Research On Robot Vision Identification And Location In Hybrid Production System

In recent years,with the rapid development and maturity of image processing and visual recognition technology,the machine vision to robot has become a hot topic in the field of industrial robot research.For the problem of variety of product models in hybrid production system and single industrial robot cannot complete the multi-model and multi-variety workpieces location fetching task,this paper presents a method for the identification of different types and types of workpieces by industrial robots based on visual recognition technology.The simulation experiment of robot kinematics and the track planning of grasping workpiece is carried out with MATLAB Robotics toolbox.Firstly,the image processing technique is used to pretreat the initial image of the workpiece.In order to improve workpiece contour feature recognition effect,use median filtering to enhance the brightness of the image,highlight the details of image,histogram equilibrium method is used to improve the contrast of image feature points and the surrounding environment,make the image of target is more likely to be identified,image segmentation technique is used to change the image is divided into a number of characteristics to independence of area,so that the target characteristics of the area and the surrounding area separated,in through graphics and processing of the workpiece image target features prominent,more easily also fit in with the next feature point extraction.Secondly,using the nonlinear processing capability of the neural network algorithm of image preprocessing of workpiece after image of target feature points extraction,by choosing a suitable neural network training function and an appropriate number of iterations,neural networks to set the appropriate network input layer,hidden layer and output layer in the number of neurons in each layer to determine the optimal threshold,the training can be more accurate to extract target feature points,and to extract the feature points,concentrated form feature set is stored as a robot used to learn the knowledge base,when identifying a specifications of the workpiece,the robot can obtain knowledge pool characteristics compared to,after successful match robots will adjust good.Thirdly,in order to achieve the target point of the robot successfully,the space coordinate relation between the robot and the target point of the image needs to be established.This paper,by using coordinate transformation formula converts the image pixel coordinates to match the robot kinematics with the physical coordinate space,through the establishment of the inverse kinematics equation of the robot robot joint coordinate transformation,the position of the robot according to the target point information to coordinate all joints movement adjust good attitude,implementation artifacts the positioning of the grasping task.Finally,the trajectory planning method of the location grab path of industrial robot is studied and simulated.For industrial robot joint space and cartesian space trajectory planning,Robotics,and by using MATLAB toolbox simulation experiment for robot localization and make the robot hand at the end of the clip according to the given capture path movement,and then simulate the robot localization grasping motion process.