基于奇异谱分析的联合去噪及规则化方法

Simultaneousseismic data de-noising and regularizationmethod based on singular spectrum analysis

在地震数据处理中,噪音压制和数据规则化研究直接影响后期的地震处理及解释效果.本文提出一种基于奇异谱分析的联合去噪及规则化方法,在迭代时自动在地震道缺失的位置进行插值,在含有地震数据的位置压制噪声,并通过分步插值改善数据相干性.通过合成地震记录和实际资料的联合去噪及规则化处理结果表明:联合方法能够在补全地震道的同时有效地压制噪声.

Good de-noising and reconstruction method leads toa better processing and interpretation result in seismic data processing.The paper presents a singular spectrum analysis algorithm that permits simultaneous noise attenuation and reconstruction of seismic data.The method can do interpolation in the position with absence of seismic traces as well as noise attenuation in the position containing seismic traces automatically through iteration.A step interpolation is applied to improve seismic data coherence.Numerical tests of 2D synthetic and field data show that the proposedmethod is powerful in complementing the seismic tracesas well asattenuating random noise at the same time.