Research On The Method Of Delta Parallel Robot Picking Up Based On Machine Vision

In the food packaging and other industries,the production efficiency is seriously affected because of a large number of repetitive manual operations,such as product sorting,picking,packing and so on.With the rise of labor costs and the improvement of production environment requirements,the manufacturing industry is gradually transforming towards a higher degree of automation and intelligence.Delta parallel robot has the advantages of light structure,fast dynamic response and high precision,which has attracted more and more attention in the field of food production and packaging.In this paper,combined with the needs of enterprises,the automatic picking and packing function of delta robot is realized based on machine vision technology and Delta parallel robot.The main research contents are as follows:(1)The structure characteristics of Delta parallel robot are studied.Structure,kinematics and dynamics analysis are the basis of Delta parallel robot control.In this paper,the inverse kinematics model is established by geometric method,and the dynamic equation of the robot is established based on the screw theory and Kane equation.The correctness of the kinematics and dynamics solution method is verified by simulation in MATLAB.(2)In order to realize the automatic picking of robot,the vision system is introduced into Delta robot.A variety of machine vision schemes are designed,analyzed and compared.The key hardware such as camera,lens and light source are calculated,analyzed and selected in detail.(3)The camera calibration method and the Robot-Camera calibration method are studied.The coordinate transformation relationship in camera calibration method is deduced in detail,and the classical coordinate transformation model solution method-Zhang Zhengyou calibration method is analyzed and explored,and the calibration effect of OpenCV and MATLAB is compared.On this basis,the realization of camera robot calibration method is introduced.(4)The image recognition algorithm is studied.The OneCut image segmentation algorithm based on ORB feature matching is proposed.The ORB feature matching is used as the initial input of OneCut segmentation,and the binary segmentation of the target image is realized.Based on the binary image,Canny edge detection is used to find the outermost contour of the object,and the centroid of the object is located by solving the moment of the contour image.(5)A parallel robot picking experimental platform based on machine vision is built.The static and dynamic picking up experiments are carried out on the platform,and the experimental results are analyzed.The experimental results show that the method is correct.