A laser welding method and laser welding apparatus based on machine vision

Abstract

This invention provides a laser welding method and apparatus based on machine vision. The laser welding method includes: after a fixture fixes the workpiece to be welded in the welding area, acquiring an initial photograph of the welding area; controlling a welding device to pre-weld the lower half of the workpiece based on the initial photograph; acquiring a pre-welding photograph of the welding area, and obtaining the actual position, orientation, and key dimensions of the welding area of ​​the workpiece based on the pre-welding photograph; determining whether the welding part meets the requirements based on the actual position, orientation, and key dimensions of the welding area of ​​the workpiece; if the requirements are met, controlling the welding device to weld the upper half of the workpiece; if the requirements are not met, reprocessing based on the actual position, orientation, and key dimensions of the welding area of ​​the workpiece. This application identifies and corrects problems before completing the entire welding process, reducing rework costs caused by welding defects and improving production efficiency.

Description

Technical Field

[0001] This invention relates to the field of automotive technology, and more specifically to a laser welding method and laser welding apparatus based on machine vision. Background Technology

[0002] In recent years, with the improvement of automation levels in manufacturing, laser welding technology has been widely used in various industrial production scenarios due to its advantages such as high energy density, high precision, and high efficiency. In traditional laser welding processes, experienced welders typically control welding robots manually or via teach pendants based on visual judgment.

[0003] However, manual operation is limited by the skill level and fatigue of workers, making it difficult to guarantee consistent quality in long-term, large-scale production, and it is also inefficient. Summary of the Invention

[0004] The problem solved by this invention is the low efficiency of existing laser welding methods.

[0005] To address the above problems, this invention provides a laser welding method based on machine vision, the laser welding method comprising:

[0006] After the fixture fixes the workpiece to be welded in the area to be welded, an initial photograph of the area to be welded is obtained;

[0007] Based on the initial photograph, the welding equipment is controlled to pre-weld the lower half of the workpiece to be welded;

[0008] Obtain a pre-welding photograph of the area to be welded, and obtain the actual position, orientation, and key dimensions of the area to be welded based on the pre-welding photograph;

[0009] Determine whether the welding part meets the requirements based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded.

[0010] If the requirements are met, the welding equipment is controlled to weld the upper half of the workpiece to be welded;

[0011] If the requirements are not met, the workpiece to be welded shall be reprocessed according to its actual position, orientation, and key dimensions of the area to be welded.

[0012] Optionally, determining whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded includes:

[0013] Determine whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded;

[0014] If satisfied, the gap width between the upper half of the workpiece to be soldered and the solder ball, and the degree of offset on the horizontal axis between the upper half of the workpiece to be soldered and the solder ball are obtained.

[0015] If the gap width or the lateral offset is within the threshold range, then the upper half of the workpiece to be welded is welded.

[0016] If the gap width or the lateral offset exceeds the threshold range, it is determined that it is not weldable.

[0017] Optionally, determining whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded includes:

[0018] The upper and lower halves of each workpiece to be welded are determined based on its actual position and orientation.

[0019] If it is determined that each workpiece to be welded has solder balls on its lower half, and each solder ball has its upper half on the side away from the lower half of the workpiece to be welded, then the welding position is determined to meet the requirements.

[0020] Optionally, determining whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded further includes:

[0021] If it is determined that no solder balls are set on the lower half of the workpiece to be soldered, then the pre-soldering is determined to be unsuccessful.

[0022] If it is determined that the upper half of the workpiece is not provided on the side of the solder ball away from the lower half of the workpiece to be soldered, then the fixture position is determined to be faulty.

[0023] Optionally, if the requirements are not met, reprocessing based on the actual position, orientation, and key dimensions of the area to be welded of the workpiece includes:

[0024] If it is determined that the pre-welding was unsuccessful, the welding equipment is controlled to re-pre-weld the upper half of the workpiece to be welded;

[0025] If the problem is determined to be due to improper fixture positioning, the welding equipment should be reloaded and reprocessed after unloading.

[0026] If it is determined that the material is not weldable, the welding equipment is controlled to stop welding.

[0027] Optionally, the welding equipment is used to weld the upper half of the workpiece to be welded, including:

[0028] The coordinates of the target weld point are obtained based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded.

[0029] Welding is performed by controlling the welding equipment according to the coordinates of the target weld point.

[0030] Optionally, obtaining the target weld point coordinates based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded includes:

[0031] Based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded, a geometric algorithm is used to obtain the initial weld point coordinates on the actual contour of the workpiece to be welded.

[0032] The initial solder joint coordinates are adjusted according to preset process parameters to obtain the target solder joint coordinates; wherein, the process parameters include lateral offset, longitudinal offset, and total adjustment ratio.

[0033] Optionally, after the welding equipment performs welding based on the target weld point coordinates, the laser welding method further includes:

[0034] Obtain post-welding photos of the area to be welded;

[0035] Obtain various dimensional parameters based on the post-weld photographs;

[0036] The dimensional parameters are compared with the process threshold to determine whether the welding is successful, and the process threshold is optimized based on the dimensional parameters.

[0037] This application also provides a machine vision-based laser welding device, the laser welding device comprising:

[0038] A fixture for fixing a workpiece to be welded in the area to be welded;

[0039] Image acquisition unit, used to acquire photographs of the area to be soldered;

[0040] Welding equipment, the welding equipment being used to weld the upper half of a workpiece to be welded to the lower half of the workpiece to be welded;

[0041] A control unit, electrically connected to the welding equipment and the image acquisition unit, is used to execute the laser welding method as described above, so as to control the welding equipment to perform welding.

[0042] Optionally, the image acquisition unit includes at least one of an industrial camera, a line laser, and a structured light camera, and the control unit includes a PLC control system.

[0043] The machine vision-based laser welding method provided in this application realizes dynamic monitoring and precise control of the welding process. On the one hand, the reference system with image data as the core greatly improves the accuracy of workpiece positioning and welding operation, effectively avoiding the overall quality problems caused by initial positioning deviation in traditional welding. On the other hand, through parameter verification and timely adjustment after pre-welding, problems can be discovered and corrected in advance before the formal completion of the entire welding process, reducing rework costs caused by welding defects, significantly improving the stability of welding operations and the finished product qualification rate, and increasing production efficiency. Attached Figure Description

[0044] Figure 1 A schematic flowchart of a laser welding method based on machine vision provided in an embodiment of this application;

[0045] Figure 2 for Figure 1 A schematic diagram of a pre-welded photograph obtained in the laser welding method shown;

[0046] Figure 3 for Figure 1 The diagram shows the process of welding the upper part of the workpiece in the laser welding method shown.

[0047] Figure 4 A schematic diagram of the frame of a machine vision-based laser welding device provided in an embodiment of this application.

[0048] Explanation of reference numerals in the attached figures:

[0049] 100. Laser welding device; 10. Fixture; 20. Image acquisition unit; 30. Welding equipment; 40. Control unit;

[0050] 200. Workpiece to be soldered; 201. Upper part; 202. Lower part; 203. Solder ball

[0051] D1, gap width; D2, lateral offset. Detailed Implementation

[0052] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below.

[0053] Traditional laser welding processes mainly rely on the following three methods:

[0054] 1. Manual teaching and manual operation: Welding is carried out by experienced welders who make visual judgments and manually or through a teach pendant to control the welding robot or actuator.

[0055] 2. Fixed path programming: A fixed welding program is pre-written for a specific workpiece, and the robotic arm strictly follows the preset trajectory to perform the welding task.

[0056] 3. Traditional visual judgment: An image of the workpiece to be welded is captured, and a pre-programmed algorithm identifies the object's outline and weld points. The weld points are then communicated to the welding program to execute the welding task.

[0057] However, existing technical solutions have one or more of the following significant drawbacks:

[0058] 1) Low production efficiency and consistency: Manual operation is limited by the skill level and fatigue of workers, making it difficult to guarantee the quality consistency of long-term, large-scale production, and the efficiency is low.

[0059] 2) Poor flexibility and adaptability: Fixed-path programming requires extremely high precision in the incoming workpiece position, necessitating expensive and complex precision fixtures for positioning. Even slight deviations in workpiece position, or inherent deformation or dimensional tolerances, can easily lead to welding failures, incomplete welds, or burn-throughs, resulting in material waste and production interruptions. Traditional visual judgment is prone to errors and shifts when the welding environment changes, leading to welding failures.

[0060] 3) Lack of intelligent decision-making capabilities: Existing automation solutions typically lack the ability to analyze and judge welding conditions (such as gap size and workpiece surface condition) in real time. Even if the workpiece position does not meet the welding requirements, the system cannot identify this and will still perform invalid welding, resulting in a high defect rate.

[0061] 4) Complex human-machine interaction and difficult adjustment: When changing workpieces of different specifications or adjusting welding processes, professional technicians are required to reprogram and debug in a complex manner, resulting in slow response speed and inability to meet the needs of flexible production.

[0062] Therefore, to solve the above-mentioned technical problems, this application provides a laser welding method and laser welding device based on machine vision. The method involves pre-welding the lower half of the workpiece from an initial photograph. After the pre-welding process is completed, another photograph of the pre-welding state of the area to be welded is acquired. Based on the pre-welding photograph, the actual position and orientation of the workpiece, as well as the key dimensions of the area to be welded, are extracted to determine whether the welding part meets the requirements. If the requirements are met, it indicates that the workpiece positioning accuracy and pre-welding quality are up to standard, and the welding equipment will be automatically controlled to continue the welding operation of the upper half of the workpiece, ensuring a continuous and efficient overall welding process. If the parameters do not meet the standards, an adjustment mechanism is immediately triggered. Based on the extracted actual position, orientation, and key dimension data, the relevant actuators are driven to reposition the workpiece or repair the pre-welding area until the process requirements are met before proceeding to the subsequent welding stage. Specific details are as follows:

[0063] Please see Figure 1 , Figure 1 This is a schematic flowchart of a machine vision-based laser welding method provided in an embodiment of this application. The laser welding method includes the following steps:

[0064] 110. After the fixture fixes the workpiece to be welded in the area to be welded, obtain an initial photograph of the area to be welded.

[0065] For example, the upper half 201 and the lower half 202 of the workpiece 200 to be welded are respectively fed by the fixture 10. The fixture 10 is controlled by the PLC program to fix the workpiece 200 to be welded in the welding area and trigger the image acquisition unit 20 to acquire the initial photo of the welding area.

[0066] In some embodiments, a customized or standardized special fixture 10 is selected based on the material, dimensions, and welding process requirements of the workpiece 200 to be welded. The material includes carbon steel, stainless steel, aluminum alloy, etc.; the dimensions include plate-shaped, tubular, and irregularly shaped structures; and the welding process requirements include the clamping force requirements corresponding to arc welding and laser welding. For example, for thin-walled tubular workpieces, a three-jaw centering fixture 10 with a flexible anti-slip pad is used to prevent workpiece deformation during clamping; for large plate-shaped workpieces, a multi-point pneumatic clamping fixture 10 is used to ensure minimal contact error between the workpiece and the reference surface of the welding area.

[0067] In some embodiments, fixed image acquisition units 20 are preset above and to the side of the area to be welded. The shooting parameters of the image acquisition units 20 are set according to the size range of the area to be welded, and each acquired photo is automatically associated with unique identification information, including workpiece number, fixture 10 number, area to be welded number, acquisition time, acquisition device number, and environmental parameters such as temperature and humidity at that time, so as to facilitate subsequent traceability. At the same time, the system automatically makes preliminary markings on key areas in the photos, such as the joint to be welded and workpiece positioning reference points, to provide clear target areas for subsequent pre-welding based on the initial photos.

[0068] The image acquisition unit 20 includes at least one of an industrial camera, a line laser, and a structured light camera. In some embodiments, the image acquisition unit 20 uses an industrial-grade high-definition CCD camera, paired with a high color rendering index ring-shaped LED supplementary light source. By uniformly illuminating the area to be welded from multiple angles, shadows or reflective dead angles are avoided, thus eliminating the impact of specular reflection on image clarity and ensuring that the workpiece outline and the details of the weld joint are clear.

[0069] 120. Based on the initial photograph, control the welding equipment to pre-weld the lower half of the workpiece to be welded.

[0070] For example, after acquiring an initial photograph of the area to be welded, an image recognition algorithm is used to perform pixel-level analysis of the initial photograph, accurately extracting key feature information of the lower half 202 of the workpiece 200 to be welded. This includes the coordinates of the start and end points of the joint, the joint width variation curve, the relative offset between the workpiece edge and the reference axis of the welding equipment 30, and the contour boundary of the pre-welding area of ​​the lower half 202. Simultaneously, combined with the positioning reference points marked in the initial photograph, the image coordinate data is converted into three-dimensional spatial coordinates (X / Y / Z axis parameters) recognizable by the welding equipment 30. Subsequently, based on the pre-welding process standards and the analyzed spatial coordinate data, the path trajectory for the pre-welding of the lower half 202 is automatically planned. After the path is determined, control commands including the welding torch starting position, movement speed, and welding parameter switching timing are sent to the welding equipment 30 to complete the pre-welding of the lower half 202.

[0071] It is important to note that throughout the process, the laser welding device 100 will also compare the reference data in the initial photograph with the actual movement position of the welding torch in real time. Through closed-loop feedback adjustment, it will ensure that the pre-welding area accurately covers the lower half 202 of the workpiece to be welded, and that the welding quality meets the strength and forming requirements of the pre-welding process. In addition, if there are slight bends or width fluctuations in the joint during the welding process, the laser welding device 100 will also adjust the movement trajectory and dwell time of the welding torch in real time to ensure that the welding torch always moves along the center line of the joint, and appropriately reduce the welding speed in the wider joint to ensure uniform penetration depth, laying a stable foundation for subsequent parameter verification and welding of the upper half 201.

[0072] 130. Obtain pre-welding photos of the area to be welded, and obtain the actual position, orientation, and key dimensions of the area to be welded based on the pre-welding photos.

[0073] The acquisition of pre-welding photographs of the area to be welded can be achieved by: after the pre-welding process of the lower half 202 of the workpiece 200 is completed, controlling the image acquisition unit 20 to simultaneously acquire pre-welding photographs of the area to be welded from multiple key perspectives, such as directly above and from a 45° side, to ensure that the photographs clearly show the shape of the pre-welded weld, the overall position of the workpiece, and the surrounding reference marks. The acquired pre-welding photographs can be as follows: Figure 2 As shown, Figure 2 for Figure 1 A schematic diagram of a pre-welded photograph obtained in the laser welding method shown.

[0074] Obtaining the actual position, orientation, and key dimensions of the welding area of ​​the workpiece 200 to be welded from a pre-welding photograph can be achieved by: importing the pre-welding photograph into an image feature extraction model in real time; the image feature extraction model quickly identifies and segments the contours of one or more workpieces 200 to be welded in the photograph; and calculating the actual position, orientation, and key dimensions of the welding area of ​​the workpiece 200 based on the segmented contours.

[0075] For example, a feature point matching algorithm is used to locate the preset positioning reference point on the workpiece surface, and the coordinate offset of the reference point in three-dimensional space is calculated. Then, the actual position (X / Y / Z axis deviation value from the initial fixed position) and posture (such as whether the workpiece is tilted or twisted, and the posture deviation is quantified by the angle change of the line connecting the reference points) of the workpiece 200 to be welded are determined. At the same time, for the area to be welded (including the pre-welded lower half 202 weld and the upper half 201 joint to be welded), key dimension parameters are accurately extracted by edge detection and pixel ranging technology.

[0076] In some embodiments, the laser welding method further includes automatically storing these actual positions, orientations, and key dimensions into a system database to provide a basis for determining whether the welded parts meet the requirements.

[0077] It should be noted that if the image acquisition unit 20 fails to acquire the pre-welding photo, it will send a failure message to trigger an alarm so that manual inspection can be carried out.

[0078] 140. Determine whether the welding part meets the requirements based on the actual position and posture of the workpiece to be welded and the key dimensions of the area to be welded.

[0079] In some embodiments, the welding position is determined to meet the requirements based on the actual position and orientation of the workpiece 200 to be welded; if it meets the requirements, the gap width D1 between the upper half 201 of the workpiece 200 to be welded and the solder ball 203, and the degree of offset on the horizontal axis between the upper half 201 of the workpiece 200 to be welded and the solder ball 203 are obtained; if the gap width D1 or the horizontal offset D2 is within the threshold range, the upper half 201 of the workpiece to be welded is welded; if the gap width D1 or the horizontal offset D2 exceeds the threshold range, it is determined that it cannot be welded.

[0080] The determination of whether the welding position meets the requirements based on the actual position and posture of the workpiece 200 includes: determining the upper half 201 and lower half 202 of each workpiece 200 based on its actual position and posture; if it is determined that solder balls 203 are provided on the lower half 202 of each workpiece 200, and the upper half 201 of the workpiece 200 is provided on the side of each solder ball 203 away from the lower half 202 of the workpiece 200, then the welding position is determined to meet the requirements. If it is determined that no solder balls 203 are provided on the lower half 202 of the workpiece 200, then the pre-welding is determined to be unsuccessful; if it is determined that the upper half 201 of the workpiece 200 is not provided on the side of the solder ball 203 away from the lower half 202 of the workpiece 200, then the fixture 10 is determined to be in a poor position.

[0081] 150. If the requirements are met, control the welding equipment 30 to weld the upper half 201 of the workpiece to be welded.

[0082] Please see Figure 3 , Figure 3 for Figure 1 The diagram illustrates the process of welding the upper half of the workpiece in the laser welding method shown below.

[0083] 151. Obtain the coordinates of the target weld point based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded.

[0084] In some embodiments, based on the actual position and orientation of the workpiece 200 to be welded and the key dimensions of the area to be welded, a geometric algorithm is used to obtain the initial weld point coordinates on the actual contour of the workpiece 200 to be welded; the initial weld point coordinates are adjusted according to preset process parameters to obtain the target weld point coordinates; wherein, the process parameters include lateral offset D2, longitudinal offset, and total adjustment ratio.

[0085] For example, a three-dimensional model of the workpiece 200 is constructed based on its actual position coordinates, three-dimensional attitude parameters (such as pitch angle and rotation angle), and key dimensional data of the area to be welded. Based on this model, geometric algorithms are used to calculate coordinates on the actual contour of the workpiece 200. For instance, using the edge of the pre-welded solder ball 203 as a reference, and combining the starting point, ending point, and transition inflection point of the area to be welded, a series of uniformly distributed initial solder joint coordinates are automatically generated to ensure that each solder joint falls precisely on the theoretical center line to be welded. Subsequently, the initial weld point coordinates are dynamically adjusted by calling the preset welding process parameters: the left and right positions of the weld point on the horizontal plane are finely adjusted according to the horizontal offset D2 parameter (the compensation value perpendicular to the welding direction) to adapt to the actual offset of the workpiece; the front and back sequence and spacing of the weld points are adjusted according to the vertical offset parameter (the compensation value along the welding direction) to optimize the continuity of the welding path; and the overall coordinates are scaled proportionally by the total adjustment ratio to finally obtain the target weld point coordinates that conform to the actual welding conditions. This ensures the accurate correspondence between the weld point and the area to be welded on the workpiece, and the compensation effect of the process parameters offsets the positioning error and equipment operation deviation, thus ensuring high-quality welding.

[0086] Among them, geometric algorithms can be least squares fitting of the joint centerline, Bézier curve generation of smooth paths, etc.

[0087] The total adjustment ratio can be a correction factor that takes into account factors such as the material's thermal deformation coefficient and the accuracy of the welding torch movement.

[0088] 152. Control the welding equipment to perform welding according to the coordinates of the target weld point.

[0089] For example, the generated sequence of target weld point coordinates is packaged into an instruction set, and the instruction set is sent to the lower-level PLC in real time. According to the received coordinate instructions, the PLC controls the welding equipment 30 to drive the laser welding gun to move sequentially to each target weld point position and perform the welding action.

[0090] In some embodiments, after welding is performed by the welding equipment 30 with the target weld point coordinates controlled, the laser welding method further includes: acquiring a post-weld photograph of the area to be welded; acquiring various dimensional parameters based on the post-weld photograph; comparing the dimensional parameters with a process threshold to determine whether the welding is successful; and optimizing the process threshold based on the dimensional parameters. For example, the median is calculated based on historical welding success parameters as a reference parameter; the most extreme 0.5% of successful samples are removed based on a preset limit parameter (0.5%); a window with the smallest failure rate increment covering 99% of successful samples is selected based on a preset confidence level (99%), and the upper and lower limits of this window are set as thresholds; the process threshold is updated every 100 successful samples or 1 failed sample.

[0091] 160. If the requirements are not met, the workpiece shall be re-processed according to its actual position, orientation, and critical dimensions of the area to be welded.

[0092] If the pre-welding is determined to be unsuccessful, the welding equipment 30 is controlled to re-pre-weld the upper half 201 of the workpiece to be welded.

[0093] If it is determined that the fixture 10 is in a bad position, then control the welding equipment 30 to unload the material and reload it for processing.

[0094] If it is determined that the material is not weldable, the welding equipment 30 is controlled to stop welding.

[0095] The process of unloading and reloading can be understood as unloading the upper half 201 and the lower half 202 of the workpiece 200 to be welded using the fixture 10 and then reloading them.

[0096] Therefore, the embodiments of this application have the following beneficial effects:

[0097] 1) Improve welding accuracy and success rate: Through visual precision positioning and dynamic generation of weld points, the incoming material error of the workpiece is effectively compensated, ensuring that the weld point falls accurately in the target position, and greatly improving the first-time welding success rate.

[0098] 2) Enhanced production flexibility: Eliminating the need to design and replace complex precision fixtures 10 for different products reduces tooling costs. When changing production varieties, simply calling or loading new visual models and process parameters allows for rapid adaptation, making it particularly suitable for flexible production modes with multiple varieties and small batches.

[0099] 3) Achieve intelligent decision-making and reduce defect rate: Automatically judge the feasibility of welding before welding, avoid ineffective welding of unqualified workpieces, and fundamentally reduce defective products and material waste.

[0100] 4) Reduced human intervention and lower costs: The entire process of “identification-decision-execution” is fully automated, which reduces reliance on operator skills, lowers labor costs, and improves production efficiency.

[0101] 5) Fast response speed to meet industrial pace: The core algorithms and models are lightweight and can be deployed on edge computing devices or industrial control computers. The system response latency can be controlled within 100ms, which fully meets the real-time requirements of industrial sites.

[0102] Please continue reading. Figure 4 , Figure 4 This is a schematic diagram of the framework of a laser welding device based on machine vision provided in an embodiment of this application. This application also provides a laser welding device 100 based on machine vision. The laser welding device 100 includes: a clamp 10 for fixing a workpiece 200 to be welded in a welding area; an image acquisition unit 20 for acquiring a photograph of the welding area; a welding device 30 for welding the upper half 201 and the lower half 202 of the workpiece to be welded; and a control unit 40 electrically connected to the welding device 30 and the image acquisition unit 20, for executing the laser welding method described above to control the welding device 30 to perform welding. Specific details are as described above and will not be repeated here.

[0103] In some embodiments, the image acquisition unit 20 includes at least one of an industrial camera, a line laser, and a structured light camera, and the control unit 40 includes a PLC control system.

[0104] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A laser welding method based on machine vision, characterized in that, The laser welding method includes: After the fixture fixes the workpiece to be welded in the area to be welded, an initial photograph of the area to be welded is obtained; Based on the initial photograph, the welding equipment is controlled to pre-weld the lower half of the workpiece to be welded; Obtain a pre-welding photograph of the area to be welded, and obtain the actual position, orientation, and key dimensions of the area to be welded based on the pre-welding photograph; Determine whether the welding part meets the requirements based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded. If the requirements are met, the welding equipment is controlled to weld the upper half of the workpiece to be welded; If the requirements are not met, the workpiece to be welded shall be reprocessed according to its actual position, orientation, and key dimensions of the area to be welded. Determining whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded includes: Determine whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded; If satisfied, then obtain the gap width between the upper half of the workpiece to be soldered and the solder ball, and the lateral offset on the horizontal axis between the upper half of the workpiece to be soldered and the solder ball. If the gap width or the lateral offset is within the threshold range, then the upper half of the workpiece to be welded is welded. If the gap width or the lateral offset exceeds the threshold range, it is determined that it is not weldable; Determining whether the welding area meets the requirements based on the actual position and orientation of the workpiece to be welded includes: The upper and lower halves of each workpiece to be welded are determined based on its actual position and orientation. If it is determined that each workpiece to be welded has solder balls on its lower half, and each solder ball has its upper half on the side away from the lower half of the workpiece to be welded, then the welding part is determined to meet the requirements. If it is determined that no solder balls are set on the lower half of the workpiece to be soldered, then the pre-soldering is determined to be unsuccessful. If it is determined that the upper half of the workpiece is not provided on the side of the solder ball away from the lower half of the workpiece to be soldered, then the fixture position is determined to be faulty.

2. The laser welding method according to claim 1, characterized in that, If the requirements are not met, reprocessing is performed based on the actual position and orientation of the workpiece to be welded, as well as the key dimensions of the area to be welded, including: If it is determined that the pre-welding was unsuccessful, the welding equipment is controlled to re-pre-weld the upper half of the workpiece to be welded; If the problem is determined to be due to improper fixture positioning, the welding equipment should be reloaded and reprocessed after unloading. If it is determined that the material is not weldable, the welding equipment is controlled to stop welding.

3. The laser welding method according to claim 1, characterized in that, The controlled welding equipment performs welding on the upper part of the workpiece to be welded, including: The coordinates of the target weld point are obtained based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded. Welding is performed by controlling the welding equipment according to the coordinates of the target weld point.

4. The laser welding method according to claim 3, characterized in that, The process of obtaining the target weld point coordinates based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded includes: Based on the actual position and orientation of the workpiece to be welded and the key dimensions of the area to be welded, a geometric algorithm is used to obtain the initial weld point coordinates on the actual contour of the workpiece to be welded. The initial solder joint coordinates are adjusted according to preset process parameters to obtain the target solder joint coordinates; wherein, the process parameters include lateral offset, longitudinal offset, and total adjustment ratio.

5. The laser welding method according to claim 3, characterized in that, After the target weld point coordinates are controlled by the welding equipment to perform welding, the laser welding method further includes: Obtain post-welding photos of the area to be welded; Obtain various dimensional parameters based on the post-weld photographs; The dimensional parameters are compared with the process threshold to determine whether the welding is successful, and the process threshold is optimized based on the dimensional parameters.

6. A laser welding device based on machine vision, characterized in that, The laser welding apparatus includes: A fixture for fixing a workpiece to be welded in the area to be welded; Image acquisition unit, used to acquire photographs of the area to be soldered; Welding equipment, the welding equipment being used to weld the upper half of a workpiece to be welded to the lower half of the workpiece to be welded; A control unit, electrically connected to the welding equipment and the image acquisition unit, is used to execute the laser welding method as described in any one of claims 1 to 5, so as to control the welding equipment to perform welding.

7. The laser welding apparatus according to claim 6, characterized in that, The image acquisition unit includes at least one of an industrial camera, a line laser, and a structured light camera, and the control unit includes a PLC control system.

Images