Intelligent deviation correction method and system for welding position

By acquiring workpiece welding characteristic information and installing visual sensing devices, and using a welding position analysis network for real-time correction, the problems of low welding position positioning accuracy and lack of real-time adjustment are solved, thereby improving welding accuracy and efficiency.

CN118371866BActive Publication Date: 2024-11-19SUZHOU SONGDE LASER TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410844951.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-06-27
Publication Date
2024-11-19
Estimated Expiration
2044-06-27

AI Technical Summary

Technical Problem

The existing welding position positioning accuracy is not high and there is a lack of real-time adjustment mechanism, resulting in unstable welding quality and inaccurate welds.

Method used

By acquiring workpiece welding characteristic information, installing vision sensing devices to monitor welding position parameters, using a welding position analysis network to determine the correction threshold, and activating an automatic correction mechanism to make adaptive adjustments when there is a real-time deviation.

Benefits of technology

This improves the accuracy and efficiency of welding, ensuring product quality.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118371866B_ABST
    Figure CN118371866B_ABST
Patent Text Reader

Abstract

The present application provides an intelligent deviation correction method and system for welding positions, and relates to the field of welding deviation correction technology. The method includes: obtaining workpiece welding characteristic information, monitoring and obtaining a set of welding position parameters through a visual sensor device, calling a welding position analysis network and performing welding position analysis, obtaining real-time welding position parameters, starting an automatic deviation correction mechanism, performing adaptive adjustment, obtaining welding position deviation correction parameters, and performing deviation correction and regulation on laser welding equipment based on the welding deviation correction parameters. The present application can solve the problems of unstable welding quality and inaccurate welds caused by the low welding position positioning accuracy and lack of a real-time adjustment mechanism in the prior art. The accuracy and efficiency of welding are improved, and product quality is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of welding deviation correction, and in particular to an intelligent deviation correction method and system for a welding position. Background Art

[0002] In the welding production process, due to factors such as workpiece processing errors, assembly errors, and thermal deformation during welding, the actual welded workpiece often differs from the design drawing, which will lead to inaccurate welding positions. Therefore, a technology that can automatically correct deviations is needed to improve the accuracy of welding positions.

[0003] At present, the existing welding position positioning technology mainly relies on the experience and skills of the operator, or adopts a relatively simple mechanical positioning method. These methods are often difficult to achieve ideal positioning accuracy and stability when dealing with complex workpieces or high-precision welding. At the same time, the existing welding equipment lacks real-time position monitoring and adjustment functions during the welding process. Once the welding position deviates, it is difficult to detect and correct it in time.

[0004] In summary, the existing technology has the technical problems of unstable welding quality and inaccurate welds due to the low welding position positioning accuracy and lack of real-time adjustment mechanism, which further affects the overall quality and performance of the product. Therefore, it is particularly important to develop a technology that can automatically monitor and adjust the welding position in real time to improve the accuracy and efficiency of welding and ensure product quality. Summary of the invention

[0005] The purpose of this application is to provide an intelligent deviation correction method and system for welding position, so as to solve the problems in the prior art such as unstable welding quality and inaccurate welds due to low welding position positioning accuracy and lack of real-time adjustment mechanism.

[0006] In view of the above problems, the present application provides an intelligent deviation correction method and system for welding position.

[0007] In a first aspect, the present application provides an intelligent deviation correction method for a welding position, which is implemented by an intelligent deviation correction system for a welding position, wherein the method includes: obtaining workpiece welding characteristic information, classifying and marking the workpiece welding characteristic information by attributes, and obtaining workpiece welding attribute calibration parameters; installing a visual sensing device on a laser welding device, and monitoring and obtaining a welding position parameter set by the visual sensing device, wherein the welding position parameter set includes a welding head position, a posture, and a workpiece size position; calling and obtaining a target welding position analysis network from a welding position analysis network library according to the workpiece welding attribute calibration parameters, performing welding position analysis on the welding position parameter set based on the target welding position analysis network, and determining a welding position deviation correction threshold parameter; acquiring welding real-time position parameters by the visual sensing device, and starting an automatic deviation correction mechanism when the welding real-time position parameters exceed the welding position deviation correction threshold parameter; adaptively adjusting the welding real-time position parameters based on the automatic deviation correction mechanism to obtain welding position deviation correction parameters, and performing deviation correction and regulation on the laser welding equipment based on the welding position deviation correction parameters.

[0008] In a second aspect, the present application further provides an intelligent deviation correction system for a welding position, which is used to execute an intelligent deviation correction method for a welding position as described in the first aspect, wherein the system comprises: a characteristic information acquisition module, which is used to acquire workpiece welding characteristic information, classify and mark the workpiece welding characteristic information, and obtain workpiece welding attribute calibration parameters; a welding position information acquisition module, which is used to install a visual sensing device on the laser welding equipment, and monitor and acquire a welding position parameter set through the visual sensing device, wherein the welding position parameter set includes the welding head position, posture, and workpiece size position; a deviation correction threshold parameter acquisition module, which is used to According to the workpiece welding property calibration parameters, a target welding position analysis network is called from the welding position analysis network library, and the welding position analysis is performed on the welding position parameter set based on the target welding position analysis network to determine the welding position correction threshold parameter; a threshold judgment module, the threshold judgment module is used to collect and obtain the welding real-time position parameters through the visual sensing device, when the welding real-time position parameters exceed the welding position correction threshold parameter, start the automatic correction mechanism; correction control module, the correction control module is used to adaptively adjust the welding real-time position parameters based on the automatic correction mechanism, obtain the welding position correction parameters, and perform correction control on the laser welding equipment based on the welding position correction parameters.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] By acquiring the welding characteristic information of the workpiece, the attribute classification and marking of the welding characteristic information of the workpiece is performed to obtain the welding attribute calibration parameters of the workpiece; a visual sensing device is installed on the laser welding equipment, and the welding position parameter set is monitored and obtained through the visual sensing device, and the welding position parameter set includes the welding head position, posture and workpiece size position; according to the welding attribute calibration parameters of the workpiece, the target welding position analysis network is called from the welding position analysis network library, and the welding position analysis of the welding position parameter set is performed based on the target welding position analysis network to determine the welding position correction threshold parameter; the welding real-time position parameter is collected and obtained through the visual sensing device, and when the welding real-time position parameter exceeds the welding position correction threshold parameter, the automatic correction mechanism is started; the welding real-time position parameter is adaptively adjusted based on the automatic correction mechanism to obtain the welding position correction parameter, and the laser welding equipment is corrected and adjusted based on the welding position correction parameter, which effectively solves the problem that the welding position positioning accuracy is not high and the real-time adjustment mechanism is lacking in the prior art, resulting in unstable welding quality and inaccurate welds. The accuracy and efficiency of welding are improved, and product quality is ensured.

[0011] 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 according to the contents of the specification, and 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 specifically cited below. It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easy to understand through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 A schematic diagram of a flow chart of an intelligent deviation correction method for a welding position according to the present application;

[0014] Figure 2 This is a structural schematic diagram of an intelligent deviation correction system for a welding position in the present application.

[0015] Description of reference numerals:

[0016] Characteristic information acquisition module 11, welding position information acquisition module 12, deviation correction threshold parameter acquisition module 13, threshold judgment module 14, deviation correction control module 15. DETAILED DESCRIPTION

[0017] This application provides an intelligent deviation correction method and system for welding positions, which solves the problems in the prior art such as unstable welding quality and inaccurate welds due to low welding position positioning accuracy and lack of real-time adjustment mechanism, improves welding accuracy and efficiency, and ensures product quality.

[0018] Below, the technical solutions in the present application will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments of the present application. It should be understood that the present application is not limited to the example embodiments described herein. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application. It should also be noted that, for the convenience of description, only the parts related to the present application are shown in the accompanying drawings, rather than all of them. Embodiment 1

[0019] Please see attached Figure 1 The present application provides an intelligent deviation correction method for a welding position, wherein the method is applied to an intelligent deviation correction system for a welding position, and the method specifically comprises the following steps:

[0020] S1: Acquire workpiece welding characteristic information, classify and mark the workpiece welding characteristic information by attributes, and obtain workpiece welding attribute calibration parameters.

[0021] Specifically, the welding characteristic information includes the physical properties of the metal, such as hardness, strength, and toughness. It is obtained through material testing reports or material storage databases. According to the workpiece material and the required welding strength, select the appropriate welding type, such as melting welding, including argon arc welding, gas shielded welding, laser welding, etc. or non-melting welding, including friction welding, pressure welding, etc. Determine the welding process, such as whether preheating, post-heating or heat treatment is required. According to the welding characteristic information, the welding properties are classified, such as weld type, welding material, welding process, weld quality, etc. Appropriate marking or coding rules are formulated for each type of attribute. For example, digital or letter codes can be used to represent different weld types, welding materials, etc. The calibration parameters of each type of welding attribute are obtained, including the strength, toughness, sealing, corrosion resistance of the weld, and the temperature and time during the welding process. It is obtained through destructive tests, such as tensile tests, impact tests, or non-destructive tests, such as X-ray testing and ultrasonic testing.

[0022] S2: Install a visual sensor device on the laser welding equipment, and monitor and obtain a welding position parameter set through the visual sensor device, wherein the welding position parameter set includes a welding head position, a posture, and a workpiece size position.

[0023] Specifically, the visual sensing device can monitor the spatial position of the welding head, that is, the coordinates in three-dimensional space, in real time. In addition to position information, the visual sensing device can also capture the posture of the welding head, including its tilt angle, rotation direction, etc. The acquisition of posture information helps to adjust the direction of the welding head to ensure the uniformity and quality of the weld. The visual sensing device can identify the contour and size of the workpiece, thereby determining the specific position of the workpiece.

[0024] S3: acquiring a target welding position analysis network from a welding position analysis network library according to the workpiece welding property calibration parameters, performing welding position analysis on the welding position parameter set based on the target welding position analysis network, and determining a welding position deviation correction threshold parameter.

[0025] Specifically, based on the obtained welding property calibration parameters, the most appropriate target welding position analysis network is called from the pre-built welding position analysis network library. This network library contains multiple analysis networks for different materials, processes and weld types. The welding position parameter set obtained by the visual sensor device, including the welding head position, posture, and workpiece size position, is input into the selected target welding position analysis network. The input welding position parameter set is processed and analyzed using the target welding position analysis network. This network may evaluate the accuracy of the welding position based on the input parameters and predict potential welding problems, such as misaligned welds and improper welding head posture. Based on the analysis results of the network, the correction threshold parameters of the welding position can be determined. Including the preset welding position and the preset degree of deviation that needs to be corrected.

[0026] S4: acquiring the real-time welding position parameter through the visual sensing device, and starting the automatic deviation correction mechanism when the real-time welding position parameter exceeds the welding position deviation correction threshold parameter.

[0027] Specifically, the visual sensing device is configured to continuously capture the relative position of the welding head and the workpiece during the welding process. These real-time position parameters include but are not limited to the three-dimensional coordinates of the welding head, attitude angles such as pitch angle, yaw angle, roll angle, and the real-time position of the weld on the workpiece. The collected real-time position parameters are compared with the previously determined welding position correction threshold parameters. If the real-time position parameters are within the allowable threshold range, the welding process will continue and no correction action will be triggered. If the real-time position parameters exceed the set correction threshold, this means that the position of the welding head or workpiece has deviated from the predetermined trajectory, which may affect the welding quality. Once it is detected that the position deviation exceeds the threshold, the automatic correction mechanism is triggered. The automatic correction mechanism includes adjusting the position of the welding head, changing the welding speed, correcting the welding trajectory, etc., to ensure that the welding head can accurately weld along the predetermined weld.

[0028] S5: Adaptively adjust the real-time welding position parameters based on the automatic deviation correction mechanism to obtain welding position deviation correction parameters, and perform deviation correction and regulation on the laser welding equipment based on the welding position deviation correction parameters.

[0029] Specifically, according to the exceeded deviation, the real-time welding position parameters are adaptively adjusted through a preset algorithm. This includes fine-tuning the welding head position, changes in welding speed, and correction of welding trajectory. After adaptive adjustment, a new set of welding position parameters, namely welding position correction parameters, is obtained. These parameters reflect the specific adjustments that need to be made to correct the deviation, such as the distance and direction the welding head needs to move, the amount the welding speed needs to be increased or decreased, etc. The welding position correction parameters are converted into specific mechanical instructions, and the laser welding equipment is controlled in real time through the drive device. For example, adjust the position and posture of the welding robot arm to return it to the correct welding trajectory; or adjust the output power and welding speed of the laser to ensure the quality and uniformity of the weld.

[0030] Furthermore, step S1 of the present application also includes:

[0031] Acquire welding characteristic factor information, wherein the welding characteristic factor information includes welding temperature, welding pressure, workpiece material and workpiece structural shape; refine the attribute content of each factor information in the welding characteristic factor information to obtain a welding factor attribute content node set; feature encode each content node in the welding factor attribute content node set to obtain a welding attribute node code set; construct a workpiece welding characteristic classifier based on the welding factor attribute content node set and the welding attribute node code set; attribute classification mark the workpiece welding characteristic information through the workpiece welding characteristic classifier to obtain the workpiece welding attribute calibration parameter.

[0032] Specifically, welding temperature records the temperature data during welding, including preheating temperature, welding temperature, and cooling rate after welding. Welding pressure is the pressure applied to the workpiece during welding. Workpiece material includes the material type, thickness, strength and other characteristics of the workpiece. The workpiece structural shape includes the location and type of the weld. Detailed attribute content refinement is performed for each welding characteristic factor. For example, welding temperature can be refined into preheating temperature range, welding temperature range, cooling rate, etc.; workpiece material can be refined into material type, thickness range, tensile strength, etc. Feature encoding is performed on the refined welding factor attribute content nodes. This usually involves assigning a unique code or identifier to each attribute content node. The code can be a combination of numbers, letters or symbols. A workpiece welding characteristic classifier is constructed using the refined welding factor attribute content node set and the corresponding feature code set. This classifier can automatically classify and mark the workpiece according to the input welding characteristic factor information. Through the workpiece welding characteristic classifier, the specific workpiece welding characteristic information is attribute classified and marked. The classification mark includes the welding difficulty level of the workpiece, the recommended welding method, the expected welding quality level, etc. Based on the classification and labeling, a set of workpiece welding property calibration parameters can be obtained.

[0033] Furthermore, step S2 of the present application also includes:

[0034] The target welding image information is captured by the visual sensing device, and the target welding image information is subjected to noise identification and wavelet denoising by Fourier transform to obtain standard welding image information; a laser welding entity label library is constructed, and a laser welding image set is subjected to semantic segmentation training based on the laser welding entity label library to obtain a welding entity semantic segmentation network; semantic segmentation and entity labeling are performed on the standard welding image information based on the welding entity semantic segmentation network to obtain a target welding labeled entity set; welding position measurement is performed on each entity information in the target welding labeled entity set to obtain the welding position parameter set.

[0035] Specifically, a visual sensing device, such as a high-resolution camera, is used to capture the target welding image information during the welding process. This image information contains key data such as the location of the welding joint, the shape and position of the workpiece, etc. Fourier transform is applied to the captured target welding image information for noise identification. Fourier transform can convert the image from the spatial domain to the frequency domain, making it easier to identify the noise components in the image. The identified noise is removed using wavelet denoising technology. A label library containing various laser welding related entities is created. This label library should cover various entities that may appear in the welding process, such as welding heads, workpieces, welds, etc., and define clear labels and attributes for each entity. Using the constructed laser welding entity label library, a set of laser welding image sets is trained for semantic segmentation. A deep learning algorithm, such as a convolutional neural network, is used to train the model so that it can accurately identify different entities in the image. Through training, a welding entity semantic segmentation network is obtained, which can automatically segment the welding image into different entity regions and label each region accordingly. The standard welding image information processed by wavelet denoising is input into the welding entity semantic segmentation network. The network will perform semantic segmentation on the image, segment each entity in the image, such as welding heads, workpieces, etc., and label them accordingly, thereby obtaining a target welding labeled entity set. The welding position of each entity in the target welding labeled entity set is measured. This is achieved by calculating the coordinates, size, and orientation of each entity in the image. The measurement results will constitute a welding position parameter set, which contains the precise position information of each key entity in the welding process.

[0036] Furthermore, the present application also includes:

[0037] The measurement elements of each entity information in the target welding labeled entity set are respectively parsed to determine the measurement entity target element information; the measurement reference mark of the target welding labeled entity set is performed based on the measurement entity target element information to establish an entity element measurement reference; pixel segmentation and identification are performed on each entity information to determine an entity measurement anchor frame set; the entity measurement anchor frame set is traversed and measured in turn according to the entity element measurement reference to determine the welding position parameter set.

[0038] Specifically, determine the measurement entity target element information, that is, the content of the entity that needs to be measured. For example, for a welding joint, its position and posture information need to be measured. Based on the measurement entity target element information, the target welding labeled entity set is marked with a measurement benchmark. A reference point or reference line is set for each entity in the image or three-dimensional space. Each entity information is segmented and identified at the pixel level, and the boundary of the entity can be accurately defined using edge detection, threshold segmentation, etc. According to the segmentation results, the measurement anchor frame set of each entity is determined. The anchor frame is a rectangular frame that tightly surrounds the entity and is used to locate the specific position and range of the entity in the image. According to the entity element measurement benchmark, the entity measurement anchor frame set is traversed and measured in sequence. Including calculating various measurement elements of the entity in the anchor frame, such as length, width, area, angle, etc. These measurement results are integrated to form a welding position parameter set. This set will contain accurate measurement data of all key entities, providing accurate position and size information for subsequent welding operations.

[0039] Furthermore, the present application also includes:

[0040] A laser welding position database is obtained, and the laser welding position database is classified and identified based on the workpiece welding characteristic classifier to obtain a welding position calibration attribute data set; the welding position calibration attribute data set is iteratively trained using a deep learning network structure to obtain an initial welding position analysis network set; the initial welding position analysis network set is verified and evaluated to obtain an analysis network verification and evaluation parameter set, and model optimization is performed based on the analysis network verification and evaluation parameter set to construct the welding position analysis network library; based on the workpiece welding attribute calibration parameters, the target welding position analysis network is matched and obtained from the welding position analysis network library.

[0041] Specifically, a laser welding position database is obtained, which contains a large amount of laser welding position data. The database is classified and identified using a workpiece welding characteristic classifier. According to the welding characteristics of the workpiece, such as material type, structural shape, etc., each record in the database is labeled accordingly. After classification and identification, a welding position calibration attribute data set is obtained, which will be used to train a deep learning network. The welding position calibration attribute data set is iteratively trained using deep learning network structures, such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs). During the training process, the weights and parameters of the network are continuously adjusted to minimize the prediction error and improve the accuracy of the network. Through multiple iterative training, a set of initial welding position analysis network sets is obtained. The initial welding position analysis network set is verified and evaluated by using an independent verification data set, which does not participate in the previous training process. The performance of each network, such as accuracy, recall rate, F1 score, etc., is evaluated, and these evaluation parameters are recorded to form a set of analysis network verification evaluation parameters. Based on the verification evaluation results, the network is tuned. This includes adjusting the network structure, increasing or decreasing the number of network layers, changing the activation function, etc. After optimization, a welding position analysis network library is constructed, which contains multiple verified and optimized deep learning network models. When a new workpiece needs to be analyzed for welding position, the welding property calibration parameters of the workpiece are first obtained. Based on these parameters, the most suitable analysis network is matched from the welding position analysis network library. The matching process can use methods such as similarity measurement or rule matching. The matched target welding position analysis network will be used to perform specific welding position analysis on the workpiece.

[0042] Furthermore, the present application also includes:

[0043] A network optimization parameter space is constructed, wherein the network optimization parameter space includes performance evaluation parameters of a historical welding position analysis network, model optimization strategy parameters, and corresponding model optimization effect data; a set of analysis network verification evaluation parameters is matched and screened with the network optimization parameter space to obtain a network optimization parameter search library set; a global comparison and optimization is performed in the network optimization parameter search library set to determine a network optimization strategy parameter set, and the initial welding position analysis network set is optimized based on the network optimization strategy parameter set to obtain the welding position analysis network library.

[0044] Specifically, the performance evaluation parameters of the historical welding position analysis network are collected. Including accuracy, recall rate, F1 score, training time, etc. The model optimization strategy parameters are sorted out. These parameters cover various strategies for improving network performance, such as learning rate adjustment, batch size, regularization method, optimizer selection, etc. The corresponding model optimization effect data are collected. These data record the changes in network performance after the implementation of different optimization strategies, and a comprehensive network optimization parameter space is constructed. The analysis network verification evaluation parameter set is used to match and screen the network optimization parameter space. According to the performance of the current network, similar or related optimization strategies and effect data are searched in the network optimization parameter space. Through matching and screening, a network optimization parameter search library set is obtained. This set contains a variety of optimization strategies that may be effective for the current network performance. In the network optimization parameter search library set, a global comparison is performed to find the best solution to improve the current network performance. A set of network optimization strategy parameter sets is determined. This set of parameters is carefully selected based on historical data and current network performance, aiming to maximize the accuracy and efficiency of the network. Based on the determined network optimization strategy parameter set, the initial welding position analysis network set is optimized, including adjusting the network structure, changing the learning rate, adding regularization terms, and other operations.

[0045] Furthermore, the present application also includes:

[0046] A welding test is performed on the laser welding equipment to obtain equipment response speed information and welding precision deviation information; a position impact analysis is performed based on the equipment response speed information and the welding precision deviation information to obtain a position deviation impact factor; a welding position deviation compensation parameter is determined according to the position deviation impact factor; and parameter compensation is performed on the welding position correction parameter based on the welding position deviation compensation parameter to determine the welding position correction parameter.

[0047] Specifically, the laser welding equipment is subjected to actual welding tests by performing multiple welding tests under standard conditions. During the test, the equipment response speed information is recorded. This includes the time required for the equipment to reach a stable welding state from the start of welding, as well as the changes in the equipment response at different welding stages. At the same time, the deviation information of the welding accuracy is measured and recorded. By comparing the difference between the actual welding position and the preset welding position. Based on the collected equipment response speed information and welding accuracy deviation information, a position influence analysis is performed. Through analysis, it is determined which factors cause the deviation of the welding position, such as equipment response delay, mechanical vibration, thermal deformation, etc. The contribution of these influencing factors to the welding position deviation is quantified to obtain the position deviation influencing factors. These factors reflect the influence of different factors on the accuracy of the welding position. According to the position deviation influencing factors, the welding position deviation compensation parameters are determined. These parameters are intended to reduce or eliminate the position deviation caused by the equipment response speed and welding accuracy deviation. The compensation parameters include adjusting the position of the welding starting point, changing the welding speed or power, optimizing the equipment response algorithm, etc. Based on the determined welding position deviation compensation parameters, the original welding position correction parameters are parameter compensated. The compensation parameters are introduced into the original correction algorithm or control strategy to adjust the welding position more accurately. After parameter compensation, the new welding position correction parameters are determined.

[0048] In summary, the intelligent deviation correction method for welding position provided by the present application has the following technical effects:

[0049] By acquiring the welding characteristic information of the workpiece, the attribute classification and marking of the welding characteristic information of the workpiece is performed to obtain the welding attribute calibration parameters of the workpiece; a visual sensing device is installed on the laser welding equipment, and the welding position parameter set is monitored and obtained through the visual sensing device, and the welding position parameter set includes the welding head position, posture and workpiece size position; according to the welding attribute calibration parameters of the workpiece, the target welding position analysis network is called from the welding position analysis network library, and the welding position analysis of the welding position parameter set is performed based on the target welding position analysis network to determine the welding position correction threshold parameter; the welding real-time position parameter is collected and obtained through the visual sensing device, and when the welding real-time position parameter exceeds the welding position correction threshold parameter, the automatic correction mechanism is started; the welding real-time position parameter is adaptively adjusted based on the automatic correction mechanism to obtain the welding position correction parameter, and the laser welding equipment is corrected and adjusted based on the welding position correction parameter, which effectively solves the problem that the welding position positioning accuracy is not high and the real-time adjustment mechanism is lacking in the prior art, resulting in unstable welding quality and inaccurate welds. The accuracy and efficiency of welding are improved, and product quality is ensured. Embodiment 2

[0050] Based on the intelligent deviation correction method of a welding position in the aforementioned embodiment, the present application also provides an intelligent deviation correction system for a welding position, please refer to the attached Figure 2 , the system comprising:

[0051] The characteristic information acquisition module 11 is used to acquire the workpiece welding characteristic information, classify and mark the workpiece welding characteristic information, and obtain the workpiece welding attribute calibration parameters.

[0052] The welding position information acquisition module 12 is used to install a visual sensor device on the laser welding equipment, and monitor and acquire a welding position parameter set through the visual sensor device. The welding position parameter set includes the welding head position, posture and workpiece size position.

[0053] The deviation correction threshold parameter acquisition module 13 is used to call and obtain the target welding position analysis network from the welding position analysis network library according to the workpiece welding property calibration parameters, perform welding position analysis on the welding position parameter set based on the target welding position analysis network, and determine the welding position deviation correction threshold parameter.

[0054] The threshold judgment module 14 is used to collect and obtain the real-time welding position parameters through the visual sensor device, and when the real-time welding position parameters exceed the welding position correction threshold parameters, the automatic correction mechanism is started.

[0055] The deviation correction and control module 15 is used to adaptively adjust the real-time welding position parameters based on the automatic deviation correction mechanism, obtain welding position deviation correction parameters, and perform deviation correction and control on the laser welding equipment based on the welding position deviation correction parameters.

[0056] Furthermore, the characteristic information acquisition module 11 in the system is also used for:

[0057] Acquiring welding characteristic factor information, wherein the welding characteristic factor information includes welding temperature, welding pressure, workpiece material and workpiece structure shape;

[0058] Refining the attribute content of each factor information in the welding characteristic factor information to obtain a welding factor attribute content node set;

[0059] Performing feature coding on each content node in the welding factor attribute content node set to obtain a welding attribute node coding set;

[0060] Constructing a workpiece welding characteristic classifier based on the welding factor attribute content node set and the welding attribute node code set;

[0061] The workpiece welding characteristic information is marked by attribute classification through the workpiece welding characteristic classifier to obtain the workpiece welding attribute calibration parameters.

[0062] Furthermore, the welding position information acquisition module 12 in the system is also used for:

[0063] The target welding image information is captured by the visual sensing device, and the target welding image information is subjected to noise identification and wavelet denoising by Fourier transform to obtain standard welding image information;

[0064] Constructing a laser welding entity label library, and performing semantic segmentation training on a laser welding image set based on the laser welding entity label library to obtain a welding entity semantic segmentation network;

[0065] Based on the welding entity semantic segmentation network, the standard welding image information is semantically segmented and entity labeled to obtain a target welding labeled entity set;

[0066] The welding position is measured for each entity information in the target welding labeled entity set to obtain the welding position parameter set.

[0067] Furthermore, the system further comprises a position parameter set determination module, wherein the position parameter set determination module is used to:

[0068] Performing measurement element analysis on each entity information in the target welding labeled entity set to determine the measurement entity target element information;

[0069] Based on the measurement entity target element information, the target welding labeled entity set is marked with a measurement reference to establish an entity element measurement reference;

[0070] Perform pixel segmentation and labeling on each entity information to determine a set of entity measurement anchor frames;

[0071] The entity measurement anchor frame set is traversed and measured in sequence according to the entity element measurement reference to determine the welding position parameter set.

[0072] Furthermore, the system further comprises a location analysis network library construction module, wherein the location analysis network library construction module is used to:

[0073] Acquire a laser welding position database, classify and identify the laser welding position database based on the workpiece welding characteristic classifier, and obtain a welding position calibration attribute data set;

[0074] Iteratively training the welding position calibration attribute data sets respectively using a deep learning network structure to obtain an initial welding position analysis network set;

[0075] Performing verification and evaluation on the initial welding position analysis network set to obtain an analysis network verification and evaluation parameter set, and performing model optimization based on the analysis network verification and evaluation parameter set to construct the welding position analysis network library;

[0076] Based on the workpiece welding property calibration parameters, the target welding position analysis network is obtained by matching from the welding position analysis network library.

[0077] Furthermore, the system further comprises a network set optimization module, wherein the network set optimization module is used to:

[0078] Constructing a network optimization parameter space, wherein the network optimization parameter space includes performance evaluation parameters of a historical welding position analysis network, model optimization strategy parameters, and corresponding model optimization effect data;

[0079] Based on the analysis network verification evaluation parameter set and the network optimization parameter space, matching and screening are performed to obtain a network optimization parameter search library set;

[0080] A global comparison and optimization is performed in the network optimization parameter search library set to determine a network optimization strategy parameter set, and the initial welding position analysis network set is optimized based on the network optimization strategy parameter set to obtain the welding position analysis network library.

[0081] Furthermore, the system further comprises a correction parameter determination module, wherein the correction parameter determination module is used to:

[0082] Performing welding tests on the laser welding equipment to obtain equipment response speed information and welding accuracy deviation information;

[0083] Performing position impact analysis based on the equipment response speed information and welding precision deviation information to obtain a position deviation impact factor;

[0084] Determining welding position deviation compensation parameters according to the position deviation influencing factors;

[0085] The welding position deviation correction parameter is compensated for based on the welding position deviation compensation parameter to determine the welding position correction correction parameter.

[0086] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. Figure 1The intelligent deviation correction method and specific examples of a welding position in the first embodiment are also applicable to the intelligent deviation correction system of a welding position in the present embodiment. Through the detailed description of the intelligent deviation correction method of a welding position, those skilled in the art can clearly know the intelligent deviation correction system of a welding position in the present embodiment, so for the sake of brevity of the specification, it will not be described in detail here. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0087] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

[0088] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application belong to the scope of the present application and its equivalent technology, the present application is also intended to include these modifications and variations.

Claims

1. An intelligent deviation correction method for welding position, characterized in that: The method comprises: Acquire workpiece welding characteristic information, classify and mark the workpiece welding characteristic information by attributes, and obtain workpiece welding attribute calibration parameters; Installing a visual sensor device on the laser welding equipment, and monitoring and acquiring a welding position parameter set through the visual sensor device, wherein the welding position parameter set includes a welding head position, a posture, and a workpiece size position; According to the workpiece welding property calibration parameters, a target welding position analysis network is acquired from a welding position analysis network library, welding position analysis is performed on the welding position parameter set based on the target welding position analysis network, and a welding position deviation correction threshold parameter is determined; The real-time welding position parameter is collected and acquired by the visual sensing device, and when the real-time welding position parameter exceeds the welding position deviation correction threshold parameter, an automatic deviation correction mechanism is started; Adaptively adjusting the real-time welding position parameters based on the automatic deviation correction mechanism to obtain welding position deviation correction parameters, and performing deviation correction and regulation on the laser welding equipment based on the welding position deviation correction parameters; The calling and obtaining of the target welding position analysis network includes: Acquire a laser welding position database, classify and identify the laser welding position database based on the workpiece welding characteristic classifier, and obtain a welding position calibration attribute data set; Iteratively training the welding position calibration attribute data sets respectively using a deep learning network structure to obtain an initial welding position analysis network set; Performing verification and evaluation on the initial welding position analysis network set to obtain an analysis network verification and evaluation parameter set, and performing model optimization based on the analysis network verification and evaluation parameter set to construct the welding position analysis network library; Based on the workpiece welding property calibration parameters, matching and acquiring the target welding position analysis network from the welding position analysis network library; The step of constructing the welding position analysis network library comprises: Constructing a network optimization parameter space, wherein the network optimization parameter space includes performance evaluation parameters of a historical welding position analysis network, model optimization strategy parameters, and corresponding model optimization effect data; Based on the analysis network verification evaluation parameter set and the network optimization parameter space, matching and screening are performed to obtain a network optimization parameter search library set; A global comparison and optimization is performed in the network optimization parameter search library set to determine a network optimization strategy parameter set, and the initial welding position analysis network set is optimized based on the network optimization strategy parameter set to obtain the welding position analysis network library.

2. An intelligent deviation correction method for welding position according to claim 1, characterized in that: The step of obtaining the workpiece welding property calibration parameters includes: Acquiring welding characteristic factor information, wherein the welding characteristic factor information includes welding temperature, welding pressure, workpiece material and workpiece structure shape; Refining the attribute content of each factor information in the welding characteristic factor information to obtain a welding factor attribute content node set; Performing feature coding on each content node in the welding factor attribute content node set to obtain a welding attribute node coding set; Constructing a workpiece welding characteristic classifier based on the welding factor attribute content node set and the welding attribute node code set; The workpiece welding characteristic information is marked by attribute classification through the workpiece welding characteristic classifier to obtain the workpiece welding attribute calibration parameters.

3. The intelligent deviation correction method for welding position according to claim 1, characterized in that: The step of obtaining a welding position parameter set includes: The target welding image information is captured by the visual sensing device, and the target welding image information is subjected to noise identification and wavelet denoising by Fourier transform to obtain standard welding image information; Constructing a laser welding entity label library, and performing semantic segmentation training on a laser welding image set based on the laser welding entity label library to obtain a welding entity semantic segmentation network; Based on the welding entity semantic segmentation network, the standard welding image information is semantically segmented and entity labeled to obtain a target welding labeled entity set; The welding position is measured for each entity information in the target welding labeled entity set to obtain the welding position parameter set.

4. The intelligent deviation correction method of welding position according to claim 3, characterized in that: The obtaining of the welding position parameter set comprises: Performing measurement element analysis on each entity information in the target welding labeled entity set to determine the measurement entity target element information; Based on the measurement entity target element information, the target welding labeled entity set is marked with a measurement reference to establish an entity element measurement reference; Perform pixel segmentation and labeling on each entity information to determine a set of entity measurement anchor frames; The entity measurement anchor frame set is traversed and measured in sequence according to the entity element measurement reference to determine the welding position parameter set.

5. The intelligent deviation correction method for welding position according to claim 1, characterized in that: The method comprises: Performing welding tests on the laser welding equipment to obtain equipment response speed information and welding accuracy deviation information; Performing position impact analysis based on the equipment response speed information and welding precision deviation information to obtain a position deviation impact factor; Determining welding position deviation compensation parameters according to the position deviation influencing factors; The welding position deviation correction parameter is compensated for based on the welding position deviation compensation parameter to determine the welding position correction correction parameter.

6. An intelligent deviation correction system for welding position, characterized in that: For implementing the steps of the method according to any one of claims 1 to 5, the system comprises: A characteristic information acquisition module, the characteristic information acquisition module is used to acquire workpiece welding characteristic information, classify and mark the workpiece welding characteristic information, and obtain workpiece welding attribute calibration parameters; A welding position information acquisition module, wherein the welding position information acquisition module is used to install a visual sensor device on the laser welding equipment, and monitor and acquire a welding position parameter set through the visual sensor device, wherein the welding position parameter set includes a welding head position, a posture, and a workpiece size position; A deviation correction threshold parameter acquisition module, the deviation correction threshold parameter acquisition module is used to call and obtain a target welding position analysis network from a welding position analysis network library according to the workpiece welding property calibration parameters, perform welding position analysis on the welding position parameter set based on the target welding position analysis network, and determine a welding position deviation correction threshold parameter; A threshold judgment module, the threshold judgment module is used to collect and obtain the real-time welding position parameters through the visual sensor device, and when the real-time welding position parameters exceed the welding position deviation correction threshold parameters, an automatic deviation correction mechanism is started; A deviation correction and control module, wherein the deviation correction and control module is used to adaptively adjust the real-time welding position parameters based on the automatic deviation correction mechanism, obtain welding position deviation correction parameters, and perform deviation correction and control on the laser welding equipment based on the welding position deviation correction parameters.