光通信技术2026，Vol.50Issue(2)：7-11,5.DOI:10.13921/j.cnki.issn1002-5561.2026.02.002
基于改进变分模态分解的光纤周界安防入侵识别研究
Research on fiber-optic perimeter security intrusion identification based on improved variational mode decomposition
摘要
Abstract
To meet the demand for accurate intrusion event recognition in the field of perimeter security,a fiber-optic vibration sensing system based on dual Mach-Zehnder interferometry is designed.The grey wolf optimization algorithm is employed to improve the variational mode decomposition,automatically optimizing the number of modal components and the penalty fac-tor.The intrinsic mode functions and kurtosis features of continuous disturbance signals from three types of intrusion events(stepping,knocking,and climbing)are extracted,and a support vector machine is used for event classification and recognition.Experimental results show that on a 2.05 km sensing fiber,the mean kurtosis values of stepping and knocking events reach their maxima in the third layer,while that of climbing events reaches its maximum in the fifth layer.The average recognition rates for the three types of events are 97.2%,98.6%,and 97.9%,respectively,demonstrating the effectiveness and practicality of the proposed method for intrusion recognition in fiber-optic perimeter security systems.关键词
光纤传感/周界安防/双马赫-曾德尔干涉/变分模态分解Key words
fiber-optic sensing/perimeter security/dual Mach-Zehnder interferometry/variational mode decomposition分类
信息技术与安全科学引用本文复制引用
彭广建,王宇,程安文,白清,王鹏飞,刘昕,靳宝全..基于改进变分模态分解的光纤周界安防入侵识别研究[J].光通信技术,2026,50(2):7-11,5.基金项目
中央引导地方科技发展资金项目(YDZJSX2024C002)资助 （YDZJSX2024C002）
山西省重点研发计划项目(202102130501021)资助 （202102130501021）
山西省科技创新团队项目(201805D131003)资助 （201805D131003）
中央引导地方科技发展资金项目(YDZJSX20231B004)资助. （YDZJSX20231B004）
