Font Size: a A A

Research On Target Pose Measurement And Robot Location Based On Monocular Vision

Posted on:2019-11-17Degree:MasterType:Thesis
Country:ChinaCandidate:J Z LeiFull Text:PDF
GTID:2428330596950991Subject:Engineering
Abstract/Summary:PDF Full Text Request
Combining computer vision with industrial robots enables industrial robots to have visual positioning and detection capabilities like human beings,which is of great significance for enhancing the perception ability of industrial robots and realizing the flexible assembly,adaptive grasping and precise positioning of industrial robots.Consequently,this paper combines vision with robot in eye-in-hand mode,and studies a target pose measurement and robot positioning technique based on monocular vision.The main research content and innovations include:1.Technical framework and pipeline of target pose measurement and robot positioning are proposed in this paper,which includes: Hardware system design and software system design.In hardware aspect,the paper mainly studies the design and selection of hardware system.In software,the software framework of target pose measurement and robot positioning system based on monocular vision is constructed,and the technology of software modularization is proposed.2.Two types of hand-eye calibration methods,which are special hand-eye calibration panel based and mirror-based calibration are presented and realized.The precision of the two methods is compared by experiment,which shows the panel-based method is more suitable to our case.3.An algorithm of target position and posture measurement and location is proposed.Firstly,image processing technology is introduced to extract the central coordinates of these circular holes of the reference plate on the workpiece target accurately.According to the projective geometry relationship,these center points are automatically and stably sorted and the reference board image coordinate system is established.Secondly,based on the principle of perspective n points(PnP),the position and pose relation of the reference plate relative to the camera is computed.Finally,according to the hand-eye calibration results,the position and posture relationship between the workpiece coordinate system and the alignment axis coordinate system can be obtained.4.A system named VAIR is established.The accuracy and efficiency of the VAIR are verified by 3D turntable experiment,the result shows our proposed method can meet the industrial robot location requirement.
Keywords/Search Tags:Monocular vision, position and posture measurement, hand-eye calibration, industrial robotics
PDF Full Text Request
Related items