摘要
奇异值分解(SVD)滤波是利用地震信号在横向上的相干性差异来实现地震波场的分离与去噪。由于P-P波和P-S波在传播特性、视速度和相干性上存在差异,本文应用视速度信息,通过两次正常时差校正(NMO),分别把P-P波和P-S波校平,使之在横向上达到最佳相干性,然后分别两次通过SVD,提取目标信号的奇异值重构信号,从而实现地震纵、横波波场分离与去噪。
As one of seismic data processing approaches,singular value decomposition(SVD) filter uses the lateral coherence difference of seismic signals to achieve wavefield separation and denoising.However,the propagation characteristics,apparent velocities and coherences are quite different between P-P wave and P-S wave.In this paper,a new idea of SVD application is proposed.Based on apparent velocities information,we first apply normal moveout(NMO) processing respectively on P-P wave and P-S wave to align P-P wave and P-S wave into the best horizontal coherence.Then we extract reconstructed signals of singular values from target signals by SVD separately on P-P wave and P-S wave,Finally P-P wave and P-S wave separations as well as noise attenuation are obtained.
出处
《石油地球物理勘探》
EI
CSCD
北大核心
2012年第5期690-697,844+675,共8页
Oil Geophysical Prospecting
基金
国家863计划项目(2007AA06Z103)
中国石油科技创新基金(2010D-5006-0303)
陕西省自然科学基础研究计划项目(2011JQ5006)
陕西省教育厅专项科研计划项目(11JK0738)
中国地质大学"构造与油气资源教育部重点实验室"开放课题(TPR-2010-28)联合资助
