现代电子技术2018，Vol.41Issue(7)：29-32,4.DOI:10.16652/j.issn.1004-373x.2018.07.008
基于改进CNN的增强现实变压器图像识别技术
Improved CNN based transformer image recognition technology in augmented reality environment
摘要
Abstract
The image recognition technology of transformer in augmented reality environment is studied. In order to solve the problem of transformer image recognition in augmented reality environment,an improved convolutional neural network (CNN)model based on two parallel structures is proposed on the basis of introduction of CNN as one of the typical deep learning models. The images obtained by scanning of an augmented reality camera are classified by means of the improved CNN to realize the transformer graphical recognition. In comparison with ordinary CNN and SIFT image recognition algorithm,the improved CNN has lower error rate,and higher accuracy for transformer image recognition. The accuracy of this method was verified with simulation experiments.关键词
增强现实/改进CNN/变压器/图像识别/识别准确度/卷积运算Key words
augmented reality/improved CNN/transformer/image recognition/recognition accuracy/convolution operation分类
信息技术与安全科学引用本文复制引用
李军锋,何双伯,冯伟夏,熊山,薛江,周青云..基于改进CNN的增强现实变压器图像识别技术[J].现代电子技术,2018,41(7):29-32,4.基金项目
广东电网虚拟现实和增强现实重点实验室资助项目(GDKJQQ20152015):基于增强现实的工作辅助课件及其支撑系统研究与开发 Project Supported by Key Laboratory of Guangdong Power Grid Virtual Reality and Augmented Reality(GDKJQQ20152015):Research and De-velopment of Work Assisted Courseware and Support System Based on Augmented Reality （GDKJQQ20152015）
