基于指数变换时频图增强的FSK信号特征提取  

Feature extraction of FSK signals based on exponential transformation time-frequency image enhancement

作　　者：郑文秀 覃艳琴 常虹 ZHENG Wenxiu;QIN Yanqin;CHANG Hong(School of communication and Information Engineering,Xi＇an University of Posts and Telecommunications,Xi＇an 710121,China)

机构地区：[1]西安邮电大学通信与信息工程学院,陕西西安710121

出　　处：《西安邮电大学学报》2018年第1期32-37,共6页Journal of Xi’an University of Posts and Telecommunications

基　　金：陕西省教育厅专项科研计划项目(15JK1649)

摘　　要：为了区分信号的载波频率,提出一种基于指数变换的图像增强技术对频移键控(frequency-shift keying,FSK)信号进行时频二维特征提取的方法。该方法根据FSK信号的离散短时频分析,将时频图转化为时频二维灰度图,并进行图像增强处理;然后基于增强的时频二维灰度图,提取出特征参数角二阶矩(angular second moment,ASM)能量,用ASM能量曲线波峰来区分载波频率。在信噪比为10dB到30dB条件下,针对载波频率为3 MHz和3.5 MHz的2FSK信号,进行特征提取和性能仿真实验,实验结果表明,提出的方法可行,能够区分出信号的载波频率,且性能较稳定。In order to distinguish the carrier frequency of signals,a time-frequency feature extraction method for frequency-shift keying（FSK）signals based on exponential transform image enhancement is proposed.In this method,the time-frequency graph is transformed into time-frequency two-dimensional grayscale and enhanced according to the discrete short-time frequency analysis of FSK signal;the characteristic parameters of angular second moment（ASM）energy is extracted based on the enhanced time-frequency two-dimensional gray image;and then the carrier frequency is distinguished by using ASM energy peaks.Under the condition of SNR of 10 dB to 30 dB,the feature extraction and performance simulation experiments are carried out for 2 FSK signals with carrier frequency of 3 MHz and3.5 MHz.Simulation results show that the proposed method is feasible,and can distinguish the carrier frequency of signals,and has stable performance.

关 键 词：离散短时频分析 指数变换 载波频率 特征提取 

分 类 号：TN911.6[电子电信—通信与信息系统]