Automatic centering device and method for hollow welded ball

Abstract

The present application belongs to the technical field of automatic welding of hollow welded ball, and the welding seam cannot be ensured to be in the central position when the hollow ball is welded, and there is no automatic centering device to calibrate the position of the ball welding seam, and only the ball body is moved by the worker to align the center, and the problem of misalignment of the welding seam caused by manual judgment is prone to occur, resulting in the problem of inability to weld or welding defects, the present application provides a kind of automatic centering device and method for hollow welded ball, which uses a front camera to shoot a ball image, determines the angle of rotation of the hollow welded ball around the X axis according to the shot ball image, determines the rotation angle of the ball around the Z axis according to the rear line laser camera, and realizes the automatic centering of the hollow welded ball welding seam through motor-driven rotating equipment, the present application has short adjustment time, reduces the influence of installation tolerance, improves the centering accuracy, greatly reduces the operation pressure in the process of welding seam image processing under the condition of ensuring high-precision welding seam identification, and is suitable for low-power equipment.

Description

Technical Field

[0001] This invention belongs to the field of automatic welding technology for space frame welded balls, specifically relating to an automatic centering device and method for hollow welded balls. Background Technology

[0002] Space frame structures, as a new type of building structure, are widely used in large-scale construction projects. Among various space frame structures, welded spherical space frame structures are favored for their advantages such as light weight, low material consumption, and high structural rigidity, offering broad application prospects and numerous advantages. They can be widely used in construction, aerospace, rail transportation, industrial equipment, stadiums, exhibition halls, and cultural and entertainment venues. Compared with traditional space frame structures, welded spherical space frame structures have several advantages: First, they allow for modular design and manufacturing, enabling rapid assembly and disassembly, accelerating construction speed and reducing costs. Second, the spherical design of the nodes allows for the connection of steel pipes of different sizes and angles through core holes in the spheres, significantly saving materials. Finally, welded spherical space frame structures have an aesthetically pleasing, modern, and stylish appearance, and the spheres exhibit excellent performance in corrosion resistance, oxidation resistance, and other properties, making them well-suited to various climatic conditions. In conclusion, welded spherical space frame structures have a very broad application prospect, and their numerous advantages make them a leader among various space frame structures.

[0003] Currently, the manufacturing method for welded spheres involves first creating two hollow hemispheres, then assembling and welding them together, and finally performing a final weld to fill in the gaps. However, due to the unpredictable posture of the spheres, existing equipment cannot control their orientation during loading, making it impossible to guarantee that the weld seam is centered after the sphere enters the equipment. Furthermore, there is no automatic alignment device to calibrate the weld seam position. Workers must manually move the spheres to align them for successful welding, resulting in high labor intensity and low production efficiency. Moreover, the need for manual judgment under high-intensity labor can easily lead to misalignment, making welding impossible or resulting in weld defects. This results in low weld quality, long adjustment times, and low welding efficiency, severely impacting overall production efficiency. Summary of the Invention

[0004] In view of the problems existing in the prior art, the purpose of the present invention is to provide an automatic centering device and method for hollow welded balls. During the welding process of hollow welded balls, the ball's posture can be automatically adjusted to achieve automatic centering of the weld seam. As a pretreatment before welding, the present invention eliminates the need for manual eye alignment. It has important practical significance for improving the welding process and eliminating production problems such as low welding quality and long welding adjustment time caused by misalignment between the welding torch and the weld seam.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0006] An automatic centering device for hollow welding ball welds includes a lifting device and a ball-holding and rotating device. It also includes a precision turntable for rotating the hollow welding ball, a front-facing camera, a rear-facing line laser camera, and a Raspberry Pi for processing weld feature images. The precision turntable is installed at the center of the automatic welding device. The front-facing camera is installed at the intersection of the ideal plane of the weld and the top support of the automatic welding device. The rear-facing line laser camera is installed at the intersection of the ideal plane of the weld and the rear support of the automatic welding device. The Raspberry Pi is installed inside the dustproof and heat dissipation control box of the automatic welding device. The two cameras transmit image data to the Raspberry Pi via USB, and each motor is connected to the Raspberry Pi via Ethernet to transmit feedback and control signals.

[0007] Furthermore, the ball-holding and rotating device includes a torque-mode servo motor, an X-axis slide rail, a ball-holding plate, and a rotary motor. The servo motor is connected to the X-axis slide rail via a coupling, and the servo motor drives the ball-holding plate to move on the X-axis slide rail. The rotary motor is connected to the ball-holding plate via a reducer, and the rotatable ball-holding plate is connected to the base via an oil-sealed bearing.

[0008] Furthermore, the lifting device includes a servo motor, a lifting slide rail, and a motion platform. The servo motor is connected to the lifting slide rail via a coupling. The upper end of the lifting slide rail is fixedly connected to the motion platform. A circular groove for placing a precision turntable is set in the center of the motion platform. The turntable of the precision turntable rotates freely in the circular groove, and the hollow ball to be welded is placed on the turntable.

[0009] A method for weld seam position identification during the automatic alignment process of hollow welded sphere weld seams, implemented using the aforementioned automatic alignment device for hollow welded sphere weld seams, wherein a front-facing camera in the device captures an image of the sphere, and the angle α of the hollow welded sphere's rotation around the X-axis is obtained based on the image captured by the front-facing camera to achieve vertical alignment of the sphere, comprising the following steps:

[0010] Step i. Take a picture of the sphere with the front camera, preprocess the image using Gaussian noise reduction and median noise reduction to generate image IMG1, and use the adaptive binary method to process image IMG1 to find the centroid position G of the sphere weld.

[0011] Step ii. Perform Hough circle detection on image IMG1, perform preliminary screening on the detected circle parameters and obtain the circle parameters representing the hollow welded sphere, repeat Hough circle detection on multiple frames captured by the camera, and perform weighted average of the circle parameters detected in multiple frames to obtain the center position C and the radius R of the circle.

[0012] Step iii. Subtract the center position C of the circle from the centroid position G of the weld. Substitute the difference, the radius R of the circle, and the centroid position G of the weld into the polar coordinate model of the sphere. Keep the relationship between the front camera and the center position of the sphere unchanged. Use the omnidirectional model to obtain the pixel coordinates of any point on the sphere projected onto the camera. Based on the pixel coordinates of the centroid position G of the weld, derive the angle α that the hollow welded sphere needs to rotate around the X-axis through reverse derivation.

[0013] The hollow welded sphere is rotated around the X-axis by an angle α controlled by the ball-holding rotation device. It is then rotated by another angle θ, where θ is the angle between the normal to the front camera lens plane and the Z-axis of the sphere. The Raspberry Pi's image source is then switched to a rear-facing line laser camera. This rear-facing line laser camera captures the image reflected back from the weld seam by the line laser beam, thus obtaining the angle β of the hollow welded sphere's rotation around the Z-axis. The specific steps are as follows:

[0014] Step I. Use the rear laser line camera to capture high-resolution images with a resolution of 1080p or higher;

[0015] Step II. After downsampling the image obtained in Step I using image pyramid, use the NCC template matching method to initially find the location of weld features. Then, enlarge the found weld features by one time based on the template size and crop them into a new ROI region.

[0016] Step III. In the new ROI region, perform NCC template matching again using the original resolution image after noise reduction to find the high-precision weld feature ordinate. Subtract the weld feature ordinate from the preset turntable center point coordinate and input it into the sphere polar coordinate model to obtain the angle β of the sphere's rotation around the Z-axis.

[0017] An automatic centering method for the weld seam of a hollow welded sphere, based on the aforementioned weld seam position recognition method, utilizes a front-facing camera to capture an image of the sphere, determines the rotation angle of the hollow welded sphere around the X-axis based on the captured image, determines the rotation angle of the sphere around the Z-axis based on a rear-facing line laser camera, and achieves automatic centering of the hollow welded sphere weld seam by driving a rotating device with a motor. Specifically, the method includes the following steps:

[0018] S1. Place the ball on the motion platform, lift it to the preset height H by the lifting device, drive the ball-holding plate with the torque mode servo motor of the ball-holding and rotating device to hold the ball tightly, and lower the lifting device to make room for rotation;

[0019] S2. Use the image captured by the front camera to determine the angle α of the hollow welding ball rotating around the X-axis. The rotary motor drives the ball-holding plate to rotate the ball counterclockwise along the X-axis so that the center of gravity G of the weld coincides with the center of the ball. Then continue to rotate around the X-axis by an angle θ, where θ is the angle between the normal of the front camera lens plane and the Z-axis of the ball.

[0020] S3. Switch the image source of the Raspberry Pi to the rear line laser camera. The lifting device rises to support the ball, and the ball-holding and rotating device is released, so that the hollow welded ball is placed on the precision turntable of the rotating hollow welded ball. The angle β of the hollow welded ball's rotation around the Z-axis is obtained by using the image captured by the rear line laser camera. The turntable drive motor drives the turntable to rotate the ball along the Z-axis, so that the vertical coordinate of the weld laser feature F1 is consistent with the vertical coordinate of the turntable's center, that is, so that the weld laser feature 1 is in the ideal plane of weld alignment.

[0021] S4. The lifting device lifts the ball, and the ball-holding and rotating device picks up the ball and rotates it by an angle γ. Angle γ is the angle between the line connecting the rear laser camera and the center of the ball and the line connecting the welding gun and the center of the ball, so that the weld laser feature F1 rotates to coincide with the Z-axis passing through the center of the turntable.

[0022] S5. Repeat the steps described in S3 again, so that the vertical coordinate of the weld laser feature F2 is consistent with the vertical coordinate of the center of the turntable, thereby realizing the automatic alignment of the hollow weld ball weld with the welding gun position.

[0023] In summary, compared with the prior art, the invention has the following beneficial effects:

[0024] (1) The present invention only requires three-axis motion to complete the arbitrary posture adjustment of the sphere, and the rotation adjustment of the sphere posture only requires a minimum of two axes, which reduces the production cost and maintenance difficulty of the device;

[0025] (2) The present invention can adjust the weld of the hollow welded ball to the center in a maximum of three steps under any posture of the hollow welded ball. The adjustment time is short, and even if the weld is far from the center position, the entire adjustment time does not exceed 70 seconds.

[0026] (3) The combined use of coarse and precise centering image sensors in this invention can effectively reduce the impact of installation tolerances and improve centering accuracy.

[0027] (4) The image processing method used in this invention can reduce the number of pixels in the image to be processed to 1 / 16 of the original image and quickly determine the initial position of the weld. On the other hand, after initially determining the initial position of the weld, a smaller ROI area is divided and the original resolution of the image is used to improve the accuracy of weld position recognition. Thus, while ensuring high-precision weld recognition, the computational pressure is greatly reduced and it can run smoothly on low-power processing devices such as Raspberry Pi. Attached Figure Description

[0028] Figure 1 This is a flowchart of the automatic weld alignment method of the present invention;

[0029] Figure 2 This is a schematic diagram of the lifting device and the precision turntable structure of the rotating hollow welded ball of the present invention;

[0030] Figure 3 This is a schematic diagram of the ball-holding and rotating device of the present invention;

[0031] Figure 4 This is a diagram showing the camera position and the sphere's position;

[0032] Figure 5 This is a schematic diagram of Hough circle detection and weld centroid determination;

[0033] Figure 6 This is a schematic diagram illustrating the process of the front camera's weld seam roughly aligning with the sphere's movement.

[0034] Figure 7 This is a schematic diagram of the sphere's movement process during the first precise alignment of the weld seam;

[0035] Figure 8 This is a schematic diagram showing the weld laser feature F1 rotated to be directly above the sphere;

[0036] Figure 9 This is a schematic diagram of the sphere's movement process during the second precise alignment of the weld seam;

[0037] Figure 10 This is a schematic diagram of the omnidirectional model coordinate transformation method;

[0038] Figure 11 This is a schematic diagram of the coordinate transformation method for the pinhole model;

[0039] Figure 12 This is a schematic diagram of the process of rotating the weld laser feature F1 to be directly above the sphere.

[0040] In the diagram: 1-Lifting device; 1.1-Servo motor; 1.2-Lifting slide rail; 1.3-Motion platform; 2-Ball-holding and rotating device; 2.1-Torque mode servo motor; 2.2-X-axis slide rail; 2.3-Ball-holding plate; 2.4-Rotation motor; 3-Precision turntable for rotating hollow welded balls; 3.1-Turntable drive motor; 3.2-Turntable; 4-Front-facing camera; 5-Rear-facing line laser camera. Detailed Implementation

[0041] The present invention will now be further described with reference to the accompanying drawings.

[0042] like Figure 1 As shown, the horizontal direction is defined as the X-axis, the front-back direction as the Y-axis, and the vertical direction perpendicular to the horizontal plane as the Z-axis.

[0043] like Figures 2-12As shown, this invention discloses an automatic centering device for hollow welded spheres, comprising a lifting device 1 and a ball-holding and rotating device 2. The lifting device 1 includes a servo motor 1.1, a lifting slide rail 1.2, and a motion platform 1.3. The servo motor 1.1 is connected to the lifting slide rail 1.2 via a coupling. The upper end of the lifting slide rail 1.2 is fixedly connected to the motion platform 1.3. The motion platform 1.3 has a circular groove at its center for placing a precision turntable 3. The turntable 3.2 of the precision turntable 3 can rotate freely within the groove. The hollow sphere to be welded should be placed on the turntable 3.2. The ball-holding and rotating device 2 includes a torque-mode servo motor. 2.1, X-axis slide rail 2.2, ball-holding plate 2.3, and rotary motor 2.4. Servo motor 2.1 is connected to X-axis slide rail 2.2 via a coupling, driving the ball-holding plate 2.3 to move on X-axis slide rail 2.2. Rotary motor 2.4 is connected to ball-holding plate 2.3 via a reducer. The rotatable ball-holding plate 2.3 is connected to the base via an oil-sealed bearing. The system also includes a precision turntable 3 for rotating the hollow welding ball, a front camera 4, a rear laser camera 5, and a Raspberry Pi for processing weld feature images. The precision turntable 3 is installed at the center of the automatic hollow welding ball welding device, and the front camera 4 is mounted on... Figure 12 The point where the ideal plane of the weld seam intersects with the top support of the automatic welding device for hollow welded balls, as shown, is... Figure 12 The right side passes through the center of the sphere and is perpendicular to both the turntable plane and the camera lens plane. Vertical centering is achieved using the front camera 4. The ideal plane of the weld is a plane that passes through the center of the hollow sphere, is perpendicular to both the turntable plane and the camera lens plane. The rear line laser camera 5 is installed at the position where the ideal plane of the weld intersects with the rear support of the automatic hollow welding sphere welding device. The rear line laser camera 5 is used to achieve lateral and precise centering. Both the front camera 4 and the rear line laser camera 5 can be fixed on the external frame structure, which is omitted in the diagram. The Raspberry Pi is installed in the dustproof and heat dissipation control box of the automatic hollow welding sphere welding device. The two cameras transmit image data to the Raspberry Pi via USB. Each motor is connected to the Raspberry Pi via Ethernet to transmit feedback and control signals.

[0044] This invention also discloses a method for weld position identification during the automatic alignment process of hollow welded sphere welds. The method utilizes the aforementioned automatic alignment device for hollow welded sphere welds to capture images of the sphere using a front-facing camera 4. Based on the images captured by the front-facing camera 4, the angle α of the hollow welded sphere's rotation around the X-axis is obtained. The method includes the following steps:

[0045] Step i. The front camera 4 captures an image of the sphere. Gaussian noise reduction and median noise reduction are used to preprocess the image to generate image IMG1. Adaptive binary method is used to process image IMG1 to find the centroid G of the sphere's weld seam. Figure 5 As shown on the right;

[0046] Step ii. Perform Hough circle detection on image IMG1, initially filter the detected circle parameters and obtain the circle parameters representing the hollow welded sphere. Repeat Hough circle detection on multiple frames captured by the camera, and perform a weighted average of the detected circle parameters from multiple frames to obtain the center position C and the radius R of the circle, as shown below. Figure 5 As shown on the left;

[0047] Step iii. Subtract the center position C of the circle from the centroid position G of the weld. Substitute the difference, the radius R of the circle, and the centroid position G of the weld into the polar coordinate model of the sphere. That is, with the center of the sphere as the origin, write the coordinates of a point on the sphere in polar coordinate form. Since the positional relationship between the camera and the center of the sphere remains unchanged, let the distance between the camera and the center of the sphere be L. Using an omnidirectional model, any point P(x,y,z) on the sphere is projected onto the spherical projection with a distance of z. The formula is:

[0048]

[0049]

[0050]

[0051] Where ||X|| is the magnitude from point P to the center of the sphere, and ζ is the distance between the center of the front camera 4 and the sphere.

[0052] like Figure 10 As shown, then substitute the pinhole model:

[0053]

[0054]

[0055] This gives the pixel coordinates p(x) of any point P projected onto the image. c ,y c ),like Figure 11 As shown, by reverse derivation, the angle α of the welding ball's rotation around axis A can be obtained using the pixel coordinates of the weld centroid position G.

[0056] After controlling the ball-holding rotation device 2 to rotate the welding ball around the X-axis by an angle α, continue to rotate it around the X-axis by an angle θ, where θ is the angle between the normal of the front camera 4 lens plane and the Z-axis of the ball. Then, switch the image source of the Raspberry Pi to the rear laser line camera 5 to obtain the angle β of the hollow welding ball's rotation around the Z-axis. The specific steps are as follows:

[0057] Step I. The rear laser line camera 5 captures a high-resolution image with a resolution of 1080p or higher, which is reflected back from the weld by the line laser beam.

[0058] Step II. After downsampling the image obtained in Step I using image pyramid, use the NCC template matching method to initially find the location of weld features. Then, enlarge the found weld features by one time based on the template size and crop them into a new ROI region.

[0059] Step III. In the new ROI region, perform NCC template matching again using the original resolution image after noise reduction to find the high-precision weld feature ordinate. Subtract the weld feature ordinate from the preset turntable center point coordinate and input it into the sphere polar coordinate model to obtain the angle β of the sphere's rotation around the C-axis.

[0060] like Figure 1 As shown, this invention also discloses an automatic centering method for hollow welded spheres. Based on the above-mentioned weld position identification method, the flow of the automatic centering method is as follows: Figure 12 As shown, the hollow welded sphere is rotated around the X-axis by the front camera 4 to capture an image of the sphere, and around the Z-axis by the rear laser camera 5. Automatic centering of the weld seam of the hollow welded sphere is achieved by using a motor-driven rotating device. The process includes the following steps:

[0061] S1. The ball is placed on the motion platform 1.3 and lifted to a preset height H by the lifting device 1. The torque mode servo motor 2.1 of the ball-holding and rotating device 2 drives the ball-holding plate 2.3 to hold the ball tightly, and the lifting device 1 descends to make room for rotation.

[0062] S2. Using the image captured by the front-facing camera 4, determine the angle α of rotation of the hollow welded sphere around the X-axis. Control the rotary motor 2.4 to drive the ball-holding plate 2.3 to rotate the sphere counterclockwise along the X-axis by the angle α, so that the center of gravity of the weld coincides with the center of the sphere. Then continue to rotate around the X-axis by the angle θ, as follows. Figure 6 As shown, θ is the angle between the normal to the plane of the front-facing camera's four lenses and the Z-axis of the sphere, as... Figure 12 As shown on the left.

[0063] S3. Switch the image source of the Raspberry Pi to the rear-mounted line laser camera 5. The lifting device 1 rises to support the sphere, and the sphere-holding and rotating device 2 releases, placing the sphere on the precision turntable 3 that rotates the hollow welded sphere. The image captured by the rear-mounted line laser camera 5 is used to obtain the angle β of the hollow welded sphere's rotation around the Z-axis. The turntable drive motor 3.1 drives the turntable 3.2 to rotate the sphere along the Z-axis, ensuring that the ordinate of the weld laser feature F1 is consistent with the ordinate of the center of the turntable 3.2, i.e., placing the weld laser feature F1 within the ideal plane of weld alignment. Figure 8 As shown;

[0064] S4. Lifting device 1 raises the sphere, while ball-holding and rotating device 2 picks up the sphere and rotates it by an angle γ. Angle γ is the angle between the line connecting the rear-mounted laser camera 5 and the center of the sphere and the line connecting the welding torch and the center of the sphere, so that the weld laser feature F1 rotates to coincide with the Z-axis passing through the center of turntable 3.2. Figure 12 As shown;

[0065] S5. Repeat the steps described in S3 again, so that the ordinate of the weld laser feature F2 is consistent with the ordinate of the center of the turntable 3.2, as follows. Figure 9 As shown, since the sphere rotates around the C-axis, the position of the weld laser feature F1, through which the C-axis passes, remains unchanged and is still within the ideal weld alignment plane. The weld laser feature F2 has now moved to the ideal weld alignment plane. Given that the center of the circle must be within the ideal weld alignment plane, the plane containing the weld ring coincides with the ideal weld alignment plane, as can be determined from three non-collinear points. Figure 12 As shown on the right, the hollow weld ball weld seam and the welding torch position can be automatically aligned.

[0066] Because the initial rotation around the X-axis is used for coarse alignment, it is necessary to consider the operating speed requirements. At the same time, due to the limitations of equipment installation accuracy, it is impossible to ensure that the weld can be moved to be completely perpendicular to the turntable platform on the first attempt. This invention can complete the arbitrary posture adjustment of the sphere through three-axis motion, which reduces the requirements for camera installation tolerance, facilitates equipment operation and maintenance, and reduces production costs.

[0067] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principle of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for weld position recognition in the automatic centering process of hollow weld ball weld seams, implemented using a hollow weld ball weld seam automatic centering device, characterized in that: The hollow welded ball weld automatic centering device, comprising a lifting device (1) and a ball holding and rotating device (2), further comprising a precision turntable (3) for rotating the hollow welded ball, a front camera (4), a rear line laser camera (5) and a Raspberry Pi for processing the weld feature image, the precision turntable (3) is installed at the center position of the hollow welded ball automatic welding device, the front camera (4) is installed at the position where the ideal weld plane intersects with the top bracket of the hollow welded ball automatic welding device, the rear line laser camera (5) is installed at the position where the ideal weld plane intersects with the rear bracket of the hollow welded ball automatic welding device, the Raspberry Pi is installed in the dustproof and heat dissipation control box of the hollow welded ball automatic welding device, the two cameras transmit image data to the Raspberry Pi through USB, and each motor is connected with the Raspberry Pi through Ethernet to transmit feedback and control signals; The front camera (4) in the hollow welded ball weld automatic centering device captures the ball image, and the angle a of the hollow welded ball rotating around the X axis is obtained according to the image captured by the front camera (4), so as to realize the vertical centering of the ball, comprising the following steps: Step i. The front camera (4) captures the ball image, generates an image IMG1 after image preprocessing by using Gaussian noise reduction and median noise reduction, and finds the gravity center position G of the ball weld by processing the image IMG1 using an adaptive binary method; Step ii. Hough circle detection is used on the image IMG1, the detected circle parameters are preliminarily screened, and the circle parameters representing the hollow welded ball are obtained, and the Hough circle detection is repeatedly performed on multiple frames of images captured by the camera, the circle parameters detected from the multiple images are weighted and averaged, and the center position C and the radius R of the circle are obtained; Step iii. The difference between the center position C and the gravity center position G of the weld is obtained, and the difference, the radius R and the gravity center position G of the weld are brought into the ball polar coordinate model, the relationship between the front camera (4) and the ball center position is unchanged, the omnidirectional model is used, the pixel coordinates of any point on the ball projected into the camera are obtained, and the angle a of the hollow welded ball rotating around the X axis is obtained according to the pixel coordinates of the gravity center position G of the weld through reverse deduction; After the ball holding and rotating device (2) is controlled to make the hollow welded ball rotate around the X axis by the angle a, the hollow welded ball continues to rotate around the X axis by the angle θ, θ is the included angle between the lens plane normal of the front camera (4) and the Z axis of the ball, then the image source of the Raspberry Pi is switched to the rear line laser camera (5), the rear line laser camera (5) captures the image reflected by the line laser irradiated on the weld, and the angle β of the hollow welded ball rotating around the Z axis is obtained, and the specific steps are as follows: Step I. The rear line laser camera (5) captures a high-definition image with a resolution of 1080p or above; Step II. After the image obtained in step I is down-sampled by using image pyramid, the NCC template matching method is used to preliminarily find the weld feature position, then the found weld feature is enlarged by one time based on the template size, and is cropped into a new ROI region; Step III. NCC template matching is performed again in the new ROI region using the denoised original resolution image to find the high-precision weld feature ordinate, and the weld feature ordinate is subtracted from the preset position coordinate of the center of the turntable and is brought into the spherical polar coordinate model to obtain the angle β of the spherical body rotating around the Z axis.

2. The weld location identification method in the automatic centering process of the weld of a hollow welded ball according to claim 1, characterized in that: The ball holding rotating device (2) comprises a torque mode servo motor (2.1), an X-axis sliding rail (2.2), a ball holding disc (2.3) and a rotating motor (2.4), the servo motor (2.1) is connected with the X-axis sliding rail (2.2) through a shaft coupling, the servo motor (2.1) drives the ball holding disc (2.3) to move on the X-axis sliding rail (2.2), the rotating motor (2.4) is connected with the ball holding disc (2.3) through a speed reducer, and the rotatable ball holding disc (2.3) is connected with the base through an oil seal bearing.

3. A method of weld location identification in a process of automatic centering of a weld of a hollow weld ball according to claim 2, characterized in that: The lifting device (1) comprises a servo motor (1.1), a lifting sliding rail (1.2) and a moving platform (1.3), the servo motor (1.1) is connected with the lifting sliding rail (1.2) through a shaft coupling, the upper end of the lifting sliding rail (1.2) is fixedly connected with the moving platform (1.3), the moving platform (1.3) is provided with a circular groove in the center for placing the precision turntable (3), the rotating disc (3.2) of the precision turntable (3) is freely rotatable in the circular groove, and the hollow welding ball to be welded is placed on the rotating disc (3.2).

4. A method for automatic centering of a hollow weld sphere weld seam, based on the weld seam position recognition method according to claim 3, characterized in that: The ball holding rotating device (2) comprises a torque mode servo motor (2.1), an X-axis sliding rail (2.2), a ball holding disc (2.3) and a rotating motor (2.4), the servo motor (2.1) is connected with the X-axis sliding rail (2.2) through a shaft coupling, the servo motor (2.1) drives the ball holding disc (2.3) to move on the X-axis sliding rail (2.2), the rotating motor (2.4) is connected with the ball holding disc (2.3) through a speed reducer, and the rotatable ball holding disc (2.3) is connected with the base through an oil seal bearing. S1. The ball is placed on the moving platform (1.3), lifted to a preset height H by the lifting device (1), and the torque mode servo motor (2.1) of the ball holding rotating device (2) drives the ball holding disc (2.3) to hold the ball tightly, and then the lifting device (1) is lowered to leave a rotating space; S2. The angle α of the hollow welding ball rotating around the X axis is determined according to the image captured by the front camera (4), the rotating motor (2.4) drives the ball holding disc (2.3) to rotate the ball along the X axis counterclockwise, so that the weld center of gravity position G coincides with the center of the ball, and then the ball is further rotated by an angle θ around the X axis, θ being the included angle between the lens plane normal of the front camera (4) and the Z axis of the ball; S3. The image source of the Raspberry Pi is switched to the rear linear laser camera (5), the lifting device (1) is lifted to hold the ball, the ball holding rotating device (2) is loosened, the hollow welding ball is placed on the precision turntable (3) for rotating the hollow welding ball, the angle β of the hollow welding ball rotating around the Z axis is obtained according to the image captured by the rear linear laser camera (5), and the rotating disc driving motor (3.1) drives the rotating disc (3.2) to rotate the ball along the Z axis direction, so that the longitudinal coordinate of the weld laser feature F1 coincides with the longitudinal coordinate of the center of the rotating disc (3.2), that is, the weld laser feature F1 is in the ideal weld centering plane. S4. The lifting device (1) drives the sphere to lift, the sphere holding and rotating device (2) holds and rotates the sphere by an angle γ, the angle γ is the included angle between the line connecting the rear line laser camera (5) and the center of the sphere and the line connecting the welding torch and the center of the sphere, so that the welding laser feature F1 is rotated to coincide with the Z axis passing through the center of the rotating disc (3.2); S5. Repeat the step S3 again, so that the longitudinal coordinate of the welding laser feature F2 coincides with the longitudinal coordinate of the center of the rotating disc (3.2), that is, the automatic centering of the hollow welded sphere welding seam and the welding torch position is realized.

Images