Vision-based CNC robot system
Through the visual module, the workpiece coordinates are calculated and the robot movement is controlled, the positioning problem of CNC machine tools is solved, the recognition rate and accuracy are improved, and rapid processing and efficient production are achieved.
Patent Information
- Application Number
- CN202110965745.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2021-08-23
- Publication Date
- 2024-12-31
- Estimated Expiration
- 2041-08-23
AI Technical Summary
Due to positioning problems during loading, feeding and unloading of existing CNC machine tools, the position recognition rate is low, the recognition error rate is high, and the recognition speed is slow.
The vision-based CNC robot system is adopted. The image of the loading platform is scanned through the visual module, the coordinates of the workpiece to be processed are calculated, and the movement of the robot is controlled by combining the CNC system. The servo drive is used to control the rotation of the motor to achieve accurate grasping and placement of the jaws. .
It improves the position recognition rate and accuracy of the robot, reduces the recognition error rate, improves processing efficiency, meets the needs of fast and large-scale processing, saves labor costs, and improves production efficiency.
Smart Images
Figure CN113635325B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a CNC machine tool manipulator control technology, and in particular to a vision-based CNC manipulator system. Background Art
[0002] The existing CNC machine tool manipulators often have low position recognition rate, high recognition error rate and slow recognition speed due to positioning problems during the process of loading, feeding and unloading materials. Therefore, it is necessary to improve the above problems. Summary of the invention
[0003] The purpose of the present invention is to provide a vision-based CNC manipulator system to solve the above-mentioned problems in the prior art.
[0004] The present invention discloses a vision-based numerical control manipulator system, which comprises: a lathe, a numerical control system, a manipulator, a vision module, a clamp and a loading platform; the lathe processes workpieces; the loading platform is used to place workpieces to be processed and processed workpieces; the clamp clamps the workpieces; the vision module scans and obtains an image of the loading platform, and calculates the coordinates of each workpiece to be processed through the image of the loading platform; the numerical control system divides the moving distance of the linear feed axis by the lead screw pitch of the feed axis according to the preset target position of the manipulator's linear feed axis movement or the coordinates of the workpiece to be processed calculated by the vision module, combined with the number of pulses required for the motor to rotate one circle, obtains the number of motor rotations, and then converts it into a corresponding number of pulses and sends it to a corresponding servo driver, and the servo driver controls the motor to rotate according to the received pulses, thereby controlling the movement of the manipulator's main shaft, and controlling the clamp to move to a corresponding position to grab the workpiece.
[0005] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the CNC system controls the Y-axis movement of the linear feed axis. During the Y-axis movement, the camera continuously takes pictures, and the vision module calculates the position of the workpiece in the XY coordinate system projected on the material tray plane by the X-axis and Y-axis based on the picture results.
[0006] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the vision module calculates the coordinates of the workpiece to be processed, including: taking any point on the XY coordinate system plane, establishing a three-dimensional space coordinate system, calculating the three-dimensional coordinates of the light spot in the three-dimensional space coordinate system through the light spot irradiated on the object to be grasped, and the three-dimensional coordinates of all light spots constitute the light spot coordinate point set on the upper surface of the object to be grasped; projecting the light spot coordinate point set on the upper surface of the object to be grasped onto the plane to obtain an image of the center of the object; converting the image coordinates of the object center through the camera calibration method in machine vision to obtain the coordinates of the workpiece to be processed.
[0007] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the light spot coordinate point set on the upper surface of the object to be grasped is projected onto a plane to obtain an image of the center of the object, specifically comprising: dividing the z coordinates of all points in the light spot coordinate point set on the upper surface of the object to be grasped by the maximum height of the object to be grasped set by the user, and then multiplying by 255 to obtain a value corresponding to the light spot coordinate point, which is a grayscale value; the light spot coordinate point set on the upper surface of the object to be grasped is a floating point number, and the coordinate XY values of the points in the light spot coordinate point set on the upper surface of the object to be grasped are multiplied by a predetermined magnification and rounded, and then an offset set by the user is added so that all coordinate XY values are positive integers, and the three-channel coordinate points composed of the coordinate XY values and the grayscale value are converted from all points in the light spot coordinate point set to obtain an image point set, create an image, set the image to pure black, and then draw the points in the image point set on the image by setting pixel values to obtain a height map of the upper surface of the object to be grasped, and then calculate the image coordinates of the center of the object.
[0008] According to an embodiment of the vision-based CNC manipulator system of the present invention, the position of the object is located and the image coordinates of the center of the object are calculated through a contour matching algorithm of two-dimensional image recognition.
[0009] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the CNC system calculates the coordinates of the workpiece on the platform according to the vision module, and controls the clamp to move to the corresponding position to grab the workpiece, including: controlling the Z-axis of the manipulator to lift to a safe height first, and then controlling the X-axis and Y-axis of the manipulator to move to the position of the workpiece to be clamped, with the center of the clamp and the center of the workpiece on the same vertical line; then controlling the Z-axis to descend to the material picking height, and then the system outputs a clamp retraction signal, and the three claws of the clamp tighten inward, thereby clamping the workpiece to complete the material picking.
[0010] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the CNC system first lifts the manipulator Z axis to the Z coordinate set for the material waiting position according to the material waiting position entered by the user, and then controls the manipulator X axis to move to the X coordinate of the material waiting position, waiting for the lathe processing to be completed.
[0011] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the CNC system controls the manipulator's spindle unloading, including: the CNC system controls the manipulator's Z axis to descend to the spindle unloading height, and the manipulator's X axis moves toward the spindle. When the X coordinate of the position to be unloaded is moved, the CNC system controls the unloading claw to retract and grasp the workpiece, and controls the chuck installed on the spindle to release, and then controls the manipulator to move in a direction away from the spindle to the X coordinate of the position to be unloaded, thereby completing the spindle unloading.
[0012] According to an embodiment of the vision-based CNC manipulator system described in the present invention, the vision module numbers all the identified workpieces during the first scan, and the CNC system first grabs the workpieces according to the numbers. When scanning the next workpiece to be grabbed, the area range where the next workpiece is located is obtained from the vision module, and then the Y-axis is controlled to move from the starting point to the end point of the area range. The camera will take pictures during the movement so that the vision module can calculate the position of the workpiece to be grabbed after the movement is completed.
[0013] According to an embodiment of the vision-based CNC manipulator system described in the present invention, after the lathe processes a workpiece, the CNC system sets the manipulator loading and unloading permission signal to 1, starts the manipulator spindle unloading and manipulator spindle loading operations, and sets the manipulator loading and unloading permission signal to 0 after the manipulator spindle loading is completed. When the lathe obtains the manipulator loading and unloading permission signal as 0, it starts workpiece processing. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 This is a schematic diagram of the CNC machine tool manipulator system of the present invention. DETAILED DESCRIPTION
[0015] In order to make the purpose, content, and advantages of the present invention more clear, the specific implementation methods of the present invention are further described in detail below in conjunction with the accompanying drawings and examples.
[0016] like Figure 1 As shown, the CNC machine tool processing system includes: a manipulator 1, a vision module 2, a clamp 3, a loading platform 4, a CNC system 5, and a lathe body 6.
[0017] like Figure 1 As shown, the lathe body 6 is a CNC machine tool used to process ordinary shaft parts. The CNC system 5 is the control core of the automatic loading and unloading lathe of the manipulator 1. It can control the automatic processing of the lathe, manually move the X-axis and Z-axis, and the spindle forward and reverse rotation, etc. It can also control the movement of each axis of the manipulator 1, realize the scanning process and the automatic loading and unloading process, etc. Manipulator 1: realizes the automatic loading and unloading process of the lathe. The XYZ of the manipulator 1 is the linear feed axis, and the spindle is the rotating axis on the lathe. The specific working process is two steps: the coordinates of each workpiece on the loading platform 4 are calculated by the visual module 2, and then the workpiece to be processed is grabbed from the loading platform 4 and then placed on the spindle chuck; after the workpiece is processed, the processed workpiece is taken from the spindle chuck and placed back on the loading platform 4. Visual module 2: The image of the loading platform 4 is obtained by scanning, and the coordinates of each workpiece to be processed are calculated by the image. The system takes the material according to the calculation results. The CNC system 5 calculates the coordinates of the workpiece on the platform according to the visual module 2, and controls the clamp 3 to move to the corresponding position to grab the workpiece. Loading platform 4: used for placing workpieces to be processed and processed workpieces.
[0018] like Figure 1 As shown, the overall material picking process of the CNC machine tool processing system can be roughly divided into: scanning → picking up materials → feeding → spindle unloading (when the spindle already has material) / scanning the next one → spindle loading → returning materials → releasing materials → picking up materials again.
[0019] like Figure 1 As shown, in the process of the manipulator 1 controlling the gripper 3 to pick up materials, it is necessary to control the movement of the three axes of the manipulator 1. The specific control process of the axis movement is: first, the system moves the axis according to the target position, which can be the position of a certain point set by the user, or the position of the workpiece calculated by the visual module 2, etc. Then, the CNC system 5 divides the moving distance by the screw pitch according to the number of pulses required for the motor to rotate one circle (the corresponding axis will move a screw pitch length) to obtain the number of motor rotations, and then converts it into the corresponding number of pulses and sends it to the corresponding servo driver. The servo driver controls the motor rotation according to the received pulses, thereby controlling the movement of the spindle.
[0020] like Figure 1 As shown, the numerical control system 5 controls the Y-axis movement of the robot 1. During the Y-axis movement, the camera installed on the X-axis of the robot 1 through the bracket will take pictures continuously. After the scanning is completed, the visual module 2 will calculate the position of each workpiece in the XY coordinate system projected on the material tray plane by the X-axis and Y-axis of the robot 1 based on the picture results.
[0021] like Figure 1 As shown, the specific working principle of the visual module 2 is as follows:
[0022] The 3D vision guidance system calculates the mechanical coordinates and guides the mechanical axis to grasp. The calculation of mechanical coordinates is mainly divided into three steps:
[0023] Step 1. Scan the plane where the workpiece to be processed is placed as the reference plane, and take the reference plane as the XY plane of the coordinate system. The Z axis is perpendicular to the XY plane. Select any point on the XY plane to establish a three-dimensional space coordinate system. The light spot irradiated on the object to be grasped is used to calculate the three-dimensional coordinates of the light spot in the above coordinate system by triangulation. The three-dimensional coordinate points of all the light spots can form the light spot coordinate point set on the upper surface of the object to be grasped;
[0024] Step 2. Project the light spot coordinate point set on the upper surface of the object to be grasped onto the plane to obtain a depth map. The specific method is: divide the z coordinates of all points in the light spot set on the upper surface of the object to be grasped by the maximum height of the object to be grasped set by the user, and then multiply by 255 to obtain the value corresponding to the point, which is called the gray value; since the light spot coordinate point set on the upper surface of the object to be grasped is often a high-precision signed floating point number, the XY coordinates of the points in the light spot coordinate point set on the upper surface of the object to be grasped can be multiplied by a magnification factor (the specific factor can be determined by the accuracy requirement) and rounded, and then add the offset set by the user to make the X and Y values of all coordinates positive integers. Use the three-channel coordinate points composed of X, Y, and gray values to transform all the points in the light spot coordinate point set. The converted point set is called the image point set. Create a sufficiently large image, set the image to pure black, and then draw the points in the image point set on the image by setting pixel values to obtain a height map of the upper surface of the object to be grasped. The position of the object can be located and the image coordinates of the center of the object can be calculated through the contour matching algorithm in two-dimensional image recognition.
[0025] Step 3. Obtain the conversion relationship between the image and the mechanical coordinates through the camera calibration method in machine vision (this method can be found in major books), and convert the image coordinates of the object center obtained in step 2 into the mechanical two-dimensional coordinates X, Y;
[0026] like Figure 1 As shown in the figure, this system is developed based on LinuxCNC. LinuxCNC comes with a HAL signal mechanism. After the visual module 2 calculates the X and Y coordinates of the workpiece on the material placement platform, it sets the coordinates to the corresponding signals agreed upon in advance. When the system needs to pick up the material, it determines the position of the workpiece on the plane by reading the value of the corresponding position signal.
[0027] Material picking: The material picking process is to control the Z-axis movement of the manipulator 1 equipped with the gripper 3 according to the X-coordinate and Y-coordinate calculated by the vision module 2 and the material picking height entered by the operator. The specific process is: the Z-axis of the manipulator 1 is first lifted to a safe height, and then the X-axis and Y-axis of the manipulator 1 are controlled to move to the X-coordinate and Y-coordinate of the workpiece to be picked up. At this time, the center of the gripper 3 and the center of the workpiece are on the same vertical line; then the Z-axis is controlled to descend to the material picking height, and then the system outputs a signal for the gripper 3 to retract. Under pneumatic control, the three claws of the gripper 3 are tightened inward to clamp the workpiece and complete the material picking.
[0028] Feeding: According to the waiting position entered by the user (this position is located above the lathe and has only X and Z coordinates, and is not related to Y), first lift the Z axis of the robot 1 to the Z coordinate set for the waiting position. If it is not lifted, it will collide with the lathe bed; then control the X axis of the robot 1 to move to the X coordinate of the waiting position and wait for the lathe processing to be completed.
[0029] Spindle unloading: When the workpiece on the lathe spindle is processed, the system will control the robot 1 Z axis to drop to the spindle unloading height, and then the robot 1 X axis will move toward the spindle. When it moves to the X coordinate of the unloading position entered by the operator, the system will control the unloading jaw 3 to retract and grab the workpiece, and then the system will control the spindle to release, and then control the robot 1 to move away from the spindle to the X coordinate of the waiting position to complete the spindle unloading. (If there is no material on the spindle, this step will be omitted)
[0030] Scan the next one: Since the robot 1 has nothing to do with the Y-axis during the process of loading and unloading the lathe spindle, when the Z-axis of the robot 1 begins to enter the lathe to load and unload materials, the other side can simultaneously identify the next material on the tray. This step is usually a local scan. The vision module 2 will number all the identified workpieces during the first scan, and grab them according to the numbers. When scanning the next workpiece to be grabbed, the system first obtains the area range where the next workpiece is located from the vision module 2, and then controls the Y-axis to move from the starting point to the end point of the local area. The camera will take pictures during the movement. After the movement is completed, the vision module 2 will calculate the position of the workpiece to be grabbed. Scanning and identifying the workpiece before grabbing will help obtain more accurate workpiece coordinates.
[0031] Spindle loading: First, rotate the clamp 3 and move the Z axis so that the workpiece on the loading clamp 3 is on the same center line as the spindle, then control the X axis of the robot 1 to move to the X axis coordinate of the spindle loading position, and release the clamp 3 after it is in place. Since the spring and spring structure on the clamp 3 are contracted, it will squeeze the workpiece toward the spindle, so the workpiece will not fall. Then control the spindle to clamp the workpiece. During this process, the spindle will shrink slightly inward, but since the spring connected to the spring is always in a contracted state, it will press the workpiece against the spindle, thereby ensuring that the workpiece is loaded to the same position each time. After the spindle shrinks, control the X axis to return to the X coordinate of the waiting position, and finally control the Z axis of the robot 1 to lift to the Z position of waiting.
[0032] Returning materials: During the returning process, the system will control the Y-axis to move the tray plane to the Y coordinate of the position of the processed workpiece on the unloading clamp 3 before processing, and the X-axis of the robot 1 will also move to the X coordinate of the position.
[0033] Unloading: The system controls the Z axis of the robot 1 to move to the unloading height entered by the operator, and then controls the unloading jaw 3 to release, and the processed workpiece will be placed back to its original position on the tray. After placement, the Z axis of the robot 1 is raised to a safe height and rotated so that the material-retrieving jaw 3 faces downward for continued material retrieving.
[0034] Re-pickup: The system re-pickups the material based on the position obtained by the local scan of the vision module 2.
[0035] The lathe automatic processing and the robot 1 loading and unloading process interact through signals. When the lathe processes a workpiece, the robot 1 loading and unloading permission signal is set to 1. When the robot 1 process obtains the signal as 1, the spindle unloading and spindle loading work starts, and after the loading is completed, the robot 1 loading and unloading permission signal is set to 0 again. When the lathe obtains the signal as 0, it starts automatic processing.
[0036] The vision-based CNC manipulator system of the present invention can meet the requirements of fast and large-scale processing rhythm, save labor costs, improve production efficiency, etc.
[0037] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A vision-based CNC manipulator system, characterized in that: include: Lathe, CNC system, manipulator, vision module, gripper and loading platform; The lathe processes the workpiece; The loading platform is used to place workpieces to be processed and processed workpieces; The clamping jaws are used to clamp the workpiece; The visual module scans and obtains the image of the loading platform, and calculates the coordinates of each workpiece to be processed through the image of the loading platform; The CNC system calculates the target position of the linear feed axis of the robot or the coordinates of the workpiece to be processed calculated by the vision module, and the number of pulses required for the motor to rotate one circle. The moving distance of the linear feed axis is divided by the screw pitch of the feed axis to obtain the number of motor rotations, which is then converted into the corresponding number of pulses and sent to the corresponding servo driver. The servo driver controls the motor to rotate according to the received pulses, thereby controlling the movement of the main axis of the robot and controlling the gripper to move to the corresponding position to grab the workpiece. The CNC robot system based on vision controls the Y-axis movement of the linear feed axis. During the Y-axis movement, the camera continuously takes pictures. The vision module calculates the position of the workpiece in the XY coordinate system projected on the tray plane by the X-axis and Y-axis according to the picture results. The vision module of the vision-based CNC manipulator system calculates the coordinates of the workpiece to be processed including: Take any point on the XY coordinate plane and establish a three-dimensional space coordinate system. Calculate the three-dimensional coordinates of the light spot in the three-dimensional space coordinate system through the light spot irradiated on the object to be grasped. The three-dimensional coordinates of all the light spots constitute the light spot coordinate point set on the upper surface of the object to be grasped. Project the light spot coordinate point set on the upper surface of the object to be grasped onto a plane to obtain an image of the center of the object; Through the camera calibration method in machine vision, the image coordinates of the center of the object are converted to obtain the coordinates of the workpiece to be processed; The vision-based CNC manipulator system projects the light spot coordinate point set on the upper surface of the object to be grasped onto the plane, and obtains the image of the center of the object, which specifically includes: Divide the z coordinates of all points on the upper surface of the object to be grasped by the maximum height of the object to be grasped set by the user, and then multiply by 255 to get the value corresponding to the light spot coordinate point, which is the grayscale value; The light spot coordinate point set on the upper surface of the object to be grasped is a floating point number. The coordinate XY values of the points in the light spot coordinate point set on the upper surface of the object to be grasped are multiplied by a predetermined magnification and rounded, and then the offset set by the user is added so that all coordinate XY values are positive integers. The three-channel coordinate points composed of the coordinate XY values and the grayscale values are converted from all the points in the light spot coordinate point set to obtain an image point set, create an image, set the image to pure black, and then draw the points in the image point set on the image by setting pixel values to obtain a height map of the upper surface of the object to be grasped, and then calculate the image coordinates of the center of the object.
2. The vision-based CNC manipulator system according to claim 1, characterized in that: Through the contour matching algorithm of two-dimensional image recognition, the position of the object is located and the image coordinates of the center of the object are calculated.
3. The vision-based CNC manipulator system according to claim 1, characterized in that: The CNC system calculates the coordinates of the workpiece on the platform based on the vision module, and controls the gripper to move to the corresponding position to grab the workpiece, including: controlling the Z-axis of the robot to lift to a safe height first, and then controlling the X-axis and Y-axis of the robot to move to the position of the workpiece to be clamped, with the center of the gripper and the center of the workpiece on the same vertical line; then controlling the Z-axis to descend to the material picking height, and then the system outputs a gripper retraction signal, and the three claws of the gripper tighten inward to clamp the workpiece and complete the material picking.
4. The vision-based CNC manipulator system according to claim 1, characterized in that: According to the waiting position entered by the user, the CNC system first lifts the robot Z axis to the Z coordinate set by the waiting position, and then controls the robot X axis to move to the X coordinate of the waiting position, waiting for the lathe processing to be completed.
5. The vision-based CNC manipulator system according to claim 1, characterized in that: The CNC system controls the manipulator's spindle unloading, including: the CNC system controls the manipulator's Z axis to descend to the spindle unloading height, the manipulator's X axis moves toward the spindle, and when the X coordinate of the position to be unloaded is moved, the CNC system controls the unloading claw to retract and grasp the workpiece, and controls the chuck installed on the spindle to release, and then controls the manipulator to move in a direction away from the spindle to the X coordinate of the position to be unloaded, thereby completing the spindle unloading.
6. The vision-based CNC manipulator system according to claim 1, characterized in that: The vision module numbers all the identified workpieces during the first scan. The CNC system first grabs the workpiece according to the number. When scanning the next workpiece to be grabbed, it obtains the area where the next workpiece is located from the vision module, and then controls the Y axis to move from the starting point to the end point of the area. The camera will take pictures during the movement so that the vision module can calculate the position of the workpiece to be grabbed after the movement is completed.
7. The vision-based CNC manipulator system according to claim 1, characterized in that: After the lathe processes a workpiece, the CNC system sets the manipulator loading and unloading permission signal to 1, starts the manipulator spindle unloading and loading operations, and sets the manipulator loading and unloading permission signal to 0 after the manipulator spindle loading is completed. When the lathe obtains the manipulator loading and unloading permission signal to 0, it starts workpiece processing.