A flexible grinding and polishing device and method based on vision and force control

Abstract

The present invention discloses a flexible grinding and polishing device and method based on vision and force control, and relates to the field of flexible grinding and polishing technology. The device comprises an industrial robot, a robot clamp is installed on the movable end of the industrial robot, a belt grinder is installed on the back of the industrial robot, a loading and unloading platform is installed on one side of the industrial robot, a visual inspection device is installed on the top of the loading and unloading platform, and a cloth wheel polisher is installed on the other side of the industrial robot. The visual inspection device comprises a profile frame, the bottom of the profile is connected to the loading and unloading platform, and the side wall of the profile frame is installed with a visual component. The present invention obtains an image of the workpiece through the visual inspection device, collects the position information of the three-dimensional surface point group of the workpiece grinding area, thereby determining the initial position and posture of the workpiece after clamping, and calculates the optimal grinding trajectory in real time according to the initial posture of the workpiece after clamping. The speed is extremely fast and the detection result is not affected by time changes and harsh environments. The operation is simple and the maintenance cost is low.

Description

Technical Field

[0001] The present invention relates to the technical field of flexible grinding and polishing, in particular to a flexible grinding and polishing device based on vision and force control. Background Art

[0002] At present, most of the metal hardware in the fields of plumbing and sanitary ware, such as faucets, door handles and other metal parts, rely on manual hand-held workpieces and grinding equipment for grinding and polishing. This is not only highly dangerous, but also has low grinding and polishing efficiency, high labor intensity, and requires a high level of technical proficiency of workers.

[0003] Although there are some automatic grinding and polishing equipment on the market that use robots to replace manual grinding, the application rate of robots is still relatively low. The main reasons are as follows: on the one hand, the initial one-time investment is large when applying robots; on the other hand, companies do not have sufficient R&D capabilities to solve the difficult problems encountered in actual application projects, especially the initial position of the workpiece in the installed state is inconsistent with the theoretical position and how to determine its initial installation position, the difficulty of teaching programming for complex parts, the real-time control of grinding force in the grinding and polishing state, and the wear compensation of tools (grinding wheels, sanding belts, etc.).

[0004] Aiming at the problems that may be encountered by the above-mentioned automated grinding and polishing equipment, such as wear of the grinding and polishing tools of the workpiece, complex trajectory planning, quality of the polished workpiece, inconsistency between the initial position and theoretical position of the workpiece, etc. Although there is also an intelligent device with six-dimensional force sensing at the end of the robot invented on the market, which feeds back the contact force detected by the force sensor at the end of the robot to the robot for real-time compensation adjustment of the six joints, this method is not only expensive, but also has a slow dynamic response of the robot, and the actual compensation effect is not ideal. The purpose of the present invention is to provide a flexible grinding and polishing equipment based on vision and force control. The equipment is equipped with a visual detection device and a force control device. The robot only needs to grab the workpiece and scan the three-dimensional information under the visual detection device. Then, it can move to the corresponding grinding and polishing station according to the grinding and polishing path planned by the visual system to perform constant torque grinding and polishing. Summary of the Invention

[0005] The purpose of the present invention is to provide a flexible grinding and polishing device based on vision and force control to solve the problems raised in the above background technology.

[0006] To achieve the above-mentioned objectives, the present invention provides the following technical solutions: a flexible grinding and polishing equipment based on vision and force control, comprising an industrial robot, a robot clamp is installed on the movable end of the industrial robot, a belt grinder is installed on the back of the industrial robot, a loading and unloading platform is installed on one side of the industrial robot, a visual inspection device is installed on the top of the loading and unloading platform, and a cloth wheel polisher is installed on the other side of the industrial robot, the visual inspection device includes a profile frame, the bottom of the profile is connected to the loading and unloading platform, and the side wall of the profile frame is installed with a visual component, and the working range of the industrial robot includes the cloth wheel polisher, visual inspection device and belt grinder.

[0007] Furthermore, the cloth wheel polishing machine includes a base, an adjustment motor is installed at the bottom of the base, the output end of the adjustment motor passes through the bottom of the base and is fixedly connected to an adjustment gear, the top of the base is slidably connected to a tooth plate that meshes with the adjustment gear, a spindle motor is installed on the top of the tooth plate, the output ends on both sides of the spindle motor are fixedly connected to the polishing cloth wheel, the top of the base is fixedly connected to an auxiliary table, and the top of the auxiliary table is in sliding contact with the spindle motor.

[0008] Furthermore, the profile frame is fixedly connected with acrylic plates except for one side and the bottom outer wall. The industrial robot is located on the side of the profile frame where the acrylic plate is not provided. The loading and unloading platform includes a support frame, the top of the support frame is fixedly connected with a plate, the top of the plate is fixedly connected with a positioning component, and the positioning component is located inside the profile frame.

[0009] Furthermore, the belt sander includes a support base, the top of the support base is fixedly connected to a mounting frame, the top and bottom of both sides of the mounting frame are rotatably connected to a driving wheel, a force control device, and an adjustable adjusting wheel, the force control device includes side frames fixed on both sides of the mounting frame, an adjusting cylinder is installed inside the side frame, the output end of the adjusting cylinder is fixedly connected to a movable plate, the top and bottom of the side of the movable plate are respectively rotatably connected to a driven wheel and an auxiliary wheel through mounting strips, the outer walls of the driving wheel, driven wheel, auxiliary wheel and adjusting wheel are provided with a closed-loop grinding belt, and the top and bottom of the back side of the mounting frame are respectively installed with driving motors corresponding to the position of the driving wheel, and the driving motors are respectively driven by the driving wheels at the corresponding positions through transmission belts.

[0010] Furthermore, a control cabinet is installed on the top of the mounting frame, and the control cabinet is electrically connected to the driving motor and the regulating wheel respectively.

[0011] Furthermore, the robot fixture includes a rotating slip ring mechanism, the top of which is fixedly connected to the industrial robot, the bottom of which is rotatably connected to a switching cylinder, and the output end of the switching cylinder is fixedly connected to a tensioning cylinder via a connecting frame.

[0012] Furthermore, the angle adjustment range of the switching cylinder is 0-90°, and the industrial robot is a rotatable six-joint robot.

[0013] Furthermore, a control unit is installed on the front of the industrial robot, and the industrial robot, robot fixture, visual inspection device, cloth wheel polisher and belt grinder are electrically connected to the control unit respectively.

[0014] A flexible grinding and polishing method based on vision and force control uses a flexible grinding and polishing device based on vision and force control. The grinding method is specifically divided into the following steps:

[0015] Step 1: Place the workpiece to be polished on the positioning assembly of the loading and unloading table, and the industrial robot grabs the workpiece through the robot fixture;

[0016] Step 2: The grasped workpiece is transported to the visual inspection device, where the visual component collects the workpiece's 3D contour information. A template point cloud is then created using a 3D reconstruction algorithm. Finally, the robot's trajectory planning software or direct teaching is used to generate a polishing program based on this 3D point cloud information.

[0017] Step 3: Use the ICP point cloud registration algorithm to calculate the faucet clamping error. The deviation translation and angle are compensated in the form of Euler angles into the tool coordinate system originally calibrated by the robot. The updated tool coordinate system is used to execute the previously taught grinding program.

[0018] Step 4: The industrial robot grinds and polishes the workpiece in sequence through the belt grinder and cloth wheel polisher according to the grinding and polishing trajectories;

[0019] Step 5: The industrial robot places the workpiece back to its original position.

[0020] Furthermore, the three-dimensional reconstruction algorithm in step 2 includes:

[0021] S1: Coordinate system construction, recorded in the default tool coordinate system of the flange end of the industrial robot , with the bottom base point of the industrial robot as , the visual measurement coordinate system on the visual component is recorded as In order to facilitate the calibration of the hand-eye matrix, the workpiece coordinate system of the industrial robot is first recorded as , when calibrating With visual measurement coordinate system The origin coincides, and the workpiece coordinate system at this time is marked as ;

[0022] S2: Turn on the visual inspection device, move the active end tip of the industrial robot to the image center of the visual component (x1, y1), and use the three-point calibration method of the industrial robot to calibrate this point as the origin of the workpiece coordinate system. Then move the tip to the image pixel (u, y1) to calibrate it as the x-axis of the workpiece coordinate system, and then move the tip to the image pixel (x1, v) to calibrate it as the y-axis of the workpiece coordinate system. Among them, u is any value greater than x1, and v is any value greater than y1. Finally, the calibration of the workpiece coordinate system is completed. At this time, the x-axis of the calibrated workpiece coordinate system coincides with the x-axis of the laser plane measurement coordinate system, and the y-axis coincides with the z-axis of the laser plane measurement coordinate system. Since the three-point calibration method of the workpiece in the robot control system is used, the workpiece coordinate system can be obtained. and the robot base coordinate system Conversion relationship ;

[0023] S3: Vision measurement coordinate system and workpiece coordinate system The origin is the same, the x-axis direction is the same, the coordinate system The z-axis and coordinate system The y-axis coincides with the visual measurement coordinate system and workpiece coordinate system Conversion relationship :

[0024] ;

[0025] S4: The industrial robot obtains the flange end coordinate system of the industrial robot during each frame scan. Relative to the robot base coordinate system The posture transformation relationship is ;

[0026] S5: Finally, the three-dimensional point cloud data in the visual measurement coordinate system is obtained Convert to the industrial robot flange end coordinate system Get the new point cloud data of the workpiece , complete the workpiece point cloud in the flange end coordinate system The specific coordinate transformation formula for 3D reconstruction is as follows:

[0027] .

[0028] Furthermore, the grinding and polishing in step 4 are both performed with constant torque.

[0029] Furthermore, the ICP point cloud registration algorithm in step 3 is specifically as follows:

[0030] M1: The scanned point cloud is converted to the industrial robot flange end coordinate system through the visual component and hand-eye calibration to form the actual measurement point cloud;

[0031] M2: Then, the actual measured point cloud is matched with the template point cloud by ICP matching, and the obtained rotation and translation matrix is compensated to the robot tool coordinate system in the form of Euler angles;

[0032] M3: Use the compensated tool coordinate system to execute the previously generated grinding program.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] This flexible grinding and polishing equipment based on vision and force control obtains the image of the workpiece through a visual detection device, collects the position information of the three-dimensional surface point group of the workpiece grinding area, and thus determines the initial position and posture of the workpiece after clamping. According to the initial posture of the workpiece after clamping, the optimal grinding trajectory is calculated in real time. It uses 3D vision instead of the human eye, is extremely fast, and the detection results are not affected by time changes and harsh environments. It is simple to operate and has low maintenance costs, which is conducive to the realization of automated product production.

[0035] At the same time, the force control device controls the end position and contact force of the grinding and polishing equipment tool, and realizes full closed-loop control of the contact force of the grinding and polishing equipment through force feedback, so that the grinding and polishing tool shows strong flexibility to the workpiece surface, improves the grinding accuracy of the robot, and avoids the huge contact force caused by small position errors of traditional robots. In addition, the force control device has low cost and fast dynamic response speed.

[0036] Furthermore, the system integrates visual inspection, force control, and the robot's front-end gripper based on specific part dimensions and geometries, along with CAD 3D models of the processed parts and theoretical grinding trajectory generation technology, to form a multi-sensor information-fusion robotic flexible grinding and polishing system. This system addresses the challenges of complex traditional teaching and contact force control, actively promoting the application of industrial robots in grinding and polishing processes, improving stability and precision during part production, reducing errors and defect rates, and ensuring product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a schematic diagram of the overall installation of the present invention;

[0038] Figure 2 This is a schematic structural diagram of the industrial robot of the present invention;

[0039] Figure 3 For the present invention Figure 2 A in the middle is an enlarged schematic diagram;

[0040] Figure 4 It is a structural schematic diagram of the cloth wheel polishing machine of the present invention;

[0041] Figure 5 This is a structural schematic diagram of the loading and unloading platform of the present invention;

[0042] Figure 6 Schematic diagram of the structure of the visual inspection device of the present invention;

[0043] Figure 7 It is a structural schematic diagram of the belt grinder of the present invention;

[0044] Figure 8 For the present invention Figure 7 The enlarged schematic diagram of point B in the middle;

[0045] Figure 9 Reconstruct the effect image of the faucet point cloud;

[0046] Figure 10 Measure the point cloud of a faucet with clamping error;

[0047] Figure 11 This is the point cloud image of the faucet template without clamping error;

[0048] Figure 12 It is the downsampled tap point cloud image;

[0049] Figure 13 This is the ICP point cloud registration effect diagram;

[0050] Figure 14 Polishing the technology roadmap for 3D vision-based faucets;

[0051] Figure 15 This is the overlap diagram of the two coordinate systems in hand-eye calibration.

[0052] Figure: 1. Industrial robot; 2. Cloth wheel polisher; 201. Base; 202. Spindle motor; 203. Polishing wheel; 204. Adjustment motor; 205. Adjustment gear; 206. Tooth plate; 207. Auxiliary table; 3. Loading and unloading table; 301. Support frame; 302. Plate; 303. Positioning component; 4. Visual inspection device; 401. Profile frame; 402. Acrylic plate; 403. Visual component; 5. Belt sander; 501. Support base; 5 02. Mounting frame; 503. Control cabinet; 504. Drive motor; 505. Driving pulley; 506. Transmission belt; 507. Driven pulley; 508. Auxiliary wheel; 509. Grinding belt; 510. Adjusting wheel; 511. Force control device; 512. Side frame; 513. Adjusting cylinder; 514. Movable plate; 515. Mounting strip; 6. Control unit; 7. Robot fixture; 701. Rotating slip ring mechanism; 702. Switching cylinder; 703. Tensioning cylinder. DETAILED DESCRIPTION

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0054] It should be noted that, in the description of the present invention, the terms "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship, are based on the orientation or position relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0055] Furthermore, it should be understood that for the sake of ease of description, the sizes of the various components shown in the drawings are not drawn according to actual proportions. For example, the thickness or width of certain layers may be exaggerated relative to other layers.

[0056] It should be noted that like numbers and letters represent similar items in the following figures, so once an item is defined or described in one figure, it will not need to be further discussed and described in detail in the description of the subsequent figures.

[0057] like Figure 1-15 As shown, the present invention provides a technical solution: a flexible grinding and polishing equipment based on vision and force control, including an industrial robot 1, a robot clamp 7 is installed at the movable end of the industrial robot 1, a belt grinder 5 is installed on the back of the industrial robot 1, a loading and unloading platform 3 is installed on one side of the industrial robot 1, a visual inspection device 4 is installed on the top of the loading and unloading platform 3, and a cloth wheel polisher 2 is installed on the other side of the industrial robot 1. The visual inspection device 4 includes a profile frame 401, the bottom of the profile is connected to the loading and unloading platform 3, and the side wall of the profile frame 401 is installed with a visual component 403. The working range of the industrial robot 1 includes the cloth wheel polisher 2, the visual inspection device 4 and the belt grinder 5.

[0058] It should be noted that the visual inspection device 4 consists of a profile frame 401 made of aluminum profile, an acrylic plate 402 and a visual component 403, wherein the acrylic plate 402 is used for "light blocking" to ensure the visual inspection effect, and the visual component 403 includes industrial cameras, lasers, filters, etc., which are mainly used for three-dimensional scanning and modeling of parts to be processed.

[0059] The cloth wheel polishing machine 2 includes a base 201, and an adjustment motor 204 is installed at the bottom of the base 201. The output end of the adjustment motor 204 passes through the bottom of the base 201 and is fixedly connected to an adjusting gear 205. The top of the base 201 is slidably connected to a tooth plate 206 that meshes with the adjustment gear 205. The top of the tooth plate 206 is installed with a spindle motor 202. The output ends on both sides of the spindle motor 202 are fixedly connected to the polishing cloth wheel 203. The top of the base 201 is fixedly connected to an auxiliary table 207, and the top of the auxiliary table 207 is in sliding contact with the spindle motor 202. The cloth wheel polishing machine 2 has two left and right stations, and can also be equipped with polishing cloth wheels 203 of corresponding specifications according to the needs of the workpiece polishing process to achieve rough polishing and fine polishing of the workpiece. The bottom of the base 201 is fixedly installed with an adjustment motor 204, and the adjustment motor 204 can make the top-mounted spindle motor 202 move back and forth through the cooperation of the adjusting gear 205 and the tooth plate 206, thereby achieving the purpose of polishing with a fixed torque.

[0060] Except for one side and the bottom outer wall, the profile frame 401 is fixedly connected with an acrylic plate 402. The industrial robot 1 is located on the side of the profile frame 401 where the acrylic plate 402 is not provided. The loading and unloading platform 3 includes a support frame 301. The top of the support frame 301 is fixedly connected with a plate 302. The top of the plate 302 is fixedly connected with a positioning component 303. The positioning component 303 is located inside the profile frame 401.

[0061] It should be noted that the loading and unloading platform 3 is composed of an aluminum alloy panel and an aluminum profile bracket, forming a support frame 301 and a plate 302, wherein the plate 302 composed of the aluminum alloy panel has 36 through holes, which can be matched with corresponding tooling fixtures according to the shape specifications of the parts.

[0062] The belt sander 5 includes a support base 501, a mounting frame 502 is fixedly connected to the top of the support base 501, and the top and bottom of both sides of the mounting frame 502 are rotatably connected to a driving wheel 505, a force control device 511, and an adjustable adjustment wheel 510. The force control device 511 includes a side frame 512 fixed to both sides of the mounting frame 502, an adjusting cylinder 513 is installed inside the side frame 512, and a movable plate 514 is fixedly connected to the output end of the adjusting cylinder 513. The top and bottom of the side of the movable plate 514 are rotatably connected by mounting strips 515. A driven wheel 507 and an auxiliary wheel 508 are connected, and a closed-loop grinding belt 509 is provided on the outer walls of the driving wheel 505, the driven wheel 507, the auxiliary wheel 508 and the adjusting wheel 510. Drive motors 504 corresponding to the positions of the driving wheel 505 are respectively installed on the top and bottom of the back side of the mounting frame 502. The drive motors 504 are respectively driven by the driving wheels 505 at the corresponding positions through transmission belts 506. A control cabinet 503 is installed on the top of the mounting frame 502, and the control cabinet 503 is electrically connected to the drive motor 504 and the adjusting wheel 510 respectively.

[0063] It should be noted that the belt grinder 5 has a total of four workstations on which four sanding belts of different mesh sizes and widths can be assembled according to the workpiece grinding process requirements. The forward and backward directions of the sanding belt wheels (including the driven wheel 507 and the auxiliary wheel 508) are controlled by a force control device 511. The force control device 511 is composed of a servo drive and an adjusting cylinder 513. Constant torque grinding is achieved by extending and retracting the adjusting cylinder 513, thereby realizing the forward and backward movement of the driven wheel 507 and the auxiliary wheel 508.

[0064] The robot fixture 7 includes a rotating slip ring mechanism 701, the top of which is fixedly connected to the industrial robot 1, and the bottom of which is rotatably connected to a switching cylinder 702. The output end of the switching cylinder 702 is fixedly connected to a tensioning cylinder 703 through a connecting frame. The angle adjustment range of the switching cylinder 702 is 0-90°. The industrial robot 1 is a rotatable six-joint robot. The six-joint robot can be equipped with the corresponding robot fixture 7 according to different workpieces to clamp the workpiece for three-dimensional contour scanning, grinding, and polishing. The control unit 6 is composed of an industrial tablet computer and an electrical control cabinet, which is used for the action logic control of the entire equipment line and the planning of the robot processing trajectory.

[0065] It should be noted that: Rotating slip ring mechanism 701: The air pipe of the control cylinder is directly inserted into the structure. The outside of the structure will not rotate with the rotation of the robot flange, and there will be no problems such as air pipe entanglement and interference. However, the inside can rotate normally to drive the movement of the fixture.

[0066] 0-90° switching cylinder 702 structure: according to the needs of each polishing surface of the faucet, it switches to a corresponding appropriate angle so that the faucet can be adjusted to a suitable spatial posture for polishing on the sanding belt.

[0067] Elastic and tightening cylinder 703: Through the principle of lever, the cylinder is extended and retracted to drive the clamp to tighten or loosen the faucet.

[0068] A control unit 6 is installed on the front of the industrial robot 1 , and the industrial robot 1 , the robot fixture 7 , the visual inspection device 4 , the cloth wheel polisher 2 and the belt sander 5 are electrically connected to the control unit 6 respectively.

[0069] The grinding method is divided into the following steps:

[0070] Step 1: placing the workpiece to be polished on the positioning assembly (303) of the loading and unloading platform (3), and the industrial robot (1) grabs the workpiece through the robot fixture (7);

[0071] Step 2: transport the grasped workpiece to the interior of the visual inspection device (4), collect the three-dimensional contour information of the workpiece through the visual component (403), and then create a template point cloud through a three-dimensional reconstruction algorithm. Finally, combine this three-dimensional point cloud information to generate a polishing program through robot trajectory planning software or directly through teaching;

[0072] Step 3: Use the ICP point cloud registration algorithm to calculate the faucet clamping error. The deviation translation and angle are compensated in the form of Euler angles into the tool coordinate system originally calibrated by the robot. The updated tool coordinate system is used to execute the previously taught grinding program.

[0073] Step 4: The industrial robot (1) grinds and polishes the workpiece in sequence through the belt grinder (5) and the cloth wheel polisher (2) according to the grinding and polishing trajectories;

[0074] Step 5: The industrial robot (1) places the workpiece back to its original position.

[0075] It should be noted that the 3D reconstruction algorithm in step 2 includes:

[0076] S1: Coordinate system construction, recorded as the default tool coordinate system of the flange end of industrial robot 1 , with the bottom base point of industrial robot 1 as , the visual measurement coordinate system on the visual component is recorded as In order to facilitate the calibration of the hand-eye matrix, the workpiece coordinate system of the industrial robot 1 is first recorded as , when calibrating With visual measurement coordinate system The origin coincides, and the workpiece coordinate system at this time is marked as ;

[0077] S2: Turn on the visual inspection device, move the active end tip of the industrial robot 1 to the image center of the visual component 401 (x1, y1), and use the three-point calibration method of the industrial robot 1 to calibrate this point as the origin of the workpiece coordinate system. Then move the tip to the image pixel (u, y1) to calibrate it as the x-axis of the workpiece coordinate system, and then move the tip to the image pixel (x1, v) to calibrate it as the y-axis of the workpiece coordinate system. Among them, u is any value greater than x1, and v is any value greater than y1. Finally, the calibration of the workpiece coordinate system is completed. At this time, the x-axis of the calibrated workpiece coordinate system coincides with the x-axis of the laser plane measurement coordinate system, and the y-axis coincides with the z-axis of the laser plane measurement coordinate system. Since the three-point calibration method of the workpiece in the robot control system is used, the workpiece coordinate system can be obtained. and the robot base coordinate system Conversion relationship ;

[0078] S3: Vision measurement coordinate system and workpiece coordinate system The origin is the same, the x-axis direction is the same, the coordinate system The z-axis and coordinate system The y-axis coincides with the visual measurement coordinate system and workpiece coordinate system Conversion relationship :

[0079] ;

[0080] S4: The flange end coordinate system of the industrial robot 1 is obtained by the industrial robot 1 during each frame scan. Relative to the robot base coordinate system The posture transformation relationship is ;

[0081] S5: Finally, the three-dimensional point cloud data in the measurement coordinate system is obtained Convert to the industrial robot 1 flange end coordinate system Get the new point cloud data of the workpiece , complete the workpiece point cloud in the flange end coordinate system The specific coordinate transformation formula for 3D reconstruction is as follows:

[0082] .

[0083] The ICP point cloud registration algorithm in step 3 is as follows:

[0084] M1: The scanned point cloud is converted to the flange end coordinate system of the industrial robot 1 through the vision component 403 and the hand-eye calibration to form the actual measurement point cloud;

[0085] M2: Then, the actual measured point cloud is matched with the template point cloud by ICP matching, and the obtained rotation and translation matrix is compensated to the robot tool coordinate system in the form of Euler angles;

[0086] M3: Use the compensated tool coordinate system to execute the previously generated grinding program.

[0087] The complete process flow is

[0088] The six-joint industrial robot 1 moves to the loading and unloading platform 3 to grab the workpiece, then moves the workpiece to the visual inspection device 4. Vision component 403 determines the initial assembly pose information of the polished workpiece in the base coordinate system of the industrial robot 1. Based on the CAD model and the 3D point group data detected online, the optimal grinding and polishing trajectory is determined. Following the optimal grinding and polishing trajectory, the industrial robot 1 moves to the corresponding workstations of the belt grinder 5 and the cloth wheel polisher 2 in sequence according to process requirements to perform grinding and polishing. During the grinding and polishing process, the force control device 511 control system of the belt grinder 5 and the cloth wheel polisher 2 collects information such as the contact pressure between the grinding and polishing tool and the workpiece, and the motor current in real time. In torque control mode, the grinding trajectory is corrected and compensated in real time, achieving precise and flexible grinding and polishing of the part. Finally, the industrial robot 1 places the polished part back to the loading and unloading platform 3 and continues to grab the next part. Example

[0089] Use visual technology to correct the faucet clamping error, and combine it with the robot's teaching trajectory to perform automatic faucet polishing.

[0090] The current faucet grinding and polishing process still relies on skilled workers holding the rough workpiece and contacting the abrasive belt to complete the grinding and polishing. Figure 1 As shown, the manual grinding and polishing environment is harsh, the production efficiency is low, the consistency of processing quality is difficult to guarantee, and the workers' experience is required to be high, and the product consistency is poor.

[0091] Compared with CNC machining centers, industrial robots used in grinding and polishing have a high level of automation and good flexibility. They can realize various postures of grinding and polishing workpieces in space and can meet the requirements of grinding and polishing of medium-complex free-form surfaces similar to faucets.

[0092] Industrial robot grinding and polishing systems integrate various general-purpose equipment onto the robot itself. These systems offer strong versatility and a flexible range of applications, making them superior to CNC machine tool grinding and polishing. Furthermore, grinding and polishing processes involve minimal cutting, resulting in minimal normal contact force between the abrasive belt and the workpiece surface, and requiring minimal precision. This avoids the shortcomings of industrial robots, such as weak rigidity and poor precision. Therefore, industrial robots are particularly suitable for grinding and polishing curved products requiring a smooth surface, as they suffer from the low efficiency of manual grinding and polishing, the limited versatility of CNC machining centers, and the inflexible processing range of CNC machining centers.

[0093] To solve the above problems:

[0094] The faucet point cloud clamped in the robot flange end coordinate system is measured in three dimensions using a monocular line structured light camera, and the ICP (Iterative Closest Point) algorithm in the third-party visual algorithm library PCL is used to align the conversion pose parameters (rotation vector, translation vector) between the template point cloud and the scanned point cloud with clamping error. This conversion pose parameter represents the rotation angle and translation amount between the template point cloud and the scanned point cloud with clamping error. The (x, y, z, a, b, c) in the robot tool coordinate system is modified according to the rotation angle and translation amount, and then the previously taught trajectory is executed using this changed tool coordinate system to complete the polishing of the faucet with clamping error. That is, the polishing trajectory of the faucet with clamping error is consistent with the polishing trajectory of the faucet without clamping error. This solves the problem of low efficiency of manual grinding and polishing.

[0095] 2. Using a robot to polish faucets significantly increases the level of automation. The robot grips the workpiece, with the faucet mounted on the end of the robot. This device can not only polish complex curved workpieces like faucets, but also relatively simple flat surfaces, demonstrating its versatility in polishing.

[0096] 3. To achieve optimal polishing results and more comprehensive polishing of all faucet surfaces, a 90° rotatable cylinder is installed at the end of the flange, facilitating the robot's swinging position. The appropriate cylinder position is selected for polishing different faucet surfaces. This device addresses the limited processing range of CNC machining centers, enabling precise polishing of every faucet surface.

[0097] pass Figure 14 A detailed explanation of the faucet polishing technology route based on 3D vision.

[0098] Solution Description:

[0099] Steps to create a 3D scanning measurement trajectory for a template point cloud

[0100] (1) Teaching programming of the initial 3D scanning trajectory, which is called TT;

[0101] (2.0) According to the initial, teaching programming, the three-dimensional scanning trajectory TT is obtained, the workpiece is three-dimensionally scanned, and the three-dimensional point cloud data of the key parts of the workpiece outer surface are obtained;

[0102] (2.1) Based on the first 3D scan, the faucet surface 3D point cloud data is obtained, and the 3D scanning trajectory is optimized to obtain a new 3D scanning trajectory, which is called PT;

[0103] (3) Perform three-dimensional scanning on the workpiece according to the three-dimensional scanning trajectory PT to obtain three-dimensional point cloud data of key parts of the workpiece outer surface;

[0104] (4) Determine whether the quality of the three-dimensional point cloud data obtained at this time meets the data quality required for use as a template; if qualified, end the work of collecting the workpiece template data; if unqualified, go to step (2.1).

[0105] The polishing plan for faucets is mainly divided into the following steps

[0106] (1) First calibrate the tool and workpiece coordinate systems, then perform teaching programming to obtain the initial grinding trajectory trace0 of the faucet.

[0107] (2) Use the initial teaching trajectory TT to scan and obtain the faucet point cloud A. Based on the judgment of the completeness of the initially scanned faucet point cloud A, optimize the teaching trajectory TT to become the trajectory PT to obtain the faucet point cloud B. Determine whether point cloud B meets the requirements for constructing the template point cloud. If not, re-optimize the teaching trajectory PT. If it does, a template point cloud D with good data quality can be obtained.

[0108] (3) Next, a 3D camera is used to scan the faucet to obtain the actual measured point cloud C. The rotation and translation matrix of the workpiece’s posture error in the clamped state is obtained through the ICP point cloud registration method.

[0109] (4) The pose error matrix obtained by ICP registration in the clamped state is compensated into the original faucet workpiece coordinate system and updated to the new faucet workpiece coordinate system.

[0110] (5) Finally, the taught grinding trajectory is executed to complete the grinding process of the faucet. Thus, the grinding of the faucet with clamping error is realized.

[0111] Specific implementation plan:

[0112] 1. Hand-eye calibration using monocular line structured light

[0113] Hand-eye calibration is used to determine the relationship between the visual coordinate system and the robot coordinate system, enabling subsequent faucet polishing operations. This can also be understood as converting the faucet point cloud scanned by the vision sensor into the robot flange end-plate coordinate system. This solution utilizes an "eye-to-hand" camera setup.

[0114] like Figure 15The figure shows the coordinate conversion diagram of the hand-eye calibration of the faucet polishing system. The two-dimensional coordinates in the picture taken by the camera are converted into three-dimensional coordinates in the visual measurement coordinate system. Hand-eye calibration converts the three-dimensional coordinates in the visual measurement coordinate system into the three-dimensional coordinates of the end of the robot flange, so that the faucet point cloud can be reconstructed in three dimensions, so that the X-axis of this workpiece coordinate system coincides with the X-axis of the visual measurement coordinate system, and the Y-axis of the workpiece coordinate system coincides with the Z-axis of the visual measurement coordinate system. In this way, there is a rotation relationship between the visual measurement coordinate system and the calibrated workpiece coordinate system. Among them, Indicates the default tool coordinate system of the robot's flange end; is the base coordinate system of the robot, which is located on the base of the robot; Represents the visual measurement coordinate system on the monocular line structured light camera, which is located on the laser plane; in order to facilitate the calibration of the hand-eye matrix, first calibrate is the workpiece coordinate system of the robot, and when calibrating, With the origin of the visual measurement coordinate system coincide; Represents the workpiece coordinate system on the faucet.

[0115] 1.1 Calibration tool coordinate system

[0116] Install the tool tip at the end of the flange and calibrate the tool coordinate system of the tip using the four-point method (The calibration of the tool tip here is only for calibrating the workpiece coordinate system using the tip ).

[0117] 1.2 Calibrate the workpiece coordinate system

[0118] The calibrated workpiece coordinate system and the laser plane measurement coordinate system coincide with each other in two-dimensional space.

[0119] Turn on the line light source laser measurement system, move the tool tip of the robot to the center of the camera image (640,518), and use the three-point calibration method of the robot teach pendant to calibrate this point as the origin of the workpiece coordinate system. Then move the tip to the image pixel (u,518) to calibrate it as the x-axis of the workpiece coordinate system, and then move the tip to the image pixel (640,v) to calibrate it as the y-axis of the workpiece coordinate system. Among them, u is any value greater than 640, and v is any value greater than 518. Finally, the calibration of the workpiece coordinate system is completed. At this time, the x-axis of the calibrated workpiece coordinate system coincides with the x-axis of the laser plane measurement coordinate system, and the y-axis coincides with the z-axis of the laser plane measurement coordinate system. Since the three-point calibration method of the workpiece in the robot control system is used, the workpiece coordinate system can be obtained. and the robot base coordinate system Conversion relationship .

[0120] Since the visual measurement coordinate system and workpiece coordinate system The origin is the same, the x-axis direction is the same, the coordinate system The z-axis and coordinate system The y-axis of the two coordinate systems coincides, so there is a certain rotation relationship between the two coordinate systems. According to calculations, the visual measurement coordinate system on the laser plane and workpiece coordinate system Conversion relationship as follows:

[0121] .

[0122] 2. 3D reconstruction renderings

[0123] The workpiece coordinate system can be obtained by the three-point calibration method of the robot control system With the robot base coordinate system Conversion relationship .

[0124] The robot flange end coordinate system is obtained by the Wasu robot type 3 interface during each frame scan. Relative to the robot base coordinate system Posture .

[0125] The transformation matrix between the visual measurement coordinate system obtained by hand-eye calibration and the calibrated workpiece coordinate system , the transformation matrix between the workpiece coordinate system and the robot base coordinate system , the inverse transformation matrix between the robot flange end coordinate system and the robot base coordinate system , the 3D point cloud data in the visual measurement coordinate system can be Convert to the robot flange end coordinate system to form a three-dimensional point cloud data , the specific conversion formula is as follows:

[0126] .

[0127] The final 3D point cloud reconstruction effect of the faucet is as follows Figure 9 As shown:

[0128] 3. Perform ICP point cloud registration

[0129] Use the 3D vision sensor to scan the point cloud of the faucet in two different installation positions. Figure 10 and Figure 11 They are the measurement point cloud of the faucet with clamping error and the point cloud of the faucet template without clamping error.

[0130] In order to improve the speed of ICP registration, a voxel filtering downsampling operation is required. This operation reduces the number of point clouds while maintaining the characteristics of the point clouds. This greatly improves the speed of point cloud registration. Figure 12 In the figure, the point cloud on the right is the faucet measurement point cloud with clamping error after downsampling, and the point cloud on the left is the faucet template point cloud without clamping error after downsampling.

[0131] According to the ICP point cloud registration algorithm, the faucet point cloud without clamping error is converted to the faucet point cloud with clamping error, and finally a rotation and translation matrix can be obtained. Figure 13 The point cloud on the left side of the center is the registered faucet point cloud. It can be observed that the registered point cloud almost overlaps with the faucet template point cloud with clamping errors, proving that the registration is ideal and the resulting rotation and translation matrix is highly reliable.

[0132] 4. Correct the faucet clamping posture

[0133] Faucet workpiece coordinate system Coordinate system relative to the flange Coordinate transformation The matrix is obtained through the robot four-point calibration method and is stored in the robot tool coordinate system tool in the form of Euler angles.

[0134] The actual faucet workpiece coordinate system is obtained through ICP registration Relative to the standard faucet workpiece coordinate system The matrix transformation relationship of the installation posture is , that is, the matrix .

[0135] pass By multiplying the matrices, we can obtain the transformation matrix of the faucet workpiece coordinate system relative to the flange end after the pose transformation, and then convert it into Euler angles to update the value of the robot tool coordinate system tool.

[0136] Voxelized grid downsampling is a filtering method that reduces the number and density of point clouds while preserving their features. Using voxelized grid downsampling can speed up ICP registration.

[0137] When a point cloud data is input, a 3D voxel grid (which can be considered as a collection of small 3D cube structures) is created on this point cloud data, and then the points in each voxel are replaced with their centroid points. Finally, the points after each voxel processing are combined into a point cloud data for output, that is, the voxel grid downsampling operation is completed, and the tap point cloud data with reduced point cloud density is obtained, such as Figure 12 The vertical direction is the faucet template point cloud after sampling, and the oblique direction is the actual measured point cloud of the faucet after downsampling.

[0138] right Figure 12 The tap point cloud in the ICP registration is performed. ICP (Iterative Closest Point Algorithm), for the two registered point clouds, first determines the corresponding point set according to certain rules. and , where the number of corresponding point sets is n pairs. Then the optimal coordinate transformation is iteratively calculated according to the least squares method, that is, the rotation matrix and the translation vector t, and minimize the error function.

[0139] Given two 3D point sets and , ICP registration steps are as follows: Step 1, calculate Every point in The corresponding nearest point in the point set; the second step is to find the rigid body transformation that minimizes the average distance between the above corresponding points, and find the rotation and translation parameters; the third step is to Use the rotation and translation parameters obtained in the previous step to obtain a new set of transformed points; in the fourth step, if the average distance between the new set of transformed points and the reference point set is less than a given threshold, the iterative calculation is stopped, otherwise the new set of transformed points is used as the new Continue iterating until the objective function is satisfied.

[0140] Finally, the faucet workpiece coordinate system without clamping error is aligned through ICP registration. Convert to the faucet workpiece coordinate system with clamping error In the figure, the pose transformation matrix is , also expressed as . Figure 13 The figure shows the effect of ICP registration of the faucet point cloud. The registered point cloud basically coincides with the point cloud of the faucet with clamping error, thus verifying the accuracy of the registration.

[0141] The rotation and translation matrix obtained by ICP registration Expressed as:

[0142] .

[0143] according to Figure 11 and Figure 10 The contrast, The calculated rotation angle of 53.6° is consistent with the actual one, which verifies the reliability of the registration from the opposite aspect.

[0144] 5. Polish the faucet.

[0145] Replace the tool coordinate system in the original teaching program with the updated tool coordinate system, and execute the previously taught trajectory to polish the faucet. This allows you to polish faucets with clamping errors.

[0146] Advantages after using the present invention:

[0147] Advantage 1: This technology can address faucet polishing problems with clamping errors. Compared to simple teach-and-play polishing, it does not constrain the workpiece's position within the robot's end-point coordinate system. This means the workpiece can rotate and translate under the robot's end-point. This allows for precise workpiece placement and allows for polishing.

[0148] Advantage 2: This technology is not limited to faucet polishing, but can also be used to polish other complex workpieces. In other words, this technology is highly versatile.

[0149] Advantage 3: Using robots to polish faucets effectively improves work efficiency.

[0150] Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is limited by the accompanying embodiments and their equivalents.

Claims

1. A flexible grinding and polishing method based on vision and force control, characterized in that: The grinding method is divided into the following steps: Step 1: Place the workpiece to be polished on the positioning assembly of the loading and unloading table, and the industrial robot grabs the workpiece through the robot fixture; Step 2: The grasped workpiece is transported to the visual inspection device, where the visual component collects the workpiece's 3D contour information. A template point cloud is then created using a 3D reconstruction algorithm. Finally, the robot's trajectory planning software or direct teaching is used to generate a polishing program based on this 3D point cloud information. Step 3: Use the ICP point cloud registration algorithm to calculate the faucet clamping error. The deviation translation and angle are compensated in the form of Euler angles into the tool coordinate system originally calibrated by the robot. The updated tool coordinate system is used to execute the previously taught grinding program. Step 4: The industrial robot grinds and polishes the workpiece in sequence through the belt grinder and cloth wheel polisher according to the grinding and polishing trajectories; Step 5: The industrial robot places the workpiece back to its original position.

2. The flexible grinding and polishing method based on vision and force control according to claim 1, characterized in that: The three-dimensional reconstruction algorithm in step 2 includes: S1: Coordinate system construction, recorded in the default tool coordinate system of the flange end of the industrial robot , with the bottom base point of the industrial robot as , the visual measurement coordinate system on the visual component is recorded as In order to facilitate the calibration of the hand-eye matrix, the workpiece coordinate system of the industrial robot is first recorded as , when calibrating With visual measurement coordinate system The origin coincides, and the workpiece coordinate system at this time is marked as ; S2: Turn on the visual inspection device, move the active end tip of the industrial robot to the image center of the visual component (x1, y1), use the three-point calibration method of the industrial robot to calibrate this point as the origin of the workpiece coordinate system, and then move the tip to the image pixel (u, y1) to calibrate it as the x-axis of the workpiece coordinate system, and then move the tip to the image pixel (x1, v) to calibrate it as the y-axis of the workpiece coordinate system, where u is any value greater than x1, and v is any value greater than y1. Finally, the calibration of the workpiece coordinate system is completed. At this time, the x-axis of the calibrated workpiece coordinate system coincides with the x-axis of the laser plane measurement coordinate system, and the y-axis coincides with the z-axis of the laser plane measurement coordinate system. Since the three-point calibration method of the workpiece in the robot control system is used, the workpiece coordinate system can be obtained. and the robot base coordinate system Conversion relationship ; S3: Vision measurement coordinate system and workpiece coordinate system The origin is the same, the x-axis direction is the same, the coordinate system The z-axis and coordinate system The y-axis coincides with the visual measurement coordinate system and workpiece coordinate system Conversion relationship : ； S4: The industrial robot obtains the flange end coordinate system of the industrial robot during each frame scan. Relative to the robot base coordinate system The posture transformation relationship is ; S5: Finally, the three-dimensional point cloud data in the visual measurement coordinate system is obtained Convert to the industrial robot flange end coordinate system Get the new point cloud data of the workpiece , complete the workpiece point cloud in the flange end coordinate system The specific coordinate transformation formula for 3D reconstruction is as follows: 。 3. The flexible grinding and polishing method based on vision and force control according to claim 1, characterized in that: The grinding and polishing in step 4 are both performed with constant torque.

4. The flexible grinding and polishing method based on vision and force control according to claim 1, characterized in that: The ICP point cloud registration algorithm in step 3 is specifically as follows: M1: The scanned point cloud is converted to the industrial robot flange end coordinate system through the visual component and hand-eye calibration to form the actual measurement point cloud; M2: Then, the actual measured point cloud is matched with the template point cloud by ICP matching, and the obtained rotation and translation matrix is compensated to the robot tool coordinate system in the form of Euler angles; M3: Use the compensated tool coordinate system to execute the previously generated grinding program.

5. A flexible grinding and polishing device based on vision and force control, using a flexible grinding and polishing method based on vision and force control according to any one of claims 1 to 4, comprising an industrial robot (1), wherein a robot fixture (7) is installed at a movable end of the industrial robot (1), and characterized in that: The industrial robot (1) is provided with a belt sander (5) on its back, a loading and unloading platform (3) is provided on one side of the industrial robot (1), a visual inspection device (4) is provided on the top of the loading and unloading platform (3), and a cloth wheel polisher (2) is provided on the other side of the industrial robot (1), the visual inspection device (4) comprises a profile frame (401), the bottom of the profile is connected to the loading and unloading platform (3), and a visual component (403) is provided on the side wall of the profile frame (401), and the working range of the industrial robot (1) includes the cloth wheel polisher (2), the visual inspection device (4) and the belt sander (5).

6. The flexible grinding and polishing equipment based on vision and force control according to claim 5, characterized in that: The cloth wheel polishing machine (2) comprises a base (201), an adjustment motor (204) is mounted on the bottom of the base (201), an output end of the adjustment motor (204) passes through the bottom of the base (201) and is fixedly connected to an adjustment gear (205), a tooth plate (206) meshing with the adjustment gear (205) is slidably connected to the top of the base (201), a spindle motor (202) is mounted on the top of the tooth plate (206), output ends on both sides of the spindle motor (202) are fixedly connected to polishing cloth wheels (203), an auxiliary table (207) is fixedly connected to the top of the base (201), and the top of the auxiliary table (207) is in sliding contact with the spindle motor (202).

7. The flexible grinding and polishing equipment based on vision and force control according to claim 5, characterized in that: The profile frame (401) is fixedly connected to an acrylic plate (402) except for one side and the bottom outer wall. The industrial robot (1) is located on the side of the profile frame (401) where the acrylic plate (402) is not provided. The loading and unloading platform (3) includes a support frame (301). The top of the support frame (301) is fixedly connected to a plate (302). The top of the plate (302) is fixedly connected to a positioning component (303). The positioning component (303) is located inside the profile frame (401).

8. The flexible grinding and polishing equipment based on vision and force control according to claim 5, characterized in that: The belt sander (5) comprises a support base (501), the top of the support base (501) is fixedly connected to a mounting frame (502), the top and bottom of both sides of the mounting frame (502) are rotatably connected to driving wheels (505), a force control device (511), and an adjustable regulating wheel (510), the force control device (511) comprises side frames (512) fixed to both sides of the mounting frame (502), an regulating cylinder (513) is installed inside the side frames (512), the output end of the regulating cylinder (513) is fixedly connected to a movable plate (514), and the movable plate ( The top and bottom of the side surface of the mounting frame (514) are rotatably connected to a driven wheel (507) and an auxiliary wheel (508) through mounting strips (515), and the outer walls of the driving wheel (505), the driven wheel (507), the auxiliary wheel (508) and the regulating wheel (510) are provided with a closed-loop grinding belt (509). The top and bottom of the back surface of the mounting frame (502) are respectively provided with driving motors (504) corresponding to the positions of the driving wheel (505), and the driving motors (504) are respectively driven by the driving wheels (505) at the corresponding positions through transmission belts (506).

9. The flexible grinding and polishing equipment based on vision and force control according to claim 8, characterized in that: A control cabinet (503) is installed on the top of the mounting frame (502), and the control cabinet (503) is electrically connected to the driving motor (504) and the regulating wheel (510) respectively.

10. The flexible grinding and polishing equipment based on vision and force control according to claim 5, characterized in that: The robot fixture (7) comprises a rotating slip ring mechanism (701), the top of the rotating slip ring mechanism (701) is fixedly connected to the industrial robot (1), the bottom of the rotating slip ring mechanism (701) is rotatably connected to a switching cylinder (702), and the output end of the switching cylinder (702) is fixedly connected to a tensioning cylinder (703) via a connecting frame.

11. The flexible grinding and polishing equipment based on vision and force control according to claim 10, characterized in that: The angle adjustment range of the switching cylinder (702) is 0-90°, and the industrial robot (1) is a rotatable six-joint robot.

12. The flexible grinding and polishing equipment based on vision and force control according to claim 5, characterized in that: A control unit (6) is installed on the front of the industrial robot (1), and the industrial robot (1), the robot fixture (7), the visual inspection device (4), the cloth wheel polisher (2) and the belt grinder (5) are electrically connected to the control unit (6) respectively.

Images