摘要
为实现陆基大气层核爆光辐射信号的有效识别,对实测核爆、模拟核爆和闪电光辐射信号进行直流剔除、平滑滤波、小波阈值去噪、插值和归一化处理;而后根据信号产生的物理过程,提取信号时域双峰波形、脉冲上升沿、最小照度时间、第二极大光照度时间与最小照度时间的比值作为特征量,将实测与模拟核爆光辐射信号合并为核爆光辐射信号,闪电光辐射信号作为主要光干扰信号,在统计分析和经验的基础上建立对应的隶属度函数;利用模糊识别方法对两种信号进行分类识别,综合识别率可以达到99.76%,并通过有效增加信号样本量提升了识别叫靠性。
In order to realize the effective identification of land-based atmospheric nuclear explosion optical pulse,optical pulse from the measured nuclear explosion,simulated nuclear explosion and lightning are subjected to DC elimination,smoothing filtering,wavelet threshold denoising,interpolation and normalization.According to the physical principle of signal generation,we extract the double optical peaks,optical pulse’s rise-time,minimum illuminance time,the ratio of second maximum illuminance time and minimum illuminance time as features.The measured and simulated nuclear explosion optical pulses are combined into nuclear explosion optical pulse,and the lightning optical pulse is the main optical interference pulse signal.The corresponding membership function is established on the basis of statistical analysis and experience;the fuzzy recognition method is used to match the two,and the comprehensive recognition rate can reach 99.78%,which effectively makes up for the shortcomings of insufficient data of the measured nuclear explosion optical pulse.
作者
黎星宇
宋立军
刘晓宏
李鹏
刘伟
LI Xing-yu;SONG Li-jun;LIU Xiao-hong;LI Peng;LIU Wei(Chemical Defense Institute,Academy of Military Sciences,Beijing 102205,China;National No.785 Factory,Taiyuan 030024,China)
出处
《核电子学与探测技术》
CAS
北大核心
2021年第4期625-630,共6页
Nuclear Electronics & Detection Technology
关键词
光干扰
核爆
闪电
光辐射
双峰波形
模糊识别
Optical Interference
Nuclear Explosion
Lightning
Optical Pulse
Double Optical Peaks
Fuzzy Recognition
