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
李军锋,何双伯,冯伟夏,熊山,薛江,周青云
摘要(Abstract)：
研究了增强现实变压器图像识别技术,为解决增强现实中变压器图像识别问题,首先在介绍深度学习的经典模型之一,即卷积神经网络CNN的基础上,提出基于两个并行结构的改进卷积神经网络模型(改进CNN),利用改进CNN模型对增强现实摄像头扫描得到的图像进行分类,实现变压器图形化识别。与普通卷积神经网络、SIFT图像识别算法等对比,改进CNN具有更低的错误率,并对变压器图像识别的准确率更高,通过仿真实验验证了此方法的准确性。
关键词(KeyWords)： 增强现实;改进CNN;变压器;图像识别;识别准确度;卷积运算
基金项目(Foundation): 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~~
作者(Author): 李军锋,何双伯,冯伟夏,熊山,薛江,周青云
DOI: 10.16652/j.issn.1004-373x.2018.07.008
参考文献(References)：
- [1]陆平.移动增强现实中的图像处理关键技术研究及应用[D].南京:东南大学,2015.LU Ping.Research on key technology of image processing in mobile augmented reality and its application[D].Nanjing:Southeast University,2015.
- [2]蔡苏,王沛文,杨阳,等.增强现实(AR)技术的教育应用综述[J].远程教育杂志,2016(5):27-40.CAI Su,WANG Peiwen,YANG Yang,et al.Review on augmented reality in education[J].China remote education magazine,2016(5):27-40.
- [3]李青,张辽东.基于增强现实的移动学习实证研究[J].中国电化教育,2013(1):116-120.LI Qing,ZHANG Liaodong.Empirical study of mobile learning based on augmented reality[J].China electrochemical education,2013(1):116-120.
- [4]RASCHMAN E,DURACKOVA D.New digital architecture of CNN for pattern recognition[C]//International Conference Mixed Design of Integrated Circuits and Systems.[S.l.]:IEEE,2009:662-666.
- [5]MU Nan,XU Xin,ZHANG Xiaolong,et al.Salient object detection using a covariance-based CNN model in low-contrast images[J].Neural computing and applications,2017(2):1-12.
- [6]LECUN Y,BOTTOU L,BENGIO Y,et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE,1998,86(11):2278-2324.
- [7]徐姗姗.卷积神经网络的研究与应用[D].南京:南京林业大学,2013.XU Shanshan.Research and application of the convolution neural network[D].Nanjing:Nanjing Agricultural University,2013.
- [8]陈先昌.基于卷积神经网络的深度学习算法与应用研究[D].杭州:浙江工商大学,2014.CHEN Xianchang.Research on algorithm and application of deep learning based on convolutional neural network[D].Hangzhou:Zhejiang Gongshang University,2014.
- [9]白廷柱,侯喜报.基于SIFT算子的图像匹配算法研究[J].北京理工大学学报,2013(6):622-627.BAI Tingzhu,HOU Xibao.Research on image matching algorithm based on SIFT operator[J].Journal of Beijing Institute of Technology,2013(6):622-627.
- [10]SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[C]//Proceedings of ICCV.[S.l.]:IEEE,2015:1-14.