基于经验模态分解和小波变换的地震瞬时频率提取方法及应用

传统的应用Hilbert变换提取地震瞬时频率的方法是基于平稳信号提出的,而野外采集的地震信号是典型的非平稳信号,因此其实际应用存在一定的局限性。本文提出了一种基于经验模态分解和小波变换的地震信号瞬时频率提取方法和流程。该方法首先对地震数据进行经验模态分解(Empirical Mode Decomposition,简称EMD),获得一组固有模态函数(Intrinsic Mode Functions,简称IMF),然后应用根据地质研究目标所划定的地层层序和声波测井数据,结合沉积旋回分析,选择地震数据中合适的固有模态函数。最后,对筛选的固有模态函数进行Morlet小波变换,提取地震瞬时频率。分别将Hilbert、Hilbert-Huang和本文方法应用于理论模型,结果表明本文方法是有效的;同时,实际地震资料应用结果表明,通过本文方法能获得更为精确的符合地质认识的地震瞬时频率剖面。

Seismic instantaneous frequency extraction based on empirical mode decomposition and wavelet transform

Seismic data acquired in the field are typical non-linear and non-stationary signals. However the conventional method such as instantaneous frequency extraction by the Hilbert transform, is properly based on linear and stationary signal. We propose in this paper a novel approach based on empirical mode decomposition (EMD) and the wavelet transform to extract seismic instantaneous frequency. In order to avoid artificial choice by experience, a method of choosing proper intrinsic mode function (IMF) decomposed from seismic traces is put forward by matching acoustic logging data and geological sequence. A selected IMF of seismic data is used to extract instantaneous frequency by the morlet wavelet transform. Tests results on both model and field data show that the proposed approach is more effective and accurate in instantaneous frequency extraction than those from both of the Hilbert transform and the Hilbert-Huang transform.