几种时频分析方法比较

The comparison of some time-frequency analysis methods

地震信号由于各种原因,往往是一种非线性、非平稳信号,基于平稳信号理论的常规傅立叶变化方法不能刻画任一时刻的频率成分.时频分析能同时保留时间与频率信息,目前已经出现了很多时频分析方法.本文介绍了Hilbert变换、Hilbert-Huang变换、正弦曲线拟合、雷克子波匹配、短时傅立叶变换、小波变换、S变换以及Cohen类这八种方法,并从时间分辨率、频率分辨率,以及对多频率成份信号适应能力等各方面阐述了各种方法的优缺点,对其中的一些方法结合了理论记录进行了试算,进一步阐述了这些方法的长处和不足之处.

The seismic data usually is a non-linear non-stationary series for many reasons, so the Fourier transform based on stationary signal can not show the news of frequency at sometime. The time-frequency analysis can give the news of time and frequency at the same time, now there have many methods of time-frequency analysis. This paper introduces eight methods of tlme-frequency analysis, such as Hilbert transform, Hilbert-Huang transform, sine curve fitting, Ricker wavelet matching, short time Fourier transform, wavelet transform, S transform and Cohen methods, This paper explains the merits and defects of those methods from the time resolution, frequency resolution and the adaptive ability of multi-frequency signals. Moreover, this paper explains the advantages and shortcomings of some methods deeply by some test cases.