A method and system for real-time monitoring of an electric welding process

Abstract

The present application relates to the technical field of electric welding process monitoring, and particularly relates to a real-time monitoring method and system for electric welding process, which comprises the following steps: acquiring dynamic working parameters and welding machine data line output parameters of the electric welding process; locking and tracking a focus point based on a camera calibration algorithm, and recording a welding path; calculating a projection matrix and a homography matrix of the tracking focus point, and obtaining welding track information; acquiring real-time dynamic images of the electric welding process, and performing image processing to obtain welding area image information; correlating the welding track information and the welding area image information to obtain comprehensive image information of the electric welding process; and synchronizing the comprehensive parameter information and the comprehensive image information in time to perform real-time monitoring on the electric welding process. Through the present application, the problem that it is difficult to monitor the electric welding effect in real time in the electric welding process and the monitoring information is single in the prior art is effectively solved, and real-time monitoring is realized by establishing a real-time monitoring model and utilizing the correlation between parameters and images.

Description

Technical Field

[0001] This invention relates to the field of electric welding process monitoring technology, and in particular to a real-time monitoring method and system for electric welding processes. Background Technology

[0002] Electric welding is a metal joining process that involves heating two or more metal workpieces to a sufficiently high temperature to melt them and then forming a strong connection after cooling. This process uses an electric current to generate heat, usually in the form of an electric arc, hence it is also known as arc welding. Electric welding technology is widely used in many fields such as the automotive industry and the construction industry.

[0003] With the continuous advancement of technology, welding technology is also constantly evolving. However, interference caused by strong noise such as spatter, high temperature, strong light and smoke not only leads to poor welding results, but also increases the difficulty of detecting problems in real time during the welding process. This results in the problem of delayed handling of deviations in the welding process. Even if there are real-time monitoring methods, the monitoring data is relatively one-dimensional. Therefore, comprehensive real-time monitoring of the welding process is particularly important.

[0004] The information disclosed in this background section is intended only to enhance the understanding of the general background of this disclosure and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0005] This invention provides a real-time monitoring method for the electric welding process, which can effectively solve the problems in the background art.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0007] The dynamic working parameters of the electric welding process and the data cable output parameters of the welding machine are obtained to obtain comprehensive parameter information of the electric welding process;

[0008] Based on the camera calibration algorithm, the tracking focus is locked, and the welding path is recorded based on the target tracking algorithm;

[0009] Calculate the projection matrix and homography matrix of the tracking focus, and obtain the welding trajectory information based on the welding path image;

[0010] Acquire real-time dynamic images of the welding process and perform image processing on the real-time dynamic images to obtain image information of the welding area;

[0011] The welding trajectory information and the welding area image information are correlated to obtain comprehensive image information of the electric welding process;

[0012] The integrated parameter information and the integrated image information are synchronized in time, and a real-time monitoring model is established to monitor the welding process in real time.

[0013] Furthermore, by acquiring the dynamic operating parameters of the welding process and the data cable output parameters of the welding machine, comprehensive parameter information of the welding process is obtained, including:

[0014] The dynamic operating parameters of the welding process are obtained by using a high-definition camera to photograph the digital display screen of the welding machine.

[0015] Connect the data cable to the welding machine and obtain the output parameters of the data cable;

[0016] The dynamic operating parameters are used to correct the output parameters of the data cable, and the comprehensive parameter information is obtained.

[0017] Furthermore, based on the camera calibration algorithm, the tracking focus is locked, and based on the target tracking algorithm, the welding path is recorded, including:

[0018] Real-time dynamic images of the welding point are acquired, and the position of the tracking focus is locked based on a camera calibration algorithm;

[0019] The target tracking algorithm is used to locate the position of the tracking focus;

[0020] Based on the positional changes of the tracking focus, an image sequence of the welding path is constructed;

[0021] The image sequence is converted into a welding path represented by 2D data point coordinates.

[0022] Further, the projection matrix and homography matrix of the tracking focus are calculated, and welding trajectory information is obtained based on the welding path, including:

[0023] Using the camera's internal and external calibration parameters, the projection matrix and the homography matrix are calculated using known world coordinate points and their corresponding points in the welding path;

[0024] Based on the projection matrix and the homography matrix, each 2D data point on the welding path is mapped to each of the world coordinate points;

[0025] The welding trajectory information is calculated based on the aforementioned world coordinate points.

[0026] The welding trajectory information includes, but is not limited to, welding path, path length, and speed.

[0027] Furthermore, real-time dynamic images of the welding process are acquired, and image processing is performed on the dynamic images to obtain welding area image information, including:

[0028] The welding process was captured in real time using a high dynamic range camera and with the assistance of a neutral density filter.

[0029] Image enhancement is performed on the real-time dynamic images;

[0030] The enhanced image is segmented to obtain the image information of the welding area.

[0031] Furthermore, the welding trajectory information and the welding area image information are correlated to obtain comprehensive image information of the welding process, including:

[0032] Match the timestamps in the welding trajectory information with the corresponding welding area image frames;

[0033] The matched welding trajectory information is integrated with the welding area image information to obtain the comprehensive image information.

[0034] Furthermore, the integrated parameter information and the integrated image information are synchronized in time, and a real-time monitoring model is established to monitor the welding process in real time, including:

[0035] The integrated parameter information is associated with the corresponding integrated image information according to the timestamp correspondence, wherein the association makes the fluctuation of the parameter information correspond to the change of the integrated image information.

[0036] Based on the real-time changing parameter information and the corresponding comprehensive image information, the real-time monitoring model is established to monitor the welding process in real time.

[0037] Furthermore, based on the real-time changing parameter information and the corresponding comprehensive image information, the real-time monitoring model is established to monitor the welding process in real time, including:

[0038] Collect the integrated parameter information and integrated image information of the welding process along different paths, and match them by timestamp;

[0039] The integrated parameter information and the integrated image information are converted into sequence data;

[0040] A real-time monitoring model for the electric welding process is established based on a recurrent neural network model, and the sequence data is used for training and verification.

[0041] For each welding process, the integrated parameter information and integrated image information, which are converted into sequence data, are used as inputs for the real-time monitoring model to monitor and issue early warnings.

[0042] A method for real-time monitoring of an electric welding process, the system comprising:

[0043] The parameter information acquisition module acquires the dynamic working parameters of the electric welding process and the data cable output parameters of the welding machine to obtain comprehensive parameter information of the electric welding process.

[0044] The path image recording module locks the tracking focus based on the camera calibration algorithm and records the welding path based on the target tracking algorithm;

[0045] The trajectory data acquisition module calculates the projection matrix and homography matrix of the tracking focus, and obtains welding trajectory information based on the welding path;

[0046] The regional image acquisition module acquires real-time dynamic images of the welding process and performs image processing on the real-time dynamic images to obtain welding area image information.

[0047] The integrated image acquisition module correlates the welding trajectory information and the welding area image information to obtain integrated image information of the electric welding process.

[0048] The real-time monitoring module synchronizes the comprehensive parameter information and the comprehensive image information in time to monitor the welding process in real time.

[0049] Furthermore, the parameter information acquisition module includes:

[0050] The dynamic parameter acquisition unit uses a high-definition camera to photograph the digital display screen of the welding machine to acquire the dynamic working parameters of the welding process;

[0051] An output parameter acquisition unit is connected to the data cable of the welding machine to acquire the output parameters of the data cable;

[0052] The integrated information acquisition unit uses the dynamic operating parameters to correct the output parameters of the data cable and obtains the integrated parameter information.

[0053] The technical solution of this invention can achieve the following technical effects:

[0054] This method effectively solves the problems of existing methods, such as difficulty in real-time monitoring of welding results and limited monitoring information. By establishing a real-time monitoring model, real-time monitoring is achieved by utilizing the correlation between parameters and images.

[0055] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating a real-time monitoring method for the electric welding process;

[0058] Figure 2 A flowchart illustrating the process of obtaining comprehensive parameter information;

[0059] Figure 3 A flowchart for recording the welding path;

[0060] Figure 4 A flowchart illustrating the process of obtaining welding trajectory information;

[0061] Figure 5 A flowchart illustrating the process of obtaining image information of the welding area;

[0062] Figure 6 A flowchart illustrating the process of using a real-time monitoring model for real-time monitoring;

[0063] Figure 7 This is a schematic diagram of a real-time monitoring method for the electric welding process. Detailed Implementation

[0064] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0066] Example 1

[0067] like Figure 1 As shown, this application provides a real-time monitoring method for an electric welding process, the method comprising:

[0068] S100: Acquires dynamic working parameters of the welding process and data cable output parameters of the welding machine to obtain comprehensive parameter information of the welding process;

[0069] S200: Based on the camera calibration algorithm, it locks the tracking focus and records the welding path based on the target tracking algorithm;

[0070] Specifically, camera calibration algorithms can effectively remove distortions introduced by camera lenses, which can affect the accurate measurement and positioning of objects in an image. Camera calibration corrects these distortions, making the position of objects in the image more accurate. Simultaneously, camera calibration helps lock the tracking focus; generally, the welding arc light is selected as the focus, allowing for effective recording of the welding path using target tracking algorithms.

[0071] S300: Calculates the projection matrix and homography matrix of the tracking focus, and obtains welding trajectory information based on the welding path image;

[0072] Specifically, projection matrices and homography matrices are commonly used matrices in computer vision and computer graphics to handle geometric transformations and mapping relationships between images and three-dimensional space.

[0073] The projection matrix consists of an intrinsic parameter matrix and an extrinsic parameter matrix. The intrinsic parameter matrix includes the camera's internal parameters, such as focal length and principal point. The extrinsic parameter matrix includes the camera's position and orientation information. The purpose of the projection matrix is ​​to convert the coordinates of points in 3D space into pixel coordinates in the camera image, which is crucial for locating and tracking the position of objects in the image.

[0074] Homography matrices are commonly used for plane-to-plane perspective transformations. They represent the mapping relationship from one plane to another, such as mapping an image of an object on one plane to another. A homography matrix is ​​a 3x3 matrix. In computer vision, homography matrices are often used in applications such as object tracking.

[0075] S400: Acquires real-time dynamic images of the welding process and performs image processing on the real-time dynamic images to obtain image information of the welding area;

[0076] Specifically, when acquiring real-time dynamic images of the welding process, it is necessary to ensure the clarity of the images. A better solution is to integrate four or more high dynamic range lenses to effectively acquire real-time dynamic images, improve the clarity of the initial images, and reduce the difficulty of subsequent processing.

[0077] S500: The welding trajectory information and the welding area image information are correlated to obtain comprehensive image information of the electric welding process;

[0078] S600: Synchronizes integrated parameter information and integrated image information in time and establishes a real-time monitoring model to monitor the welding process in real time.

[0079] Specifically, by first associating welding trajectory information and welding area image information on the image, the welding path is displayed on the welding area image to obtain comprehensive image information of the welding process. Then, by synchronously comparing the comprehensive image information and the parameter change information of the welding process over time, the synchronous changes of images and parameters during the welding process can be clearly obtained.

[0080] The technical solution of this invention effectively solves the problems of difficulty in real-time monitoring of welding effect and limited monitoring information in existing methods. By establishing a real-time monitoring model, real-time monitoring is achieved by utilizing the correlation between parameters and images.

[0081] The key to this solution lies in obtaining comprehensive parameter data and clear real-time dynamic images, and establishing the relationship between the two. Therefore, comprehensive parameter data acquisition is crucial, and clear real-time dynamic images are essential for subsequent image processing. Based on this premise, a strong correlation between the two is established by analyzing and basing the temporal relationship of the welding process. This ensures comprehensive real-time monitoring of the entire welding process. Furthermore, for problematic parameter data or real-time dynamic images, corresponding analysis of problematic parameters or images at the same time point enhances the understanding of the source and nature of the problem.

[0082] As a preferred embodiment of the above, such as Figure 2 As shown, in step S100, the dynamic working parameters of the welding process and the data cable output parameters of the welding machine are obtained to acquire comprehensive parameter information of the welding process, including:

[0083] S110: Uses a high-definition camera to photograph the digital display screen of the welding machine to obtain dynamic working parameters of the welding process;

[0084] S120: Connects the data cable to the welding machine and obtains the data cable output parameters;

[0085] S130: Use dynamic operating parameters to correct the output parameters of the data cable and obtain comprehensive parameter information.

[0086] Specifically, by acquiring data from the welding machine's digital display screen, the Tesseract optical character recognition engine is used to analyze the images in real time and record the dynamic electrical parameters of the welding process. These parameters are then compared with the welding machine's output parameters. This is because it is difficult to obtain comprehensive parameter information solely by acquiring the output parameters of the data cable. Acquiring dynamic electrical parameters can effectively compensate for the incompleteness of data output from the data cable, such as the inability of the data cable to display the working curves of some welding machines.

[0087] As a preferred embodiment of the above, such as Figure 3As shown, step S200 involves locking the tracking focus based on the camera calibration algorithm and recording the welding path based on the target tracking algorithm, including:

[0088] S210: Acquires real-time dynamic images of the welding point and locks the position of the tracking focus based on the camera calibration algorithm;

[0089] S220: Uses a target tracking algorithm to locate the position of the tracking focus;

[0090] S230: Construct an image sequence of the welding path based on the positional changes of the tracking focus;

[0091] S240: Convert the image sequence into a welding path represented by 2D data point coordinates.

[0092] Specifically, the camera is used to capture real-time dynamic images of the welding area. A camera calibration algorithm determines the camera's internal and external parameters to map pixel coordinates in the image to real-world coordinates. This step ensures that the acquired image accurately represents the welding area. Camera calibration can be achieved by photographing a calibration target or marker with known geometry and then using the appropriate calibration algorithm. Generally, the welding arc light is preferred as the tracking focus because it is usually one of the most obvious features in the welding process. It is typically a high-brightness area in the image, easy to detect and track, and has high real-time performance, allowing for continuous monitoring during the welding process. The target tracking algorithm detects the focus position in the image and continuously updates it over time. Based on the changes in the focus position, an image sequence of the welding path is constructed. As the focus moves during the welding process, each frame records the new position of the focus. The focus position in each frame of the image sequence is converted into 2D data point coordinates to represent the welding path. These data point coordinates can be used for subsequent analysis.

[0093] As a preferred embodiment of the above, such as Figure 4 As shown, in step S300, the projection matrix and homography matrix of the tracking focus are calculated, and welding trajectory information is obtained according to the welding path, including:

[0094] S310: Using the camera's internal and external calibration parameters, known world coordinate points and their corresponding points in the welding path, calculate the projection matrix and homography matrix;

[0095] S320: Based on the projection matrix and homography matrix, map each 2D data point on the welding path to each world coordinate point;

[0096] S330: Welding trajectory information is calculated based on various world coordinate points.

[0097] The welding trajectory information includes, but is not limited to, the welding path, path length, and speed.

[0098] Specifically, using the camera calibration algorithm in S200 to obtain the camera's internal and external calibration parameters, known world coordinate points, and their corresponding points in the welding path, the DLT (Direct Linear Transform) method is used to calculate the projection matrix and homography matrix. Using the calculated projection and homography matrices, each 2D data point on the welding path can be mapped to its corresponding world coordinate point. This step is used to correlate the welding path information in the image with the actual world coordinate points. Once the 2D data points are mapped to world coordinate points, these coordinate points can be used to calculate the welding trajectory information. This information includes, but is not limited to, the shape of the welding path, the path length, and the speed of the welding torch during welding. For example, the path length can be obtained by calculating the distance between world coordinate points, while the speed can be calculated as the ratio of path length to time.

[0099] As a preferred embodiment of the above, such as Figure 5 As shown, in step S400, real-time dynamic images of the welding process are acquired, and image processing is performed on the dynamic images to obtain image information of the welding area, including:

[0100] S410: Uses a high dynamic range camera and an optical neutral density filter to capture real-time dynamic images of the welding process;

[0101] S420: Image enhancement for real-time dynamic images of the region;

[0102] S430: Perform image segmentation on the enhanced image to obtain image information of the welding area.

[0103] Specifically, a high dynamic range (HDR) camera is used. This type of camera can capture more brightness information under a wide range of lighting conditions to preserve details in the welding process. In practice, four or more HDR camera lenses can be installed to capture images from multiple angles simultaneously. When acquiring images, a neutral density (ND) filter is used to help reduce exposure in strong light conditions to avoid overexposure and ensure image quality. After obtaining a relatively clear real-time dynamic image, homomorphic filtering can be used to enhance the image. The core idea of ​​homomorphic filtering is to decompose the image into two components: a reflection component and an illumination component. The reflection component contains the structural information of the image, while the illumination component contains the lighting information of the image. The illumination component is usually a low-frequency signal, while the reflection component is a high-frequency signal. Homomorphic filtering improves the image contrast and reduces illumination unevenness by processing these two components separately. The enhanced image is then segmented, and image information of the welding area is extracted, while interference from other areas is eliminated.

[0104] As a preferred embodiment of the above, step S500 involves data association between the welding trajectory information and the welding area image information to obtain comprehensive image information of the welding process, including:

[0105] S510: Match the timestamp in the welding trajectory information with the corresponding welding area image frame;

[0106] S520: Integrate the matched welding trajectory information with the welding area image information to obtain comprehensive image information.

[0107] Specifically, each welding trajectory information point should have a corresponding timestamp, indicating the time of that information point during the welding process. These timestamps need to match the frame timestamps of the welding area image. This can be achieved by recording the timestamp of each frame of the welding area image and associating the welding trajectory information with the corresponding timestamps. Once the timestamps are successfully matched, the matched welding trajectory information can be integrated with the corresponding welding area image information to create a new data structure. Each timestamp corresponds to a comprehensive data point, including welding trajectory information and the corresponding welding area image. Therefore, the comprehensive data can be considered as welding process information containing the time dimension. This comprehensive image information can be used for subsequent real-time monitoring and analysis to enable real-time monitoring of the welding process.

[0108] As a preferred embodiment of the above, step S600 involves synchronizing the integrated parameter information and integrated image information over time, and establishing a real-time monitoring model to monitor the welding process in real time, including:

[0109] S610: Associate the integrated parameter information with the corresponding integrated image information according to the correspondence of timestamps, wherein the association makes the fluctuation of parameter information correspond to the change of integrated image information;

[0110] S620: Based on real-time changing parameter information and corresponding comprehensive image information, a real-time monitoring model is established to monitor the welding process in real time.

[0111] Specifically, similar to the matching of welding trajectory information and welding area image frames, the timestamps in the comprehensive parameter information are matched with the timestamps of the corresponding comprehensive image information to ensure that each data point of the parameter information corresponds to the corresponding image information frame. By associating the parameter information with the corresponding image frames, a data structure is created, in which each timestamp corresponds to a comprehensive data point, including parameter information and corresponding image information. A real-time monitoring model is then established, which can analyze the fluctuations of the comprehensive parameter information and the changes of the corresponding comprehensive image information, thereby realizing real-time monitoring of the welding process.

[0112] As a preferred embodiment of the above, such as Figure 6As shown, step S620 involves establishing a real-time monitoring model based on the real-time changing parameter information and the corresponding integrated image information to monitor the welding process in real time, including:

[0113] S621: Collect comprehensive parameter information and comprehensive image information of welding processes along different paths, and match them according to timestamps;

[0114] S622: Convert the integrated parameter information and integrated image information into sequence data;

[0115] S623: Based on a recurrent neural network model, a real-time monitoring model for the electric welding process is established, and sequence data is used for training and verification.

[0116] S624: For each welding process, the integrated parameter information and integrated image information converted into sequence data are used as inputs, and the real-time monitoring model monitors and issues early warnings.

[0117] Specifically, the matched integrated parameter information and integrated image information are transformed into sequence data, where each sequence represents a welding path or time period. A recurrent neural network (RNN) model is used as the real-time monitoring model because RNNs are designed to process temporally sequential data, thus effectively capturing the temporal changes in the welding path. This is crucial for analyzing the evolution of integrated parameter information and integrated image information during the welding process. Furthermore, RNNs can handle long-term dependencies, enabling them to capture the relationships between different time steps, which is essential for understanding historical information and trends in the welding process. RNNs can handle sequence data of varying lengths, making them suitable for welding path lengths that may change during welding. Because RNNs can learn patterns and features, they can adapt to different welding methods and routes. For each welding process, the integrated parameter information and integrated image information, transformed into sequence data, are used as input to the established RNN model. The model monitors the changes in parameters and image information in real time during the welding process, following the learned patterns to detect deviations from normal conditions and issuing real-time warnings as needed.

[0118] Example 2:

[0119] Based on the same inventive concept as the real-time monitoring method for an electric welding process described in the foregoing embodiments, this invention also provides a real-time monitoring system for an electric welding process, such as... Figure 7 As shown, the system includes:

[0120] The parameter information acquisition module acquires the dynamic working parameters of the electric welding process and the data cable output parameters of the welding machine to obtain comprehensive parameter information of the electric welding process.

[0121] The path image recording module locks the tracking focus based on the camera calibration algorithm and records the welding path based on the target tracking algorithm;

[0122] The trajectory data acquisition module calculates the projection matrix and homography matrix of the tracking focus, and obtains welding trajectory information based on the welding path;

[0123] The regional image acquisition module acquires dynamic images of the welding process and performs image processing on the dynamic images to obtain welding area image information.

[0124] The integrated image acquisition module correlates welding trajectory information and welding area image information to obtain integrated image information of the electric welding process;

[0125] The real-time monitoring module synchronizes comprehensive parameter information and comprehensive image information in time to monitor the welding process in real time.

[0126] The adjustment system described above in this invention can effectively realize the real-time monitoring method of the electric welding process, and the technical effects it can achieve are as described in the above embodiments, which will not be repeated here.

[0127] As a preferred embodiment of the above, the parameter information acquisition module includes:

[0128] The dynamic parameter acquisition unit uses a high-definition camera to capture images of the welding machine's digital display screen to obtain dynamic working parameters of the welding process.

[0129] The output parameter acquisition unit is connected to the welding machine's data cable and acquires the data cable's output parameters.

[0130] The integrated information acquisition unit uses dynamic operating parameters to correct the output parameters of the data cable and obtain integrated parameter information.

[0131] Similarly, the above-mentioned optimization schemes for the system can also achieve the optimization effects corresponding to the methods in Embodiment 1, which will not be repeated here.

[0132] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the spirit and scope of this application. Accordingly, this specification and drawings are merely exemplary illustrations of the application as defined herein, and are to be considered as covering any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A method for real-time monitoring of an electric welding process, characterized in that, The method includes: The dynamic working parameters of the electric welding process and the data cable output parameters of the welding machine are obtained to obtain comprehensive parameter information of the electric welding process; Based on the camera calibration algorithm, the tracking focus is locked, and the welding path is recorded based on the target tracking algorithm; Calculate the projection matrix and homography matrix of the tracking focus, and obtain the welding trajectory information based on the welding path image; Acquire real-time dynamic images of the welding process and perform image processing on the real-time dynamic images to obtain image information of the welding area; The welding trajectory information and the welding area image information are correlated to obtain comprehensive image information of the electric welding process; Synchronize the integrated parameter information and the integrated image information in time, and establish a real-time monitoring model to monitor the welding process in real time, including: The integrated parameter information is associated with the corresponding integrated image information according to the timestamp correspondence, wherein the association makes the fluctuation of the parameter information correspond to the change of the integrated image information. Based on the real-time changing parameter information and the corresponding comprehensive image information, the real-time monitoring model is established to monitor the welding process in real time, including: Collect the integrated parameter information and integrated image information of the welding process along different paths, and match them by timestamp; The integrated parameter information and the integrated image information are converted into sequence data; A real-time monitoring model for the electric welding process is established based on a recurrent neural network model, and the sequence data is used for training and verification. For each welding process, the integrated parameter information and integrated image information, which are converted into sequence data, are used as inputs for the real-time monitoring model to monitor and issue early warnings.

2. The real-time monitoring method for the electric welding process according to claim 1, characterized in that, By acquiring the dynamic operating parameters of the electric welding process and the data cable output parameters of the welding machine, comprehensive parameter information of the electric welding process can be obtained, including: The dynamic operating parameters of the welding process are obtained by using a high-definition camera to photograph the digital display screen of the welding machine. Connect the data cable to the welding machine and obtain the output parameters of the data cable; The dynamic operating parameters are used to correct the output parameters of the data cable, and the comprehensive parameter information is obtained.

3. The real-time monitoring method for the electric welding process according to claim 1, characterized in that, Based on the camera calibration algorithm, the tracking focus is locked, and based on the target tracking algorithm, the welding path is recorded, including: Real-time dynamic images of the welding point are acquired, and the position of the tracking focus is locked based on a camera calibration algorithm; The target tracking algorithm is used to locate the position of the tracking focus; Based on the positional changes of the tracking focus, an image sequence of the welding path is constructed; The image sequence is converted into a welding path represented by 2D data point coordinates.

4. The real-time monitoring method for the electric welding process according to claim 3, characterized in that, Calculate the projection matrix and homography matrix of the tracking focus, and obtain welding trajectory information based on the welding path, including: Using the camera's internal and external calibration parameters, the projection matrix and the homography matrix are calculated using known world coordinate points and their corresponding points in the welding path; Based on the projection matrix and the homography matrix, each 2D data point on the welding path is mapped to each of the world coordinate points; The welding trajectory information is calculated based on the aforementioned world coordinate points. The welding trajectory information includes the welding path, path length, and speed.

5. The real-time monitoring method for the electric welding process according to claim 1, characterized in that, Acquire real-time dynamic images of the welding process and perform image processing on the dynamic images to obtain welding area image information, including: The welding process was captured in real time using a high dynamic range camera and with the assistance of a neutral density filter. Image enhancement is performed on the real-time dynamic images; The enhanced image is segmented to obtain the image information of the welding area.

6. The real-time monitoring method for the electric welding process according to claim 1, characterized in that, By associating the welding trajectory information and the welding area image information, comprehensive image information of the welding process is obtained, including: Match the timestamps in the welding trajectory information with the corresponding welding area image frames; The matched welding trajectory information is integrated with the welding area image information to obtain the comprehensive image information.

7. A real-time monitoring system for an electric welding process, characterized in that, The system includes: The parameter information acquisition module acquires the dynamic working parameters of the electric welding process and the data cable output parameters of the welding machine to obtain comprehensive parameter information of the electric welding process. The path image recording module locks the tracking focus based on the camera calibration algorithm and records the welding path based on the target tracking algorithm; The trajectory data acquisition module calculates the projection matrix and homography matrix of the tracking focus, and obtains welding trajectory information based on the welding path image; The regional image acquisition module acquires real-time dynamic images of the welding process and performs image processing on the real-time dynamic images to obtain welding area image information. The integrated image acquisition module correlates the welding trajectory information and the welding area image information to obtain integrated image information of the electric welding process. The real-time monitoring module synchronizes the comprehensive parameter information and the comprehensive image information in time, and establishes a real-time monitoring model to monitor the welding process in real time, including: The integrated parameter information is associated with the corresponding integrated image information according to the timestamp correspondence, wherein the association makes the fluctuation of the parameter information correspond to the change of the integrated image information. Based on the real-time changing parameter information and the corresponding comprehensive image information, the real-time monitoring model is established to monitor the welding process in real time, including: Collect the integrated parameter information and integrated image information of the welding process along different paths, and match them by timestamp; The integrated parameter information and the integrated image information are converted into sequence data; A real-time monitoring model for the electric welding process is established based on a recurrent neural network model, and the sequence data is used for training and verification. For each welding process, the integrated parameter information and integrated image information, which are converted into sequence data, are used as inputs for the real-time monitoring model to monitor and issue early warnings.

8. The real-time monitoring system for the electric welding process according to claim 7, characterized in that, The parameter information acquisition module includes: The dynamic parameter acquisition unit uses a high-definition camera to photograph the digital display screen of the welding machine to acquire the dynamic working parameters of the welding process; An output parameter acquisition unit is connected to the data cable of the welding machine to acquire the output parameters of the data cable; The integrated information acquisition unit uses the dynamic operating parameters to correct the output parameters of the data cable and obtains the integrated parameter information.

Images