A bevel cutting method based on a robot bevel cutting workstation
The robot bevel cutting workstation method addresses fixture limitations and posture challenges by using edge detection and point cloud processing to adjust robot arm and gun head positions, ensuring precise and efficient bevel cutting across various workpiece types and sizes.
Patent Information
- Application Number
- CN202211354469.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-01
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2042-11-01
AI Technical Summary
Existing bevel cutting processes require precise fixture positioning, limiting workpiece type and bevel position, and struggle with robot end cutting gun posture adjustments due to bevel angle and workpiece size, leading to inefficiencies and material waste.
A robot bevel cutting workstation method that uses edge detection and recognition to determine workpiece position without fixtures, employing point cloud image processing and robot arm posture adjustments based on feature points for precise cutting.
Enables accurate bevel cutting without fixture constraints, allowing for diverse workpiece types and sizes, and optimizing robot arm and gun head angles for efficient material utilization.
Smart Images

Figure CN115556113B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the field of visual recognition tooling, and in particular to a bevel cutting method based on a robot bevel cutting workstation. Background Art
[0002] In the existing bevel cutting process, the bevel process is often performed directly on the steel plate, which will waste more steel than beveling the workpiece. There are the following problems when beveling the workpiece directly: 1. The position and placement angle of the workpiece may deviate when not using a fixture for positioning, and the use of a fixture for positioning has restrictions on the workpiece type and bevel position. 2. Affected by the bevel angle, bevel direction and workpiece size, the posture of the robot's end cutting gun is difficult to handle. Traditional processes generally use 6-axis rotation to reach the bevel angle, and the size of the workpiece that can be processed by the cutting gun will be greatly limited. Summary of the invention
[0003] The present invention aims to provide a bevel cutting method based on a robot bevel cutting workstation, which no longer requires precise positioning of a fixture for workpieces placed on a material table, and can directly determine the position of the workpiece through edge detection and recognition, without limiting the accuracy of placement. And through the acquisition of point cloud images and calculation and adjustment of the gun head, the workstation can optimize the robot arm posture and gun head angle corresponding to the cutting angle when preparing to cut the bevel.
[0004] The present invention is achieved through the following technical solutions:
[0005] A bevel cutting method based on a robot bevel cutting workstation comprises the following steps:
[0006] S1: placing the workpiece on the material table, detecting its edge, scanning to determine the first anchoring profile, and packing the first anchoring profile to the workstation; the workstation pre-processes the packed first anchoring profile and generates a second anchoring profile;
[0007] S2: guiding the robot arm to move the image acquisition device to the correct position to acquire the edge of the workpiece, aligning the edge acquired image information with the second anchoring contour, and opening the grating after determining the matching degree to generate point cloud image information of the workpiece;
[0008] S3: performing noise filtering on the point cloud image information, and fusing and extracting the point cloud image information after the noise filtering by using the first preset value and the second preset value to reconstruct the workpiece point cloud image; the workpiece point cloud image includes at least one feature point associated with the workpiece, and the distribution of the feature point is transmitted to the workstation;
[0009] S4: The workpiece point cloud image obtained in the above step S3 is transmitted to the robot arm according to the workpiece type, groove angle, groove direction and related feature point information in the order of outer contour clockwise and inner hole counterclockwise. The robot arm is divided into the first type and the second type according to the degree of difference of the feature points, and the groove cutting posture is prepared according to the transmission order of the feature points;
[0010] S5: Adjust the gun head posture at the end of the robot arm according to the position of the workpiece and the direction of the pre-cut groove.
[0011] Furthermore, in the above step S1, edge detection includes:
[0012] The camera takes an image of the workpiece placed on the material table, performs Gaussian filtering on the image to remove noise, and then passes it into the coordinate model. The coordinate model uses the YOLOV5 algorithm to identify and iteratively learn the workpiece, and identify the coordinate parameters of the workpiece center in the material table coordinate system;
[0013] The camera calibration is used to determine the intrinsic parameter matrix, and the hand-eye calibration is used to obtain the extrinsic parameter matrix. The above matrix is then multiplied with the coordinate parameters of the workpiece center in the above coordinate system to convert it into the physical position of the workpiece in the coordinate system.
[0014] Furthermore, in the above step S1, the preprocessing is to perform local annotation point correction on the first anchor contour, and multiple annotation points are arbitrarily selected on the edge of the first anchor contour, and the multiple annotation points are moved to the gradient peak of the image, and local linear fitting is performed to fit the multiple annotation points into a curve, and then through interpolation and correction sampling, the optimized second anchor image is output.
[0015] Furthermore, in the above step S3, the first preset value is the height, and the second preset value is the size; the noise reduction process includes:
[0016] A Kd tree is generated according to the workpiece point cloud image. A point is randomly selected as the base point on the Kd tree. The average value of any other point in the neighborhood of the base point within a certain radius is calculated, and it is determined whether the average value exceeds the threshold. If the average value is greater than the threshold, the point in the neighborhood is regarded as a noise point and is removed.
[0017] Furthermore, in the above step S3, the fusion extraction is to generate a workpiece point cloud image that matches the workpiece in three dimensions by using the height and size thresholds in combination with the denoised point cloud image information.
[0018] Further, in the above step S4, the first type is a straight line, and the second type is an arc; the judgment process of posture preparation is to compare the rotation angles of the current posture, the target posture and the intermediate vertical posture around the X and Y axes, and then calculate the difference of the rotation of each vertical posture around the Z axis; it includes:
[0019] After receiving the feature point, the robot arm obtains the current position of the gun head and determines whether the current position is a feature point. If it is a feature point, it determines whether the gun head and the feature point are horizontal. If it is not a feature point, it replaces the starting position with the current position for determination.
[0020] According to the above horizontal and vertical judgment results, the inverse tangent value of the slope is calculated as the rotation angle at the Z axis, and a corresponding rotation is performed around the Z axis, and then the height of the feature point is judged whether it is consistent;
[0021] According to the above height results, the inverse tangent of the slope is calculated as the rotation angle about the X-axis, and a corresponding rotation is performed around the X-axis. Then, the first type and the second type are distinguished, and the rotation direction along the Z-axis is selected through the feature point information in the groove direction, and the Euler angle and quaternion conversion is used to update the posture that meets the feature point.
[0022] Furthermore, the workstation uses an ABB robot, which records angles through quaternions and drives the robotic arm to rotate around the Z axis to track the workpiece at different heights.
[0023] Furthermore, the position of the robot arm is monitored during actual operation to determine whether it has reached the critical point between the preset posture and the changed posture. The monitoring process includes:
[0024] Establish a socket protocol communication channel between the robot arm and the workstation terminal, and send an information acquisition request to the terminal;
[0025] Clear the original data group and wait for the data to be sent;
[0026] Receive the original data and store it in an array;
[0027] Unpack the original data according to the ASCII code, and determine whether the end character meets the preset value. If it does not meet the preset end character value, store the value at the end of the temporary string until the current data is the same as the preset end character value, and save the temporary string to the data array.
[0028] Furthermore, it also includes an input display device, which is used to manually input information such as the workpiece type, groove angle, groove direction, etc. of the workpiece.
[0029] Furthermore, the image acquisition device is a 3D camera.
[0030] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0031] The bevel cutting method based on the robot bevel cutting workstation provided by the present invention can extract the workpiece information and transmit it to the robot to adjust the cutting posture, solving the problem of tracking cutting after the workpiece is placed randomly. In addition, the robot arm cuts according to the cutting direction based on the acquired feature point information, and handles different types of workpiece cutting progress separately. The composite posture adjustment cutting of the robot arm and the gun head has application potential in some application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The drawings described herein are used to provide a further understanding of the embodiments of the present invention, constitute a part of this application, and do not constitute a limitation of the embodiments of the present invention. In the drawings:
[0033] Figure 1 A schematic flow chart of a bevel cutting method based on a robot bevel cutting workstation provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0034] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.
[0035] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0036] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0037] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, or the positions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific position, be constructed and operated in a specific position, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description, and cannot be understood as indicating or implying relative importance.
[0038] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "set", "install", "connect", and "connect" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a direct connection, or an indirect connection through an intermediate medium, or it can be the internal communication of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0039] Example
[0040] See also Figure 1 This embodiment provides a bevel cutting method based on a robot bevel cutting workstation, comprising the steps of:
[0041] S1: Place the workpiece on the material table, detect its edge, scan and determine the first anchoring profile, and pack the first anchoring profile to the workstation; the workstation pre-processes the packed first anchoring profile and generates a second anchoring profile. In this embodiment, the first anchoring profile is the workpiece profile of the preliminary search, and the second anchoring profile is the optimized and cleaner workpiece profile. Among them, the pre-processing is to correct the local annotation points of the first anchoring profile, select multiple annotation points arbitrarily on the edge of the first anchoring profile, move the multiple annotation points to the gradient peak of the image, perform local linear fitting on it to fit the multiple annotation points into a curve, and then output the optimized second anchoring image through interpolation and correction sampling. Specifically, first manually mark multiple annotation points on the first anchoring profile one by one, connect them to form a closed loop, and then fit them. However, since the manually marked standard points are not necessarily uniform and smooth, interpolation processing is also required for places with larger gaps. The workstation uses an ABB robot, which records angles through quaternions and drives the mechanical arm to rotate around the Z axis to achieve tracking of workpieces at different heights.
[0042] S2: After guiding the robot arm to move the image acquisition device to the correct position, the edge of the workpiece is acquired, and the image information acquired at the edge is overlapped and positioned with the second anchor contour. After determining the matching degree, the grating is turned on to generate the point cloud image information of the workpiece. In this embodiment, the image acquisition device uses a 3D camera for image acquisition. The photo of the 3D camera is overlapped and matched with the second anchor contour to determine the degree of overlap between the actual image and the acquired image. In this embodiment, if the grayscale overlap of the two is compared at the same midpoint, if the grayscale overlap area reaches more than 95%, it can be used, and the point cloud image information of the workpiece is generated.
[0043] S3: Perform noise filtering on the point cloud image information, and fuse and extract the point cloud image information after the noise filtering through the first preset value and the second preset value to reconstruct the workpiece point cloud image; the fusion extraction is to generate a workpiece point cloud image that matches the workpiece in three dimensions by using the height and size thresholds in combination with the point cloud image information after noise reduction.
[0044] The workpiece point cloud image includes at least one feature point associated with the workpiece. The distribution of the feature point is transmitted to the workstation to adjust the posture of the workstation robot arm. The noise reduction process includes: generating a Kd tree based on the workpiece point cloud image, taking any point on the Kd tree as the base point, and calculating the average value of any other point in the neighborhood of the base point within a certain radius of the base point, and judging whether the average value exceeds the threshold. If the average value is greater than the threshold, the point in the neighborhood is regarded as a noise point and is removed.
[0045] S4: Since the cutting scheme is a plasma power supply, in order to reduce the thermal deformation of the workpiece, according to whether the groove position of the workpiece is an outer contour or an inner hole, the upper computer transmits the workpiece type, groove angle, groove direction and other information and feature points to the robot arm in a clockwise order for the outer contour and a counterclockwise order for the inner hole. The robot arm is divided into the first type and the second type according to the degree of difference of the feature points, and the groove cutting posture preparation is performed for the two types in turn according to the transmission order of the feature points. In this embodiment, the first type is a straight line and the second type is an arc. The judgment process of posture preparation is to compare the current posture with the target posture and the intermediate vertical posture around the X and Y axis rotation angles, and then calculate the difference between the rotation of each vertical posture around the Z axis. It includes:
[0046] After receiving the feature point, the robot arm obtains the current position of the gun head and determines whether the current position is a feature point. If it is a feature point, it determines whether the gun head and the feature point are horizontal. If it is not a feature point, it replaces the starting position with the current position for determination.
[0047] According to the above horizontal and vertical judgment results, the inverse tangent value of the slope is calculated as the rotation angle at the Z axis, and a corresponding rotation is performed around the Z axis, and then the height of the feature point is judged whether it is consistent;
[0048] According to the above height results, the inverse tangent of the slope is calculated as the rotation angle at the X-axis, and a corresponding rotation is performed around the X-axis; then, the first type and the second type are distinguished, and the rotation direction along the Z-axis is selected through the characteristic point information of the groove direction, and the transformation of Euler angles and quaternions is used to update the posture that conforms to the characteristic points. In this way, the robot arm of the workstation cuts in the cutting direction according to the characteristic point information given by the terminal, and processes the straight line and arc segments separately according to the workpiece information provided by the workstation terminal. Especially at the intersection of the turn, the cutting method at the actual intersection is determined according to the type of adjacent edges and whether the angle between the two sides is an acute angle or an obtuse angle.
[0049] S5: Adjust the gun head posture at the end of the robot arm according to the position of the workpiece and the direction of the pre-cut groove.
[0050] The angle calculation formula first compares the current posture and target posture with the middle vertical posture around the X and Y rotation angles, and then calculates the difference in the rotation around the Z axis of each vertical posture. The final robot execution effect is multi-axis synchronous motion. In the case of excess material, multi-axis rotation will cut out the workpiece to adjust the posture. The purpose of adjusting the posture in this way is to meet the needs of parametric programming. In addition, since the string receiving method of the ABB robot can only receive 50 characters, it is only enough for 2 to 3 points of information in actual use. For this reason, the original binary data is directly received in array form and unpacked into ASCII code. When extracting information such as coordinate points, the identifier is automatically detected. If the element in the current array does not end with the identifier, the information of the next element is spliced to ensure the complete transmission of the feature point information.
[0051] Therefore, in order to ensure the position monitoring of the robot arm in actual operation and determine whether it has reached the critical point between the preset posture and the changed posture, the monitoring process includes: establishing a socket protocol communication channel between the robot arm and the workstation terminal, and sending an information acquisition request to the terminal; clearing the original data group and waiting for the data to be sent; receiving the original data and storing it in an array; unpacking the original data according to the ASCII code, and determining whether the end character meets the preset value. If it does not meet the preset end character value, the value is stored at the end of the temporary string until the current data is the same as the preset end character value, and the temporary string is saved in the data array.
[0052] Through the above technical solution, the workstation terminal controls the camera to extract the workpiece information and transmit it to the robot arm to adjust the cutting posture, which includes workpiece information extraction, data transmission, posture calculation, etc., solving the problem of tracking cutting after the workpiece is placed randomly. And from the perspective of the movement of the robot arm alone, during multilateral continuous operation, the adjustment of the compound posture has less restrictions on the joints of the robot body. Single Z-axis rotation is simpler in robot program control, and the convex arc actually cut is more practical and convenient for subsequent welding. However, due to the physical limitations of the robot joints, the actual application scenario can only be used in small-sized U-shapes. The inner arc cut by compound posture adjustment may have certain application potential in some application scenarios.
[0053] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A bevel cutting method based on a robot bevel cutting workstation, characterized in that: Includes steps: S1: Place the workpiece on the material table, detect its edge, scan and determine the first anchoring profile, and pack the first anchoring profile to the workstation; the workstation pre-processes the packed first anchoring profile and generates a second anchoring profile; S2: guiding the robot arm to move the image acquisition device to the correct position to perform edge acquisition on the workpiece, aligning the edge acquired image information with the second anchoring contour, and opening the grating after determining the matching degree to generate point cloud image information of the workpiece; S3: performing noise filtering on the point cloud image information, and fusing and extracting the point cloud image information after the noise filtering by using the first preset value and the second preset value to reconstruct a workpiece point cloud image; the workpiece point cloud image includes at least one feature point associated with the workpiece, and the distribution of the feature point is transmitted to the workstation; S4: The workpiece point cloud image obtained in the above step S3 is transmitted to the robot arm according to the workpiece type, groove angle, groove direction and related feature point information in the order of outer contour clockwise and inner hole counterclockwise. The robot arm is divided into the first type and the second type according to the degree of difference of the feature points, and the groove cutting posture is prepared according to the transmission order of the feature points; S5: adjusting the posture of the gun head at the end of the robot arm according to the position of the workpiece and the direction of the pre-cut groove; In the above step S1, the edge detection includes: taking an image of the workpiece placed on the material table by a camera, performing noise removal on the image by Gaussian filtering, and transferring the image into a coordinate model, wherein the coordinate model recognizes and iteratively learns the workpiece by a YOLOV5 algorithm, and recognizes the coordinate parameters of the center of the workpiece in the coordinate system with the material table as the coordinate system; The camera calibration is used to determine the intrinsic parameter matrix, the hand-eye calibration is used to obtain the extrinsic parameter matrix, and then the matrix is multiplied by the coordinate parameters of the center of the workpiece in the coordinate system to be converted into the physical position of the workpiece in the coordinate system; In the above step S1, the preprocessing is to perform local annotation point correction on the first anchor contour, arbitrarily select multiple annotation points on the edge of the first anchor contour, move the multiple annotation points to the gradient peak of the image, perform local linear fitting on them to fit the multiple annotation points into a curve, and then output the optimized second anchor image through interpolation and correction sampling; In the above step S3, the first preset value is the height, and the second preset value is the size; the noise reduction process includes: generating a Kd tree according to the workpiece point cloud image, taking any point on the Kd tree as a base point, and calculating the average value of any other point in the neighborhood of the base point within a certain radius of the base point, and judging whether the average value exceeds a threshold value, if the average value is greater than the threshold value, the point in the neighborhood is regarded as a noise point and is removed; In the above step S3, the fusion extraction is to generate a workpiece point cloud image that matches the workpiece three-dimensionally by using the height and size thresholds and the denoised point cloud image information; In the above step S4, the first type is a straight line, and the second type is an arc; the judgment process of the posture preparation is to compare the current posture with the target posture and the intermediate vertical posture around the X and Y axis rotation angles, and then calculate the difference between the respective vertical postures around the Z axis rotation; it includes: after the robot arm receives the feature point, it obtains the current position of the gun head, determines whether the current position is a feature point, and if it is a feature point, determines whether the gun head and the feature point are horizontal; if it is not a feature point, replaces the starting position with the current position for judgment; According to the above horizontal numerical judgment results, the inverse tangent value of the slope is calculated as the rotation angle when the Z axis is rotated, and the corresponding rotation is performed around the Z axis, and then the height of the feature point is judged whether it is consistent; According to the height judgment result, the inverse tangent value of the slope is calculated as the rotation angle about the X-axis, and a corresponding rotation is performed around the X-axis; then, the first type and the second type are distinguished, and the rotation direction along the Z-axis is selected through the feature point information in the groove direction, and the Euler angle and quaternion conversion is used to update the posture that meets the feature point.
2. A bevel cutting method based on a robot bevel cutting workstation according to claim 1, characterized in that: The workstation uses an ABB robot, which records angles through quaternions and drives the robot arm to rotate around the Z axis to track the workpiece at different heights.
3. A bevel cutting method based on a robot bevel cutting workstation according to claim 2, characterized in that: It also includes monitoring the position of the robot arm during actual operation to determine whether it has reached a critical point between a preset posture and a changed posture, wherein the monitoring process includes: establishing a socket protocol communication channel between the robot arm and the workstation terminal, and sending an information acquisition request to the terminal; Clear the original data group and wait for the data to be sent; Receive the original data and store it in an array; Unpack the original data according to the ASCII code, and determine whether the end character meets the preset value, and store the value that does not meet the preset end character at the end of the temporary string until the current data is the same as the preset end character value, and save the temporary string to the data array.
4. The bevel cutting method based on a robot bevel cutting workstation according to claim 1, characterized in that: It also includes an input display device, which is used to manually input the workpiece type, groove angle, and groove direction information of the workpiece.
5. The bevel cutting method based on a robot bevel cutting workstation according to claim 1, characterized in that: The image acquisition device is a 3D camera.
