Research On Binocular Stereoscopic Vision Eye Calibration

With the rapid development of China's manufacturing transformation and upgrading,industrial robots with visual functions are increasingly used in traditional manufacturing.Visually functional robots can meet the needs of modern production quickly and efficiently.To accomplish these operations,hand-eye calibration must be performed between the visual system and industrial robots in advance.This paper focuses on the calibration of the binocular stereo vision hand-eye system.The research content and progress made include the following aspects:1.For the binocular vision system,the internal parameters and distortion parameters of the two cameras are obtained by visual calibration of the binocular stereo vision system;the binocular stereo vision is stereoscopically calibrated based on the obtained internal parameters and distortion parameters,and the left and right cameras are obtained.The pose relationship,and then discussed the binocular stereo vision mathematical model and binocular stereo vision triangulation principle.2.In this paper,based on the Tsai hand-eye calibration algorithm and steps,the preliminary hand-eye calibration of the binocular stereo vision hand-eye system is completed.In order to overcome the shortcomings of the conventional step-by-step calibration algorithm for the first-to-finish rotation matrix and then the translation vector,a hand-eye model integrating the binocular stereo vision system with the robotic arm control system was established.For the solution of the two unknown optimal transformation matrices in the hand-eye model,this paper translates them into a model parameter optimization process that minimizes the re-projection error of all corner points in the calibration data.For the two unknown transformation matrices in the hand-eye model,this paper uses Loop iterations based on the levenberg-marquardt algorithm.The optimized hand-eye calibration algorithm in this paper makes full use of the calibration observation data,which can well suppress the influence of noise on the experimental results,and can obtain better hand-eye calibration results.3.For a workpiece with a regular geometric shape on an actual industrial production line,this article uses a three-dimensional scanner to obtain the three-dimensional shape of the workpiece and extracts the three-dimensional position of the center axis normal vector of the workpiece and the center of the top circle on the cylindrical surface.The hand-eye calibration results complete the positioning and grabbing of the workpiece.