针对持续入侵事件的分级识别算法
DOI:
CSTR:
作者:
作者单位:

(北方工业大学 电子信息工程学院,北京 100144)

作者简介:

盛智勇(1979-),男,实验师,研究方向为电子电路、机器学习.

通讯作者:

中图分类号:

基金项目:

国家自然科学基金(61571014)和北京自然科学基金(4172017)资助项目 (北方工业大学 电子信息工程学院,北京 100144)


A hierarchical recognition algorithm for continuous intrusion event
Author:
Affiliation:

(College of Electronic and Information Engineering,North China University of Tec hnology,Beijing 100144,China)

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:

    提出了一种基于应激反应过程的光纤预警分级 识别算法。该 算法受启发于人体受到外界刺激时机体分阶段调用组织系统和能量进行抵抗的机制,针对持 续振动的光纤 信号设计了分级识别算法。对于持续入侵事件引起的光纤振动信号,首先用高识别精度、高 时间消耗的算 法进行短时间的识别,以确定当前入侵事件类型;后续振动信号用低识别率、低时间消耗的 算法进行识别, 以监测是否新入侵事件产生。当发现新入侵事件后,需用高识别率的算法再次识别以查正。 实验结果表明,本算法能在一定识别精度下识别速度提升为原来的3.67倍,保证了保证了系统实时性监测的要求。

    Abstract:

    A hierarchical recognition algorithm for optical fiber pre-warming system (OFPS ) based on stress reaction process is proposed.The algorithm is inspired by the mechanism of the body calling system and energy in stage when under stress,and we use the hierarchical algorithm for recognizing continuous invasi on.For the optical fiber vibration signal caused by the continuous intrusion event,the hig h recognition accuracy and high time consumption algorithm is used to identify the current int rusion event types in a short time.The following vibration signals are identified with low recogni tion rate and low time consumption algorithm to monitor whether the new intrusion event is gen erated.When the new intrusion is found,the signal will be re-identified by the high recognition rate algorithm.The experimental results show that the recogniti on speed of the algorithm is 3.67times faster than that of the complicated algo rithm while the recognition accuracy keeps the same level.Thus the requirements of real-time monitoring and high-accuracy recognition can be ensured simultane cously.

    参考文献
    相似文献
    引证文献
引用本文

盛智勇,曲洪权,王彦平,苑世娇.针对持续入侵事件的分级识别算法[J].光电子激光,2018,29(8):884~892

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:2017-09-22
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2018-09-11
  • 出版日期:
文章二维码