Battery cell module grabbing and feeding method based on visual compatible overexposure

Abstract

The application discloses a kind of based on visual compatible overexposure battery module grabbing feeding method, it is related to the method of lithium battery module turnover feeding positioning.To existing explosion-proof valve on the top surface of battery as core, based on 2D industrial camera realizes the accurate grabbing of 5 to 8 batteries.The following steps are carried out: unify the coordinate system of camera and manipulator;Template library is generated and imported;First battery incoming, obtain the position of the explosion-proof valve of the first battery;Obtain the line segment as a reference;Record each reference value;Formally incoming, grab the target battery of predetermined quantity.This application realizes the completion of module turnover large package feeding compatible with a variety of number of battery grabbing mode feeding mode, and compatible with a variety of metal battery overexposure identification, this way breaks the traditional double camera feeding and can only be whole grabbing whole problem, optimizes the influence of overexposure on battery grabbing, and sets a variety of foolproof detection, more stable and more efficient.

Description

Technical Field

[0001] This invention relates to a method for loading and positioning lithium battery modules, and more particularly to a method for gripping and controlling battery cells in groups. Background Technology

[0002] Currently, the common grasping and loading process for new energy battery cell module loading stations using 2D machine vision positioning is as follows: two cameras take pictures of the first and last ends of a row of 16 cells, with the first camera photographing the first cell and the second camera photographing the last cell. The selected feature points from each pair of cells are then connected, and the robot is guided by the angle and offset of this line segment. However, existing control methods can only grasp and release entire rows at a time. If the grasping is not done in a row, misidentification can occur, and overexposure can also lead to the failure to recognize feature points.

[0003] Especially when current processing requirements change and require multiple grabbing of a row of battery cells, the existing grabbing and feeding methods cannot accurately position or grab the corresponding number of battery cells when grabbing 5 to 8 battery cells. Summary of the Invention

[0004] To address the above problems, this invention proposes a method for gripping and loading battery cell modules based on vision-compatible overexposure. It uses the explosion-proof valve on the top surface of existing battery cells as the core and a 2D industrial camera to accurately grip 5 to 8 battery cells.

[0005] The technical solution of this invention is as follows: It is carried out according to the following steps:

[0006] Step 1: Unify the coordinate systems of the camera and the robotic arm;

[0007] Step 2: Generate and import the template library;

[0008] The light source and camera are fixedly mounted on the robotic arm. The robotic arm is controlled to take a battery cell with the explosion-proof valve facing upwards. The battery cell is photographed from multiple different angles to form a set of battery cell photos with the explosion-proof valve in multiple angles. These photos are then used as a template library and added to the CogMultiPMAlignTool tool in the VisionPro software so that the software can learn autonomously.

[0009] Step 3: Upon initial arrival of battery cells, locate the explosion-proof valve of the first battery cell.

[0010] After the battery cells arrive, the robot operates a camera to photograph the explosion-proof valve of the first battery cell from left to right. Then, the CogMultiPMAlignTool is used to find the feature points of the explosion-proof valve, and the CogFixture tool of the VisionPro software is used to correct the position to obtain the location of the explosion-proof valve of the first battery cell.

[0011] Step 4: Obtain the line segment used as the reference;

[0012] Step 4.1: Based on the location of the explosion-proof valve of the first battery cell, use the FindLineTool to find the left and top edges of the first battery cell and the right and top edges of the fourth battery cell within the field of view.

[0013] Step 4.2: Use the CogIntersectLineLineTool to find the intersection of the left and top edges of the first cell, and the intersection of the right and top edges of the fourth cell.

[0014] Step 4.3: Use the CogCreatSegmentTool to fit the two intersection points into a line segment;

[0015] Step 5: Record all baseline values;

[0016] The midpoint of the fitted line segment in step 4.2 is derived as the reference position, and the angle between the line segment and the X-axis of the coordinate system is derived as the reference angle. Then, the operator controls the robot to grab a predetermined number of target cells and records the grabbing position of the robot as the reference grabbing position.

[0017] Step 6: Formal material receipt, grabbing the predetermined number of target battery cells;

[0018] When the material arrives again, the intersection of the left and top edges of the first cell from left to right and the intersection of the right and top edges of the fourth cell are obtained in the VisionPro software according to steps 3 and 4. The material is then fitted again to obtain the line segment obtained from the refit. The reference position and reference angle from step 5 are used to calculate the difference, and the XY offset of the midpoint of the refitted line segment and the angle offset of the line segment are obtained.

[0019] The XY offset and angle offset are exported and input to the PLC via the S7 communication protocol. Based on the XY offset and angle offset, the PLC adjusts the gripping position of the robot arm by an equal amount based on the reference gripping position, and then grips the target battery cell that is receiving material again.

[0020] Furthermore, after fitting a line segment using the CogCreatSegmentTool in step 4, the angle of that line segment is obtained through the following steps:

[0021] The angle between the line segment and the X-axis is obtained using the CogCreatSegmentTool tool. The code for converting the radian to the angle and defining positive and negative values ​​is written using a C# simple script. The logic for defining positive and negative values ​​is as follows: determine whether the reference angle obtained in step 5 is positive or negative. If the reference angle is positive, the value obtained in subsequent step 6 is the absolute value; if the reference angle is negative, the value obtained in subsequent step 6 is the negative absolute value.

[0022] Furthermore, in step 3, the battery cell being photographed is determined to be the first battery cell from left to right through the following steps:

[0023] The judgment code is written in C# simple script. The judgment logic is as follows: obtain the coordinate value of the center point of the explosion-proof valve output by the CogMultiPMAlignTool tool. When only one coordinate value is obtained, the cell where the coordinate value is located is selected as the first cell. When two coordinate values ​​are obtained, the cell where the larger coordinate value is located is selected as the first cell.

[0024] This invention realizes a feeding method that can complete the loading of large-package modules and is compatible with multiple cell grabbing modes. It is also compatible with the recognition of overexposure of various metal cells. This method breaks the problem of traditional dual-camera feeding and grabbing which can only grab and place whole cells. It optimizes the impact of overexposure on cell grabbing and sets up multiple error-proof detections, making it more stable and efficient.

[0025] This invention takes into account that the battery cells have a 3mm movement range within the foam padding at the loading station, meaning they are not completely jammed. Therefore, if the first battery cell grasped is positioned slightly higher within the foam padding and the last slightly lower, the grasped battery cell will be skewed. Thus, the angle of the straight line at the upper edge of the first battery cell cannot be used as the overall grasp angle. Furthermore, considering the inconsistent number of cells grasped each time, the minimum number is 5. Since a 30mm shift near the edge might exceed the field of view and result in poor performance, this invention ultimately adopts the mode of grasping the first and fourth battery cells. Considering that as long as the battery cells are not completely jammed within the foam tray, the angle of any single battery cell's edge cannot be used as a reference for the offset angle. Unless the robot gripper is designed to grasp batteries in any polarity position within the foam padding, the angle given by vision must be based on the angle of the entire row of battery cells. Therefore, this invention ultimately uses a fitted line segment as the benchmark. Attached Figure Description

[0026] Figure 1 This is the workflow diagram for this case. Detailed Implementation

[0027] To clearly illustrate the technical features of this patent, the following detailed description is provided through specific embodiments and in conjunction with the accompanying drawings.

[0028] Large package loading is the first process on the module repackaging line, and its functional stability and efficiency directly affect the operation of subsequent production processes. Industrial cameras play an indispensable role in the positioning of large package loading, and their stability directly impacts the stability of the loading process. Manual loading begins at the roller shutter station. The wooden pallet contains foam with three rows of battery cells arranged in varying numbers, primarily based on three templates. The number of cells to be picked up is 88, 775, and 776 respectively. The foam must not extend beyond the edge of the pallet, with a dimensional difference of ±30mm. After the wooden pallet is conveyed to its position via a slide rail, the PLC identifies the layer number using sensors and sends this information to the robot, guiding it to a fixed position to take a picture. The camera, on the robotic arm, transmits the offset to the PLC, which in turn transmits it to the robot, guiding the robot to pick up the cells in batches and place them on the transfer table. The specific layout is as follows... Figure 1 As shown.

[0029] This invention employs single-camera positioning, and based on VisionPro software, eliminates overexposure, ensures compatibility with feature points of metal battery cells, and completes the function of grasping irregular battery cells. Its main steps are as follows: Figure 1 As shown:

[0030] Step 1: Unify the coordinate systems of the camera and the robotic arm;

[0031] First, the camera is calibrated using a feature circle. The camera and the robot arm are simultaneously translated by nine coordinates, corresponding to the camera's (x, y, r) coordinates and the robot's (X, Y, R) coordinates, to obtain... The value of A is determined and the rotation center is confirmed by rotating the object by three angles. After saving the calibration file, it is imported into the CogNPointToNPointTool in the main program to unify the coordinate systems of the camera and the robot arm.

[0032] Step 2: Generate and import the template library;

[0033] The light source and camera are fixedly mounted on the robotic arm. The robotic arm is controlled to take a battery cell with the explosion-proof valve facing upwards. The battery cell is photographed from multiple different angles to form a set of battery cell photos with the explosion-proof valve in multiple angles. These photos are then used as a template library and added to the CogMultiPMAlignTool tool in the VisionPro software so that the software can learn autonomously.

[0034] Step 3: Upon initial arrival of battery cells, locate the explosion-proof valve of the first battery cell.

[0035] After the battery cells arrive, the robot operates a camera to photograph the explosion-proof valve of the first battery cell from left to right. Then, the CogMultiPMAlignTool is used to find the feature points of the explosion-proof valve, and the CogFixture tool of the VisionPro software is used to correct the position to obtain the location of the explosion-proof valve of the first battery cell.

[0036] Step 4: Obtain the line segment used as the reference;

[0037] Step 4.1: Based on the location of the explosion-proof valve of the first battery cell, use the FindLineTool to find the left and top edges of the first battery cell and the right and top edges of the fourth battery cell within the field of view.

[0038] Step 4.2: Use the CogIntersectLineLineTool to find the intersection of the left and top edges of the first cell, and the intersection of the right and top edges of the fourth cell.

[0039] Step 4.3: Use the CogCreatSegmentTool to fit the two intersection points into a line segment;

[0040] Step 5: Record all baseline values;

[0041] The midpoint of the fitted line segment in step 4.2 is derived as the reference position, and the angle between the line segment and the X-axis of the coordinate system is derived as the reference angle. Then, the operator controls the robot to grab a predetermined number of target cells and records the grabbing position of the robot as the reference grabbing position.

[0042] Step 6: Formal material receipt, grabbing the predetermined number of target battery cells;

[0043] When the material arrives again, the intersection of the left and top edges of the first cell from left to right and the intersection of the right and top edges of the fourth cell are obtained in the VisionPro software according to steps 3 and 4. The material is then fitted again to obtain the line segment obtained from the refit. The reference position and reference angle from step 5 are used to calculate the difference, and the XY offset of the midpoint of the refitted line segment and the angle offset of the line segment are obtained.

[0044] The XY offset and angle offset are exported and input to the PLC via the S7 communication protocol. Based on the XY offset and angle offset, the PLC adjusts the gripping position of the robot arm by an equal amount based on the reference gripping position, and then grips the target battery cell that is receiving material again.

[0045] In the above:

[0046] Since the sign of the included angle is random, after fitting a line segment using the CogCreatSegmentTool in step 4, the angle of that line segment is obtained through the following steps:

[0047] The angle between the line segment and the X-axis is obtained using the CogCreatSegmentTool tool; the code for converting radians to angles and defining positive and negative values ​​is written using a C# simple script. The logic for defining positive and negative values ​​is as follows:

[0048] Determine whether the reference angle obtained in step 5 is positive or negative. If the reference angle is positive, the value obtained in subsequent step 6 is the absolute value; if the reference angle is negative, the value obtained in subsequent step 6 is the negative absolute value.

[0049] In addition, to prevent mistakes:

[0050] In step 3, the following steps are used to determine that the battery cell being photographed is the first battery cell from left to right:

[0051] Write conditional code using C# simple scripts, with the following logic:

[0052] Obtain the coordinates of the center point of the explosion-proof valve output by the CogMultiPMAlignTool tool. When only one coordinate value is obtained, the cell at which that coordinate value is located is selected as the first cell. When two coordinate values ​​are obtained, the cell at which the larger coordinate value is located is selected as the first cell.

[0053] There are many specific ways to implement this invention. The above description is only a preferred embodiment of this invention. It should be noted that for those skilled in the art, several improvements can be made without departing from the principle of this invention, and these improvements should also be considered within the scope of protection of this invention.

Claims

1. A method for gripping and loading battery cell modules based on vision-compatible overexposure, characterized in that, Follow these steps: Step 1: Unify the coordinate systems of the camera and the robotic arm; Step 2: Generate and import the template library; The light source and camera are fixedly mounted on the robotic arm. The robotic arm is controlled to take a battery cell with the explosion-proof valve facing upwards. The battery cell is photographed from multiple different angles to form a set of battery cell photos with the explosion-proof valve in multiple angles. These photos are then used as a template library and added to the CogMultiPMAlignTool tool in the VisionPro software so that the software can learn autonomously. Step 3: Upon initial arrival of battery cells, locate the explosion-proof valve of the first battery cell. After the battery cells arrive, the robot operates a camera to photograph the explosion-proof valve of the first battery cell from left to right. Then, the CogMultiPMAlignTool is used to find the feature points of the explosion-proof valve, and the CogFixture tool of the VisionPro software is used to correct the position to obtain the location of the explosion-proof valve of the first battery cell. Step 4: Obtain the line segment used as the reference; Step 4.1: Based on the location of the explosion-proof valve of the first battery cell, use the FindLineTool to find the left and top edges of the first battery cell and the right and top edges of the fourth battery cell within the field of view. Step 4.2: Use the CogIntersectLineLineTool to find the intersection of the left and top edges of the first cell, and the intersection of the right and top edges of the fourth cell. Step 4.3: Use the CogCreatSegmentTool to fit the two intersection points into a line segment; Step 5: Record all baseline values; The midpoint of the fitted line segment in step 4.2 is derived as the reference position, and the angle between the line segment and the X-axis of the coordinate system is derived as the reference angle. Then, the operator controls the robot to grab a predetermined number of target cells and records the grabbing position of the robot as the reference grabbing position. Step 6: Formal material receipt, grabbing the predetermined number of target battery cells; When the material arrives again, the intersection of the left and top edges of the first cell from left to right and the intersection of the right and top edges of the fourth cell are obtained in the VisionPro software according to steps 3 and 4. The material is then fitted again to obtain the line segment obtained from the refit. The reference position and reference angle from step 5 are used to calculate the difference, and the XY offset of the midpoint of the refitted line segment and the angle offset of the line segment are obtained. The XY offset and angle offset are exported and input to the PLC via the S7 communication protocol. Based on the XY offset and angle offset, the PLC adjusts the gripping position of the robot arm by an equal amount based on the reference gripping position, and then grips the target battery cell that is receiving material again.

2. The method for gripping and loading battery cell modules based on visual compatibility overexposure according to claim 1, characterized in that, After fitting a line segment using the CogCreatSegmentTool in step 4, obtain the angle of the line segment through the following steps: The angle between the line segment and the X-axis is obtained using the CogCreatSegmentTool tool. The code for converting the radian to the angle and defining positive and negative values ​​is written using a C# simple script. The logic for defining positive and negative values ​​is as follows: determine whether the reference angle obtained in step 5 is positive or negative. If the reference angle is positive, the value obtained in subsequent step 6 is the absolute value; if the reference angle is negative, the value obtained in subsequent step 6 is the negative absolute value.

3. The method for gripping and loading battery cell modules based on visual compatibility overexposure according to claim 1, characterized in that, In step 3, the following steps are used to determine that the battery cell being photographed is the first battery cell from left to right: The judgment code is written in C# simple script. The judgment logic is as follows: obtain the coordinate value of the center point of the explosion-proof valve output by the CogMultiPMAlignTool tool. When only one coordinate value is obtained, the cell where the coordinate value is located is selected as the first cell. When two coordinate values ​​are obtained, the cell where the larger coordinate value is located is selected as the first cell.

Images