基于形态成分分析的压缩感知和局部相似性校正的面波压制方法

A surface wave attenuation method based on MCA compressed sensing and local similarity correction

基于压缩感知稀疏反演框架,采用形态成分分析理论(MCA)对含面波的单点采集地震数据进行信噪分离。对于低信噪比的深层弱反射信号来说,面波的稀疏表征中会含有部分残余反射能量。进一步采用局部相似性去噪技术,用分离的面波和压制面波后的反射信息计算相似性因子,提取出稀疏表征面波中的反射波成分,再将与稀疏表征的反射信号加权叠加,获得最优的分离掉面波的结果。典型模型和实际数据测试验证了本文方法的有效性和适用性。结果表明,该处理流程既分离了面波,又提高了反射信号的保真度。

Based on the compressive sensing framework,the morphological component analysis(MCA)theory is utilized to separate the signal and noise of seismic data.For deep weak reflections with low signal-to-noise ratios,the sparse representation of surface waves may stillcontain residual reflections.To address this issue,the paper proposes the use of the local similarity method.The similarity factor is computedusing the separated surface wave and reflections,and the residualeffective reflections are extracted from the separated surface waves.The proposed method is verified using both synthetic and field data,demonstrating its effectiveness in removing surface waves and improving the fidelity of the reflected signal.