Narrow gap weld deviation detection method based on arc visual information and application

Abstract

The application discloses a narrow-gap weld deviation detection method based on arc visual information and application, collects a frame of welding image of an arc located at a left sidewall of a groove during a staying period, and a front frame of welding image is collected at a right sidewall of the groove; the frame of welding image is processed to obtain an arc contour fitting line, column scanning is performed on the arc contour fitting line to obtain coordinates of upper and lower limit points, and the upper limit point coordinates are the highest point coordinates of the arc; a difference between the highest point longitudinal coordinates of the frame of arc and the highest point longitudinal coordinates of the front frame of arc is obtained to obtain a weld height deviation; the arc contour center horizontal coordinates and the average of the front frame of arc center horizontal coordinates are used as the torch center line horizontal coordinates, a difference between the torch center line horizontal coordinates and a set torch center line horizontal coordinate threshold value before welding is obtained to obtain a left and right deviation value of the weld, synchronous detection with the arc position is realized, detection real-time performance and precision are improved, and the application is applied to detection of a torch center, left and right deviation of a weld, weld height deviation and real-time tracking control of a welding process.

Description

Technical Field

[0001] This invention belongs to the field of welding technology and relates to narrow gap gas metal arc welding, specifically a narrow gap weld deviation detection technology based on arc vision information. Background Technology

[0002] Currently, data-driven intelligent manufacturing in welding is developing. As a novel welding process, the oscillating arc narrow-gap gas metal arc welding (GMAW) method adjusts the arc heat distribution through periodic arc oscillation, improving the fusion of the narrow-gap groove sidewalls. It boasts advantages such as good arc directionality, less wear on the contact tip, and applicability to all-position welding. Combined with sensing technology, it can realize intelligent manufacturing of thick plate welding. However, factors such as groove machining errors, assembly errors, welding deformation, and groove shape can cause the welding torch center to deviate from the groove center and uneven weld penetration on both sides of the groove during the welding process. Therefore, real-time detection of weld deviations during narrow-gap welding is a crucial factor in achieving a more integrated and intelligent new industrialized modern industrial system.

[0003] Various methods for detecting narrow-gap weld deviations have been proposed both domestically and internationally, including arc sensing and vision sensing methods, primarily focusing on solving weld tracking problems. Among these, vision sensing methods have received widespread attention due to their advantages of being non-contact, providing rich information, and enabling synchronous detection with the arc position. However, during narrow-gap welding with a wiggling arc, not only are welding fumes, spatter, and arc light interference prominent, but there is also a lack of synchronization in weld deviation detection, posing a challenge to vision sensing. There is an urgent need to innovate vision sensing methods and anti-interference detection mechanisms to improve the applicability of vision sensing methods. However, most existing passive vision detection methods for narrow-gap weld deviations with wiggling arcs rely on obtaining weld deviation values ​​based on the bevel edge and the welding wire outline position.

[0004] The detection method disclosed in Chinese Patent Application No. 2024106592389, entitled "Passive Visual Sensing Anti-interference Detection Method and Application for Narrow Gap Weld Deviation," utilizes extracted arc feature points to locate the origin coordinates of the groove edge ROI window and the welding wire ROI window, and then detects weld deviation by using the groove center and welding torch center positions. Its disadvantages are: 1) After determining the arc feature points, this method extracts the groove edge and welding wire edge, increasing the difficulty of the detection algorithm and reducing real-time performance; 2) This method only detects the left and right deviations of the weld, failing to detect the weld height deviation, thus limiting the engineering application of the narrow gap weld deviation detection method in thick plates. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing methods for detecting weld deviations in narrow gap welding processes, such as poor synchronization, insufficient real-time performance, low detection efficiency, and limited information on weld deviations. This invention proposes a new method for detecting narrow gap weld deviations based on arc vision information, which offers better real-time performance and environmental adaptability. This method improves the synchronization, real-time performance, and multi-dimensional detection of weld deviations, enabling efficient detection in the narrow gap welding process and enhancing the applicability of narrow gap welding methods.

[0006] To achieve the above objectives, the present invention provides a narrow gap weld deviation detection method based on arc vision information, which employs the following technical solution, comprising the following steps:

[0007] Step 1): Acquire a frame of welding image showing the arc dwelling on the left side of the bevel. The previous frame of welding image was acquired from the right side of the bevel.

[0008] Step 2): Process the welding image of this frame to obtain the arc profile fitting line, and perform row and column scanning on the arc profile fitting line to obtain the coordinates of the upper, lower, left and right limit points of the arc profile fitting line. The coordinates of the upper limit point are the coordinates of the highest point of the arc. Connect the left and right limit points and connect the upper and lower limit points. The intersection of the two connecting lines is the center of the arc profile.

[0009] Step 3): Subtract the ordinate of the highest point of the arc in this frame from the ordinate of the highest point of the arc in the previous frame to obtain the weld height deviation; take the average of the abscissa of the arc contour center and the abscissa of the arc center in the previous frame as the abscissa of the welding torch centerline, and subtract the abscissa of the welding torch centerline from the abscissa of the welding torch centerline set before welding to obtain the left and right deviation values ​​of the weld.

[0010] The technical solution adopted in this invention for the narrow gap weld deviation detection method based on arc vision information is: to detect the center of the welding torch, the left and right deviations of the weld, the height deviation of the weld, and to track and control the welding process in real time.

[0011] Compared with the prior art, the advantages and beneficial effects of the present invention are mainly:

[0012] 1. In order to detect weld height deviation, this invention proposes two deviation detection methods for narrow gap welds in the left and right and height directions based on arc vision information. This solves the problem that the detection method disclosed in the document with application number 2024106592389 requires calculation of the groove center and welding torch center to obtain the left and right deviation of the weld. At the same time, it further optimizes the sensing detection synchronized with the arc position, realizes synchronous detection with the arc position, and improves the real-time performance and accuracy of the detection.

[0013] 2. This invention establishes an anti-interference extraction method for electric arc contours. Based on the acquired binary image of the electric arc, mathematical morphology operation is used to eliminate the influence of interference data on the extraction accuracy of the electric arc contour. Furthermore, by using a fitting method, the fitting line of the electric arc contour is obtained, thereby improving the recognition accuracy of the electric arc contour.

[0014] 3. The weld left and right deviation detection proposed in this invention obtains the weld left and right deviation based on the difference in arc contour width between two adjacent welding images, thus realizing the detection of weld left and right deviation synchronized with the arc position.

[0015] 4. The weld height deviation detection proposed in this invention can obtain the weld height deviation based on the difference in the position of the highest point of the arc or the difference in the height of the arc profile between two adjacent welding images, thus improving the applicability of monocular vision sensing methods.

[0016] 5. The dynamic detection of the welding torch center position proposed in this invention realizes the real-time dynamic detection of the welding torch center position by using the arc contour center position of two adjacent welding images, which optimizes the pre-welding static detection or dynamic feedback of the welding wire center position disclosed in the document with application number 2024106592389.

[0017] 6. This invention establishes a synchronous detection mechanism for weld deviations in different dimensions, including left and right and height, enabling real-time detection of weld deviations synchronized with the arc position. Starting from the extracted arc profile, synchronous, real-time, and efficient detection of weld deviations in both left and right and height can be achieved solely through the extracted welding arc profile, thus improving the real-time performance and adaptability of weld deviation detection. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of a captured frame of a narrow-gap welding image;

[0019] Figure 2 This is a flowchart of the narrow gap weld deviation detection method of the present invention;

[0020] Figure 3 An example of extracting the welding arc profile when weld deviation exists;

[0021] Figure 4 This is an example of the dynamic migration of the arc profile when weld deviation exists;

[0022] Figure 5 An example diagram of the dynamic migration of the arc profile when there is no weld deviation;

[0023] Figure 6 This is an example diagram showing the dynamic migration of the vertical coordinate of the highest point of the electric arc when weld deviation exists.

[0024] Figure 7This is an example diagram showing the dynamic migration of the vertical coordinate of the highest point of the electric arc when there is no weld deviation.

[0025] Figure 8 This is an example diagram of weld deviation detection;

[0026] Figure 9 This is an example of the detection results for the center position of the electric arc profile.

[0027] Figure 1 In the middle: 1—Welding image; 2—Electric arc; 3—Welding wire; 4—Left edge of the bevel; 5—Right edge of the bevel; 6—End of the molten pool; 7—Lead of the molten pool; 8—Centerline of the welding torch; O i —The origin of welding image 1, which is also the x-axis of the rectangular coordinate system. i -O i -y i The origin; w1—width of welding image 1; h1—height of welding image 1; A a1 —The electric arc is located at the midpoint of the arc profile when it stops on the left side wall of the bevel; A a2 —The electric arc is located at the midpoint of the arc profile when it stops on the right side wall of the bevel; x a1 —A a1 In the rectangular coordinate system x i -O i -y i x-coordinate; y a1 —A a1 In the rectangular coordinate system x i -O i -y i The ordinate; x a2 —A a2 In the rectangular coordinate system x i -O i -y i x-coordinate; y a2 —A a2 In the rectangular coordinate system x i -O i -y i The ordinate of d; L —A a1 The distance from the x-coordinate to the left edge of the bevel (4); d R —A a2 The distance from the x-coordinate to the right edge of the bevel (5); A i1 —The electric arc is located at its highest point when it stops on the left side wall of the bevel; A i2 —The electric arc is located at the highest point of the arc profile when it stops on the right side wall of the bevel; A L1 —The width of the arc profile when the arc is positioned on the left side wall of the bevel; A S1 —The height of the arc profile when the arc is stationary on the left side wall of the bevel. Detailed Implementation

[0028] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. However, the scope of protection of the present invention is not limited to the following embodiments. All technical solutions obtained by equivalent substitution or equivalent transformation are within the scope of protection of the present invention.

[0029] like Figure 1 As shown, during welding, the welding wire 3 extends between the left edge 4 and the right edge 5 of the narrow-gap groove. The electric arc 2, located at the end of the welding wire 3, periodically vibrates at a certain frequency within the groove encompassed by the left edge 4 and the right edge 5 of the narrow-gap groove. An infrared camera is used to acquire images, with the infrared camera at a certain angle α to the horizontal plane of the welding specimen. The angle α ranges from 0 ≤ α ≤ 180°.

[0030] An infrared camera captures a single frame of welding image 1, showing the arc 2 remaining on the left side wall of the bevel during its dwell period. Welding image 1 includes the end of the molten pool 6 and the front of the molten pool 7. The height of welding image 1 is h1, the width is w1, and the x-axis of welding image 1 is... i -O i -y i The coordinate system is based on the width direction x. i The axis direction, with the height direction as y i Axial direction. Welding image 1 is the image of this frame. The arcs in the preceding and following frames of welding images adjacent to welding image 1 are located on the right side wall of the bevel. For ease of understanding of the invention, in Figure 1 The welding wire 3 and the electric arc 2 in the previous frame welding image are shown in dashed lines.

[0031] Combination Figure 2 As shown, welding image 1 is processed using image processing methods. Image filtering methods, including median filtering, mean filtering, and Gaussian filtering, are used to obtain the filtered welding image 1. A segmentation threshold is obtained through grayscale histogram analysis. Using the segmentation threshold, a binary image of welding image 1 is obtained. Mathematical morphology methods, including morphological erosion, morphological dilation, morphological opening, and morphological closing operations, are used to further denoise the binary image. Based on the denoised binary image, edge extraction operators, including Roberts, Prewitt, Sobel, Laplacian, and Canny edge operators, are used to obtain the contour image of arc 2. Arc contour judgment methods, including arc diameter and arc circumference judgment methods, are used to determine the presence of interfering data, which is then removed to accurately identify the contour of arc 2. Finally, fitting methods, including curve fitting and ellipse fitting, are used to obtain the arc contour fitting line.

[0032] The arc profile fitting line is scanned horizontally to obtain the coordinates of the upper and lower limit points of the arc profile fitting line, where the upper limit point is also the highest point A of arc 2.i1 , the highest point A i1 The coordinates are (x i1 ,y i1 The highest point in the previous frame welding image is A. i2 .

[0033] The arc profile fitting line is scanned in columns, and the horizontal and vertical coordinates of the left and right limit points in the arc profile fitting line are searched in a certain vertical sequence.

[0034] Connect the left and right extreme points, and connect the top and bottom extreme points; the intersection of the two connecting lines is the center A of the electric arc profile. a1 A a1 The coordinates are (x a1 ,y a1 The center of the arc contour in the previous frame welding image is A. a2 A a2 The coordinates are (x a2 ,y a2 ).

[0035] According to the center A of the electric arc contour in this frame a1 x-coordinate a1 The center of the electric arc profile A can be obtained. a1 The distance from the left edge 4 of the bevel is d L Based on the arc contour center A in the previous frame welding image a2 x-coordinate a2 The center A of the arc profile in the previous frame can be obtained. a2 The distance from the right edge 5 of the bevel is d R .

[0036] Calculate the center A of the arc profile in this frame. a1 x-coordinate a1 With respect to the center A of the arc profile in the previous frame a2 x-coordinate a2 The average value is used to obtain the position of the welding torch centerline 8 at this time, which is also the center position of the welding torch.

[0037] For sampling of the center positions of multiple consecutive welding torches, a numerical filtering method is used to filter the sampling values ​​of adjacent N1 (3≤N1≤number of images acquired) welding torch center positions to obtain the center position of the welding torch.

[0038] Based on the highest point A of the electric arc i1 x-coordinate i1 Determine the direction in which arc 2 rests on the bevel sidewall. When arc 2 is located at the left edge 4 of the bevel, obtain the arc profile width A based on the difference in the abscissas of the left and right limit points of the arc profile fitting line. L1It can also be obtained from the projection of the straight line formed by the left and right extreme points onto the horizontal axis. The arc profile height A is obtained from the difference in the ordinates of the upper and lower extreme points of the fitted arc profile line. S1 .

[0039] Calculate the highest point A of the electric arc in this frame. i1 The ordinate y i1 The vertical coordinate y of the highest point of the arc in the previous frame i2 The difference is used to obtain the weld height deviation Δx. v The value can also be the arc profile height A. S1 The change in the arc profile height A of this frame is used to obtain the weld height deviation value. S1 The arc profile height A of the previous frame S2 The difference is calculated to obtain the weld height deviation Δx of the welding image in this frame. v The value of .

[0040] When the weld height deviation Δx v When Δx = 0, it indicates that there is no deviation in weld height; when Δx = 0, it indicates that there v When Δx < 0, it indicates that the weld height has decreased and there is a negative deviation in the weld height; when Δx v A value greater than 0 indicates an increase in weld height and a positive deviation in weld height.

[0041] The center A of the arc contour in this frame a1 x-coordinate a1 With the arc center A of the previous frame a2 x-coordinate a2 The mean value, i.e., the abscissa of the welding torch centerline 8, is subtracted from the threshold value of the abscissa of the welding torch centerline set before welding to obtain the left and right deviation value Δx of the weld. h Alternatively, the arc contour width A of this frame can be adjusted. L1 The arc contour width A of the previous frame and the current frame L2 By subtracting the values, we obtain the left and right deviations of the weld, Δx. h Or based on the center A of the arc contour in this frame. a1 x-coordinate a1 The distance d from the left edge of the bevel 4 L The center of the arc contour in the previous frame, A a2 The distance d between the horizontal coordinate of the location and the right edge 5 of the bevel. R Calculate d L and d R The difference is used to obtain the left and right deviations of the weld, Δx. h .

[0042] When the weld seam deviates left and right by Δx h When Δx = 0, it indicates that there is no deviation in the left or right of the weld; when Δx = 0 h When Δx < 0, it indicates that the arc is deflected to the right, i.e., the weld is deflected to the left; when .... hA value >0 indicates that the arc is deflected to the left, meaning the weld is deflected to the right.

[0043] The narrow gap weld deviation detection method is applied to multi-dimensional detection of welding torch center deviation, weld left and right deviation, weld height deviation, and real-time tracking and control of the welding process.

[0044] The following are seven specific embodiments of the present invention.

[0045] Example 1

[0046] Figure 3 The image shown is an extraction diagram of the welding arc profile when weld deviation exists. Figure 3 Figures (a1) and (a2) are two consecutive welding images captured by an infrared camera, showing the arc stopping at the right and left walls of the bevel, respectively. The base material of the welded specimen is Q370qE steel plate, the bevel depth is 20mm, and the gap is 14mm; the DC arc current is 315A, the arc voltage is 30V, and the welding speed is V. w The welding speed is 204 mm / min, the wire extension of welding wire 3 is 20 mm, the diameter of welding wire 3 is 1.2 mm, the welding shielding gas is Ar-20% CO2, and the flow rate is 20 L / min. The arc oscillation frequency of arc 2 is 2.5 Hz, the oscillation angle is 82°, and the dwell time on both sides of the bevel is 100 ms.

[0047] Figure 3 Figures (b1) and (b2) in the figure are the edge extraction images of the welding images shown in Figures (a1) and (a2), respectively. A segmentation threshold of 250 was used to obtain the binary images of the welding images. Then, the Canny edge extraction method was applied to extract the edge images as shown in Figures (b1) and (b2), respectively. Due to the presence of arc reflection and spatter, the extracted edge lines not only include the arc edge but also other welding interference edge lines, as shown in the area indicated by the red circle. The diameter of each contour was calculated using the arc diameter judgment method. When the searched diameter was less than 20 pixels, the proposed contour was considered invalid. For the obtained binary images, a 3×3 elliptical morphological kernel was used to perform three mathematical morphological opening operations, resulting in the obtained arc contours as shown in Figures (b1 and (b2)). Figure 3 As shown in Figures (c1) and (c2), the method of arc diameter determination and morphological opening operation effectively eliminates the influence of welding interference on arc contour recognition, and can obtain a real arc contour with a smooth and interference-free shape, providing a reliable technical foundation for the next step of arc feature recognition and weld deviation calculation.

[0048] Example 2

[0049] Figure 4 The image shown is a dynamic migration diagram of the arc profile when weld deviation exists. 301 consecutive frames of welding images were acquired using an infrared camera, with welding conditions and... Figure 3 consistent. At the start of welding, the arc is offset 1mm toward the right side wall of the groove. As welding proceeds, the arc gradually deviates from the right side wall of the groove and shifts 1mm toward the left side wall of the groove. Using the above image processing method, 301 obtained dynamic diagrams of the arc profile are shown in Figure 4 . As can be seen from Figure 4 , due to the existence of weld deviation, when the arc swings toward the side wall of the groove, the side wall squeezes the arc shape, resulting in a reduction in the width of the arc profile; when the arc deviates from the side wall of the groove, the arc shape gradually expands, and the width of the arc profile also gradually increases. Specifically, when the arc is offset toward the right side wall of the groove, the extracted width of the right arc profile is relatively small, and the extracted width of the left arc profile is relatively large, that is, the distance d1 from the left limit point of the arc profile to the left edge of the welding image is greater than the distance d2 from the right limit point of the arc profile to the right edge of the welding image; as the arc gradually shifts toward the left side wall of the groove, the extracted width of the left arc profile gradually decreases, and the width of the right arc profile gradually increases, that is, d1<d2. During the movement of the arc from being offset toward the right side wall of the groove to being offset toward the left side wall of the groove, the arc profile migrates from right to left, and the horizontal displacement of the arc gradually increases as the weld deviation increases.

[0050] Example 3

[0051] Figure 5 shows the dynamic migration diagram of the arc profile when there is no weld deviation. An infrared camera is used to collect 248 consecutive frames of welding images, the welding current and voltage are slightly increased compared with Figure 3 , there is no weld deviation, and the rest of the welding conditions are consistent with Figure 3 . At the start of welding, the arc is located at the center of the groove, with the same distance to both side walls, and the horizontal arc displacement of the 248 obtained arc profiles is significantly smaller than that in the case with weld deviation in Figure 4 . Therefore, the weld deviation value can be characterized by the parameter change of the arc profile.

[0052] Example 4

[0053] Figure 6 shows the dynamic migration diagram of the ordinate of the highest point of the arc when there is a weld deviation. According to the 301 frames of welding images collected in Figure 4 , the sampling values of the ordinate of the highest point of the arc are obtained by a row search method, as shown in Figure 6 . The positions of the highest point of the arc when the arc is located at the right side wall and the left side wall of the groove are respectively indicated by the solid circle and the hollow circle in Figure 6 .

[0054] The welding process reaches stability when the weld length is ~10mm. Therefore, between approximately 10 and 88mm in weld length, as the arc deviates from right to left, the arc reaches its highest point on the right side of the bevel wall, which is higher than when the arc resides on the left side of the bevel wall. Between approximately 88 and 112mm in weld length, the arc is in a centered state, and the changes in the highest points of the extracted arc on the left and right sides are relatively small. However, between approximately 112 and 205mm in weld length, the arc gradually deviates towards the left side of the bevel wall, and the arc reaches its highest point on the right side of the bevel wall, which is lower than when the arc resides on the left side of the bevel wall. Figure 6 In the figure, the range of the change in the position of the highest point of the electric arc is 74 pixels.

[0055] Example 5

[0056] Figure 7 The image shows the dynamic migration of the vertical coordinate of the highest point of the electric arc when there is no weld seam deviation. Figure 5 The 248 welding images collected showed stable fluctuations in the arc's highest point position during its dwell period on the right side of the bevel and the arc's highest point position during its dwell period on the left side of the bevel, because there was no weld deviation.

[0057] Therefore, by calculating the position of the highest point of the electric arc, the change in arc height can be characterized, and the weld height deviation can be obtained. Similarly, the weld height deviation can be obtained by the difference in the arc profile height.

[0058] Example 6

[0059] Figure 8 The image shown is an example of weld deviation detection. The left-right and height deviations of the weld are obtained by analyzing two adjacent welding images. Figure 4 Of the 301 welding images acquired, two adjacent images were selected, located approximately 25 mm in length from the weld. The effectiveness of the proposed method was demonstrated by detecting the lateral deviation and height deviation of the weld.

[0060] Figure 8 The middle image (a1) shows a welding image during the dwell period of the electric arc on the right side wall of the bevel. By performing median filtering, histogram analysis, and threshold segmentation on the image, with the segmentation threshold set to 240, the resulting binary image is shown below. Figure 8 The white area in Figure (a2) is used to extract the arc contour from the thresholded binary image using the Canny algorithm, followed by ellipse fitting. The resulting arc contour fitting line is shown below. Figure 8The elliptical region in Figure (a2) is analyzed. The upper and lower limit points of the arc profile fitting line are obtained by searching the rows. The coordinates of the upper limit point, i.e., the highest point of the arc, are 403 x-coordinate and 208 y-coordinate. The arc profile height is obtained by calculating the y-coordinates of the upper and lower limit points. The left and right limit points of the arc profile fitting line are obtained by searching the columns. The arc profile width is obtained by calculating the difference between the x-coordinates of the left and right limit points. The center coordinates of the arc profile are obtained as (398, 149) by calculating the intersection of the arc profile width and arc profile height.

[0061] Figure 8 Figure (b1) shows a welding image of the arc during its dwell period on the left side wall of the bevel, which is consistent with... Figure 8 The image in Figure (a1) is a welding image adjacent to the previous frame. By analyzing... Figure 8 Figure (b1) in the middle is compared with Figure 8 The same image processing as Figure (a1) yields a binary image as shown below. Figure 8 The white area in Figure (b2) is used to extract the arc contour from the thresholded binary image using the Canny algorithm, followed by ellipse fitting. The resulting arc contour fitting line is shown below. Figure 8 The elliptical region in Figure (b2) is obtained by fitting the arc profile using a row scan. The x-coordinate of the highest point of the arc is 159, and the y-coordinate is 200. The center coordinates of the arc profile are obtained as (176, 145) by the intersection of the arc profile width and the arc profile height.

[0062] from Figure 8 It can be seen that by obtaining the horizontal coordinate of the center position of the arc contour, the center position of the welding torch can be obtained as (398+176) / 2 pixels. Subtracting this from the set center position of the welding torch (310 pixels), the difference is -23 pixels, which means that the left and right deviation of the weld is approximately △x1=-0.92mm.

[0063] from Figure 8 It can be seen that by obtaining the difference in the vertical coordinate of the highest point of the electric arc, the weld height can be reduced by 8 pixels, with a difference of approximately -0.32mm.

[0064] Alternatively, a curve fitting method can be used to obtain the arc profile fitting line. Then, by calculating the arc profile width and arc profile height, the left-right deviation and weld height deviation can be obtained, respectively.

[0065] Example 7

[0066] Figure 9 The image shown is an example of arc profile center position detection. According to... Figure 4 The 301 captured welding images were used to obtain the arc profile fitting line through ellipse fitting, and the center position of the arc profile was calculated. The obtained arc profile center position data is as follows: Figure 9 As shown. Because the weld seam deviates to the right first and then to the left, the offset of the arc profile center position also initially deviates to the right, and then gradually deviates to the left as welding progresses. Therefore, the arc profile center position can be used to detect weld seam deviations.

[0067] The above description is only a specific embodiment of some aspects of the present invention. Of course, the present invention may have many other embodiments. Without departing from the spirit and essence of the present invention, any person skilled in the art can make various corresponding equivalent changes and modifications based on the present invention, and all such changes and modifications should fall within the protection scope of the present invention.

Claims

1. A method for detecting narrow gap weld deviations based on arc vision information, characterized in that: Includes the following steps: Step 1): Acquire a frame of welding image showing the arc dwelling on the left side of the bevel. The previous frame of welding image was acquired from the right side of the bevel. Step 2): Process the welding image of this frame to obtain the arc profile fitting line, and perform row and column scanning on the arc profile fitting line to obtain the coordinates of the upper, lower, left and right limit points of the arc profile fitting line. The coordinates of the upper limit point are the coordinates of the highest point of the arc. Connect the left and right limit points and connect the upper and lower limit points. The intersection of the two connecting lines is the center of the arc profile. Step 3): Subtract the ordinate of the highest point of the arc in this frame from the ordinate of the highest point of the arc in the previous frame to obtain the weld height deviation; take the average of the abscissa of the arc contour center and the abscissa of the arc contour center in the previous frame as the abscissa of the welding torch centerline, and subtract the abscissa of the welding torch centerline from the abscissa of the welding torch centerline set before welding to obtain the left and right deviation values ​​of the weld.

2. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: In step 3), the method for obtaining the left and right deviation values ​​of the weld is replaced by: obtaining the distance d between the center of the arc profile and the left edge of the bevel based on the abscissa of the arc profile center in this frame. L The distance d between the center of the arc profile in the previous frame and the right edge of the bevel is obtained from the x-coordinate of the center of the arc profile in the previous frame welding image. R , will d L With d R The difference is taken as the left and right deviation of the weld.

3. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: In step 3), the method for obtaining the left and right deviation values ​​of the weld is replaced by: determining the direction in which the arc stops on the side wall of the groove based on the abscissa of the highest point of the arc; when the arc is located at the left edge of the groove, taking the difference between the abscissas of the left and right limit points of the arc profile fitting line as the arc profile width; and subtracting the arc profile width of the current frame from the arc profile width of the previous frame to obtain the left and right deviation values ​​of the weld.

4. The method for detecting narrow gap weld deviation according to claim 3, characterized in that: In step 3), the arc profile width is obtained by projecting the straight line connecting the left and right extreme points onto the horizontal coordinate axis.

5. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: In step 3), the method for obtaining the weld height deviation is replaced by: obtaining the arc profile height based on the difference between the ordinates of the upper and lower limit points of the arc profile fitting line, and obtaining the weld height deviation by subtracting the arc profile height of the current frame from the arc profile height of the previous frame.

6. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: When the weld height deviation △x v When Δx = 0, there is no deviation in weld height; when Δx = 0, there is no deviation in weld height v When Δx < 0, the weld height decreases, and there is a negative deviation in the weld height; when Δx v When the value is greater than 0, the weld height increases, and there is a positive deviation in the weld height.

7. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: When the left and right deviation of the weld is △x h When Δx = 0, there is no deviation in the left and right sides of the weld; when Δx = 0, there is no deviation in the left and right sides of the weld. h When Δx < 0, the arc deflects to the right, and the weld deflects to the left; when .... h When the value is greater than 0, the arc deflects to the left and the weld deflects to the right.

8. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: In step 1), an infrared camera is used to acquire images. The infrared camera is at an angle α of 0 ≤ α ≤ 180 degrees with the horizontal plane of the narrow gap weldment.

9. The method for detecting narrow gap weld deviation according to claim 1, characterized in that: In step 2), the filtered welding image is first obtained by image filtering method, the binary image of the welding image is obtained by segmentation threshold, the binary image is then denoised by mathematical morphology method, the arc contour image is obtained by edge extraction operator, the arc contour is then identified by arc contour judgment method, and finally the arc contour fitting line is obtained by fitting method.

10. An application of the narrow gap weld deviation detection method as described in any one of claims 1-9, characterized in that: It is used for detecting the center deviation of the welding torch, the left and right deviation of the weld, the height deviation of the weld, and for real-time tracking and control of the welding process.

Images