Design And Implementation Of Software For Intelligent Visual Inspection Equipment

Machine vision is a branch of artificial intelligence that is developing rapidly.In short,it uses machines to replace human eyes for measurement and judgment.In the large-scale repetitive industrial production process,machine vision technology can greatly improve the efficiency and automation of production.Based on the actual industrial production needs,thesis designs and implements an embedded intelligent visual inspection software.This system is mainly used in product quality inspection.It can take images for the workpieces waiting to be tested,and calculate the score by calling intelligent processing algorithm.And this score can be used to distinguishing product quality.The entire intelligent visual inspection system mainly includes three modules:client software,embedded software,and core processing algorithms.Thesis mainly introduces the design and implementation of parameter management module,I/O detection module and firmware upgrade module.The parameter management module maintains and manages all the system parameters.This module improves the reliability of the parameters by using various strategies such as backing up parameters,verifying parameters in parameter reading and writing to avoid the abnormalities in system caused by power off.The I/O detection module is mainly responsible for all hardware-related system controls.The firmware upgrade module is mainly responsible for the remote upgrade function of the embedded system software,including:U-Boot image upgrade,kernel image upgrade,root filesystem image upgrade and application upgrade.In order to avoid the negative impact on the system caused by power failure during upgrade,thesis designs a backup mechanism,so that the system can always choose a normal image to load during the startup process,ensuring the reliability of the embedded operating system.The embedded vision sensor control system is a multi-module distributed system.In addition to these functional modules,thesis additionally introduces the design of the system daemon and the design of the image creation tool,which improves the overall design of the system.At the end of thesis,different test cases are designed for each module and used to test system function.According to the test result,each module of the intelligent visual inspection software functions normally.All functions of the system satisfied the requirements.