Design And Implementation Of Industrial Vision Interactive Development Platform

With the development of technology and industrial production,more and more factories use industrial vision systems to realize automated production.Improving the development efficiency of industrial vision applications can advance the industrial automation process.The current popular industrial vision development platform has the problems of high user requirements and low development efficiency,and limited support for deep learning.The research goal of this thesis is to design and implement an industrial vision development platform with a low knowledge requirement and high development efficiency,and provide the platform with deep learning tool set.This thesis first introduces the industrial vision application scenarios and the research status of domestic and international industrial vision development platforms.In view of the shortcomings of the current industrial vision development platform,the Industrial Vision Interactive Development Platform is designed and implemented.The platform supports development using Python language,and provides machine vision algorithm libraries,camera calibration tools,image labeling tools and other commonly used tools in the development of industrial vision applications.In order to improve the ease of the platform and the development efficiency of industrial vision applications,in addition to basic functions such as code syntax highlighting and code block debugging and operation,the platform is also designed to implement interactive programming.Users can adjust function parameters by pulling the scroll bar,and the processing results of the function are immediately visible to the users;due to the large number of function parameter adjustment operations required during the development of industrial vision applications,the function parameter settings and performance effects during the adjustment process are the key for developers to analyze the function effect.This thesis designs and implements the snapshot recording function,which assists developers to record function parameter settings and performance effects,and analyzes these records at any time to improve the functionality of the platform.This thesis studies the problem of applying deep learning to industrial vision application development,and designs a deep learning tool set and graphical interface for industrial vision application development.Traditional machine vision algorithms cannot solve industrial vision problems that require high generalization ability.Due to the strong generalization ability of deep learning,applying deep learning to industrial vision applications is currently a hot research topic.This thesis designs and implements a deep learning tool set,which guides users to develop industrial vision applications in a graphical interface,and provides sufficient deep learning pre-training models to ensure the flexibility of the tool set.Finally,the Industrial Vision Interactive Development Platform is used for workpiece counting and workpiece positioning application development,using machine vision algorithm library and deep learning tool set for development.Then this thesis deploys workpiece positioning applications in the workpiece sorting system.These cases verify the functionality and effectiveness of the platform.