电子元器件产品外观质量智能检测技术研究  

Research on Intelligent Inspection Technology of Appearance Quality of Electronic Components

作　　者：石锦成 刘鑫 刘蜜 刘超 SHI Jin-cheng;LIU Xin;LIU Mi;LIU Chao(Guizhou Space Appliance Co.Ltd,Guiyang 550009,China)

机构地区：[1]贵州航天电器股份有限公司,贵州贵阳550009

出　　处：《仪表技术与传感器》2020年第12期101-104,109,共5页Instrument Technique and Sensor

基　　金：国防基础科研计划项目(JCKY2018204B044)。

摘　　要：为了解决电子元器件导线产品外观质量检测环节中存在的现实困难以及人工筛检时的弊端,将经典机器视觉与深度神经网络相结合,构建产品失效样本数据集,构建产品筛选方案整体实施流程,研制出与产品特性相适应的筛检机构,并通过对比试验与结果分析证实了方法的可行性。In order to solve the practical difficulties in the inspection of electronic component conductor product appearance quality and the disadvantages of manual screening,the classical machine vision and deep neural network are combined,the product failure sample data set is built,the overall implementation process of product screening scheme is constructed,and a set of screening mechanism adapted to the product characteristics is developed.The feasibility of the method is verified by comparative test and result analysis.

关 键 词：机器视觉 深度学习 产品外观 质量检测 

分 类 号：TK267[动力工程及工程热物理—动力机械及工程]