基于神经网络的室内外场景识别方法
DOI:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

P228.1

基金项目:

国家重点研发计划项目(2016YFB0502102)第一作者:胡贤贤(1996—),男,汉,陕西石泉,硕士研究生。研究方向:室内定位与导航。E-mail:64112690@qq.com。*通信作者:汪云甲(1960—),男,汉,江苏建湖,博士,教授。研究方向:室内定位与导航。E-mail:wyjc411@163.com。第三作者:孙 猛(1995—),男,汉,河南沈丘,博士研究生。研究方向:室内定位与导航。E-mail:msun@cumt.edu.cn。第四作者:齐红霞(1984—),女,汉,河北石家庄,博士研究生。研究方向:室内定位与导航。E-mail:hongxiaqi@yeah.net。


Indoor and outdoor scene recognition method based on the neural network
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    针对目前室内外场景识别方法所面临的低精度、低可靠性和低稳定性的问题,提出了一种基于神经网络算法的高精度室内外场景识别的方法。该方法利用智能手机内置的光传感器、磁传感器和GNSS模块采集训练数据,根据卫星数量、高度角、信噪比数据在室内外具有不同的特性,将其划分成不同的区间并结合室内外的地磁数据与光照强度数据构成场景识别的特征,最后将不同时间的特征数据输入神经网络模型进行训练室内外场景识别模型。大量的测试实验结果表明,本文提出的基于神经网络的室内外场景识别方法在不同的场景下识别准确率可以达到96%,能有效地识别室内外场景,具有较强的稳定性,可为室内外无缝定位系统的构建提供参考。

    Abstract:

    This manuscript aims at the problems of low accuracy, low reliability and low stability faced by current indoor and outdoor scene recognition methods and propose s a high precision scene recognition method based on the neural network. This method utilized the built-in light sensor, magnetic sensor and GNSS module of the smartphones to collect training data. Since the number of satellites, altitude angles and the signal noise ratio data have different characteristics in indoor and outdoor environments, they were divided into different sections and combined with the geomagnetic data and illumination data to build the characteristics data. Finally, the scene recognition model was obtained by using the neural network to train the characteristics data of different times. A large number of experimental results show that the proposed method of indoor and outdoor scene recognition based on neural network can achieve the recognition accuracy of 96% in different scenes, which means that this method can effectively identify the indoor and outdoor scenes and has strong stability, and can provide reference for the construction of indoor and outdoor seamless positioning system.

    参考文献
    相似文献
    引证文献
引用本文

胡贤贤,汪云甲,孙猛,等. 基于神经网络的室内外场景识别方法[J]. 科学技术与工程, 2021, 21(3): 1091-1096.
Hu Xianxian, Wang Yunjia, Sun Meng, et al. Indoor and outdoor scene recognition method based on the neural network[J]. Science Technology and Engineering,2021,21(3):1091-1096.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2020-05-12
  • 最后修改日期:2020-06-10
  • 录用日期:2020-07-04
  • 在线发布日期: 2021-02-09
  • 出版日期:
×
律回春渐,新元肇启|《科学技术与工程》编辑部恭祝新岁!
亟待确认版面费归属稿件,敬请作者关注