基于混合遗传卷积神经网络的智能制造系统图像分类  

Image Classification of Intelligent Manufacturing System Based on Hybrid Genetic Convolutional Neural Network

作　　者：陶沙 司伟 王奎 TAO Sha;SI Wei;WANG Kui(Tongling University,Tongling 244000,China;China United Network Communication Co Ltd,Beijing 100032,China)

机构地区：[1]铜陵学院电气工程学院,安徽铜陵244000 [2]中国联合网络通信有限公司,北京100032

出　　处：《洛阳理工学院学报（自然科学版）》2021年第2期64-68,共5页Journal of Luoyang Institute of Science and Technology：Natural Science Edition

基　　金：安徽省自然科学重点项目(KJ2018A0484).

摘　　要：传统制造业系统检测主要借助显微工具进行观察。当前智能制造零件尺寸越来越小,加上传统方法受主观性影响,往往存在一定偏差,需要设计一种效率和精度均更高的检测技术。随着高分辨率相机以及计算机技术的发展,利用一些改进的算法模型对智能制造系统图像进行识别分类是可行的。基于混合遗传卷积神经网络(GA-CNN)技术分类准确率高,比较适合应用于智能制造系统零件检测中。In the traditional manufacturing system,microscopic tools are often used for observations.At present,the size of manufacturing parts is getting smaller and the traditional measurement method has deviations caused by subjective factors,therefore,it is necessary to design a more efficient and accurate detection method.With the development of high-resolution camera and computer technology,some improved algorithm models are proposed to recognize and classify the image of intelligent manufacturing system.The results show that based on hybrid genetic convolutional neural network(GA-CNN)technology,classification accuracy rate is high,which is suitable for parts detection in intelligent manufacturing system.

关 键 词：机器视觉 遗传算法 卷积神经网络 工件测量 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]