A vision-based narrow gap weld seam tracking control method

Abstract

The present application provides a kind of visual-based narrow gap weld tracking control method, comprising the following steps: S1, the welding image when the narrow gap weld is welded is acquired;S2, based on the position information of groove edge, the position information of tungsten electrode edge is obtained based on the welding image;S3, based on the position information of groove edge, the weld center line is calculated, and based on the position information of tungsten electrode edge, the welding torch head position is calculated;S4, based on the weld center line and the welding torch head position, the relative position relationship of welding torch head and weld center line is calculated;S5, based on the relative position relationship of welding torch head and weld center line, welding torch head position is adjusted.The present application uses monocular passive vision sensor, realizes real-time acquisition of welding area image, so that the timeliness of narrow gap weld tracking control is good, and the robustness is high, effectively reduces position error and time lag, thereby avoiding the advanced detection error caused by factors such as thermal deformation, and the accuracy of tracking control is significantly improved.

Description

Technical Field

[0001] This invention belongs to the field of weld seam tracking and image processing technology, and more specifically relates to a vision-based narrow gap weld seam tracking control method. Background Technology

[0002] Narrow-gap oscillating tungsten inert gas welding (NGS-GTAW) is widely used in the production of large structural components with high comprehensive mechanical performance requirements due to its excellent welding characteristics. It is especially suitable for single-layer, multi-pass narrow-gap bevel welding of medium and thick plates. To ensure sidewall fusion, the welding torch is generally oscillating. However, dimensional and positional deviations often occur in the processing of bevels for welding large components. Existing welding equipment lacks intelligent sensing and control technology for the oscillating welding process and lacks the self-adaptive capability to deviations in bevel processing. It cannot achieve self-adaptive control of the welding process under conditions such as torch oscillation, changing bevel form, and heat accumulation deformation. Therefore, welders need to make corrections and interventions for a long time to ensure good welding quality. Insufficient welding experience, differences in technical level, and work fatigue pose significant hidden dangers to the welding quality of engineering equipment.

[0003] Under varying welding conditions, the key to stabilizing welding quality and improving welding efficiency by reducing reliance on manual experience during the welding process lies in adopting intelligent weld seam tracking control technology with good real-time performance and strong robustness. Monocular passive vision sensing systems directly acquire images of the welding area, obtaining more comprehensive and intuitive information, which is highly beneficial for adaptive control of welding.

[0004] The key to realizing weld seam tracking control technology based on passive vision sensing lies in the stable and reliable acquisition and extraction of image information. However, due to factors such as arc light interference and electromagnetic effects during image data transmission and conversion, image quality deteriorates, affecting the extraction of feature information. This greatly limits the application of passive vision sensing technology in narrow gap weld seam tracking control.

[0005] In summary, the setup strategy for the passive vision-based narrow-gap weld seam tracking control method needs to be optimized. Summary of the Invention

[0006] Based on the aforementioned shortcomings and deficiencies in the prior art, one of the objectives of this invention is to at least solve one or more of the aforementioned problems in the prior art. In other words, one of the objectives of this invention is to provide a vision-based narrow gap weld seam tracking control method that meets one or more of the aforementioned requirements.

[0007] To achieve the above-mentioned objectives, the present invention adopts the following technical solution:

[0008] This invention provides a vision-based narrow-gap weld seam tracking and control method, comprising the following steps:

[0009] S1. Obtain welding images during narrow gap weld welding;

[0010] S2. Based on the welding image, obtain the position information of the bevel edge and the position information of the tungsten electrode edge;

[0011] S3. Calculate the weld centerline based on the position information of the bevel edge, and calculate the welding torch position based on the position information of the tungsten electrode edge;

[0012] S4. Based on the weld centerline and the position of the welding torch, calculate the relative positional relationship between the welding torch and the weld centerline;

[0013] S5. Adjust the position of the welding torch head based on the relative positional relationship between the welding torch head and the center line of the weld.

[0014] As a preferred option, in step S2:

[0015] The location information of the bevel edge includes the straight line information of the right bevel edge and the straight line information of the left bevel edge;

[0016] The position information of the tungsten electrode edge includes the straight line information of the right tungsten electrode edge and the straight line information of the left tungsten electrode edge.

[0017] As a preferred embodiment, in step S3, calculating the weld centerline based on the position information of the bevel edge includes:

[0018] Calculate the equation of the right bevel edge line based on the coordinates of any two points located on the right bevel edge line;

[0019] Calculate the equation of the left bevel edge line based on the coordinates of any two points located on the left bevel edge line;

[0020] Based on the equations of the right bevel edge line and the left bevel edge line, the angle bisector of the angle formed by the right bevel edge line and the left bevel edge line is calculated to obtain the weld centerline.

[0021] As a preferred embodiment, in step S3, the equation of the weld centerline is:

[0022]

[0023] Among them, (x C1 y C1 (x) represents a point on the center line of the weld. C2 y C2 ) represents another point on the center line of the weld.

[0024] As a preferred embodiment, in step S3, calculating the welding torch tip position based on the position information of the tungsten electrode edge includes:

[0025] Calculate the equation of the right tungsten electrode edge line based on the coordinates of any two points located on the right tungsten electrode edge line;

[0026] Calculate the equation of the left tungsten electrode edge line based on the coordinates of any two points located on the left tungsten electrode edge line;

[0027] Based on the equations of the right tungsten electrode edge line and the left tungsten electrode edge line, the intersection point of the reverse extension of the right tungsten electrode edge line and the reverse extension of the left tungsten electrode edge line is calculated to obtain the position of the welding torch head.

[0028] As a preferred embodiment, in step S4, the equation for the relative positional relationship between the welding torch tip and the weld centerline is:

[0029]

[0030] Where D represents the positional difference between the welding torch tip and the center of the weld, (x C1 y C1 (x) represents a point on the center line of the weld bead. C2 y C2 (x0, y0) represents another point on the center line of the weld bead, and (x0, y0) represents the coordinates of the welding torch head.

[0031] As a preferred embodiment, the positional difference D between the welding torch tip and the weld center represents the relative positional relationship between the welding torch tip and the weld centerline, specifically as follows:

[0032] When the position difference D between the welding torch tip and the weld center is 0, it indicates that the welding torch tip is on the weld centerline;

[0033] When the position difference D between the welding torch tip and the weld center is less than 0, it indicates that the welding torch tip is located to the left of the weld centerline;

[0034] When the position difference D between the welding torch tip and the weld center is greater than 0, it indicates that the welding torch tip is located to the right of the weld centerline.

[0035] As a preferred embodiment, in step S5, adjusting the position of the welding torch head based on the relative positional relationship between the welding torch head and the weld centerline includes:

[0036] A preset correction accuracy K is used to compare the position difference D between the welding torch head and the weld center with the preset correction accuracy K, and the position of the welding torch head is adjusted based on the comparison result.

[0037] When -K < D < K, do not adjust the welding torch position;

[0038] When D < -K, adjust the welding torch head position to the right;

[0039] When D > K, adjust the welding torch head position to the left.

[0040] As a preferred embodiment, in step S1, acquiring the welding image during narrow gap weld welding includes:

[0041] When the welding torch swings to the left limit position and stops, the extracted welding image contains the straight line information of the right bevel edge and the straight line information of the right tungsten electrode edge;

[0042] When the welding torch swings to its right limit and stops, the extracted welding image contains the straight line information of the left bevel edge and the straight line information of the left tungsten electrode edge.

[0043] As a preferred embodiment, the following step is further included between step S1 and step S2:

[0044] The average grayscale value of the welding image is obtained, and the average grayscale value is subtracted from the grayscale value of the welding image to enhance the contrast of the bevel edge and the contrast of the tungsten electrode edge in the welding image.

[0045] Set the parameters of interest, and locate the region of interest in the welding image based on the parameters of interest.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] This invention employs monocular passive vision sensing to achieve real-time acquisition of welding area images, resulting in good timeliness and robustness of narrow gap weld seam tracking control. It effectively reduces positional errors and time lag, thereby avoiding advance detection errors caused by factors such as thermal deformation, and significantly improving the accuracy of tracking control.

[0048] Further or more detailed beneficial effects will be described in conjunction with specific embodiments in the detailed implementation. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a schematic diagram illustrating an application scenario of the vision-based narrow gap weld seam tracking control method described in this invention.

[0051] Figure 2This is a flowchart of a vision-based narrow gap weld seam tracking control method according to the present invention.

[0052] Figure 3 This is a schematic diagram illustrating the calculation of the welding torch head position in the vision-based narrow gap weld seam tracking control method described in this invention.

[0053] Figure 4 This is a schematic diagram illustrating the calculation of the relative positional relationship between the welding torch head and the weld centerline in the vision-based narrow gap weld tracking control method described in this invention.

[0054] Figure 5 This is an image showing the enhanced contrast of the bevel edge after processing the welding image in the vision-based narrow gap weld tracking control method described in this invention.

[0055] Figure 6 This is an image showing the enhanced contrast of the tungsten electrode edge after processing the welding image in the vision-based narrow gap weld seam tracking control method described in this invention.

[0056] Figure 7 This is a schematic diagram of the process of using the vision-based narrow gap weld seam tracking control method described in this invention in actual use. Detailed Implementation

[0057] The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0058] In the following description, several embodiments of this application are provided. Different embodiments can be substituted or combined. Therefore, this application can also be considered to include all possible combinations of the same and / or different embodiments described. Thus, if one embodiment includes features A, B, and C, and another embodiment includes features B and D, then this application should also be considered to include embodiments containing one or more other possible combinations of A, B, C, and D, even if such embodiments are not explicitly described in the following text.

[0059] The following description provides examples and does not limit the scope, applicability, or examples set forth in the claims. Changes may be made to the function and arrangement of the described elements without departing from the scope of this application. Various processes or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with respect to some examples may be combined into other examples.

[0060] To facilitate a better understanding of the embodiments of this application, the application scenarios will be explained before providing a detailed explanation of the specific implementation methods.

[0061] The narrow gap weld tracking control method provided in the embodiments of this specification can be applied to the welding process of medium and thick plates of large components in the field of intelligent welding. In these scenarios, the application of the tracking control method aims to reduce the dependence of the welding process on human experience, stabilize welding quality, and improve welding efficiency.

[0062] Please see Figure 1 , Figure 1 This diagram illustrates an application scenario of the tracking control method. The welding torch moves along the welding direction with the main motion mechanism, oscillating within the bevel range. The image acquisition card acquires weld information sensed by the CCD camera through the oscillation signal. This information is then processed by the software system in the embedded control system to obtain the welding torch deviation information. A control command is then issued in a specific format. The control card generates corresponding pulse instructions, and the servo driver, upon receiving the instructions, generates a corresponding number of pulses, driving the servo motor in the actuator to perform the corresponding action. The actuator includes a motor that moves the welding torch perpendicular to the weld direction and a servo motor that oscillates around the weld direction. In this embodiment, to clearly sense the bevel, molten pool, and tungsten electrode information, the welding torch and CCD camera maintain a fixed relative position. Specifically, the angle between the camera and the direction perpendicular to the workpiece is set between 75° and 85°, and the distance between the camera and the welding torch is maintained at approximately 0.35m to keep the field of view constant.

[0063] The following is a brief explanation of the narrow gap welding, beveling, weld centerline, welding torch, tungsten electrode, welding torch tip, and welding images involved in the embodiments of this specification:

[0064] Narrow gap welding is a welding method suitable for welding thicker workpieces or large structural components. The gap between the weld seams is relatively small, typically ranging from a few millimeters to tens of millimeters.

[0065] A bevel is a groove or flange formed by cutting or grooving on a workpiece that needs to be welded. The shape of the bevel can be determined according to the welding requirements; common shapes include V-grooves, U-grooves, and J-grooves. It should be noted that the bevel in the embodiments of this specification is a U-grooved bevel.

[0066] The weld centerline is the center line of the weld during the welding process; it is the midpoint between the weld edges on both sides of the welding area.

[0067] A welding torch is a handheld tool used for welding, typically consisting of a handle, cable, and electrodes. It guides the electric arc to the welding area and controls the stability and position of the arc during the welding process.

[0068] Tungsten electrodes are a commonly used electrode material, frequently employed in TIG (Tungsten Inert Gas) welding or argon arc welding. Tungsten electrodes possess a high melting point, good electrical conductivity, corrosion resistance, and stable arc characteristics. They are used to generate and maintain the electric arc, heating the welding area to a molten state through the heat transfer of the arc in narrow-gap welding.

[0069] The welding torch tip, also known as the torch nozzle or tip, is a part of the welding torch. It is the front end of the torch and comes into direct contact with the tungsten electrode and the welding area. The design and shape of the welding torch tip have a significant impact on weld quality and weld appearance.

[0070] Understandably, during the welding process, it is necessary to keep the positional difference between the welding torch tip and the center of the weld within the preset correction accuracy range in order to ensure the quality and accuracy of the weld, and to ensure the strength and stability of the weld.

[0071] Welding images refer to visual information related to welding obtained through image recording or image processing techniques. They can be images of the welding process captured in real time by cameras or other image acquisition devices.

[0072] Please see Figure 2 , Figure 2 A schematic flowchart of a vision-based narrow gap weld seam tracking control method provided in an embodiment of this specification is shown.

[0073] like Figure 2 As shown, this embodiment provides a vision-based narrow-gap weld seam tracking control method, including the following steps:

[0074] S1. Obtain welding images during narrow gap weld welding;

[0075] S2. Based on the welding image, obtain the position information of the bevel edge and the position information of the tungsten electrode edge;

[0076] S3. Calculate the weld centerline based on the position information of the bevel edge, and calculate the welding torch position based on the position information of the tungsten electrode edge;

[0077] S4. Based on the weld centerline and the position of the welding torch, calculate the relative positional relationship between the welding torch and the weld centerline;

[0078] S5. Adjust the position of the welding torch head based on the relative positional relationship between the welding torch head and the center line of the weld.

[0079] Specifically, this embodiment provides a preferred method for step S2, in which: the position information of the bevel edge includes the straight line information of the right bevel edge and the straight line information of the left bevel edge; the position information of the tungsten electrode edge includes the straight line information of the right tungsten electrode edge and the straight line information of the left tungsten electrode edge.

[0080] Specifically, this embodiment provides a preferred method for step S3. In step S3, calculating the weld centerline based on the position information of the bevel edge includes: calculating the equation of the right bevel edge line based on the coordinates of any two points on the right bevel edge line; calculating the equation of the left bevel edge line based on the coordinates of any two points on the left bevel edge line; and calculating the angle bisector of the angle formed by the right bevel edge line and the left bevel edge line based on the equation of the right bevel edge line and the equation of the left bevel edge line, so as to obtain the weld centerline.

[0081] It is understandable that, due to perspective distortion in the two-dimensional field of view of visual sensing, the straight lines of the left and right bevels of the weld are not parallel. According to the angle bisector theorem, a point on the angle bisector is equidistant from the two sides of the angle. That is, the weld centerline lies on the angle bisector of the straight lines at the edges of the left and right bevels. Therefore, this invention obtains the weld centerline information through the straight line information of the two bevels.

[0082] Specifically, this embodiment provides a preferred method for step S3, in which the equation of the weld centerline is:

[0083]

[0084] Among them, (x C1 y C1 (x) represents a point on the center line of the weld. C2 y C2 ) represents another point on the center line of the weld.

[0085] Please see Figure 3 , Figure 3 This diagram illustrates the calculation of the welding torch head position in a vision-based narrow gap weld seam tracking control method provided in an embodiment of this specification.

[0086] Specifically, this embodiment provides a preferred method for step S3. In step S3, calculating the welding torch position based on the position information of the tungsten electrode edge includes: calculating the equation of the right tungsten electrode edge line based on the coordinates of any two points on the right tungsten electrode edge line; calculating the equation of the left tungsten electrode edge line based on the coordinates of any two points on the left tungsten electrode edge line; and calculating the intersection of the reverse extension of the right tungsten electrode edge line and the reverse extension of the left tungsten electrode edge line based on the equation of the right tungsten electrode edge line and the equation of the left tungsten electrode edge line, so as to obtain the welding torch position.

[0087] like Figure 3As shown, regardless of whether the tungsten electrode swings in a pendulum or horizontal manner, it always oscillates symmetrically. Therefore, the backward extension lines of the tungsten electrode edges at the left and right positions are always at the center of the welding torch. It is understandable that during narrow-gap welding, the bevel is stationary while the welding torch oscillates symmetrically within the bevel range, with the fastest linear velocity at the weld center. To directly represent the welding torch position by acquiring image information of the torch swinging to the weld center, two conditions are required: a high frame rate from the CCD camera and high accuracy in detecting the encoder's zero-point position. Furthermore, the camera triggering requires response time. Therefore, acquiring image information of the welding torch swinging to the center position involves many uncertainties, severely affecting the accuracy of the acquisition system. However, to ensure sidewall fusion during welding, the welding torch needs to pause for 0.2s-0.5s at each of the left and right sidewall positions (left and right bevels). This meets the time requirement for welding image acquisition. Under these conditions, the hardware requirements for acquiring welding images are not high, and it is easy to achieve encoder signal triggering of the camera to acquire accurate position information. There are fewer cases where the trigger signal and image position do not correspond. Furthermore, the weld image information at the left and right pause positions fully exposes one edge of the tungsten electrode, avoiding the welding wire from obscuring the tungsten electrode information when the swing angle is large. Therefore, this invention obtains the welding torch head position by acquiring the right edge of the tungsten electrode when the welding torch swings to the left limit position and the left edge of the tungsten electrode when the welding torch swings to the right limit position, and calculating the intersection point of the reverse extension line of the right tungsten electrode edge line and the reverse extension line of the left tungsten electrode edge line.

[0088] Specifically, this embodiment provides a preferred method for step S4, in which the equation for the relative positional relationship between the welding torch head and the weld centerline is:

[0089]

[0090] Where D represents the positional difference between the welding torch tip and the center of the weld, (x C1 y C1 (x) represents a point on the center line of the weld bead. C2 y C2 (x0, y0) represents another point on the center line of the weld bead, and (x0, y0) represents the coordinates of the welding torch head.

[0091] Please see Figure 4 , Figure 4 This diagram illustrates the calculation of the relative positional relationship between the welding torch head and the weld centerline in a vision-based narrow gap weld tracking control method provided in an embodiment of this specification.

[0092] like Figure 4As shown, the calculation process for the relative positional relationship between the welding torch tip and the weld centerline includes establishing a rectangular coordinate system with the upper left corner of the welding image as the origin, the length direction of the welding image as the X-axis, and the width direction of the welding image as the Y-axis. Based on the welding image, the positional information of the bevel edge and the positional information of the tungsten electrode edge are obtained. The positional information of the bevel edge includes two coordinate points (x, y, y) on the straight line of the right bevel edge. R1 y R1 ) and (x R2 y R2 ), the two coordinate points (x) on the straight line at the left bevel edge L1 y L1 ) and (x L2 y L2 The positional information of the tungsten electrode edge includes two coordinate points (x, y, y) on the straight line of the right tungsten electrode edge. R0 y R0 ) and (x R01 y R01 ), the two coordinate points (x) on the straight line of the left tungsten electrode edge L0 y L0 ) and (x L01 y L01 It should be noted that the embodiments in this specification have undergone effective information filtering for the position information of the bevel edge and the tungsten electrode edge obtained from the welding images. Specifically, during the welding process, the changing trajectories of the left and right bevel edge positions, the left and right tungsten electrode edge positions, and the target detection position are continuously extracted in real time. Differential calculations are performed on each position information, and if the rate of change exceeds a certain standard, the information is excluded. This operation can prevent the detection system from issuing erroneous control signals, affecting the normal welding process, and further ensures the stability of weld seam tracking control.

[0093] Specifically, this embodiment provides a preferred method in which the positional difference D between the welding torch tip and the weld center represents the relative positional relationship between the welding torch tip and the weld centerline. Specifically, when the positional difference D between the welding torch tip and the weld centerline is 0, it indicates that the welding torch tip is on the weld centerline; when the positional difference D between the welding torch tip and the weld centerline is <0, it indicates that the welding torch tip is on the left side of the weld centerline; and when the positional difference D between the welding torch tip and the weld centerline is >0, it indicates that the welding torch tip is on the right side of the weld centerline.

[0094] Specifically, this embodiment provides a preferred method for step S5, in which adjusting the position of the welding torch head based on the relative positional relationship between the welding torch head and the weld centerline includes:

[0095] A preset correction accuracy K is set, and the position difference D between the welding torch head and the weld center is compared with the preset correction accuracy K. The position of the welding torch head is adjusted based on the comparison result. When -K < D < K, the position of the welding torch head is not adjusted. When D < -K, the position of the welding torch head is adjusted to the right. When D > K, the position of the welding torch head is adjusted to the left.

[0096] Understandably, in order to ensure the stability of the control system, a reasonable control precision K is set so that the D value is kept within a certain range. This means that the welding torch is considered to be in the center of the weld. This position judgment method is robust and has a wide range of applications. Even if there are different degrees of perspective distortion in the camera's field of view or if the field of view direction deviates from the weld direction, it can still make an accurate judgment.

[0097] Specifically, this embodiment provides a preferred method for step S1, in which obtaining the welding image during narrow gap welding includes: when the welding torch swings to the left limit position and stops, extracting a welding image containing the straight line information of the right bevel edge and the straight line information of the right tungsten electrode edge; and when the welding torch swings to the right limit position and stops, extracting a welding image containing the straight line information of the left bevel edge and the straight line information of the left tungsten electrode edge.

[0098] More specifically, the embodiments in this specification use a PLC system to acquire the welding images. It can be understood that the PLC system acquires angle information at a frequency of approximately 15Hz. During the 0.5s dwell time at the left and right extreme positions, it can acquire approximately 7-8 identical angle values. Therefore, using the characteristic of equal angle information at the extreme positions as a judgment criterion can well adapt to image acquisition of different bevel sizes. When multiple equal angles are acquired, it indicates that the welding torch is at the left and right extreme positions. However, only one frame is needed at each extreme position. A 30fps camera can acquire approximately 10-15 image information at the extreme positions using the equal angle feature. To avoid this situation, the equal angle signal needs to be judged. Only the change in angle value from unequal to equal is used as the trigger signal to ensure that only one frame is acquired at the extreme position. Therefore, comparing the equal angle signals, the trigger signal is used when the equal angle signal changes from unequal to equal, ensuring accurate acquisition of one frame at each of the left and right extreme positions.

[0099] Specifically, this embodiment provides a preferred method, which further includes the following steps between steps S1 and S2: obtaining the average gray value of the welding image, subtracting the average gray value from the gray value of the welding image to enhance the contrast of the bevel edge and the contrast of the tungsten electrode edge in the welding image; setting a parameter of interest, and locating the region of interest in the welding image based on the parameter of interest.

[0100] Understandably, without accurate localization of the welding area, edge detection across the entire field of view would not only be time-consuming but also generate significant interference, hindering the stable operation of the detection system. Therefore, this embodiment employs the SSD target detection algorithm based on the Tensorflow framework for detection and localization, and uses the Openvino tool to optimize the detection and recognition model, significantly reducing the inference time on the CPU and meeting the requirements for real-time weld seam tracking and detection. After locating the welding area, a reference coordinate system is established. The region of interest (ROI) pre-set in the base coordinate system is transformed to the reference coordinate system, forming an adaptive edge detection ROI placement technique based on the depth target detection algorithm. This greatly improves the effectiveness and robustness of the edge detection algorithm.

[0101] More specifically, in order to adapt to the welding of workpieces with different plate thicknesses and different bevel sizes, a variety of region of interest distribution modes are pre-made in the base coordinate system for selection. If all modes cannot meet the detection requirements of the actual conditions, the region of interest can be adjusted by controlling the software system through the touch screen, so that the detection area meets the needs of tracking control and can extract feature information stably and accurately.

[0102] Please see Figure 5 and 6 , Figure 5 This image shows the effect of enhanced contrast at the bevel edge after processing the welding image, as described in the vision-based narrow gap weld tracking control method provided in this specification. Figure 6 The image shown is an illustration of the enhanced contrast of the tungsten electrode edge after processing the welding image described in the vision-based narrow gap weld tracking control method provided in the embodiments of this specification.

[0103] Understandably, during visual sensing, interference from various complex factors such as arc light and signal transmission severely affects the sensing of weld area feature information. To improve the accuracy and robustness of various image analysis and processing algorithms, it is generally necessary to convert the entire image to grayscale and then perform arithmetic and logical operations on the pixels to eliminate noise or improve image contrast. The purpose of these preliminary image preprocessing steps is to improve the effectiveness of subsequent information detection algorithms. Based on the requirements of subsequent edge detection algorithms, image processing algorithms need to be performed from the following aspects:

[0104] Firstly, histogram analysis revealed that subtracting the average gray value of the entire image from its gray value resulted in most pixels outside the weld bevel having zero gray values, while most pixels inside the bevel had gray values ​​greater than zero, and the gray values ​​of pixels in the tungsten electrode and tungsten electrode tip regions were much greater than zero.

[0105] Secondly, the weld image processed above was enhanced by multiplication. It was found that the pixel grayscale value of the tungsten electrode and the tungsten electrode tip area was greatly enhanced. Pixels with grayscale values ​​greater than zero inside the bevel were enhanced to a certain extent, while pixels with grayscale values ​​equal to zero outside the bevel continued to have grayscale values ​​of zero.

[0106] Through the above processing, the grayscale information of the processed welding image is roughly divided into three parts: the pixel grayscale value of the tungsten electrode and the tungsten electrode tip area is close to 255, the pixel grayscale value inside the bevel remains at a medium level, and the pixel grayscale value outside the bevel is close to zero. This maximizes the intensity of the bevel and the edge of the tungsten electrode.

[0107] Finally, there is adaptive image enhancement: During the welding process, the oscillation of the welding torch and the changes in welding parameters will cause the arc to contract or rise. When the arc contracts, the difference in gray values ​​between pixels inside and outside the bevel decreases, and the gray values ​​of the tungsten electrode and tungsten tip areas are relatively small, making the whole image darker. When the arc rises, the difference in gray values ​​between pixels inside the bevel and tungsten electrode pixels is small, and the gray values ​​of pixels inside the bevel are larger, making the whole image brighter. If an enhancement coefficient suitable for a darker image is used to enhance a brighter image, the gray values ​​of pixels inside the bevel and pixels in the tungsten electrode area will reach saturation, making it impossible to identify the edge information of the tungsten electrode. Conversely, if an enhancement coefficient suitable for a brighter image is used to enhance a darker image, it will not be able to widen the difference in gray values ​​inside and outside the bevel, thus making it difficult for subsequent edge detection algorithms to detect. Therefore, an adaptive image enhancement method is adopted. Based on the histogram analysis algorithm, the method starts from the pixel with the highest gray value in each image and calculates the number of pixels stacked sequentially downwards. When the number of stacked pixels exceeds 3000, the stacking stops, and the gray value t corresponding to the pixel at this point is recorded. The enhancement coefficient is set to n = 255 / t. This ensures that weld images from different sources only enhance approximately 3000 pixels in the arc region of the tungsten electrode tip to a saturated state. Darker images will not result in weak bevel edge strength due to a small enhancement coefficient, and brighter images will not result in oversaturated pixels within the bevel due to a large enhancement coefficient, preventing the edge detection algorithm from failing to detect tungsten electrode edge information. An appropriate image enhancement coefficient is adaptively selected based on the maximum gray value characteristic of the image, enhancing weld images acquired under each condition to the same degree.

[0108] For practical operation of the tracking control method provided in the embodiments of this specification, please refer to... Figure 7 Welding parameters are configured according to the actual working conditions. Once the sensor camera position and all hardware systems are ready, welding begins. The embedded industrial controller processes image information and controls the motion mechanism through the software system, converting the deviation signal D into pulse commands that the servo driver can recognize, and performing adaptive weld seam tracking control.

[0109] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0110] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0111] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of embodiments of this disclosure upon considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described herein. The specification and embodiments are to be considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A vision-based narrow-gap weld seam tracking and control method, characterized in that, Includes the following steps: S1. Obtain welding images during narrow gap weld welding; S2. Based on the welding image, obtain the position information of the bevel edge and the position information of the tungsten electrode edge; the position information of the bevel edge includes the straight line information of the right bevel edge and the straight line information of the left bevel edge, and the position information of the tungsten electrode edge includes the straight line information of the right tungsten electrode edge and the straight line information of the left tungsten electrode edge; S3. Calculate the weld centerline based on the position information of the bevel edge, and calculate the welding torch position based on the position information of the tungsten electrode edge; S4. Based on the weld centerline and the position of the welding torch, calculate the relative positional relationship between the welding torch and the weld centerline; S5. Adjust the position of the welding torch head based on the relative positional relationship between the welding torch head and the center line of the weld; In step S1, acquiring welding images during narrow-gap weld seam welding includes: When the welding torch swings to the left limit position and stops, the extracted welding image contains the straight line information of the right bevel edge and the straight line information of the right tungsten electrode edge; When the welding torch swings to its right limit position and stops, the extracted welding image contains the straight line information of the left bevel edge and the straight line information of the left tungsten electrode edge; In step S3, calculating the weld centerline based on the position information of the bevel edge includes: Calculate the equation of the right bevel edge line based on the coordinates of any two points located on the right bevel edge line; Calculate the equation of the left bevel edge line based on the coordinates of any two points located on the left bevel edge line; Based on the equations of the right bevel edge line and the left bevel edge line, the angle bisector of the angle formed by the right bevel edge line and the left bevel edge line is calculated to obtain the weld centerline. In step S3, the equation of the weld centerline is: Among them, (x) C1 y C1 (x) represents a point on the center line of the weld. C2 y C2 () indicates another point on the centerline of the weld; In step S3, calculating the welding torch tip position based on the position information of the tungsten electrode edge includes: Calculate the equation of the right tungsten electrode edge line based on the coordinates of any two points located on the right tungsten electrode edge line; Calculate the equation of the left tungsten electrode edge line based on the coordinates of any two points located on the left tungsten electrode edge line; Based on the equation of the right tungsten electrode edge line and the equation of the left tungsten electrode edge line, the intersection point of the reverse extension line of the right tungsten electrode edge line and the reverse extension line of the left tungsten electrode edge line is calculated to obtain the position of the welding torch head. In step S4, the equation for the relative positional relationship between the welding torch head and the weld centerline is: Where D represents the positional difference between the welding torch tip and the center of the weld, (x C1 y C1 (x) represents a point on the center line of the weld. C2 y C2 (x0, y0) represents another point on the center line of the weld, and (x0, y0) represents the coordinates of the welding torch head.

2. The vision-based narrow gap weld seam tracking control method according to claim 1, characterized in that, The positional difference D between the welding torch tip and the weld center represents the relative positional relationship between the welding torch tip and the weld centerline, specifically: When the position difference D between the welding torch tip and the weld center is 0, it indicates that the welding torch tip is on the weld centerline; When the position difference D between the welding torch tip and the weld center is less than 0, it indicates that the welding torch tip is located to the left of the weld centerline; When the position difference D between the welding torch tip and the weld center is greater than 0, it indicates that the welding torch tip is located to the right of the weld centerline.

3. The vision-based narrow gap weld seam tracking control method according to claim 2, characterized in that, In step S5, adjusting the position of the welding torch head based on the relative positional relationship between the welding torch head and the weld centerline includes: A preset correction accuracy K is used to compare the position difference D between the welding torch head and the weld center with the preset correction accuracy K, and the position of the welding torch head is adjusted based on the comparison result. When -K < D < K, do not adjust the welding torch position; When D < -K, adjust the welding torch head position to the right; When D > K, adjust the welding torch head position to the left.

4. A vision-based narrow-gap weld tracking control method according to any one of claims 1-3, characterized in that, Between step S1 and step S2, the following steps are also included: The average grayscale value of the welding image is obtained, and the average grayscale value is subtracted from the grayscale value of the welding image to enhance the contrast of the bevel edge and the contrast of the tungsten electrode edge in the welding image. Set the parameters of interest, and locate the region of interest in the welding image based on the parameters of interest.

Images