Research On The Visual Servo Technology Of The 6-Dof Manipulator

With the rapid development of robotics,people are demanding more and more intelligent robots.As vision is a manifestation of highly intelligent robots,robotic vision servo technology has also become one of the core contents of the study.At present,the visual servo technology has been widely used in the control of industrial robots,accounting for a large proportion of industrial production.At the same time,the particularity of the industrial scene determines the working environment information is relatively simple,so the visual detection and positioning are not difficult to achieve.However,in the future,visual robots will not be limited to industrial sites,so the visually-perceived environmental information will become more complicated.In a complex working environment,to make the robot have a smart resolution and a clear ability to act,which requires the robot has a visual ability,which also reflect the degree of visual servo technology.This article is precisely for this purpose as a starting point,combined with laboratory research projects to focus on visual servo technology of a 6-DOF manipulator.First of all,based on a large amount of relevant documents,this paper clarifies the composition of visual servo control system.Then,the related mathematical modeling and analysis of the visual servo system are combined with the coordinate transformation in the robot system of this paper to establish the visual servo form of the eye on the hand,which is the relationship between the imitation of the human eye and the coordination of the arm.Then,according to the inherent mechanical properties of the manipulator,the kinematics is modeled and analyzed by classical D-H method without considering the dynamics,and the related modeling analysis and experimental simulation are completed.Secondly,combined with the difficulties of visual servo technology,this paper focuses on the feature extraction and recognition algorithms.Under complex background information,the SIFT algorithm is proposed for feature extraction and matching,and the matching error is eliminated by RANSAC algorithm so that the target object can be accurately identified.In the process,the OpenCV visual library is used to complete the preparation of the relevant algorithm,and the feasibility of the algorithm is verified through experiments.At the same time,calculate the target object in the image coordinate system pixel coordinate area.Third,combined with the calibrated binocular stereo camera,aiming at the target object of this experiment,the design method of actual position and attitude measurement is put forward.The GrabCut algorithm is used to process the complex background image,and the target contour and contour center are extracted.Then according to the geometric characteristics of objects and binocular vision calculation to determine the actual position and attitude of the space,so as to complete the function of the entire visual system.Finally,based on the standard three-dimensional coordinates of the target object,the end effector of the manipulator can grab the target object through the kinematic inverse kinematics of the 6-DOF manipulator.Through experimental test and data analysis,it shows the calibration of the hand-eye system and the accuracy of the positioning of the space object meets the expected design requirements.