Research On Identification And Grasping Of Vessels In Automatic Proportioning Process Of Chemical Reagents

With the development of intelligent industry,the intelligent degree of chemical reagent ratio may be further improved.With the combination of machine vision and robot,the ability which can perceive environment is endowed to robots and applied to the ratio of chemical reagent.The mode not only avoids the risk of injury during the experiments,but also improves the intelligent degree of chemical reagent ratio.Recognition and orientation technology of chemical reagents is the first step in the whole matching experiment,so this paper studies binocular stereo vision.Firstly,This paper establishes the coordinate system of the visual system and solve depth of images by analyzing the imaging principle and distortion principle of the camera.Zhang’s calibration method is used to calibrate the stereo camera by analyzing three calibration methods.The intrinsic and extrinsic parameters and distortion parameters of the binocular camera are determined by comparing the experimental results of Matlab and Open CV.Secondly,Spatial filtering and Gamma correction are used to preprocess the collected images.Template matching based on feature points and color threshold segmentation method are used to implement object identification.With the introduction to the traditional SIFT based matching,Brute Force and Fast Library for Approximate Nearest Neighbors are combined with SIFT algorithm to perform matching experiments in bright,dim,clutter,occlusion and other environments where the experimental matching rate and matching time are analyzed.With stereo vision for images,epipolar error is 0.49 px,which ensures alignment of two image rows meets the basic experimental requirements.SGBM algorithm is used to solve the disparity map of a single beaker and the actual experimental scene.By looking for the contour edge of the target object,the maximum connected region is selected to solve the pixel center point of the left and right view of the target object,and then the spatial pose of the center point is determined.Finally,the binocular vision grasping system is built,and the key components such as binocular camera,mechanical claw,compressor and manipulator are reasonably selected.The kinematics model of the robot is established,and the forward and inverse kinematics of the UR3 robot are solved.According to the principle of hand-eye calibration,the lower position coordinate of the camera coordinate system is transformed into the lower position coordinate of the base coordinate system,and the grasping experiment is completed.The experimental results are analyzed which meets the requirements of robot grasping.