Research On Positioning And Motion Control Of Scaffolding Grape Picking Manipulator

China's grape production ranks first in the world,and its cultivation area is the third,but its harvest is still dominated by labor.The harvesting process has high labor intensity,high cost,low mechanization and low degree of automation,which seriously restricts the development of the grape industry.With the maturity of technology,how to achieve automatic and intelligent grape harvesting has become the focus of current research.This paper takes scaffolding grapes as the research object,and studies the positioning and motion control of the scaffolding grape picking manipulator,aiming to improve the picking efficiency of the scaffolding grape and provide reference and reference for the development of harvesting robot.This paper has completed the following four parts of the work:(1)Construct a binocular vision positioning system and preprocess the collected grape images.Firstly,the camera linear model is established to analyze and correct the tangential distortion and radial distortion existing in the linear model of the camera.Then,using Zhang Zhengyou's plane calibration method to calibrate the camera,determine the internal parameters,external parameters and the conversion relationship between the coordinate systems.Finally,in order to improve the recognition success rate of the scaffolding grape picking point,the image acquired by the camera is preprocessed.The steps of image preprocessing mainly include mathematical morphology,boundary representation,image representation and feature extraction.(2)Research on the position of picking points for scaffolding grapes.Image-Pro Plus was used to analyze the relationship between the channels of the images collected by the camera.According to the analysis results,the color difference method in RGB space was proposed to identify the position of the 2-D plane picking point.In order to solve the influence of illumination and threshold selection on the recognition results,an adaptive threshold method based on HSI space is proposed to identify the position of 2-D plane grapes and determine the coordinates of grape positioning points.In order to determine the position of the target grape stalk,a detection device was designed by using the infrared pair tube,and the grape stalk was detected by the on-off of the infrared tube during the ascending process to determine the position of the final picking point.Aiming at the problem of multiple target grapes in the image,a method based on the number of pixel values is proposed to identify the test results.(3)Design of motion control algorithm for picking manipulator.After obtaining the coordinates of the scaffolding grape positioning point,the trajectory planning of the picking manipulator is carried out by using the fifth-order polynomial interpolation method.In order to avoid the chattering during the movement of the manipulator,a new sliding mode variable structure algorithm is proposed to track the picking manipulator trajectory based on the traditional sliding mode variable structure control.The simulation results show that the design of the sliding surface and the approach law is beneficial to eliminate the chattering phenomenon during the movement of the picking manipulator,further improve the picking success rate and shorten the picking time.(4)Construction and test verification of the scaffolding grape harvesting platform.Based on the 6-DOF manipulator to build a scaffolding grape picking test platform.The picking manipulator and the binocular camera model are determined according to the experimental requirements.Then,an end shear clamping device is designed to ensure that the mechanical arm can grip the grape stem during the cutting process.Finally,based on the test platform,the success rate of the positioning,the success rate of picking and the picking time were tested and the experimental results were statistically analyzed.In summary,this paper studies the positioning and motion control of the scaffolding grape picking robotic arm.Constructing a visual positioning system model for the scaffolding grape planting mode.The color difference method and the adaptive threshold method are proposed to determine the position of the 2-D plane grape,and the coordinates of the positioning point are solved by the visual model.Designing an infrared pair tube detecting device to detect the grape stalk,and finally obtaining the picking point position.A method based on the number of pixel values is proposed to identify and locate the single target grape and multiple target grape picking points.Using the fifth-order polynomial interpolation to trajectory planning of the picking manipulator and proposing a new sliding mode variable structure control to track the picking manipulator trajectory.Finally,build a test platform to verify the feasibility of the proposed algorithm.