基于广义S变换的地震资料谱分解技术研究

Spectral Decomposition Technique for Seismic Data Based on Generalized S Transform(GST)

谱分解技术一般采用短时Fourier变换和连续小波变换。短时Fourier变换的分析时窗是恒定的,不随频率的变化而自动调节,连续小波变换的尺度与频率的关系并不确定。广义S变换具有优良的时频聚集性和时频分辨率,它综合了短时Fourier变换和连续小波变换的一些优点,而克服了它们的不足,同时通过调节时窗宽度因子可以获得更好的时频分辨率。研究将其应用于地震资料的谱分解中,给出了基于广义S变换的谱分解业务流程,并对实际地震资料进行了分析。结果表明该方法是可行的。

Short-time Fourier Transform(STFT) and continous wavelet transform(CWT) were generally used in spectral decomposition.Time-window analysis of STFT was constant,it was not automatically adjusted with frequency changes;while the relation between size and frequency of CWT was not certain.Generalized S transform(GST) had high time-frequency resolution and good time-frequency localization.It had the advantages of STFT and CWT and overcame their shortcomings,thus better time-frequency resolution could be obtained by adjusting the factors of time-window width.GST is used for spectral decomposition of seismic data and its flow chart is provided,it is used for actual analysis of seismic data.Application shows that it is feasible.