基于数学形态学的短时傅里叶分析  

Short Time Fourier Analysis Based on Mathematical Morphology

作　　者：罗伟强[1] 黄家栋[1] 于建辉[2] 于建涛[1] 

机构地区：[1]华北电力大学电气工程学院,河北省保定市071003 [2]国家开发银行河北省分行评审处,河北省石家庄市050051

出　　处：《电力系统自动化》2007年第23期38-41,共4页Automation of Electric Power Systems

摘　　要：为了提取恰当的短时傅里叶分析(STFT)时间窗宽度和中心,进而利用STFT分析暂态信号成分,提出了基于数学形态学的STFT方法。针对电力系统信号特点,讨论了结构元素的选择问题,推导了结构元素大小公式及离散化的结构元素方程。同时提出了奇异点定位算法,进而得到了基于数学形态学的STFT算法。在MATLAB中的仿真结果表明该方法能够快速准确地得到暂态信号中主要频率成分及其幅值,可提高STFT的精度和效率。In order to determine the appropriate time window width and center of short-time Fourier analysis （STFT）, a method combining STFT and mathematical morphology is presented to analyze the components of a transient signal. According to the characteristics of power system signals, the identification of suitable structure elements and their discrete formulas, such as structural elements size formula and structural elements equations are investigated. Singular point location algorithm is used in STFT with mathematical morphology. The MATLAB simulation examples show that the method can rapidly and accurately access the main frequency and amplitude components of a transient signal compared to the traditional STFT.

关 键 词：短时傅里叶分析 数学形态学 奇异点定位 

分 类 号：TM744[电气工程—电力系统及自动化]