Micrograph visual laser AI weld seam tracking method and system

Abstract

The present application relates to a kind of micrograph visual laser AI weld seam tracking method and system, it is related to the field of weld seam tracking technology, it includes response to welding signal, obtains workpiece video;Determine welding initial route based on workpiece video;Obtain workpiece category to determine initial welding point and execute welding operation to obtain welding starting point;Determine required welding point based on welding starting point;Combining required welding point, welding operation and adjustment rule determine welding fixed point and collect welding connection point;Welding connection point is defined as required welding point and repeat step 5 to obtain welding point group and welding endpoint;Determine actual welding route based on welding starting point, welding point group and welding endpoint and analyze welding deviation value in combination with welding initial route;If welding deviation value does not fall into safe deviation range, obtain welding quantity;Combining welding deviation value and welding quantity analyzes deviation situation and finds corresponding solution output.The present application has the effect of improving the safety and reliability of welding.

Description

Technical Field

[0001] This invention relates to the field of weld seam tracking technology, and in particular to a micro-image vision laser AI weld seam tracking method and system. Background Technology

[0002] In industrial welding scenarios such as engineering machinery, automobile manufacturing, pressure vessels, and ship steel structures, the real-time performance, accuracy stability, and adaptive correction capability of weld seam tracking are the core keys to ensuring the quality of automated welding, improving the welding yield, and reducing the impact of workpiece assembly and thermal deformation.

[0003] Currently, laser vision weld seam tracking equipment and control systems have emerged in the industry. Their core relies on line laser emission combined with industrial camera acquisition structure, and basic image processing algorithms to achieve weld seam feature extraction and position recognition. Then, welding robots or special welding machines are linked to complete the simple trajectory following of the welding torch, replacing the traditional manual teaching fixed-point welding and mechanical limit guidance operation mode, laying the technical foundation for realizing the automation of welding process and reducing human intervention.

[0004] Regarding the aforementioned technologies, in the actual welding process, AI only analyzes the next welding point based on the current welding point to ensure that the welding points are continuous and stable. Therefore, there is still room for improvement in the safety and reliability of AI automatic welding. Summary of the Invention

[0005] To improve the safety and reliability of welding, this invention provides a micro-image vision laser AI weld seam tracking method and system.

[0006] In a first aspect, the present invention provides a micro-image vision laser AI weld seam tracking method, which adopts the following technical solution: A micro-image vision laser AI weld seam tracking method includes: Step 1: In response to a preset welding signal, acquire a video of the workpiece; Step 2: Determine the initial welding path based on the workpiece video; Step 3: Obtain the workpiece category to determine the initial welding point and perform the welding operation to obtain the welding start point; Step 4: Determine the required welding points based on the welding start point; Step 5: Determine the welding points and collect the welding connection points by combining the required welding points, welding operations and preset adjustment rules; Step 6: Define the weld joint as the required weld joint and repeat step 5 to obtain the weld joint group and the weld endpoint; Step 7: Determine the actual welding route based on the welding start point, welding point group, and welding end point, and analyze the welding deviation value in conjunction with the initial welding route; Step 8: If the welding deviation value does not fall within the preset safety deviation range, obtain the welding quantity; Step 9: Analyze the deviation by combining the welding deviation value and the number of welds, and find the corresponding solution output.

[0007] By adopting the above technical solution, the initial welding route is obtained through workpiece video, and then the initial welding point is obtained through workpiece category and welded as the starting point. The next welding point is determined, and this step is repeated in combination with adjustment rules to obtain the welding point group and welding endpoint. Finally, the actual welding route is formed together with the welding starting point and compared. Then, the deviation and corresponding solution are obtained by combining the number of welds. This avoids the situation where only adjusting parameters is focused on in the weld seam tracking process, while ignoring the reasons for adjusting the parameters, thus improving the reliability and safety of weld seam tracking during welding.

[0008] Optionally, methods for analyzing deviations and finding corresponding solutions by combining welding deviation values ​​with welding quantity include: Step 90: If the number of welds is equal to 1, obtain the micro-image acquisition device number; Step 91: Control the micro-image acquisition device corresponding to the micro-image acquisition device number to perform an anomaly detection operation to obtain the anomaly situation; Step 92: Define the abnormal situation as a deviation and find the corresponding solution output; Step 93: If the number of welds is greater than 1, the deviation is obtained by comparing multiple weld deviation values; Step 94: When the deviation is within the preset uniform condition, generate and output a device abnormality signal based on the deviation. Step 95: Based on the deviation, find the corresponding device repair plan and output it; Step 96: When the deviation is not uniform, generate and output a workpiece abnormality signal based on the deviation.

[0009] By adopting the above technical solution, the method of determining the analysis situation by the number of welds, and then finding the corresponding solution output for subsequent use, the cause of the anomaly can be determined based on the deviation of the weld tracking, reducing the situation where the weld tracking fails to detect anomalies, resulting in non-compliance of equipment or welded workpieces, and improving the reliability and safety of weld tracking.

[0010] Optionally, methods for generating and outputting workpiece abnormality signals based on deviation conditions include: Step 960: Obtain welding workpiece parameters; Step 961: Find the corresponding standard welding parameters based on the workpiece category; Step 962: When the standard welding parameters are inconsistent with the welding workpiece parameters, generate and output a workpiece abnormality signal based on the deviation. Step 963: When the standard welding parameters are consistent with the welding workpiece parameters, no operation is performed.

[0011] By adopting the above technical solution, the welding workpiece parameters are compared with the standard welding parameters. When they are inconsistent, an abnormal workpiece signal is generated and output. When they are consistent, it indicates that the workpiece is normal. This avoids misjudgment caused by analyzing the deviation of the route alone, thus improving the reliability and accuracy of abnormal workpiece identification.

[0012] Optionally, when the standard welding parameters are inconsistent with the welding workpiece parameters, the method for generating and outputting a workpiece abnormality signal based on the deviation includes: Step 9620: Obtain the abnormal workpiece number based on the welding workpiece parameters; Step 9621: Count the number of abnormal workpieces corresponding to the abnormal workpiece numbers in the welding quantity; Step 9622: Determine the critical value of the abnormality by combining the number of welds with the preset abnormality rules; Step 9623: When the number of abnormal workpieces reaches the abnormal threshold, a workpiece abnormality signal is generated and output based on the deviation. Step 9624: When the number of abnormal workpieces does not reach the abnormal threshold, an abnormal workpiece group is formed and output based on the abnormal workpiece number and deviation.

[0013] By adopting the above technical solution, the abnormal workpiece number is obtained through the welding workpiece parameters and the number of abnormal workpieces is counted. Then, the abnormal threshold value is determined by combining the abnormality rules. Whether to form a corresponding workpiece abnormality signal or an abnormal workpiece group depends on whether the threshold value is reached. This reduces the possibility of rejecting all workpieces to be welded when there is a workpiece abnormality and improves the reliability of workpiece abnormality determination.

[0014] Alternatively, other methods for finding corresponding solutions include: Step 97: Obtain images of the welded workpiece and the standard workpiece; Step 98: Extract the corresponding weld seam information from the welded workpiece image; Step 99: Determine standard weld information using standard workpiece images; Step 100: Compare the workpiece weld information with the standard weld information to determine the weld condition; Step 101: Determine and output the corresponding solution based on the weld condition.

[0015] By adopting the above technical solution, the weld seam information of the workpiece and the standard weld seam information are obtained by using images of the welded workpiece and images of the standard workpiece, respectively. Then, the weld seam information is compared to determine the corresponding solution. This allows for a more accurate analysis of the actual situation of the weld seam and improves the reliability of weld seam tracking.

[0016] Optional, also includes: Step 910: Extract the weld width from the workpiece weld information; Step 911: Extract the standard weld width from the standard weld information; Step 912: When the weld width is greater than the standard weld width, define the weld condition as oxidation and output it. Step 913: When the weld width is not greater than the standard weld width, define the weld condition as normal and output it.

[0017] By adopting the above technical solution, the weld width is compared with the standard weld width to obtain the oxidation status or normal status, making the analysis of the weld condition more detailed and improving the safety and reliability of weld tracking and identification.

[0018] Optionally, methods for defining and outputting weld condition as oxidation condition include: Step 914: Extract the grayscale value and color value of the weld from the image of the welded workpiece and form a weld color value group; Step 915: Extract the grayscale standard values ​​and color standard values ​​from the standard workpiece image and form a standard color value group; Step 916: Compare the weld color value group with the standard color value group to obtain the color group difference; Step 917: If the color group difference is greater than the preset allowable threshold, define the weld condition as oxidation and output it; Step 918: If the color group difference is not greater than the allowable threshold, define the weld condition as normal and output it.

[0019] By adopting the above technical solution, a weld color value group is obtained by using the weld gray value and weld color value. Then, a standard color value group is obtained by using the gray standard value and color standard value. The two color value groups are compared to obtain the color group difference. The oxidation status or normal status is further determined based on the color group difference. This reduces the possibility of misjudgment caused by relying solely on the width to determine whether there is an anomaly, and improves the reliability of weld tracking and anomaly identification.

[0020] Optional, also includes: Step 102: Extract the corresponding actual weld parameters based on the workpiece weld information; Step 103: Extract the corresponding standard weld parameters from the standard weld information; Step 104: Calculate the difference in weld parameters by combining the actual weld parameters with the standard weld parameters; Step 105: Calculate the air value of the weld based on the weld parameter difference; Step 106: Combine the air value of the weld, the abnormal workpiece number and the deviation to form abnormal workpiece information and output it.

[0021] By adopting the above technical solution, the difference between weld parameters is calculated by comparing the actual weld parameters with the standard weld parameters. Then, the weld air value between the welded workpieces is estimated based on the difference in weld parameters. Finally, abnormal workpiece information is generated and output by combining the abnormal workpiece number and deviation. This allows the staff to have a further understanding of the abnormal workpieces while outputting data, thus improving the flexibility of weld tracking.

[0022] Optionally, methods for determining the initial welding path based on workpiece video include: Step 20: Upon receiving a preset normal signal, obtain the deviation area corresponding to the welding deviation value; Step 21: Integrate the deviation area with the welding deviation value to form the modified deviation information; Step 22: Obtain the welding modification route based on the modified deviation information and workpiece video; Step 23: Define the modified welding path as the initial welding path and output it.

[0023] By adopting the above technical solution, the deviation area is determined by the welding deviation value, and the two are combined to form the modified deviation information. Then, the welding modification route is obtained by combining the workpiece video, and finally it is defined as the initial welding route output. This avoids the possibility of considering it abnormal when the equipment or workpiece is normal, and improves the reliability and flexibility of abnormal identification in the welding tracking process.

[0024] Secondly, this invention provides a micro-image vision laser AI weld seam tracking system, which adopts the following technical solution: A micro-image vision laser AI weld seam tracking system includes: The acquisition module is used to acquire workpiece videos, workpiece categories, welding quantities, micro-image acquisition device numbers, welding workpiece parameters, abnormal workpiece numbers, welding workpiece images, standard workpiece images, and deviation areas. The memory is used to store the program of the micro-image vision laser AI weld seam tracking method as described above; The processor loads and executes programs from memory.

[0025] By adopting the above technical solution, basic data such as workpiece video are collected uniformly by the acquisition module, and the complete tracking method program is stored in the memory. Then, the processor loads and runs the program to complete the entire process of automation. This avoids the drawback of traditional weld seam tracking, which only adjusts parameters and ignores the root cause of deviation, and improves the automation, reliability and safety of laser weld seam tracking.

[0026] In summary, the present invention has at least one of the following beneficial technical effects: The initial welding route is obtained by taking the workpiece video as the starting point, and then the initial welding point is obtained by taking the workpiece category as the starting point. The next welding point is determined and this step is repeated in combination with the adjustment rules to obtain the welding point group and the welding endpoint. Finally, the actual welding route is formed together with the welding starting point and compared. Then, the deviation and corresponding solutions are obtained by combining the number of welds. This avoids the situation where only the adjustment parameters are focused on during the weld tracking process, while ignoring the reasons for the adjustment parameters. This improves the reliability and safety of weld tracking during welding. By comparing weld width, weld grayscale value, and weld color value, and combining them with the standard weld width, standard grayscale value, and standard color value, a dual-limit analysis is performed to determine whether the weld is in an oxidized state, i.e. whether the shielding gas is insufficient or leaking. This avoids misjudging the weld condition due to relying on only one condition analysis and improves the reliability of weld tracking and corresponding weld condition judgment. By calculating the difference between actual weld parameters and standard weld parameters, and then further estimating the weld air value, staff can obtain more anomaly-related data, improving the reliability and flexibility of weld tracking. Attached Figure Description

[0027] Figure 1 This is a flowchart of a micro-image vision laser AI weld seam tracking method according to an embodiment of this application; Figure 2 This is a schematic diagram illustrating the scenario where the required welding points are determined according to an embodiment of this application. Detailed Implementation

[0028] The present invention will now be described in further detail with reference to the accompanying drawings and embodiments.

[0029] This invention discloses a micro-image vision laser AI weld seam tracking method. (Refer to...) Figure 1 A micro-image vision laser AI weld seam tracking method includes: Step 1: In response to the preset welding signal, acquire the workpiece video.

[0030] A welding signal is a start signal indicating that welding is about to begin on the workpiece. The welding signal is activated by an electrical button on the actual welding equipment; pressing this button triggers the response.

[0031] Workpiece video refers to video footage taken around the area of ​​the workpiece that needs to be welded. This workpiece video is acquired by a camera installed on the welding system, which captures continuous frames around the workpiece at the welding point.

[0032] Step 2: Determine the initial welding path based on the workpiece video.

[0033] The initial welding path refers to the ideal weld path for welding a workpiece. Here, the initial welding path is determined by the system using AI image algorithms to perform target recognition and weld feature extraction on the workpiece video, locating all points to be welded on the workpiece. Then, path fitting smoothly connects these points to generate a standard baseline-form initial welding path. For example, first, a deep learning-based instance segmentation model is used to segment each frame of the workpiece video at the pixel level, identifying the Region of Interest (ROI) of the weld area. Then, within the ROI, the Canny edge detection algorithm is used to extract the weld edge contour, and a skeletonization algorithm is used to extract the weld centerline. Next, the three-dimensional coordinates of the points to be welded are obtained by sampling at equal intervals along the centerline. Finally, a cubic B-spline curve is used to smoothly fit these discrete points, generating a continuous initial welding path.

[0034] Step 3: Obtain the workpiece category to determine the initial welding point and perform the welding operation to obtain the welding start point.

[0035] The workpiece category refers to the type of workpiece to be welded. The workpiece category is obtained by having skilled personnel input all workpiece types and related information into the system. The system then extracts corresponding workpiece features from the workpiece video to retrieve and match the corresponding workpiece category. The initial welding point refers to the first welding position on the workpiece. The initial welding point is determined by the system using the workpiece video to identify the welding area corresponding to the welding equipment, and then determining the welding position at the center of that area based on the workpiece category. The welding operation refers to the set of operations performed to weld the workpieces together to achieve a tight connection. The welding operation is executed by the system controlling the welding equipment to align with the initial welding point, and then driving the welding torch or laser welding equipment to perform the fusion operation on the initial welding point. The welding start point refers to the first position after welding is completed. The welding start point is obtained by the system defining this point as the welding start point after completing the welding operation on the initial welding point.

[0036] Step 4: Determine the required welding points based on the welding start point.

[0037] The required welding point refers to the next location where welding operations are needed. Here, the required welding point is determined by the system extracting weld features from the image, then determining the welding path, and then controlling the corresponding rotating device of the welding apparatus to rotate. According to the welding intervals set by those skilled in the art, the next location to be welded is obtained within the welding path. For details, please refer to... Figure 2 After rotating the workpiece, select the next point that meets the welding interval at a fixed interval and define it as the required welding point. For example, the required welding point needs to be 15mm away from the welding start point.

[0038] Step 5: Determine the welding points and collect the welding connection points by combining the required welding points, welding operations and preset adjustment rules.

[0039] The adjustment rules refer to the rules for fine-tuning the required welding points. Specifically, it is a real-time correction algorithm based on visual feedback. For example, during the movement of the welding torch to the required welding point, the control system calls a laser vision sensor to synchronously acquire real-time contour images of the weld area. An image processing algorithm extracts the coordinates of the true center point of the weld in the current frame and compares them with the required welding point to calculate the real-time deviation. This deviation is then used as a compensation amount and added to the next interpolation cycle command of the robot motion controller, i.e., the actual welding point. This adjustment process is executed cyclically at a frequency of no less than 50Hz to ensure that the welding torch tip always tracks the real-time corrected position. The adjustment rules are obtained by those skilled in the art through relevant experiments and their own experience, and are input into the system. The adjustment method is common knowledge in the field.

[0040] The welding point refers to the point where welding is required but has already been done. The welding point is determined by the system controlling the welding equipment. After fine-tuning the specific welding position according to adjustment rules, the system performs the welding operation on the point, obtaining the completed welding point and defining it as the welding point. The welding connection point refers to the next position to be welded, connected to the welding point. The method for acquiring the welding connection point is the same as that described in step 4. The system rotates the workpiece based on the welding requirement point and, according to the welding interval and the required welding path, obtains the next position to be welded, defining it as the welding connection point. The welding connection point is essentially the same as the welding point, the difference being that the welding point is found based on the welding start point, while the welding connection point is found based on the welding point.

[0041] Step 6: Define the weld joint as the required weld joint and repeat step 5 to obtain the weld joint group and the weld endpoint.

[0042] A welding point group refers to the collection of welding points formed by combining all welding points. Here, the welding point group is obtained by the system integrating all welding points to obtain a group, which is then defined as the welding point group. The welding endpoint refers to the last position to be welded. Here, the welding endpoint is obtained by the system repeatedly rotating the workpiece and welding according to requirements. When the next required welding point is a point that has already been welded, the current welding point is defined as the welding endpoint.

[0043] Step 7: Determine the actual welding route based on the welding start point, welding point group, and welding end point, and analyze the welding deviation value in conjunction with the initial welding route.

[0044] The actual welding route refers to the actual welding path obtained by the welding equipment performing welding operations on the workpiece. The actual welding route is determined by the system retrieving the welding start point, all welding points in the welding point group, and the welding end point, and performing smooth path fitting and trajectory concatenation according to the welding sequence to generate a complete and coherent actual welding route.

[0045] Welding deviation refers to the difference in position and trajectory coordinates between the initial welding path and the actual welding path. The analysis method for this welding deviation involves the system extracting the characteristic coordinate points of the initial and actual welding paths, calculating the lateral and longitudinal coordinate offsets point by point, and then integrating the offset data from each point to obtain the welding deviation value.

[0046] Step 8: If the welding deviation value does not fall within the preset safety deviation range, obtain the welding quantity.

[0047] The safety deviation range refers to the normal error threshold interval within which the actual welding path deviates from the initial welding path. This safety deviation range is obtained by those skilled in the art based on welding process accuracy requirements and workpiece assembly tolerance experimental data, setting the maximum permissible error interval for path deviation, and inputting it into the system to obtain the safety deviation range.

[0048] The welding quantity refers to the total number of workpieces that have been welded. This welding quantity is obtained by the system counting the number of workpieces that have completed full-trajectory welding during the current continuous welding process in real time.

[0049] If the welding deviation value does not fall within the safe deviation range, it indicates that the actual workpiece weld is too different from the initial welding path, which may be due to an abnormality in the workpiece or welding equipment. Therefore, the welding quantity is obtained.

[0050] Step 9: Analyze the deviation by combining the welding deviation value and the number of welds, and find the corresponding solution output.

[0051] Deviation refers to the comprehensive state of trajectory offset magnitude, offset location, and deviation distribution characteristics. The analysis method for deviation here involves the system integrating welding deviation values ​​exceeding the safe deviation range with the offset area, and then combining this with the number of welds to form the deviation situation. For example: a deviation situation is that the overall weld trajectory offset of the workpiece corresponding to the number of welds exceeds the standard by 0.5mm, belonging to multiple workpieces with unified unidirectional offset; a deviation situation is that the local weld offset of a single workpiece or a few workpieces exceeds the standard by 0.6mm, occurring only in the weld starting section. Solutions refer to the equipment adjustment, parameter correction, or workpiece rectification plans for different deviation situations. The solution search method here involves different solutions corresponding to different deviation situations. Experts in the field conduct experiments under various deviation conditions to obtain the corresponding solutions, which are then linked to the deviation situation and input into the system. When the system receives a deviation situation, it automatically retrieves the corresponding solution. For example, if the deviation situation is that the uniform trajectory offset of multiple workpieces exceeds the standard by 0.5mm, it is determined that the fixing screws of the laser vision acquisition bracket are loose, and the corresponding solution is to tighten and calibrate the bracket fixing screws. If the deviation situation is that the local offset of the welding starting point of a single workpiece exceeds the standard by 0.6mm, it is determined that the workpiece clamping and positioning is off, and the corresponding solution is to readjust the position of the workpiece tooling fixture and reset the clamping. The solution is output by the system either to the device's existing anomaly resolution module or to the control terminal controlled by the operator for display.

[0052] The methods for analyzing deviations and finding corresponding solutions by combining welding deviation values ​​and welding quantity include: Step 90: If the number of welds is equal to 1, obtain the micro-image acquisition device number.

[0053] The micro-image acquisition device number refers to the unique identifier of the micro-image acquisition device. The micro-image acquisition device number is obtained by assigning a unique identifier to each device, which is set and entered into the system by those skilled in the art. When the system receives a welding signal, it automatically retrieves the welding device that triggered the signal and then obtains its corresponding micro-image acquisition device number.

[0054] If the number of welds is equal to 1, it means that the welding has been completed and there are no other workpieces to be welded at the moment. Therefore, the micro-image acquisition device number is obtained.

[0055] Step 91: Control the micro-image acquisition device corresponding to the micro-image acquisition device number to perform an anomaly detection operation to obtain an anomaly.

[0056] Anomaly detection refers to the process of performing a full-area scan of the entire welding apparatus and then analyzing the equipment's operating condition and installation misalignment. This anomaly detection is performed by the system calling the corresponding acquisition device based on its micro-image acquisition device number, controlling it to perform a comprehensive scan of the entire welding apparatus, including the machine body, welding torch station, laser emission components, and workpiece clamping position. Anomalies refer to abnormal conditions identified after the micro-image acquisition device's full-area scan, such as welding apparatus positional misalignment, installation misalignment, lens obstruction, and laser focusing deviation. These anomalies are obtained by comparing the scanned image of the entire machine with a standard equipment reference image input by the operator, identifying and integrating issues such as structural deviations and positional misalignments.

[0057] Step 92: Define the abnormal situation as a deviation and find the corresponding solution output.

[0058] After receiving an abnormal situation, the system treats it as a deviation and then searches for a corresponding solution. This allows the system to accurately determine the device's abnormality, identify the correct solution, and output it.

[0059] Step 93: If the number of welds is greater than 1, the deviation is obtained by comparing multiple weld deviation values.

[0060] The deviation is obtained by the system retrieving the welding deviation values ​​of multiple workpieces, and then performing horizontal pairwise comparisons and trajectory offset trend comparisons to obtain the corresponding deviation.

[0061] If the number of welds is greater than 1, it means that welding needs to continue. Since the abnormal situation cannot be obtained by direct scanning and imaging, the deviation situation is obtained by comparing multiple welding deviation values.

[0062] Step 94: When the deviation is within the preset uniform condition, generate and output a device abnormality signal based on the deviation.

[0063] A "uniform situation" refers to a situation where the deviation occurs at the same point. This uniform situation is obtained by having those skilled in the art define it as a situation where the deviation occurs at the same location through experimentation and inputting it into the system. Specifically, the system can be used to sequentially rotate and weld welding workpieces 1, 2, and 3. If workpiece 1, after determining its welding start point, needs to be adjusted 1mm to the right on the 10th welding attempt, and workpieces 2 and 3 also need to be adjusted 1mm to the right on the 10th welding attempt, then this is considered a uniform situation. Conversely, if the deviations are not at the same "point," it is not considered a uniform situation. A device abnormality signal refers to a signal indicating a fixed fault or parameter deviation in the welding or micro-image acquisition device. This device abnormality signal is generated when the system determines that the deviation meets the uniform situation and then generates a warning command indicating an equipment malfunction. The device abnormality signal is output by the system through a display screen, an audio-visual module, or a backend communication interface.

[0064] When the deviation is uniform, it means that the location with the excessive deviation value is in the same place. Therefore, an abnormal signal is generated and output based on the deviation.

[0065] Step 95: Find and output the corresponding device repair plan based on the deviation.

[0066] Equipment maintenance plans refer to the handling procedures for equipment repair, calibration, and debugging. The method for finding equipment maintenance plans here is as follows: each deviation caused by equipment malfunction corresponds to a specific equipment maintenance plan, which is obtained by personnel in the field through experiments and input into the system. When the system receives a deviation report, it automatically retrieves and matches the corresponding equipment maintenance plan. The output method for the equipment maintenance plan is that the system displays the matched plan in text on the equipment display screen and the operating terminal interface, and can also simultaneously push it to the backend management terminal for output completion.

[0067] Step 96: When the deviation is not uniform, generate and output a workpiece abnormality signal based on the deviation.

[0068] A workpiece anomaly signal refers to a signal indicating a problem with the specifications, shape, or installation posture of the workpiece to be welded. This workpiece anomaly signal is generated automatically by the system when it determines that the deviation is not uniform, resulting in an alert command indicating individual workpiece differences, clamping misalignment, or out-of-tolerance dimensions. The workpiece anomaly signal is output directly from the system via the equipment operation terminal, audible and visual prompt module, or backend system.

[0069] When the deviation is not uniform, it means that the deviation position of each workpiece is inconsistent, which may be due to a problem in the production and manufacturing of the workpiece. Therefore, an abnormal workpiece signal is generated and output based on the deviation.

[0070] The methods for generating and outputting workpiece abnormality signals based on deviation conditions include: Step 960: Obtain the parameters of the workpiece to be welded.

[0071] Welding workpiece parameters refer to parameters such as the overall dimensions, weld morphology, deformation deviation, and assembly fit of the workpiece after welding. These parameters are obtained by the system taking images of the welded workpiece using a micro-image acquisition device, performing visual recognition and feature extraction on the images, and then analyzing the images to obtain the various parameters of the welded workpiece.

[0072] Step 961: Find the corresponding standard welding parameters based on the workpiece category.

[0073] Standard welding parameters refer to the ideal standard dimensions, weld morphology, allowable deformation, and appearance tolerances of the workpiece after welding. The method for finding standard welding parameters here is that each workpiece corresponds to specific standard parameters after welding. Personnel skilled in the art input the relevant parameters or parameter ranges of the welded workpiece into the system through experiments. When the system obtains the welding workpiece parameters, it automatically retrieves the corresponding standard welding parameters based on the workpiece category.

[0074] Step 962: When the standard welding parameters are inconsistent with the welding workpiece parameters, a workpiece abnormality signal is generated and output based on the deviation.

[0075] When the standard welding parameters are inconsistent with the welding workpiece parameters, it indicates that there is an abnormality in the workpiece corresponding to the welding workpiece parameters. Therefore, a workpiece abnormality signal is generated and output based on the deviation.

[0076] Step 963: When the standard welding parameters are consistent with the welding workpiece parameters, no operation is performed.

[0077] When the standard welding parameters are consistent with the welding workpiece parameters, it means that the workpiece corresponding to the welding workpiece parameters meets the standard and there is no abnormality, so no operation is performed.

[0078] Among them, the methods for generating and outputting workpiece abnormality signals based on deviations when standard welding parameters are inconsistent with the parameters of the welded workpiece include: Step 9620: Obtain the abnormal workpiece number based on the welding workpiece parameters.

[0079] The abnormal workpiece number refers to the unique number of a workpiece that fails the welding test. The abnormal workpiece number is obtained by assigning a unique code to each workpiece during welding, binding it to its relevant information, and storing it. The system then compares the identified welding workpiece parameters with the corresponding standard welding parameters item by item, filtering out unqualified workpieces with parameters exceeding tolerances, automatically retrieving their numbers, and defining them as abnormal workpiece numbers.

[0080] Step 9621: counting the number of abnormal workpieces corresponding to abnormal workpiece numbers in the welding quantity.

[0081] The number of abnormal workpieces refers to the total number of workpieces with unqualified parameters and unqualified welding. The counting method for the number of abnormal workpieces herein is that the system searches all abnormal workpiece numbers, and counts and summarizes the unqualified workpieces corresponding to the numbers one by one to obtain the number of abnormal workpieces.

[0082] Step 9622: determining an abnormality critical value in combination with the welding quantity and a preset abnormality rule.

[0083] The abnormality rule refers to a conditional rule for determining whether workpieces of the same batch have problems. The obtaining method of the abnormality rule herein is that the abnormality rule is set and input into the system by a person skilled in the art according to actual requirements, for example: it is set that when the proportion of abnormal workpieces of the same batch exceeds 35%, it is determined as abnormal. The abnormality critical value refers to a critical value for determining whether the whole batch of workpieces is qualified. The determining method of the abnormality critical value herein is that the system obtains the abnormality critical value by multiplying the proportion threshold in the abnormality rule by the total welding quantity of the current batch.

[0084] Step 9623: when the number of abnormal workpieces reaches the abnormality critical value, generating and outputting a workpiece abnormality signal based on the deviation condition.

[0085] When the number of abnormal workpieces reaches the abnormality critical value, it indicates that most of the welded workpieces have abnormalities, and therefore the workpiece abnormality signal is generated and output based on the deviation condition.

[0086] Step 9624: when the number of abnormal workpieces does not reach the abnormality critical value, generating and outputting an abnormal workpiece group based on the abnormal workpiece numbers and the deviation condition.

[0087] The abnormal workpiece group refers to a set formed by combining abnormal workpiece numbers. The forming method of the abnormal workpiece group herein is that the system summarizes all abnormal workpiece numbers, associates the welding deviation conditions corresponding to each number, and obtains the abnormal workpiece group after collection and arrangement. The output method of the abnormal workpiece group herein is that the system displays the arranged abnormal workpiece group in a list on a control terminal interface.

[0088] When the number of abnormal workpieces does not reach the abnormality critical value, it indicates that only a few welded workpieces are abnormal, and problems may have occurred during the welding process, and therefore the abnormal workpiece group is generated and output based on the abnormal workpiece numbers and the deviation condition.

[0089] Wherein, the method for finding a corresponding solution further comprises: Step 97: acquiring a welded workpiece image and a standard workpiece image.

[0090] A welded workpiece image refers to an image captured by a micro-image acquisition device after welding. The method for acquiring this image is as follows: the system invokes the micro-image acquisition device to take a fixed-point photograph of the currently welded workpiece, acquiring the image in real time. A standard workpiece image refers to an ideal, standard image of the welded workpiece. The method for acquiring this image is as follows: each workpiece corresponds to a different standard workpiece image. Personnel skilled in the art input qualified images of different welded workpieces into the system. When the system receives a welded workpiece image, it automatically retrieves the corresponding standard workpiece image based on the workpiece category.

[0091] Step 98: Extract the corresponding weld seam information from the welded workpiece image.

[0092] Workpiece weld information refers to weld information such as weld location, weld width, and weld length. The method for extracting this workpiece weld information involves the system performing visual image processing, edge detection, and feature segmentation on the welded workpiece image. This automatically identifies the weld area and calculates various morphological parameters, while also identifying surface defects, to obtain the workpiece weld information.

[0093] Step 99: Determine standard weld information using standard workpiece images.

[0094] Standard weld information refers to information such as the ideal standard weld position, standard weld width, and standard weld length on the workpiece. This standard weld information is determined by the system performing visual analysis and feature calibration on the standard workpiece image, extracting the weld morphology parameters from the standard template.

[0095] Step 100: Compare the workpiece weld information with the standard weld information to determine the weld condition.

[0096] Weld condition refers to the quality and deviation of the weld, such as misalignment, excessive width, incorrect length, and surface defects. The method for determining weld condition here is as follows: the system extracts parameters such as the actual weld location, width, length, and surface defects, compares each item with standard weld information, and performs threshold judgments to summarize whether the actual weld is compliant and the specific type of deviation to determine the weld condition.

[0097] Step 101: Determine and output the corresponding solution based on the weld condition.

[0098] The solution here is determined by the fact that different weld conditions correspond to different solutions. Similarly, those skilled in the art obtain solutions for different weld conditions through experiments and then input them into the system. When the system receives a weld condition, it automatically retrieves and matches the corresponding solution. The output method of the solution here is consistent with that described in step 9.

[0099] This also includes: Step 910: Extract the weld width from the workpiece weld information.

[0100] Weld width refers to the width of the weld perpendicular to the welding path after the workpiece is welded. The weld width is extracted directly from the workpiece weld information by the system.

[0101] Step 911: Extract the standard weld width from the standard weld information.

[0102] The standard weld width refers to the standard weld width dimension perpendicular to the welding path. The standard weld width is extracted here by the system directly extracting the standard width value perpendicular to the welding path from the standard weld information.

[0103] Step 912: When the weld width is greater than the standard weld width, define the weld condition as oxidation and output it.

[0104] Oxidation condition refers to an abnormal weld state where the actual weld width is too wide, possibly indicating insufficient gas or shielding gas. The output method for oxidation condition here is that the system defines the weld condition as oxidation and then directly outputs it to the control terminal for display.

[0105] When the weld width is greater than the standard weld width, it indicates that there may be gas leakage or insufficient shielding gas during the welding process. Therefore, the weld condition is defined as oxidation and output.

[0106] Step 913: When the weld width is not greater than the standard weld width, define the weld condition as normal and output it.

[0107] "Normal condition" refers to a weld condition where the actual weld width does not exceed the standard weld width, there is no gas infiltration during the welding process, and the shielding gas condition meets the standards. The normal condition is output by the system defining the weld condition as normal and then directly outputting it to the control terminal for display.

[0108] When the weld width is not greater than the standard weld width, it indicates that the weld of the workpiece is normal. Therefore, the weld condition is defined as normal and output.

[0109] The methods for defining and outputting the weld condition as an oxidation condition include: Step 914: Extract the grayscale value and color value of the weld from the image of the welded workpiece and form a weld color value group.

[0110] The weld grayscale value refers to the brightness value of the pixel corresponding to the weld in the weld image. Here, the weld grayscale value is extracted by locating the weld area in the image and collecting the brightness information of that area to obtain the corresponding grayscale value. The weld color value refers to the color value of the pixel corresponding to the weld in the weld image. Here, the weld color value is extracted by locking the weld area in the image and collecting the RGB color channel information of the pixels in that area to obtain the corresponding color value. The weld color value group refers to the parameter set composed of the weld grayscale value and the weld color value. Here, the weld color value group is formed by the system classifying and integrating the collected weld grayscale value and weld color value.

[0111] Step 915: Extract the grayscale standard values ​​and color standard values ​​from the standard workpiece image and form a standard color value group.

[0112] Grayscale standard values ​​refer to the brightness values ​​of pixels corresponding to qualified welds in a standard workpiece image. Here, grayscale standard values ​​are extracted by locating the standard weld area in the standard workpiece image and collecting the brightness information of that area. Color standard values ​​refer to the color values ​​of pixels corresponding to qualified welds in a standard workpiece image. Here, color standard values ​​are extracted by locking the standard weld area in the standard workpiece image and collecting the RGB color channel information of the pixels in that area. Standard color value groups refer to a set of reference parameters composed of grayscale standard values ​​and color standard values. Here, standard color value groups are formed by the system classifying and integrating the extracted grayscale and color standard values.

[0113] Step 916: Compare the weld color value group with the standard color value group to obtain the color group difference.

[0114] Color group difference refers to the set of numerical differences obtained by comparing the weld grayscale value and weld color value with the grayscale standard value and color standard value, respectively. The color group difference is obtained by the system performing a difference calculation on each item of the weld color value group and the standard color value group, and then integrating all the grayscale and color deviation data to obtain the color group difference.

[0115] Step 917: If the color group difference is greater than the preset allowable threshold, define the weld condition as oxidation condition and output it.

[0116] The allowable threshold refers to the reference value for determining whether the grayscale and color deviation of the weld meets the standard. The allowable threshold is obtained by those skilled in the art by subtracting the ideal standard color value from the extreme color value of the actual qualified weld obtained through experiments. The absolute value of this value is defined as the allowable threshold and input into the system. For example, the grayscale allowable threshold is 15 and the color allowable threshold is 10.

[0117] If the color group difference is greater than the allowable threshold, it means that the color of the weld is not a normal color, which may indicate that there is a gas leak or insufficient protective gas. Therefore, the weld condition is defined as an oxidation condition and output.

[0118] Step 918: If the color group difference is not greater than the allowable threshold, define the weld condition as normal and output it.

[0119] If the color group difference is not greater than the allowable threshold, it means that the weld color meets the standard. If the weld is too wide, it may be due to other reasons. Therefore, the weld condition is defined as normal and output.

[0120] This also includes: Step 102: Extract the corresponding actual weld parameters based on the workpiece weld information.

[0121] Actual weld parameters refer to the set of parameters such as actual weld width, weld length, and weld position. The actual weld parameters are extracted by the system breaking down and analyzing the workpiece weld information, separating and reading weld feature data from each dimension, and then integrating them to obtain the actual weld parameters.

[0122] Step 103: Extract the corresponding standard weld parameters from the standard weld information.

[0123] Standard weld parameters refer to the set of standard parameters such as ideal and qualified weld width and weld length. The extraction method of standard weld parameters here is that the system decomposes and analyzes the standard weld information, separates and reads the weld feature data of each dimension, and then integrates them to obtain the standard weld parameters.

[0124] Step 104: Calculate the difference in weld parameters by combining the actual weld parameters with the standard weld parameters.

[0125] The weld parameter difference refers to the set of numerical deviations of each parameter after comparing the actual weld parameters with the standard weld parameters item by item. The weld parameter difference is calculated by the system calculating the difference between the actual weld parameters and the standard weld parameters one by one, and then summarizing and integrating the deviation data of all dimensions to obtain the weld parameter difference.

[0126] Step 105: Calculate the air value of the weld based on the weld parameter difference.

[0127] The weld air value refers to a comprehensive assessment of the gas content and vacuum level within the weld. The weld air value is calculated by substituting various deviation data of the weld parameter differences into the system and then performing a weighted calculation using a multiple linear regression algorithm model. For example, a regression model for the weld air value can be constructed as follows: Weld Air Value = First Weighting Coefficient + Second Weighting Coefficient * Weld Width Deviation + Third Weighting Coefficient * Color Value Deviation + Fourth Weighting Coefficient * Grayscale Value Deviation. The first to fourth weighting coefficients are obtained by fitting calibration test data under vacuum conditions using the least squares method.

[0128] Step 106: Combine the air value of the weld, the abnormal workpiece number and the deviation to form abnormal workpiece information and output it.

[0129] Abnormal workpiece information refers to a complete set of information on welding abnormalities in a workpiece. This abnormal workpiece information is generated by the system associating, collecting, and uniformly encapsulating the detected weld air value, the corresponding abnormal workpiece number, and the deviations of various weld parameters to produce complete abnormal workpiece information. The abnormal workpiece information is output by the system sending the generated information to the terminal interface for display.

[0130] The methods for determining the initial welding path based on workpiece video include: Step 20: Upon receiving a preset normal signal, obtain the deviation area corresponding to the welding deviation value.

[0131] A normal signal indicates that the device or workpiece is functioning normally and without problems. The normal signal is received by the control terminal or device, where a corresponding electrical signal exists. This signal is generated and transmitted after the user presses a button, and the system automatically receives it.

[0132] The deviation region refers to the weld location on the welded workpiece where welding deviations exist and do not conform to the standard weld parameters. The deviation region is obtained by the system locating the pixel area in the workpiece image where the weld parameters exceed the standard, based on the abnormal dimension of the parameter corresponding to the welding deviation value.

[0133] When a normal signal is received, it indicates that the abnormality of the device or workpiece is actually due to the information not being updated in time or misjudgment. Therefore, the deviation area corresponding to the welding deviation value is obtained.

[0134] Step 21: Integrate the deviation area with the welding deviation value to form the modified deviation information.

[0135] Modified deviation information refers to the set of deviation data for correcting welding parameters. Here, the modified deviation information is generated by the system associating and binding the located deviation area with the corresponding welding deviation value, and then collecting and integrating this information to obtain the modified deviation information.

[0136] Step 22: Obtain the welding modification route based on the modified deviation information and the workpiece video.

[0137] The modified welding path refers to the initial welding path after modification. The modified welding path is obtained by the system modifying and adjusting the path in the workpiece video based on the deviation type and amount of the deviation information, and then generating the modified initial welding path, which is then used as the modified welding path.

[0138] Step 23: Define the modified welding path as the initial welding path and output it.

[0139] The initial welding route is output by the system defining the modified welding route as the initial welding route and storing it for later comparison with the actual welding route.

[0140] Based on the same inventive concept, embodiments of the present invention provide a micro-image vision laser AI weld seam tracking system.

[0141] One example is a micro-image vision laser AI weld seam tracking system, which includes: The acquisition module is used to acquire workpiece videos, workpiece categories, welding quantities, micro-image acquisition device numbers, welding workpiece parameters, abnormal workpiece numbers, welding workpiece images, standard workpiece images, and deviation areas. Memory for storing a program for a micro-image vision laser AI weld seam tracking method; The processor loads and executes programs from memory.

[0142] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional modules is used as an example. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the system, device, and unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0143] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A micro-image vision laser AI weld seam tracking method, characterized in that, include: Step 1: In response to a preset welding signal, acquire a video of the workpiece; Step 2: Determine the initial welding path based on the workpiece video; Step 3: Obtain the workpiece category to determine the initial welding point and perform the welding operation to obtain the welding start point; Step 4: Determine the required welding points based on the welding start point; Step 5: Determine the welding points and collect the welding connection points by combining the required welding points, welding operations and preset adjustment rules; Step 6: Define the weld joint as the required weld joint and repeat step 5 to obtain the weld joint group and the weld endpoint; Step 7: Determine the actual welding route based on the welding start point, welding point group, and welding end point, and analyze the welding deviation value in conjunction with the initial welding route; Step 8: If the welding deviation value does not fall within the preset safety deviation range, obtain the welding quantity; Step 9: Analyze the deviation by combining the welding deviation value and the number of welds, and find the corresponding solution output.

2. The micro-image vision laser AI weld seam tracking method according to claim 1, characterized in that, Methods for analyzing deviations and finding corresponding solutions by combining welding deviation values ​​and welding quantity include: Step 90: If the number of welds is equal to 1, obtain the micro-image acquisition device number; Step 91: Control the micro-image acquisition device corresponding to the micro-image acquisition device number to perform an anomaly detection operation to obtain the anomaly situation; Step 92: Define the abnormal situation as a deviation and find the corresponding solution output; Step 93: If the number of welds is greater than 1, the deviation is obtained by comparing multiple weld deviation values; Step 94: When the deviation is within the preset uniform condition, generate and output a device abnormality signal based on the deviation. Step 95: Based on the deviation, find the corresponding device repair plan and output it; Step 96: When the deviation is not uniform, generate and output a workpiece abnormality signal based on the deviation.

3. The micro-image vision laser AI weld seam tracking method according to claim 2, characterized in that, Methods for generating and outputting workpiece abnormality signals based on deviation conditions include: Step 960: Obtain welding workpiece parameters; Step 961: Find the corresponding standard welding parameters based on the workpiece category; Step 962: When the standard welding parameters are inconsistent with the welding workpiece parameters, generate and output a workpiece abnormality signal based on the deviation. Step 963: When the standard welding parameters are consistent with the welding workpiece parameters, no operation is performed.

4. The micro-image vision laser AI weld seam tracking method according to claim 3, characterized in that, When standard welding parameters are inconsistent with the parameters of the workpiece being welded, methods for generating and outputting workpiece abnormality signals based on the deviation include: Step 9620: Obtain the abnormal workpiece number based on the welding workpiece parameters; Step 9621: Count the number of abnormal workpieces corresponding to the abnormal workpiece numbers in the welding quantity; Step 9622: Determine the critical value of the abnormality by combining the number of welds with the preset abnormality rules; Step 9623: When the number of abnormal workpieces reaches the abnormal threshold, a workpiece abnormality signal is generated and output based on the deviation. Step 9624: When the number of abnormal workpieces does not reach the abnormal threshold, an abnormal workpiece group is formed and output based on the abnormal workpiece number and deviation.

5. The micro-image vision laser AI weld seam tracking method according to claim 2, characterized in that, Other methods for finding corresponding solutions include: Step 97: Obtain images of the welded workpiece and the standard workpiece; Step 98: Extract the corresponding weld seam information from the welded workpiece image; Step 99: Determine standard weld information using standard workpiece images; Step 100: Compare the workpiece weld information with the standard weld information to determine the weld condition; Step 101: Determine and output the corresponding solution based on the weld condition.

6. The micro-image vision laser AI weld seam tracking method according to claim 5, characterized in that, Also includes: Step 910: Extract the weld width from the workpiece weld information; Step 911: Extract the standard weld width from the standard weld information; Step 912: When the weld width is greater than the standard weld width, define the weld condition as oxidation and output it. Step 913: When the weld width is not greater than the standard weld width, define the weld condition as normal and output it.

7. The micro-image vision laser AI weld seam tracking method according to claim 6, characterized in that, Methods for defining and outputting weld condition as oxidation condition include: Step 914: Extract the grayscale value and color value of the weld from the image of the welded workpiece and form a weld color value group; Step 915: Extract the grayscale standard values ​​and color standard values ​​from the standard workpiece image and form a standard color value group; Step 916: Compare the weld color value group with the standard color value group to obtain the color group difference; Step 917: If the color group difference is greater than the preset allowable threshold, define the weld condition as oxidation and output it; Step 918: If the color group difference is not greater than the allowable threshold, define the weld condition as normal and output it.

8. The micro-image vision laser AI weld seam tracking method according to claim 5, characterized in that, Also includes: Step 102: Extract the corresponding actual weld parameters based on the workpiece weld information; Step 103: Extract the corresponding standard weld parameters from the standard weld information; Step 104: Calculate the difference in weld parameters by combining the actual weld parameters with the standard weld parameters; Step 105: Calculate the air value of the weld based on the weld parameter difference; Step 106: Combine the air value of the weld, the abnormal workpiece number and the deviation to form abnormal workpiece information and output it.

9. The micro-image vision laser AI weld seam tracking method according to claim 1, characterized in that, Methods for determining the initial welding path based on workpiece video include: Step 20: Upon receiving a preset normal signal, obtain the deviation area corresponding to the welding deviation value; Step 21: Integrate the deviation area with the welding deviation value to form the modified deviation information; Step 22: Obtain the welding modification route based on the modified deviation information and workpiece video; Step 23: Define the modified welding path as the initial welding path and output it.

10. A micro-image vision laser AI weld seam tracking system, characterized in that, include: The acquisition module is used to acquire workpiece videos, workpiece categories, welding quantities, micro-image acquisition device numbers, welding workpiece parameters, abnormal workpiece numbers, welding workpiece images, standard workpiece images, and deviation areas. A memory for storing a program for a micro-image vision laser AI weld seam tracking method as described in any one of claims 1 to 9; The processor loads and executes programs from memory.

Images