Bolt tightening method and device based on visual recognition

Abstract

The application provides a bolt tightening method and device based on visual identification, which comprises the following steps: step S1: reading model information of a bolt component to be tightened; step S2: calling a corresponding tightening program according to the model information; step S3: controlling a conveying roller bed to convey the component to a stationary operation station; step S4: mobilizing a visual device to collect an in-situ image of the component at the operation station; step S5: identifying bolt position information on the in-situ image obtained in step S4, and calculating a moving track of a robot according to the bolt position information; and step S6: controlling the robot to move to a position according to the moving track obtained in step S5, and tightening the bolt according to the tightening program called in step S2. The bolt tightening method and system of the application adopts a combination of a visual identification device and a model identification device, improves the flexibility of the equipment, and can well accommodate bolt tightening of various components without additional modification of a clamping mechanism.

Description

Technical Field

[0001] This invention belongs to the technical field of component manufacturing methods and equipment, specifically relating to a bolt tightening method and device based on visual recognition. Background Technology

[0002] In traditional automobile assembly lines, bolt tightening is mostly done manually using handheld electric guns or assisted robotic arms. This is because: 1. Assembly workshops often employ a pull-based production line model, where tightening operations mostly occur during component movement, making automatic tightening difficult; 2. Precise positioning and process monitoring are required for bolt tightening, and automatic tightening cannot guarantee tightening quality or equipment uptime.

[0003] In today's rapidly evolving trend of automation and intelligence, OEMs urgently need to implement fully automated bolt tightening processes, replacing manual labor with automated equipment to continuously reduce labor costs. Existing automated bolt tightening solutions mostly use mechanical structures to position and clamp the parts and tightening shaft at stationary workstations. One clamping mechanism holds and positions the parts to be tightened, while another clamping mechanism holds the tightening shaft. The tightening shaft is moved to the tightening position of the parts by mechanical moving parts such as cylinders, and the bolt is inserted and automatically tightened through program settings. However, such technical solutions usually have the following disadvantages: (1) The mechanical structure of the part or component clamping mechanism is relatively complex, resulting in high manufacturing and maintenance costs; (2) High positioning accuracy is required for the parts or components, leading to high design and manufacturing accuracy requirements for the clamping mechanism; (3) Poor compatibility and flexibility with different parts or components, often requiring a completely new positioning and clamping mechanism to be replaced if the shape of the part changes slightly.

[0004] Invention Patent Content

[0005] To address the aforementioned problems, this invention provides a visual recognition-based bolt tightening method. This method utilizes visual recognition equipment to identify the bolt position and control a robot to perform the tightening operation, thus automating the bolt tightening process in a streamlined, automated manner. The objective of this invention is achieved through the following technical solution:

[0006] This invention provides a bolt tightening method based on visual recognition, which includes the following steps:

[0007] Step S1: Read the model information of the component whose bolts need to be tightened;

[0008] Step S2: Retrieve the corresponding tightening program based on the model information;

[0009] Step S3: Control the conveyor roller bed to transport the component to the stationary operating position;

[0010] Step S4: Use the vision device to collect in-situ images of the component at the operating station. Generally, for components with only one or two bolts that need to be tightened, only one in-situ image can be collected. For components with three or more bolts that need to be tightened, it is preferable to collect two or more in-situ images to ensure the clarity of the collected in-situ images and facilitate the identification of bolt positions and the positioning of work reference points.

[0011] Step S5: Identify the bolt position information on the in-situ image obtained in step S4, and calculate the robot's working trajectory based on the bolt position information;

[0012] Step S6: Control the robot to move into position according to the work trajectory obtained in step S5 and tighten the bolts according to the tightening program retrieved in step S2.

[0013] Furthermore, step S3 specifically includes:

[0014] Step S3.1 - Receive the signal from the operating station that allows the vehicle to enter;

[0015] Step S3.2 - Send a permission command to the conveyor roller bed to move the part to the operating position;

[0016] Step S3.3 - After the vehicle is in place, control the clamping device on the operating station to lock the skid or the component itself carrying the bolt to be tightened, so that the component to be tightened remains stationary. The clamping device has an X-axis centering mechanism and a Y-axis clamping mechanism, which enables coarse positioning and locking of the component.

[0017] Furthermore, step S5 includes the following specific steps:

[0018] Step S5.1 - Identify the location of the bolt on the captured in-situ image and locate the coordinates of the reference point for tightening the bolt on the in-situ image. Here, the bolt can be identified by pixel aggregation features, image contour recognition, or image comparison recognition, depending on the performance of the image acquisition equipment used. The reference point can be set as the center point of the bolt or the edge point of the bolt. For each bolt to be tightened, multiple reference points can be set to increase the accuracy of the data.

[0019] Step S5.2 - Compare the coordinates of the located work reference point with the coordinates of the standard reference point for the bolt tightening operation of the corresponding model component, and calculate the coordinate deviation value of the bolt to be tightened;

[0020] Step S5.3 - Calculate the actual target position that the robot should move to based on the coordinate deviation value calculated in step S5.2, and define the robot's working trajectory based on the actual target position.

[0021] Furthermore, it also includes step S7: controlling the robot to reset based on the received signal indicating that the bolt tightening is complete.

[0022] Furthermore, in step S6, the force data and / or displacement data of the tightening shaft are read in real time when the robot tightens the bolt; step S7 includes the following specific steps:

[0023] Step S7.1 - Receive the signal that the robot has completed the tightening operation;

[0024] Step S7.2: Generate the actual tightening curve based on the force data and / or displacement data of the tightening shaft read in step S6;

[0025] Step S7.3: Compare the actual tightening curve obtained in step S7.2 with the standard tightening curve to generate a tightening judgment result. In specific implementation methods, the tightening judgment result can be generated by comparing the difference in line shape between the actual tightening curve and the standard tightening curve, or by comparing the fluctuation range of the actual tightening curve, or by calculating the offset or deviation value of the actual tightening curve and the standard tightening curve, and judging whether the tightening process is qualified based on whether the offset or deviation value exceeds the preset threshold.

[0026] Step S7.4 - Control the robot to reset;

[0027] Step S7.5 - Unlock the clamping device on the control operating station;

[0028] Step S7.6 - Send a command to the conveyor roller bed to remove the component, so that the conveyor roller bed removes the component from the current operating position.

[0029] Furthermore, when the judgment result generated in step S7.3 is unqualified, a maintenance notification is sent to the subsequent workstations of the operation station.

[0030] Furthermore, the process of continuously tightening the component bolts is performed. While performing any of the operations from steps S3 to S6 on the first component, steps S1 and S2 are performed on the second component located in the standby position.

[0031] Furthermore, the method includes step S7.6 - sending a command to the conveyor roller bed to remove the component, causing the conveyor roller bed to remove the component from the current operating station; in this method, while performing step S7.6 on the first component located at the operating station, step S3 is performed on the second component located at the waiting station.

[0032] The present invention further provides a component bolt tightening device based on vision recognition, which includes a main controller and a conveyor roller bed, a robot, and a vision device electrically connected to the main controller. The main controller has a storage module, an information acquisition module, a recall module, a calculation module, and an output control module.

[0033] The storage module is used to store tightening programs corresponding to different models and standard tightening curve information for tightening bolts of different models; preferably, it is also used to store in-situ image information of the component at the operating station acquired by the signal acquisition module.

[0034] The information acquisition module is used to collect the model information of the bolt to be tightened, the signal of whether the operating station allows the vehicle to enter, the robot working status signal, the in-situ image of the part at the operating station, and the force data and / or displacement data of the tightening shaft when the robot tightens the bolt.

[0035] The module is used to retrieve the corresponding tightening program based on the model information;

[0036] The calculation module is used to identify bolt position information on the image and calculate the robot's working trajectory based on the bolt position information;

[0037] The output control module is used to control the component to move to the operating station, move the vision device to the set position to collect the image of the component at the operating station; control the robot to move to the position according to the set work trajectory and tighten the bolt according to the tightening program; after the tightening operation is completed, control the component to move out of the operating station; and transmit the tightening judgment result to the subsequent station.

[0038] Furthermore, in continuous production operations, the conveyor roller bed includes at least two skids that move in tandem. These two skids reciprocate between a waiting station and an operating station. Specifically, when two skids are used, while the first skid is performing a bolt tightening operation at the operating station with a first component, the second skid moves to the waiting station to load a second component and wait for an entry command. When the first skid receives an instruction to remove a component and leaves the operating station with the first component, the second skid receives an entry command and enters the operating station with the second component. When the second component is subjected to a bolt tightening operation, the first skid moves to the waiting station to load a third component and wait for an entry command. This cycle continues continuously to complete the continuous tightening operation of multiple components on the production line.

[0039] The beneficial effects of this invention are as follows:

[0040] (1) For component clamping, a simple X-axis centering mechanism and Y-axis clamping mechanism can be set up directly at the operating station. The clamping accuracy does not need to be too high, and the initial coarse positioning of the component can be achieved.

[0041] (2) For positioning, a visual recognition device is used to accurately identify the positional deviation of the tightening bolt and guide the robot to perform deviation compensation. There is no need to have high precision requirements for the position of the clamping mechanism and component parts.

[0042] (3) The combination of visual recognition device and model recognition device improves the flexibility of the equipment and can be well compatible with various models of parts without the need for additional modification of the clamping mechanism.

[0043] (4) Integrating the tightening shaft and visual recognition device into the industrial robot can achieve stable and consistent tightening operations, and can also maximize compatibility with the manufacturing and positioning deviations of different vehicles. Attached Figure Description

[0044] Figure 1 The diagram shown is a flowchart of Embodiment 1 of the present invention;

[0045] Figure 2 The diagram shown is a structural schematic of Embodiment 2 of the present invention. Detailed Implementation

[0046] The technical solutions of the preferred embodiments of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely a part of the embodiments of the present invention, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0047] Example 1:

[0048] This embodiment uses the bolt tightening process of a vehicle body transmission as an example to provide a bolt tightening method based on vision recognition. However, this method is also applicable to the bolt tightening process of other components outside the vehicle body transmission. The implementation process of this method is as follows: Figure 1 As shown, it includes the following steps:

[0049] Step S1: Read the vehicle model information of the bolts to be tightened. In the specific implementation process, an RFID reader can be set up at the standby station to read and write the vehicle model data stored in the RFID of the skid carrying the vehicle body. The reader will transmit the read vehicle model information to the main controller PLC.

[0050] Step S2: Call the corresponding tightening program according to the vehicle model information. The tightening program should at least include the tightening axis turning action parameters of the robot. In the preferred embodiment, it should also include the standard movement trajectory of the telescopic cylinder in each direction during the movement of the robot's tightening axis from the origin to the bolt tightening position.

[0051] Step S3: Control the conveyor roller bed to transport the car body to the stationary operating position. In the specific implementation process, this can be broken down into the following steps:

[0052] Step S3.1 - Receive the signal from the operating station that allows the vehicle to enter;

[0053] Step S3.2 - Send a permission command to the conveyor roller bed to control the conveyor roller bed to transport the skid carrying the car body with tightening bolts to the position. Specifically, the main controller PLC sends a permission command to the conveyor roller bed, and the conveyor roller bed transports the skid carrying the car body to be tightened to the operating position. The conveyor roller bed automatically decelerates and stops by relying on the position sensors set on the skid and the operating position.

[0054] Step S3.3 - After the vehicle is in place, the clamping device on the control operation station clamps the skid or the vehicle body. The clamping device usually has an X-axis centering mechanism and a Y-axis clamping mechanism. The X-axis centering mechanism and the Y-axis clamping mechanism lock the skid or the vehicle body to keep the vehicle body to be tightened stationary, and at the same time realize the secondary coarse positioning of the vehicle body to ensure that the state of the vehicle body when the in-situ image is captured in step S4 is the same as the state of the vehicle body when the robot tightens the bolts.

[0055] Step S4: Use the vision device to collect in-situ images of the vehicle body at the operating station. In a specific implementation, the vision device can be a camera, video camera, or scanner, as long as it can collect images with a certain degree of clarity to facilitate the identification of bolts.

[0056] Step S5: Identify the bolt position information on the in-situ image obtained in step S4, and calculate the robot's movement trajectory based on the bolt position information. The specific implementation process can be broken down into the following steps:

[0057] Step S5.1 - Identify the location of the bolt on the captured image and calculate the coordinates of the reference point for tightening the bolt. In a specific implementation, the bolt location can be identified by pixel feature recognition, graphic contour recognition, image comparison recognition, etc. The reference point can be set as one or more points on the bolt axis or the outer contour of the bolt as needed.

[0058] Step S5.2 - Compare the calculated coordinates of the working reference point with the tightening bolt with the coordinates of the standard reference point of the corresponding bolt model, and calculate the coordinate deviation between the current actual position and the standard position of the bolt to be tightened;

[0059] Step S5.3 - Based on the coordinate deviation value calculated in step S5.2 and the standard movement trajectory of the robot stored in the tightening program called in step S2, calculate the actual target position that the robot should move to, and define the actual operation trajectory of each telescopic cylinder of the robot according to the actual target position.

[0060] Step S6: Control the robot to move to the position according to the work trajectory obtained in step S5 and complete the bolt tightening operation according to the tightening shaft turning action parameters in the tightening program retrieved in step S2. In step S6, read the force data and / or displacement data of the tightening shaft when the robot tightens the bolt in real time.

[0061] Step S7: Based on the received signal indicating that the bolt tightening is complete, control the robot to reset and control the conveyor rollers to move the tightened car body out of the operating station. The specific implementation process may include the following steps:

[0062] Step S7.1 - Receive the signal that the robot has completed the tightening operation;

[0063] Step S7.2: Generate the actual tightening curve based on the force data and / or displacement data of the tightening shaft read in step S6;

[0064] Step S7.3: Compare the actual tightening curve obtained in step S7.2 with the standard tightening curve. When the generated judgment result is unqualified, send maintenance notification information to the subsequent workstations of the operation station. For a car body with multiple bolts that need to be tightened, as long as the judgment result of tightening one bolt is unqualified, a maintenance notification information is sent to the subsequent workstations. The maintenance notification information includes information on whether each bolt of the car body is qualified.

[0065] Step S7.4 - Control the robot to reset. Here, robot reset can be to return to the origin or to a preset positioning point that is convenient for subsequent vehicle operations.

[0066] Step S7.5 - Unlock the X-axis centering mechanism and Y-axis clamping mechanism on the control operation station;

[0067] Step S7.6 - Send a command to the conveyor roller bed to remove the component, causing the conveyor roller bed to move the car body out of the current operating position.

[0068] Step S5.3 can also be implemented as follows: Calculate the deviation compensation the robot should make on the standard movement trajectory based on the coordinate deviation value calculated in step S5.2 (usually including deviations in the X and Y directions, and possibly also the Z direction deviation), thereby defining the actual operating trajectory of each telescopic cylinder of the robot. In a specific implementation, the overall process of steps S4, S5, and S6 can also be implemented as follows: The robot, equipped with a vision recognition device, moves to the top of the bolts to be tightened, takes two photos, compares the position of each bolt with the standard bolt positions of the corresponding vehicle model stored in the main controller, and feeds the position deviation back to the main controller PLC; the main controller PLC calculates the position deviation of each bolt and feeds the information back to the robot, which then obtains the actual coordinates after position compensation and moves to the top of each bolt. For example, taking a vehicle with four bolts as an example, a fast calculation method for robot position compensation can be: take two photos of the four bolts of a vehicle using a vision device, and identify two bolts in each photo. If both bolts in each photo are correctly identified, then each of the four bolts is assigned its own positional deviation. If only one bolt in each photo is identified, then the deviation of the identified bolt is assigned to both bolts. If only one bolt is identified in two photos, then the deviation of that bolt is assigned to all four bolts. This calculation logic reduces the number of photos taken, increases cycle time, and reduces downtime caused by visual recognition failures, greatly improving equipment uptime.

[0069] In other implementations, the position information of the standard reference point of the bolt can be not stored in advance, but the coordinate position of the bolt can be directly located. In actual operation, the robot can be controlled to move directly from the origin or starting point to the position of the bolt to be tightened to complete the tightening operation. However, this operation method may produce large errors in the operation process due to the lack of a relative positioning reference.

[0070] On industrial production lines, in order to improve the working cycle of the operating station, it is necessary to perform the process of continuously tightening component bolts. That is, while performing any of the operations S3 to S6 on the front machine, the rear machine located at the standby station is performed with steps S1 and S2. Furthermore, while performing step S7.5 on the front machine, step S3 can be performed on the rear machine at the same time, so that the entry and exit of the operating station are synchronized, thereby improving work efficiency.

[0071] Example 2:

[0072] This embodiment provides a vision-based component bolt tightening device that can implement the method of Embodiment 1. It includes a main controller and a conveyor roller bed, a robot, and a vision device electrically connected to the main controller. Position sensors are provided on the conveyor roller bed, the operating station, and the standby station. The operating station is equipped with a clamping device for clamping and fixing a skid or vehicle body. The clamping device has an X-axis centering mechanism and a Y-axis clamping mechanism. The main controller includes a storage module, an information acquisition module, a recall module, a calculation module, and an output control module.

[0073] The storage module is used to store the tightening program corresponding to different models, the standard tightening curve information of different models of body bolts, and the preset threshold of the allowable offset or key node deviation value of the actual tightening curve; preferably, it is also used to store the in-situ image information of the body at the operating station acquired by the signal acquisition module.

[0074] The information acquisition module is used to collect the model information of the parts to be tightened, the signal of whether the operating station allows the vehicle to enter, the robot working status signal, the in-situ image of the vehicle body at the operating station, and the force data and / or displacement data of the tightening shaft when the robot tightens the bolts.

[0075] The module is used to retrieve the corresponding tightening program based on the model information;

[0076] The calculation module is used to identify bolt position information based on pixel features in the in-situ image, calculate the robot's working trajectory based on the bolt position information, generate the actual tightening curve based on the force data and / or displacement data of the tightening shaft, and compare the deviation of the actual tightening curve with the standard tightening curve.

[0077] The output control module is used to control the conveyor roller bed to move the car body to the operating station, adjust the vision device to the set position to collect images of the parts at the operating station, and control the robot to move to the position according to the set work trajectory and tighten the bolts according to the tightening program.

[0078] In a preferred embodiment, the vision device can be directly mounted on one of the robot's telescopic cylinders, and the movement trajectory of the vision device can be directly controlled by controlling the robot's telescopic cylinder movement, thus ensuring the consistency and coordination of the movement trajectories of the robot and the vision device.

[0079] In a preferred embodiment, the conveyor roller bed in continuous production includes at least two skids that move in tandem. These skids reciprocate between a waiting station and an operating station. Specifically, when two skids are used, while the first skid, carrying the preceding car, is performing bolt tightening operations at the operating station, the second skid moves to the waiting station to load the following car and await entry commands. When the first skid, carrying the preceding car, leaves the operating station, the second skid receives an entry command and carries the following car into the operating station. When the following car enters the operating station to perform bolt tightening operations, the first skid moves to the waiting station to load the third car and await entry commands (at this point, the following car at the operating station becomes the preceding car, and the third car becomes the following car). This cycle continues continuously, completing the continuous tightening operations on multiple components on the production line. This design allows for the movement of the following car into the operating station while the preceding car, having completed bolt tightening, is being moved out during continuous operation, ensuring production cycle time. In the above steps, the stationary operating station is preferably located in the buffer zone between two production lines. Utilizing the existing rotary roller bed station to construct the stop station for fully automated tightening operations ensures that the vehicle has sufficient stationary operating time while preventing congestion and chain stoppages on the preceding and following production lines. Because other vehicles are transporting components to the next production line, even if the automatic tightening equipment experiences a short-term malfunction, it will not immediately cause the downstream production line to stop, providing a certain buffer time for troubleshooting. The components are carried and transported via the roller bed, separated from the robot column, avoiding direct contact that could transmit roller bed vibrations to the robot and affect the tightening process quality.

[0080] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention; the dimensions described in the drawings and embodiments are not related to the specific physical object and are not used to limit the protection scope of the present invention. The physical dimensions can be selected and changed according to actual needs.

Claims

1. A bolt tightening method based on visual recognition, characterized in that, Includes the following steps: Step S1: Read the model information of the component whose bolts need to be tightened; Step S2: Retrieve the corresponding tightening program according to the model information. The tightening program should at least include the robot's tightening axis turning action parameters and the standard movement trajectory of the telescopic cylinders in each direction during the movement of the robot's tightening axis from the origin to the bolt tightening position. Step S3: Control the conveyor roller bed to transport the component to the stationary operating position; Step S4: Use the vision device to capture in-situ images of four bolts on a vehicle component at the work station in two separate steps, identifying two bolts in each in-situ image. Step S5: Identify the bolt position information on the in-situ image obtained in step S4, and calculate the robot's working trajectory based on the bolt position information; Step S5.1: Identify the location of the bolt in each captured in-situ image, and locate the coordinates of the reference point for tightening the bolt in each in-situ image; Step S5.2: Compare the coordinates of the located work reference point with the coordinates of the standard reference point for the bolt tightening operation of the corresponding model component, and calculate the coordinate deviation value of the bolt to be tightened; if both bolts in each in-situ image are correctly identified, then assign the positional deviation to each of the four bolts; if only one bolt is identified in each in-situ image, then assign the deviation of the identified bolt to both bolts; if only one bolt is identified in two in-situ images, then assign the deviation of the identified bolt to all four bolts. Step S5.3: Calculate the deviation compensation that the robot should make on the standard movement trajectory based on the coordinate deviation value calculated in step S5.2, thereby defining the actual operation trajectory of each telescopic cylinder of the robot. Step S6: Control the robot to move into position according to the work trajectory obtained in step S5 and tighten the bolts according to the tightening program retrieved in step S2.

2. The bolt tightening method based on vision recognition according to claim 1, characterized in that, Step S3 specifically includes: Step S3.1: Receive the signal from the operating station that allows the vehicle to enter; Step S3.2: Send a permission command to the conveyor roller bed to move the part to the operating station; Step S3.3: After the vehicle is in place, control the clamping device on the operating station to lock the skid of the transport component or the component itself so that the component to be tightened remains stationary.

3. A bolt tightening method based on vision recognition according to any one of claims 1 to 2, characterized in that, It also includes step S7: controlling the robot to reset based on the received signal that the bolt tightening is complete.

4. The bolt tightening method based on vision recognition according to claim 3, characterized in that, In step S6, the force data and / or displacement data of the tightening shaft are read in real time when the robot tightens the bolt; step S7 includes the following specific steps: Step S7.1: Receive the signal that the robot has completed the tightening operation; Step S7.2: Generate the actual tightening curve based on the force data and / or displacement data of the tightening shaft read in step S6; Step S7.3: Compare the actual tightening curve obtained in step S7.2 with the standard tightening curve to generate a tightening judgment result; Step S7.4: Control the robot to reset; Step S7.5: Unlock the clamping device on the control operating station; Step S7.6: Send a command to the conveyor roller bed to remove the component, so that the conveyor roller bed removes the component from the current operating position.

5. The bolt tightening method based on vision recognition according to claim 4, characterized in that, When the judgment result generated in step S7.3 is unqualified, a maintenance notification message is sent to the subsequent workstations of the operation station.

6. The bolt tightening method based on vision recognition according to claim 5, characterized in that, The process of continuously tightening component bolts involves performing any of the operations from step S2 to step S6 on the first component located at the operating station, while simultaneously performing step S1 on the second component located at the standby station.

7. A bolt tightening method based on vision recognition according to claim 6, characterized in that, The method includes step S7.6: sending a command to the conveyor roller bed to remove the component, causing the conveyor roller bed to remove the component from the current operating station; in this method, while performing step S7.6 on the first component located at the operating station, step S3 is performed on the second component located at the waiting station.

8. A component bolt tightening device based on vision recognition, employing the bolt tightening method based on vision recognition as described in any one of claims 1 to 7, characterized in that, The system includes a main controller and a conveyor roller bed, a robot, and a vision device electrically connected to the main controller. The main controller has a storage module, an information acquisition module, a recall module, a calculation module, and an output control module. The storage module is used to store the tightening programs corresponding to different models, as well as the standard tightening curve information for tightening bolts of different models. The information acquisition module is used to collect the model information of the bolt to be tightened, the signal of whether the operating station allows the vehicle to enter, the robot working status signal, the in-situ image of the part at the operating station, and the force data and / or displacement data of the tightening shaft when the robot tightens the bolt. The module is used to retrieve the corresponding tightening program based on the model information. The calculation module is used to identify bolt position information on the in-situ image and calculate the robot's working trajectory based on the bolt position information; The output control module is used to control the component to move to the operating station, move the vision device to the set position to collect the image of the component at the operating station, and control the robot to move to the position according to the set work trajectory and tighten the bolts according to the tightening program.

9. A component bolt tightening device based on vision recognition according to claim 8, characterized in that, The conveyor roller bed includes at least two skids that move in a front-to-back linkage, and the at least two skids reciprocate between the waiting station and the operating station to transfer the component.

Images