基于CEEMDAN和小波阈值的地震信号去噪方法

Seismic Signal Denoising Method based on CEEMDAN and Wavelet Threshold

采用自适应噪声完备集合经验模态分解(CEEMDAN)方法,对地震数据进行分析预处理.将地震信号分解成一系列具有不同特征时间尺度的固有模态函数IMF,对各IMF分量进行自相关计算,筛选出含噪分量,使用小波硬阈值对含噪IMF分量进行去噪处理.将去噪后分量与无噪分量进行重构,从而降低噪声干扰,有效提取特征成分,发现有用的地层信息,获取高质量的数据,为地震监测服务.

In this paper,the adaptive noise complete set empirical mode decomposition(CEEMDAN)method is used to analyze and preprocess seismic data,and decompose the seismic signal into a series of intrinsic mode functions IMF with different characteristic time scales;Perform autocorrelation calculation on each IMF component,filter out noisy components,and use wavelet hard threshold to denoise noisy IMF components;Reconstruct the denoised and non noisy components to reduce noise interference,effectively extract feature components,discover useful stratigraphic information,obtain high-quality data,and better serve seismic monitoring.