Welding bead identification method of laser sealing welding machine

Abstract

The invention discloses a welding bead identification method for a laser sealing welding machine. The method comprises the following steps that S1, a product is fixed to a platform; s2, obtaining the overall contour of the image; s3, preprocessing the image; s4, preliminarily selecting a candidate welding bead point; s5, further screening and confirming the welding bead points to ensure that the selected welding bead points are accurate; s6, the selected welding bead two side points are sequentially connected, a welding bead outer side edge visual simulation track and a welding bead inner side edge visual simulation track are formed, and the welding seam track width is calculated; and S7, selecting a visual simulation track offset pixel point to reach a proper position, and performing coordinate conversion to form a real motion welding track. According to the method, a machine vision detection algorithm is adopted, accurate, stable and anti-interference vision positioning of a real welding seam can be achieved, meanwhile, the width of the welding seam can be effectively detected, welding bead out-of-tolerance defective products can be removed, and the production cost is saved.

Description

Technical Field

[0001] The invention relates to a method for identifying a weld bead of a laser sealing welding machine, and belongs to the technical field of identifying a weld bead of a laser sealing welding machine. Background Art

[0002] The weld bead recognition system is the result of the integration of related technologies in the fields of computers and machinery. It has a wide range of applications in manufacturing, aerospace, construction and other fields. In recent years, with the rapid development of new generation information technologies such as artificial intelligence, big data, and cloud computing, welding technology is moving towards intelligence. Intelligent welding uses advanced technologies such as intelligent sensors and intelligent control systems to achieve intelligent perception, intelligent decision-making and intelligent execution of the welding process. At this stage, the weld bead recognition system has also ushered in an intelligent upgrade. Weld bead recognition algorithms based on machine vision have begun to be applied in actual production. These algorithms can identify welds more accurately and stably, significantly improving welding efficiency and weld quality. Weld bead recognition is easily affected by factors such as surrounding impurities and unstable imaging. Most existing weld bead recognition systems use pre-written fixed trajectories and correct the trajectory by identifying the deviation of a stable feature of the template product. This method does not really identify the weld bead. When the weld bead of the test product differs from that of the template product, it cannot be captured in time. Problems such as inaccurate welding position and incomplete welding may occur, resulting in an increase in the defective product rate. Summary of the invention

[0003] The technical problem to be solved by the present invention is to provide a method for identifying welds of a laser welding machine. The method adopts a machine vision detection algorithm, which can achieve accurate, stable and interference-resistant visual positioning of the real weld. At the same time, it can effectively detect the weld width, eliminate defective welds with out-of-tolerance properties, and save production costs.

[0004] The present invention is implemented by the following scheme: A method for identifying a weld bead of a laser welding machine comprises the following steps:

[0005] S1. The product is fixed on the platform;

[0006] S2. Use an image acquisition mechanism to capture an image of the workpiece perpendicular to the product surface to obtain the overall contour of the image;

[0007] S3, preprocessing the image to further highlight the weld features, making the weld easier to extract;

[0008] S4. The weld recognition tool automatically generates several detection lines perpendicular to the weld, and preliminarily selects the candidate weld points by calculating the grayscale change points of the detection lines;

[0009] S5. Further screening and confirmation of weld points to ensure that the selected weld points are accurate;

[0010] S6, connecting the selected points on both sides of the weld in sequence to form a visual simulation track of the outer edge of the weld and a visual simulation track of the inner edge of the weld, and calculating the width of the weld track;

[0011] S7, select the visual simulation trajectory to offset the pixel point to the appropriate position, and perform coordinate transformation to form a real motion welding trajectory.

[0012] The weld point selection process in step S4 includes:

[0013] S41, dividing the product into slices with a width of 1 pixel and a height of a set value along one circumference of the product;

[0014] S42, calculating the grayscale mean of all pixels in each slice, and projecting the mean onto the midline of the measurement object. If the projection is not in the horizontal direction or the vertical direction, an interpolation calculation is performed when calculating the grayscale mean of each slice;

[0015] S43, smoothing the projection value on the midline of the measurement object using a Gaussian filter;

[0016] S44, calculating the first-order derivative of the value on the midline of the measured object after filtering, and taking all extreme value positions of the first-order derivative as candidate points;

[0017] S45, preliminarily selecting candidate points whose calculated values ​​are within the set threshold range as candidate welding points.

[0018] The weld point selection process in step S5 includes:

[0019] S51, using the least squares fitting method, performing straight line fitting on the five points before and after the points initially screened; the fitted straight line can be regarded as the preliminary trajectory line of the weld bead;

[0020] S52, calculating the vertical distance between the screened point and the fitting straight line, and the distance reflects the degree to which the screened point deviates from the weld track;

[0021] S53. Confirm the weld points.

[0022] In step S53, points with a large distance from the fitting straight line are traversed longitudinally to detect whether there is a significant change in the grayscale values ​​around these points. On the basis of the grayscale value traversal, if a point is still determined to be a normal point, that is, its grayscale value change is in line with expectations, it is further checked whether the point is a continuous trend point; if not, it is excluded, and the intersection of the fitting curve of the 5 points before and after it and the detection line is selected as the new weld point.

[0023] The beneficial effects of the present invention are:

[0024] 1. The laser welding machine weld identification method of the present invention can flexibly set the point selection mode, and can select a single point, that is, "the outer edge of the weld" or "the inner edge of the weld"; it can also select points on both sides of the weld edge, which can effectively detect the weld width, eliminate welds with out-of-tolerance defective products, and save production costs;

[0025] 2. In a method for identifying weld bead of a laser welding machine of the present invention, when the weld bead point is accurately selected, the selected point will be compared with the surrounding area for grayscale value, and further screened if it meets the set range. The further screening algorithm is to detect the "selected point" and the "distance between the 5 points before and after the fitting straight line". When the distance is abnormal, the grayscale value will be traversed in the longitudinal direction to detect whether there is a large change in the grayscale value; if the change is not large, the point is the actual weld edge; if the change is too large, the traversal in the longitudinal direction will continue until the best weld edge is found, thereby eliminating the interference of other factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 The present invention is a schematic diagram of a weld bead identification process of a weld bead identification method for a laser welding machine.

[0027] Figure 2 This is a collection diagram of an image collection mechanism of a method for identifying weld bead of a laser welding machine according to the present invention.

[0028] Figure 3 This is a diagram before image preprocessing of a method for identifying welds of a laser welding machine according to the present invention.

[0029] Figure 4 This is a diagram after image preprocessing of a method for identifying weld bead of a laser welding machine according to the present invention.

[0030] Figure 5 The present invention is a schematic diagram of weld bead points of a method for identifying weld bead of a laser welding machine.

[0031] Figure 6 The present invention provides a visual simulation trajectory diagram of a weld bead recognition method for a laser welding machine.

[0032] Figure 7 This is a real welding trajectory diagram of a laser welding machine weld bead identification method according to the present invention. DETAILED DESCRIPTION

[0033] Combine the following Figure 1-7 The present invention is further described, but the protection scope of the present invention is not limited to the described contents.

[0034] For the sake of clarity, not all features of an actual embodiment are described. In the following description, well-known functions and structures are not described in detail because they would obscure the invention with unnecessary detail. It should be recognized that in the development of any actual embodiment, a large number of implementation details must be made to achieve the specific goals of the developer, such as changing from one embodiment to another according to relevant system or relevant business limitations. In addition, it should be recognized that such development work may be complex and time-consuming, but it is merely routine work for those skilled in the art.

[0035] Example 1

[0036] A method for identifying a weld bead of a laser welding machine comprises the following steps:

[0037] S1. The product is fixed on the platform;

[0038] S2. Use an image acquisition mechanism to capture an image of the workpiece perpendicular to the product surface to obtain the overall contour of the image;

[0039] S3, preprocessing the image to further highlight the weld features, making the weld easier to extract;

[0040] S4. The weld recognition tool automatically generates several detection lines perpendicular to the weld, and preliminarily selects the candidate weld points by calculating the grayscale change points of the detection lines;

[0041] S5. Further screening and confirmation of weld points to ensure that the selected weld points are accurate;

[0042] S6, connecting the selected points on both sides of the weld in sequence to form a visual simulation track of the outer edge of the weld and a visual simulation track of the inner edge of the weld, and calculating the width of the weld track;

[0043] S7, select the visual simulation trajectory to offset the pixel point to the appropriate position, and perform coordinate transformation to form a real motion welding trajectory.

[0044] The weld point selection process in step S4 includes:

[0045] S41, dividing the product into slices with a width of 1 pixel and a height of a set value along one circumference of the product;

[0046] S42, calculating the grayscale mean of all pixels in each slice, and projecting the mean onto the midline of the measurement object. If the projection is not in the horizontal direction or the vertical direction, an interpolation calculation is performed when calculating the grayscale mean of each slice;

[0047] S43, smoothing the projection value on the midline of the measurement object using a Gaussian filter;

[0048] S44, calculating the first-order derivative of the value on the midline of the measured object after filtering, and taking all extreme value positions of the first-order derivative as candidate points;

[0049] S45, preliminarily selecting candidate points whose calculated values ​​are within the set threshold range as candidate welding points.

[0050] The weld point selection process in step S5 includes:

[0051] S51, using the least squares fitting method, performing straight line fitting on the five points before and after the points initially screened; the fitted straight line can be regarded as the preliminary trajectory line of the weld bead;

[0052] S52, calculating the vertical distance between the screened point and the fitting straight line, and the distance reflects the degree to which the screened point deviates from the weld track;

[0053] S53. Confirm the weld point.

[0054] In step S53, points with a large distance from the fitting straight line are traversed longitudinally to detect whether there is a significant change in the grayscale values ​​around these points. On the basis of the grayscale value traversal, if a point is still determined to be a normal point, that is, its grayscale value change is in line with expectations, it is further checked whether the point is a continuous trend point; if not, it is excluded, and the intersection of the fitting curve of the 5 points before and after it and the detection line is selected as the new weld point.

[0055] Example 2

[0056] To facilitate understanding of this embodiment, the application scenarios and design concepts of the embodiments of the present application are briefly introduced below.

[0057] A device for identifying weld bead of a laser welding machine, the device applies the method for identifying weld bead of a laser welding machine in Example 1, comprises an image acquisition mechanism, the image acquisition mechanism is connected to a control mechanism, a weld bead recognition model is stored in the control mechanism, and the weld bead recognition model comprises: a detection calibration module, an image data preprocessing module, a detection line generation module, a visual simulation trajectory module, and a weld bead conversion module;

[0058] The detection and calibration module is used to calibrate the detection area, set the standard workpiece in the detection area, and collect image data of the standard workpiece;

[0059] Image data preprocessing module, including denoising, contrast enhancement, grayscale correction, etc., aims to highlight the weld features, suppress image noise and the influence of slight impurities on the product surface, and provide a clear and stable image basis for subsequent weld point identification;

[0060] Detection line generation module: generates a detection line perpendicular to the weld. The detection line identifies two points at the junction of light and dark on both sides;

[0061] Visual simulation track module: several detection lines are connected to the same side point to form two visual simulation tracks;

[0062] Weld bead conversion module: offset the visual simulation trajectory by a certain number of pixels to reach the appropriate position, and perform coordinate conversion to form a real motion welding trajectory.

[0063] The identification process is as follows:

[0064] 1. The product is fixed on the platform;

[0065] 2. Use the image acquisition mechanism to capture the image of the workpiece perpendicular to the product surface to obtain the overall outline of the image. The collected image is referenced Figure 2 ;

[0066] 3. Preprocess the image to further highlight the weld features, making it easier to extract the weld. Figure 3 , after preprocessing, refer to Figure 4 ;

[0067] 4. The weld identification tool automatically generates several detection lines perpendicular to the weld. Figure 5 ,Through calculating the points where the grayscale of the detection line changes, they are preliminarily selected as candidate weld points;

[0068] 5. Further screening and confirmation of weld points to ensure that the selected weld points are accurate;

[0069] 6. Reference Figure 6 , connect the selected points on both sides of the weld bead in sequence to form a visual simulation track of the outer edge of the weld bead and a visual simulation track of the inner edge of the weld bead, and calculate the weld track width;

[0070] 7. Reference Figure 7 , select the outer edge visual simulation trajectory to offset 5 pixels to reach the appropriate position, and perform coordinate transformation to form a real motion welding trajectory.

[0071] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0072] Although the technical solutions of the present invention have been described and listed in detail, it should be understood that it is obvious to those skilled in the art to make modifications to the above embodiments or adopt equivalent alternatives. These modifications or improvements made without departing from the spirit of the present invention are within the scope of protection required by the present invention.

Claims

1. A method for identifying weld bead of laser welding machine, characterized in that: The steps include: S1. The product is fixed on the platform; S2. Use an image acquisition mechanism to capture an image of the workpiece perpendicular to the product surface to obtain the overall contour of the image; S3, preprocessing the image to further highlight the weld features, making the weld easier to extract; S4. The weld recognition tool automatically generates several detection lines perpendicular to the weld, and preliminarily selects the candidate weld points by calculating the grayscale change points of the detection lines; S5. Further screening and confirmation of weld points to ensure that the selected weld points are accurate; S6, connecting the selected points on both sides of the weld in sequence to form a visual simulation track of the outer edge of the weld and a visual simulation track of the inner edge of the weld, and calculating the width of the weld track; S7, select the visual simulation trajectory to offset the pixel point to the appropriate position, and perform coordinate transformation to form a real motion welding trajectory.

2. A method for identifying weld bead of a laser welding machine according to claim 1, characterized in that: The weld point selection process in step S4 includes: S41, dividing the product into slices with a width of 1 pixel and a height of a set value along one circumference of the product; S42, calculating the grayscale mean of all pixels in each slice, and projecting the mean onto the midline of the measurement object. If the projection is not in the horizontal direction or the vertical direction, an interpolation calculation is performed when calculating the grayscale mean of each slice; S43, smoothing the projection value on the midline of the measurement object using a Gaussian filter; S44, calculating the first-order derivative of the value on the midline of the measured object after filtering, and taking all extreme value positions of the first-order derivative as candidate points; S45, preliminarily selecting candidate points whose calculated values ​​are within the set threshold range as candidate welding points.

3. A method for identifying weld bead of laser welding machine according to claim 1, characterized in that: The weld point selection process in step S5 includes: S51, using the least squares fitting method, performing straight line fitting on the five points before and after the points initially screened; the fitted straight line can be regarded as the preliminary trajectory line of the weld bead; S52, calculating the vertical distance between the screened point and the fitting straight line, and the distance reflects the degree to which the screened point deviates from the weld track; S53. Confirm the weld points.

4. A method for identifying weld bead of laser welding machine according to claim 3, characterized in that: In step S53, points with a large distance from the fitting straight line are traversed longitudinally to detect whether there is a significant change in the grayscale values ​​around these points. On the basis of the grayscale value traversal, if a point is still determined to be a normal point, that is, its grayscale value change is in line with expectations, it is further checked whether the point is a continuous trend point; if not, it is excluded, and the intersection of the fitting curve of the 5 points before and after it and the detection line is selected as the new weld point.

Images