Laser line contour extraction method capable of avoiding welding smoke interference

Abstract

The invention discloses a laser line contour extraction method capable of avoiding welding smoke interference, and relates to the technical field of laser line measurement, and the method comprises the following steps: 1, creating and initializing a template contour T; step 2, acquiring a laser line image; step 3, extracting a laser contour line P; 4, judging whether the anti-smoke function is started or not; 5, traversing points in the laser contour line P, and judging the accuracy of the current point Ai; step 6, if the current point Ai is an abnormal point, finding a point Ti with the same serial number in the template contour T; 7, correcting an abnormal point Ai according to the information around the point Ti; step 8, updating the point Ai which is not abnormal in the step 5 or the abnormal point Ai in the step 7 into the template contour T after correction, and traversing the laser contour line P through the methods in the step 5, the step 6 and the step 7; and 9, outputting the contour P, and returning to the step 2. According to the method, the contour interfered by the smoke is corrected by using the contour information before the smoke appears, so that the influence of the welding smoke on laser line extraction can be quickly and effectively reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of laser line measurement, and in particular to a laser line profile extraction method that avoids interference from welding smoke. Background Art

[0002] Laser line measurement technology is a high-precision measurement method widely used in the welding field. Using laser line scanning to perform three-dimensional reconstruction of the groove can obtain accurate groove size information, thereby calculating groove width, groove depth, groove inclination angle and other morphological features. It is used for pre-weld planning and in-weld adjustment of welding parameters, as well as post-weld weld quality inspection.

[0003] In practice, the grooves before and after welding have little effect on laser line measurement. However, welding often produces a large amount of smoke. This smoke, composed of metal vapor, dust, and gases, has a high degree of uncertainty and strongly scatters the propagation of the laser line, resulting in reduced brightness and a weaker Gaussian distribution. In the captured images, the smoke from the welding process often obscures the laser line, weakening the laser line feature and causing the laser contour points to be mistakenly identified as being on the smoke.

[0004] While numerous image dehazing methods have been proposed in the field of computer vision, they are not applicable to laser line extraction during welding. In recent years, deep learning technology has demonstrated strong feature extraction capabilities in the image field, making it suitable for visual tasks in complex, dynamic, and interfering environments, and holds great promise for solving this problem. However, deep learning relies on large-scale, high-quality annotated samples, which, due to the scarcity and difficulty of annotating welding scene data, presents challenges for rapid deployment in real-world production environments.

[0005] Therefore, how to quickly and effectively reduce the impact of welding smoke on laser line extraction is a key technical issue that needs to be solved urgently. Summary of the Invention

[0006] The object of the present invention is to provide a laser line profile extraction method that avoids interference from welding smoke, so as to solve the above-mentioned problem.

[0007] To achieve the purpose of the present invention, the technical solution adopted is: a laser line profile extraction method avoiding the interference of welding smoke, comprising the following steps: Step 1: Create and initialize the template outline T; Step 2: Acquire laser line image; Step 3: Extract the laser contour line P; Step 4: Determine whether the anti-smoke function is turned on; Step 5: Traverse the points in the laser contour line P and determine the accuracy of the current point Ai; Step 6: If the current point Ai is an outlier, find the point Ti with the same sequence number in the template contour T; Step 7: Correct the abnormal point Ai based on the information around point Ti; Step 8: update the non-abnormal point Ai in step 5 or the abnormal point Ai in step 7 after correction to the template contour T, and traverse the laser contour line P through the methods in steps 5, 6 and 7; Step 9: Output the contour P and return to step 2.

[0008] Furthermore, when it is determined in step 4 that the anti-smog function is not enabled, the points in the laser contour line P are updated to the template contour T, and the process proceeds to step 9.

[0009] Furthermore, when the laser contour line P is not traversed in step eight, return to step five.

[0010] Furthermore, the method for extracting the laser contour line P in step 3 is edge method, grayscale centroid method or Steger algorithm.

[0011] Furthermore, the method for determining the accuracy of the current point Ai in step five is: by comparing the difference between the current point Ai and its previous and next points Ai-n and Ai+n, if the difference exceeds a threshold, the current point Ai is determined to be an abnormal point.

[0012] Furthermore, the method for determining the accuracy of the current point Ai in step five is: by comparing the current point Ai with the same point Ti in the template contour T, if the difference exceeds a threshold, the current point Ai is determined to be an abnormal point.

[0013] Furthermore, the method for judging the accuracy of the current point Ai in step 5 is: taking the contour point on the laser contour line P as the center, within a neighborhood with a radius of r, the variance of the pixel values ​​in the neighborhood is counted. If the variance is less than a preset value, it is judged as an abnormal point.

[0014] Furthermore, the correction method of the abnormal point Ai in step seven is: by searching for the maximum value point Mi within the preset range above and below Ti in the laser line image obtained in step two, and replacing the abnormal point Ai with the coordinates of point Mi.

[0015] Furthermore, the correction method for the abnormal point Ai in step seven is: by calculating the grayscale centroid Gi of the pixels within the preset range above and below the point Ti in step six, and replacing the abnormal point Ai with the coordinates of the point Gi.

[0016] The beneficial effects of the present invention are: In the present invention, the contour information before the smoke appears is used to correct the contour interfered by the smoke, so that the influence of welding smoke on laser line extraction can be quickly and effectively reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings illustrate exemplary embodiments of the present invention and together with the description serve to explain the principles of the present invention. These drawings are included to provide a further understanding of the present invention and are incorporated in and constitute a part of this specification.

[0018] Figure 1 A schematic flow chart of a laser line profile extraction method for avoiding welding fume interference provided by an embodiment of the present invention; Figure 2 Schematic diagram of a laser line image obtained in an embodiment of the present invention; Figure 3 Schematic diagram of the result of extracting the laser contour line P in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It will be understood that the specific embodiments described herein are intended only to explain the relevant content and are not intended to limit the present invention. It should also be noted that, for ease of description, only portions relevant to the present invention are shown in the accompanying drawings.

[0020] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0021] like Figure 1 As shown, the present invention provides a laser line profile extraction method that avoids interference from welding smoke, comprising the following steps: Step 1: Create the template outline T through code. Template outline T is a two-dimensional array containing a series of points (x, y). x represents the number of the outline point (initialized to the image column number), so the value of x increases sequentially. y represents the position of the outline point in template outline T (that is, the row number of the current outline point), and the initial value of y is 0. It should be noted that template outline T is completed during program initialization.

[0022] Step 2: Use laser vision sensor or laser profilometer to obtain laser line image, such as Figure 2 shown.

[0023] Step 3: Use edge methods, grayscale centroid methods, Steger algorithms, etc. to extract the laser contour line P of the laser line image. Each point on the laser contour line P corresponds to a coordinate. The composition of the coordinates is the same as the two-dimensional array of points in step 1, that is, the coordinates include the contour coordinate x and the contour coordinate y. As shown in the following table, the coordinates corresponding to each point on the laser contour line P are as follows: Table 1 shows the coordinates of each point on the laser contour line P. Coordinate x 1 2 3 ... 100 ... 129 130 ... 156 157 ... Coordinate y 10 11 9 9 11 10 19 30 51 55 53 55 Coordinate x 200 ... 252 253 ... 300 301 ... 400 401 402 ... Coordinate y 54 53 38 17 10 12 11 10 11 12 13 14 like Figure 3 As shown, the red line in the figure is the extracted laser contour line P.

[0024] Step 4: By receiving an external signal, determine whether the anti-smoke function is turned on according to the signal value; if the external signal value is 1, it is determined that the anti-smoke function is turned on, and then proceed to step 5; if the external signal value is 0, it is determined that the anti-smoke function is not turned on, and then proceed to step 9.

[0025] Since the welder will produce a lot of smoke after welding starts, the anti-smoke function of the welder needs to be turned on at the beginning of welding or before welding, that is, the anti-smoke function is turned on before the welder starts the arc; since there is no smoke after welding is completed, the anti-smoke function of the welder needs to be turned off after welding or after the welder is turned off, that is, it is turned off after the welder extinguishes the arc.

[0026] Meanwhile, the source of the external signal may be a manual determination of whether the external smoke treatment device is activated, or a system detection of whether the external smoke treatment device is activated to make the determination.

[0027] Step 5: traverse the points in the laser contour line P and determine whether the current point Ai is an abnormal point; if the current point Ai is an abnormal point, then proceed to step 6; if the current point Ai is not an abnormal point, then proceed to step 10.

[0028] Here, there are several ways to determine whether the current point Ai is an outlier, as follows: Method 1: The judgment can be made based on the spatial continuity of the contour (that is, the changes between adjacent contour points are very small). By comparing the difference between the current point Ai and its previous and next points Ai-n and Ai+n (n is a preset value, indicating the number of previous and next points), if the difference exceeds the threshold, the current point Ai is judged as an outlier.

[0029] Method 2: Based on the temporal continuity of the contour (i.e., the change at the same point between the two previous and subsequent contours is minimal), the current point Ai is compared with the same point Ti in the template contour T. If the difference exceeds a threshold, the current point Ai is considered an outlier. The method for finding the same point Ti in the template contour T is to use the number x of the contour point where the current point Ai is located to find the point Ti in the template contour T with the same sequence number.

[0030] Method 3: Judge by the pixel values ​​around the contour point, with the contour point on the laser contour line P as the center and within a neighborhood with a radius of r; when not in the smoke area, the pixel value of the laser contour line P changes significantly, and the difference between each pixel is large; when in the smoke area, the pixel value of the laser contour line P changes smoothly; and by counting the variance of the pixel values ​​in the area, if the variance is less than the preset value, the current point Ai is judged to be an abnormal point.

[0031] Step 6: Find the point Ti with the same sequence number in the template contour T through the contour point number x where the current point Ai is located.

[0032] Step 7: Correct the abnormal point Ai based on the information around point Ti. The correction method is as follows: Method 1: By searching for the maximum value point Mi within the preset range above and below point Ti in the laser line image, the coordinates of point Mi are used to replace the current point Ai, thereby correcting the abnormal current point Ai.

[0033] Method 2: By calculating the grayscale centroid Gi of the pixels within the preset range above and below point Ti (that is, multiplying the grayscale value of each pixel by the row number of each pixel, dividing the result by the sum of the grayscale values ​​of all pixels in the range, taking the result as the row number, the column where point Ti is located as the column number, and the point formed by the row number and the column number as the grayscale centroid Gi), the coordinates of the grayscale centroid Gi are used to replace the previous point Ai, and the abnormal current point Ai is corrected. Here, the pixels within the preset range above and below point Ti can be within the range of five pixels above and below, ten pixels above and below, etc., and the specific selection can be based on actual conditions.

[0034] Step 8: Determine whether the points in the laser contour line P have been traversed. If the points in the laser contour line P have been traversed, proceed to step 11. If the points in the laser contour line P have not been traversed, proceed to step 5.

[0035] Step 9: Update the points in the laser contour line P to the template contour T and proceed to step 11.

[0036] Step 10: Update the non-abnormal points Ai in step 6 or the abnormal points Ai in step 7 after correction to the template contour T, and proceed to step 8.

[0037] Step 11: Output the contour P and return to step 2.

[0038] It should be understood by those skilled in the art that the above embodiments are merely for the purpose of illustrating the present invention clearly, and are not intended to limit the scope of the present invention. For those skilled in the art, other changes or modifications may be made based on the above disclosure, and these changes or modifications are still within the scope of the present invention.

Claims

1. A laser line profile extraction method that avoids interference from welding smoke, characterized in that: The steps include: Step 1: Create and initialize the template outline T; Step 2: Acquire laser line image; Step 3: Extract the laser contour line P; Step 4: Determine whether the anti-smoke function is turned on; Step 5: Traverse the points in the laser contour line P and determine the accuracy of the current point Ai; Step 6: If the current point Ai is an outlier, find the point Ti with the same sequence number in the template contour T; Step 7: Correct the abnormal point Ai based on the information around point Ti; Step 8: update the non-abnormal point Ai in step 5 or the abnormal point Ai in step 7 after correction to the template contour T, and traverse the laser contour line P through the methods in steps 5, 6 and 7; Step 9: Output the contour P and return to step 2.

2. The laser line profile extraction method for avoiding welding fume interference according to claim 1 is characterized in that: When it is determined in step 4 that the anti-smog function is not turned on, the points in the laser contour line P are updated to the template contour T, and the process proceeds to step 9.

3. The laser line profile extraction method for avoiding welding fume interference according to claim 1 or 2, characterized in that: When the laser contour line P is not traversed in step eight, return to step five.

4. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The method for extracting the laser contour line P in step 3 is the edge method, the grayscale centroid method or the Steger algorithm.

5. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The method for judging the accuracy of the current point Ai in step 5 is: by comparing the difference between the current point Ai and its previous and next points Ai-n and Ai+n, if the difference exceeds the threshold, the current point Ai is judged to be an abnormal point.

6. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The method for judging the accuracy of the current point Ai in step 5 is: by comparing the current point Ai with the same point Ti in the template contour T, if the difference exceeds the threshold, the current point Ai is judged to be an abnormal point.

7. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The method for judging the accuracy of the current point Ai in step 5 is: taking the contour point on the laser contour line P as the center, within a neighborhood with a radius of r, statistics the variance of the pixel values ​​in the neighborhood. If the variance is less than the preset value, it is judged as an abnormal point.

8. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The correction method of the abnormal point Ai in step 7 is: by searching the maximum value point Mi within the preset range above and below Ti in the laser line image obtained in step 2, the coordinates of point Mi are used to replace the abnormal point Ai.

9. The laser line profile extraction method for avoiding welding fume interference according to claim 1, characterized in that: The correction method for the abnormal point Ai in step 7 is: by calculating the grayscale centroid Gi of the pixels within the preset range above and below the point Ti in step 6, the coordinates of the point Gi are used to replace the abnormal point Ai.

Images