| This paper proposes a kind of method based on machine vision,which aims at controlling the welding quality of white body roof effectively.This method is helpful to ensure the accuracy of the roof's initial position before welding,in other words,the welding quality of the roof is improved.For a type of car,the 3D pose of the white body roof region is derived by machine vision technology,and the pose is fed back to the robot to form a closed control loop,therefore,it can guide robot to place the roof quickly and accurately,which has been grasped,in the corresponding position of the body in white.Firstly,the calibration process and method of the visual control system for the welding quality of the white body roof are studied,at the same time,the camera and hand-eye system are calibrated;Secondly,in order to extract the feature points of the white body roof,the image processing flow and method are analyzed and studied;Thirdly,the geometric mapping model between pixel feature points and actual feature points of the white-body's roof is constructed,on the basis of the model,the pose of the actual feature points are solved.Then,a three-dimensional pose transformation model based on multi feature points constraint is proposed,so that the three-dimensional poses of these feature points can be transformed into a point's three-dimensional pose,which represents the position and orientation of the white body roof region and the manipulator's target.Finally,the operation principle and function requirement of the visual control system for the welding quality of the white body roof are analyzed,then the hardware and software architecture of the system are constructed.On the basis of above,the visual control prototype system for the welding quality of the white body roof is designed and developed.We hope that it can provide a new reference model and a new realization way for automobile enterprises while designing welding technology of white body roof. |
PDF Full Text Request