基于HHT和频带能量比特性的人与动物雷达微动信号的辨识  

Micro-Vibration Identification between Humans and Animals Based on HHT and Frequency Energy Ratio Features Using UWB Radar

作　　者：殷悦 吕昊 祁富贵 于霄 YIN Yue;LV Hao;QI Fugui;YU Xiao(Department of Medical Engineering,The General Hospital of Western Theater Command PLA,Chengdu Sichuan 610083,China;School of Military Biomedical Engineering,Air Force Medical University,Xi’an Shaanxi 710032,China)

机构地区：[1]中国人民解放军西部战区总医院医学工程科,四川成都610083 [2]空军军医大学军事生物医学工程学系,陕西西安710032

出　　处：《中国医疗设备》2021年第6期26-30,58,共6页China Medical Devices

基　　金：国家自然科学基金(81601567);陕西省重点研发计划(2018-SF170);空军军医大学珠峰工程(2020ZFB009)。

摘　　要：目的提出一种基于希尔伯特-黄变换(Hilbert-Huang Transform,HHT)的生命体雷达微动信号辨识方法。方法构建生命体微动特征频带,通过提取边际谱和计算能量占比这两类能反映微动信号频带特性的信息实现人与兔、犬的辨识。结果根据人与两种动物的微动信号经HHT处理得到的边际谱和能量占比信息具有显著差异的特性(P<0.01),本方法区分人与动物的准确度达91.67%,精准识别人与兔、犬目标的准确度达83.33%。结论本文提出的辨识方法为静止状态下障碍物后人与多种动物的辨识提供了一种通用的新方法,对于灾害救援、反恐维稳等多种应用场合下提高搜救效率,优化救援资源等具有重要的社会价值和实际意义。Objective Proposed an identification method of stationary living bodies based on micro-vibration radar signal using Hilbert-Huang transform(HHT).Methods The identification of human,rabbit and dog was realized by extracting marginal spectrum and calculating energy ratio,two kinds of information which can reflect the frequency band characteristics of fretting signal.Results Based on the statistically significant difference of marginal spectrum and energy ratio information(P<0.01)after HHT analysis among humans and animals,the accuracy of this method to distinguish humans and animals identification was 91.67%,and the accuracy of humans,rabbits and dogs precise identification was 83.33%.Conclusion The identification method proposed in this paper provides a new general approach to recognize motionless humans and animals and has great social value and realistic significance for increasing rescue efficiency and optimizing rescue resources in disasters and anti-terrorism situations.

关 键 词：超宽谱生物雷达 希尔伯特-黄变换 边际谱 能量占比 人与动物辨识 

分 类 号：R197.39[医药卫生—卫生事业管理] TN957[医药卫生—公共卫生与预防医学]