Reserch On Robot Calibration Method Based On Landmark Location

Due to its low cost and simple structure,the use of micro robots is becoming more and more popular.However,because of its low cost,micro-robots also have problems with installation position accuracy and parallelism.At the same time,micro-robots are easy to move,but the mutual positions of multiple robots need to be re-calibrated after moving.This paper proposes a robot parameter calibration method based on artificial landmarks and machine vision technology,which can conveniently calibrate the horizontality,verticality,zero position and other parameters of a three-degree-of-freedom robot,providing a basis for the application of micro-robots.The research work of this article mainly includes the following aspects:(1)Research on the design and generation methods of landmarks.In order to quickly obtain the position coordinates of the robot,this article first designed two kinds of coding slices,block and strip,and coded the coding slices by shape and color.There are 125 and 972 types of landmarks for block and strip respectively.Different coded code slices;the stitching method between two coded slices and three coded slices is studied;the single-window M array is used to improve the backtracking generation algorithm to generate a landmark with the global uniqueness of the code,and a 31*31 single-window M is realized Array landmark generation.(2)Research on landmark decoding and visual positioning methods.The landmark image is preprocessed in terms of multi-value processing,connected block formation,denoising,etc.,and then the extraction method of coding components such as the center calibration ring of the coding piece and six coding blocks is studied,and the centroid coordinates of the coding components are extracted.According to the extracted coding components and coding code rules,the codeword is identified to determine the plane position of the coding slice,and the decoding success rate is 98.7%;the precise positioning method of the camera center point in the coding slice is studied,and the calibration ring is first extracted The centroid and the center line of each code block are fitted using the least square method to fit the six center lines to obtain six feature points.The six feature points are used to establish a quadric surface equation to fit the calibration plane to eliminate the distortion of the camera.Influence;Interpolate positioning of the center point of the camera in the quadric surface.Get the position coordinates of the camera center in the calibration plane.Experiments show that the maximum position error of the obtained camera center in the X and Y directions in the image is 3 pixels and 2 pixels,respectively,and the corresponding real space coordinates are 0.6mm and 0.4mm,respectively.(3)Research on robot parameter detection based on landmark positioning.Modeling and kinematics analysis of the three-degree-of-freedom 3R robot;researched methods such as the zero position detection of the robot base joint,the level detection of the large and small arms and the verticality detection of the base to the ground,and the changes of the X and Y coordinates of the visual center point The amount does not exceed 0.003 mm.The zero position of the large and small arms is detected through the two states of natural suspension and energization.It is found that there is an error of 0.0388 mm in the X direction and an error of0.0375 mm in the Y direction;the relative position of multiple robots is studied.The detection and calibration method of,derives the multi-robot closed kinematic chain equation,and designs the calibration process.(4)Robot parameter detection experiment.The detection method of the forearm installation position is studied,and the position of the forearm of the micro-manipulator used on the link shaft has an installation error of 0.3mm through experimental analysis;the installation parallelism of the forearm motor shaft is tested,and the experiment shows that the small arm is installed.The parallelism between the arm axis and the calibration plane has an inclination angle of 0.0504°;the relative position of the two micro-robots was tested and calibrated.The experiment showed that the calibration errors of the relative positions of the multi-robots in the X and Y directions were 0.1497 mm and 0.1602 mm,respectively.