基于数字化车间的改进图像识别系统  

An Improved Image Recognition System Based on Digital Workshop

作　　者：徐崇良 胡毅[2,3] XU Chong-liang;HU Yi(University of Chinese Academy of Sciences,Beijing 100049,China;National Engineering Research Center for High-end CNC,Shenyang Institute of Computing Technology,Chinese Academy of Sciences,Shenyang 110168,China;不详)

机构地区：[1]中国科学院大学,北京100049 [2]中国科学院沈阳计算技术研究所高档数控国家工程研究中心,沈阳110168 [3]沈阳高精数控技术有限公司,沈阳110168

出　　处：《组合机床与自动化加工技术》2020年第2期70-73,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：2017年智能制造综合标准化项目:数控装备故障信息数据字典标准研制及试验验证(20172150299)

摘　　要：数字化车间的主要特点是对于普通车间的重新定义,是通过虚拟的方式将传统车间构建到虚拟环境中,对整个生产过程进行虚拟仿真推进生产。针对数字化车间虚拟信息化的识别精度问题,传统的图像识别系统对于数字化车间中精密仪器的识别准确率偏低,文章中的图像识别系统是通过针对数字化车间与卷积神经网络[1]的特点进行改进,通过将卷积神经网络与支持向量机相结合,将分类模型加入到图像识别中,最终提高识别准确率,稳定高效的帮助用户进行仿真从而推进生产。The main feature of the digital workshop is the redefinition of the ordinary workshop, which is to facilitate the production by placing the traditional workshop into the virtual environment in a virtual way. Aiming at the recognition accuracy of digital workshop virtual information, the traditional image recognition system has low recognition accuracy for precision instruments in the digital workshop. In this study, the image recognition system is improved by the characteristics of digital workshop and convolutional neural network. Adding a classification model to image recognition by combining a convolutional neural network with a support vector machine,Increase recognition accuracy, stabilize and efficiently help users to simulate and improve production.The main characteristic of digital workshop is to redefine the common workshop. It constructs the traditional workshop into the virtual environment by virtual way, and carries on the virtual simulation to the whole production process to promote production. Aiming at the problem of recognition accuracy of virtual information in digital workshop, the traditional image recognition system has a low recognition accuracy for precision instruments in digital workshop. The image recognition system in this paper is improved by aiming at the characteristics of digital workshop and convolution neural network. By combining convolution neural network with support vector machine, the classification model is added to the image. In recognition, ultimately improve the accuracy of recognition, stable and efficient help users to carry out simulation to promote production.

关 键 词：图像识别 卷积神经网络 计算机应用 数字化车间 支持向量机 

分 类 号：TH162[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]