基于总体经验模式分解的地震信号随机噪声消除

Random noise attenuation based on ensemble empirical mode decomposition

地震资料去噪是地震数据处理非常重要的步骤,现代地震勘探对地震资料信噪比的要求越来越高.总体经验模式分解(ensemble empirical mode decomposition,简写为EEMD)是一种新的时域信号处理方法 ,它是对经验模式分解(empirical mode decomposition,简写为EMD)的一种改进.EEMD将目标信号经验地分解为几个被称为本征模态函数(intrinsic mode function,简写为IMF)的子信号,它是一个自适应的带通滤波器组.本文介绍了EMD和EEMD分解的基本原理,提出了一种基于EEMD分解的地震信号随机噪声消除的方法 .本文利用含噪信号EEMD分解后其有效信号和随机噪声在IMF中差异分布的特点,给出一种地震信号随机噪声消除的新方法 .

Noise attenuation is a very important step for seismic data processing. It is forward demanded to improve the seismic data signal-to-noise rate in modern seismic exploration. The ensemble empirical mode decomposition （EEMD）, which is the betterment of empirical mode decomposition （EMD）, is proposed as a highly effective timedomain filter. The EEMD is an adaptive band-pass filter, which empirically reduces a signal to several subsignals defined as intrinsic mode function （IMF）. The paper introduces the basic principles of EMD and EEMD, and proposes a noise attenuation method based on EEMD. Using the different characteristics of significant signal and noise in IMFs, the method provides a new solution for random noise attenuation of seismic data.