Research On Shaft-Hole Clearance Assembly Technology Based On Robot Vision

Intelligent manufacturing and automated assembly have become the main developmental of the current manufacturing industry,while the traditional manufacturing industry,which is mainly operated by manual operation,and is facing numerous huge challenges and changes.Therefore,this thesis,taking the typical clearance fit shaft-hole assembly as the research object,based on the robot vision technology,and setting up an experimental platform for the robot vision shaft-hole clearance assembly,discussed and studied the target feature recognitional position and trajectory planning in the shaft-hole clearance assembly.Firstly,this thesis analyzes the research status of robot vision shaft-hole assembly at home and abroad.Next,according to the functional requirements and performance index requirements of the robot vision assembly system,based on the Eye-in-hand vision measurement theory,the thesis started to design the overall scheme of the assembly system,and then completed the system hardware,next built the experimental platform of the assembly system,completed the camera calibration and the hand-eye calibration by Zhang’s calibration method and Tsai-Lenz algorithm respectively at last.Secondly,through image pre-processing,the thesis completed the coarse extraction of the target round holes and auxiliary marks,and in terms of Canny operator completed the detection and extraction of ellipse edges.The thesis proposed a direct least squares algorithm based on random sampling consistency,which solved the problem of large data errors caused by uneven natural illumination and defects in the extracted target image,and improved the positioning accuracy and system of circular holes and auxiliary markers.robustness.Based on the Pn P principle,the thesis,applying the DLT method,solved the transformation matrix between the marker point and solved the problem of insensitivity to monocular vision depth.And taking use of the relation of coplanar constraint between the mark point and the end face of the round hole,the thesis figured out the plane normal vector of the mark point,which solved the problem of the assembly pose of the round hole.Next,according to the relationship between the two adjacent connecting rods and joints of the robot,the D-H parameter connecting rod coordinate system is established,which further worked out the conversion relationship among the connecting rods of the robot.Then the simulation model in ROS is established,and through comparing and analyzing three kinds of trajectories:RRT、RRT*、RRT-Connect,the RRT-Connect algorithm is determined as the trajectory planning algorithm to study the trajectory planning of the assembly robot in the Cartesian space,which realized the spatial attitude interpolation algorithm,the construction of the linear trajectory and the arc trajectory in the Cartesian space.Finally,the thesis analyzed software requirements of the assembly system and designed the overall software architecture.Meanwhile,the design and development of data communication are completed.The experimental platform had set up to verify the systemrepeat-positioning accuracy and visual-positioning error.The experimental results showed that by designing robot vision,the shaft-hole clearance assembly system has highpositioning accuracy and accurate results,which meets the requirements of assembly matching accuracy.