摘要
针对大地电磁信号具有非线性、非平稳和非最小相位的特点,提出了一种基于经验模态分解法结合小波变换的联合信号去噪方式,将时间序列信号通过经验模态分解,利用连续均方误差准则确定原始信号能量转折点,进而再使用小波阈值去噪法对剩余固有模态函数分量进行去噪,最后重构出消噪信号。通过对实测信号处理前后结果的对比,表明了本方法能够有效地应用于信号时域去噪。
In consideration of the nonlinear,non-stationary and non-minimum phase characteristics of the magnetotelluric signal,this paper has proposed a denoising method based on empirical mode decomposition and wavelet transform,in which the time series signal is decomposed by EMD transform,the continuous mean square error criteria are used to determine a turning point in the original signal energy,then the wavelet threshold denoising method is employed to denoise the residual amount of intrinsic mode function,and finally the signal is reconstructed.Tests of signal processing show that the method can be effectively applied to signal denoising process.
出处
《物探与化探》
CAS
CSCD
2012年第3期452-456,共5页
Geophysical and Geochemical Exploration
基金
国家自然科学基金项目"扬子地台西缘深部地质结构与油气赋存背景研究"(40839909)
国家自然科学基金项目"汶川大地震余震综合监测与龙门山前山断裂带活动性分析"(40841012)
国家自然科学基金项目"龙门山中段地壳与上地幔电性结构及其动力学特征"(40674035)
关键词
长周期大地电磁测深
经验模态分解
小波分析
时间域去噪
long-period magnetotelluric sounding
EMD
wavelet analysis
noise suppression in time domain
