Research On Grinding Technology Of Faucet Robot Based On Vision Localization And Compensation

Faucet grinding is mainly divided two kinds into manual grinding and industrial robot automatic grinding.Manual grinding efficiency is low,and harm the health of workers,the use of industrial robots for automatic grinding is increasingly widespread.However,in the automatic polishing process of the robot,the blank of the faucet needs to be placed neatly on the bracket in advance,The positioning accuracy of the blank directly determines the clamping accuracy of the robot when grasping the faucet,which is easy to cause problems such as large clamping error and poor consistency of grinding quality.Therefore,this paper proposes a robot automatic grinding system based on vision,which adopts vision localization and compensation technology to improve the clamping accuracy and the consistency of grinding quality of water faucet during automatic processing.Its main functions are,before grinding,the positioning of the faucet in the coordinate system at the end of the robot flange can be realized by vision scanning,then the algorithm is used to compensate the position and pose of the blank in the robot tool coordinate system,finally realize automatic grinding of faucet in arbitrary clamping state.The main research contents of this paper are as follows.(1)Construction of vision localization compensation grinding system.To deal with the difficulties of automatic grinding of faucets,a system with robot grasping faucets,vision processing and actual grinding of faucets is constructed.Conduct hardware platform selection and construction.Developed a faucet polishing software system.(2)Three-dimensional vision measurement system.According to the actual faucet grinding scene,the light path and structure design of monocular structured light 3D camera are carried out.The relation between image pixel coordinates and 3d world coordinates based on homography matrix is established by light plane calibration.An improved algorithm for extracting pixels from the center of light strip is proposed.(3)Point cloud reconstruction matching and optimization.Through the hand-eye matrix and coordinate transformation,the faucet point cloud is reconstructed in the robot flange terminal coordinate system.Pretreatment of faucet template point cloud and measured point cloud was carried out.ICP point cloud matching was performed between the two point clouds after pretreatment,a fast iterative AAICP point cloud matching optimization algorithm is applied.(4)Vision localization compensation faucet grinding experiment.The vision calibration experiment was designed and completed,the whole system error is analyzed.When the cylinder working Angle is 0° and 90° respectively,the tap grinding experiment with different position is carried out.After grinding the workpiece surface roughness detection,meet the actual needs.