specification standards and welding process conditions as constraints; based on the size information of various welding defects in the defect database, a third welding quality detection sub-cycle is set, and the termination condition corresponding to the third welding quality detection sub-cycle is associated with a third preset quality standard, and the third preset quality standard is determined by industry specification standards and welding material characteristics as constraints; The first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle are connected to perform welding quality detection on the connector to be inspected.

2. A method for detecting the welding quality of a connector according to claim 1, characterized in that: The method comprises: Configuring 3D scanner parameters, including resolution and scanning speed; Based on the 3D scanner parameters, a comprehensive solder joint coverage scan is performed on the connector to be inspected, and the coverage positions corresponding to the point cloud information include M solder joints and M peripheral areas, where the peripheral area refers to the surrounding area that closely surrounds each solder joint and has slight soldering differences; The M welding points and the M peripheral areas are combined to perform synchronous adjustment on the 3D scanner during the scanning process.

3. A method for detecting the welding quality of a connector as claimed in claim 2, characterized in that: The M welding points and the M perimeter regions are combined, and the 3D scanner is synchronously adjusted during scanning. The method includes: According to the shape structure of the connector to be inspected, P position states are added, and the P position states correspond to the relative positions of the connector to be inspected and the 3D scanner; Based on the P position states, and in comparison with the blocked area, a position adjustment instruction is issued; After the connector to be inspected receives the position adjustment instruction, the rotation action, tilt action, and translation action corresponding to the position adjustment instruction are synchronously executed.

4. A method for detecting the welding quality of a connector as claimed in claim 3, characterized in that: The P position states correspond to the relative positions of the connector to be inspected and the 3D scanner, and the method further includes: In a first position state, q scanning angles are added, and the first position state is any one of the P position states; Based on the q scanning angles, the numbers of solder joints in each scanning angle are arranged in descending order to obtain a first scanning angle sequence; Based on the q scanning angles, the perimeter area coverage ratio in each scanning angle is arranged in descending order to obtain a second scanning angle sequence; According to the first scanning angle sequence and the second scanning angle sequence, q′ optimal scanning angles in a first position state are determined.

5. A method for detecting the welding quality of a connector as claimed in claim 4, characterized in that: Determining q′ optimal scanning angles in a first position state according to the first scanning angle sequence and the second scanning angle sequence, the method further includes: Comparing the first scanning angle sequence and the second scanning angle sequence, screening out N candidate scanning angles for the same scanning angle whose serial number in the first scanning angle sequence and the serial number in the second scanning angle sequence are consistent; receiving a first verification signal, performing solder joint coverage integrity verification on the N candidate scanning angles, and if the verification passes, issuing a second verification signal; The coverage integrity of the peripheral area of ​​the N candidate scanning angles is verified through the second verification signal. If the verification passes, q′ optimal scanning angles in the first position state are determined.

6. A method for detecting the welding quality of a connector as claimed in claim 5, characterized in that: Receiving a first verification signal, and performing solder joint coverage integrity verification on the N candidate scanning angles, the method further comprising: If it fails, the uncovered solder joints are determined based on N alternative scanning angles; Recording multiple scanning angles except N candidate scanning angles in the first scanning angle sequence as remaining scanning angles, and arranging the uncovered solder joints in descending order according to the number of solder joints in the remaining scanning angles, to obtain a first scanning angle rearrangement sequence; Based on the first scanning angle re-arrangement sequence, N candidate scanning angles are supplementarily verified until the solder joint coverage integrity verification is passed.

7. A method for detecting the welding quality of a connector as claimed in claim 6, characterized in that: Based on the first scanning angle re-arrangement sequence, N candidate scanning angles are supplementarily verified until the solder joint coverage integrity verification is passed, and the method further includes: According to the shape structure of the connector to be inspected, determining U key solder joints and V regular solder joints among the M solder joints; According to the U key solder joints, a solder joint coverage integrity verification strategy is set, wherein the solder joint coverage integrity verification strategy includes a minimum scanning number threshold of the U key solder joints; When performing supplementary verification, a list of key solder joints is generated by the solder joint coverage integrity verification strategy, and the number of scans of each key solder joint is counted; If the scanning times of the key solder joints do not reach the corresponding minimum scanning times threshold, then the supplementary scanning is continued from the first scanning angle rearrangement sequence, and the scanning angles in the first scanning angle rearrangement sequence are only allowed to be supplemented once.

8. A method for detecting the welding quality of a connector as claimed in claim 7, characterized in that: According to the shape structure of the connector to be inspected, U key solder joints and V regular solder joints among the M solder joints are determined, and the method further includes: If all the scanning angles in the first scanning angle rearrangement sequence are supplemented with one scan, an auxiliary tool is used to perform supplementary scanning in combination with the V conventional welding points, and the auxiliary tool includes a reflector.

9. A method for detecting the welding quality of a connector as claimed in claim 1, characterized in that: The first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle are connected, and the method includes: Define the termination condition corresponding to the first welding quality detection sub-cycle as the starting condition corresponding to the second welding quality detection sub-cycle, define the termination condition corresponding to the second welding quality detection sub-cycle as the starting condition corresponding to the third welding quality detection sub-cycle, and connect the first welding quality detection sub-cycle, the second welding quality detection sub-cycle, and the third welding quality detection sub-cycle.

Images