Application Research Of Machine Vision In Industrial Robot Hard Pen Calligraphy

With the development of computer technology,machine vision is becoming more and more widely used in various industries.This paper takes text pictures as the research object and designs three kinds of text recognition schemes.The optimal solution is selected by comparing the correctness of the identification.This paper establishes an industrial robot simulation platform and control platform to analyze the robot trajectory.And the robot realizes writing function.The main research contents are as follows:Firstly,the artificial neural network,convolutional neural network structure and training algorithm,network parameter optimization method,long-term and short-term memory network(LSTM),CTC loss function were introduced.The concepts of related to networks include optimizing neurons,activation functions,and softmax regression.Network parameter optimization methods include gradient descent method,Dropout,regularization constraints,and so on.Secondly,three text image recognition schemes were designed.The scheme includes the K-nearest neighbor rule scheme,convolutional neural network scheme,convolutional neural network combined with long-term and short-term memory network scheme.And the optimal scheme was selected by comparing the recognition accuracy of the three schemes.Thirdly,the simulation movement of industrial robots was researched.Taking industrial robots as research objects,a three-dimensional model of industrial robots,a simulation platform based on MATLAB and a robot control platform based on Visual Studio 2015 were established.Fourthly,implementation of robot writing function.The robot simulates the writing process on the simulation platform and is was verified on the robot entity.