Design And Implementation Of Distributed Intelligent Visual Inspection System Based On EIS

With the successive proposals of Industry 4.0,Industrial Internet and Made in China 2025,smart factories and smart manufacturing have become an important development trend in the transformation and upgrading of the manufacturing industry,and the importance of machine vision in the practice of industrial intelligence is self-evident..As one of the fastest growing areas of artificial intelligence,machine vision is rapidly iterating whether it is visual algorithms or landing products.Based on the actual industrial production requirements,this paper designs and implements a distributed intelligent visual inspection system based on the architecture mode and basic components of Intel's Industrial Edge Insight Platform(EIS).The system realizes data collection,data analysis,data storage and dynamic configuration management through distributed deployment.The intelligent visual inspection system can be used for the identification,inspection,measurement,positioning and matching of products in industrial production.The distributed intelligent visual inspection system mainly includes seven modules: client,data acquisition,data processing,data storage,data transmission,configuration management and image processing algorithms.This thesis mainly introduces the architecture design of the entire system and the core implementation logic of each module,but does not discuss specific image processing algorithms.The data acquisition module collects real-time images through the Aravis vision platform and industrial cameras;the data processing module constructs image processing procedures and dynamically creates image processing tools based on the configuration parameter information,and processes the images to be inspected in real time;the data storage module can be selected according to the user Persistent storage of core data;configuration management module uses ETCD as a distributed key-value storage system to realize the dynamic management function of configuration parameters in the intelligent detection system.The data transmission module adopts the standard component MsgBus of EIS as the network communication bus of each module in the system.The bottom layer of MsgBus is implemented based on the message queue ZeroMQ,which has the characteristics of large data throughput and low latency.The publish and subscribe mechanism of ZeroMQ can easily realize the implementation of any module.data communication.The client mainly provides the functions of system control,display of image processing results and configuration of image processing tool parameters.The intelligent visual inspection system adopts distributed and independent deployment.Each module can be deployed in the cloud or at the edge.At the same time,it can be considered to cooperate with Intel's industrial edge insight platform to use containerized deployment.The development environment of the system is Ubuntu 18.04,the programming languages used are C++ and QML,and the client module uses the Qt 5.8 development framework.The system has good portability and can run stably in a Linux-like environment.At the end of the thesis,different test cases are designed for the data processing module,configuration management module and client module,and the overall function and performance of the system are tested.The test results show that the functions of the various modules of the intelligent vision inspection system are normal,and all the functions of the system have reached the demand expectations.