多次波压制的研究现状与进展  

Research status and progress in multiple elimination

作　　者：宋家文[1] D.J.Verschuur 陈小宏[1] 

机构地区：[1]中国石油大学(北京)油气资源与探测国家重点实验室,北京102249 [2]荷兰代尔夫特理工大学

出　　处：《地球物理学进展》2014年第1期240-247,共8页Progress in Geophysics

基　　金：国家科技重大专项(2011ZX05023-005-005);国家自然科学基金项目(U1262207)联合资助

摘　　要：目前地震资料处理技术主要利用一次波对地下地质体进行成像或反演,多次波的存在会严重影响地震成像、反演与解释结果.因此,多次波通常作为噪声在地震数据叠前处理阶段进行压制.为了合理选择不同的多次波压制方法,提高地震资料信噪比和保真度,本文系统综述了多次波压制的国内外研究现状,简要介绍了滤波法、预测相减法、稀疏反演法的基本原理,并围绕理论假设条件与实际数据要求,分析比较了各种方法的适用性与优缺点.基于波动方程的预测相减法克服了滤波法关于地震信号统计特征的假设,增强了对复杂地震数据的适应性,但最小二乘自适应相减会损害有效信号.稀疏反演法采用全波形反演的方法估计一次波与多次波,避免了自适应相减步骤,能够更加保真地恢复有效信号.本文通过实际算例重点阐述了稀疏反演法相对于反馈迭代法在多次波压制中的突破性进展,同时指明了稀疏反演法存在的问题并展望了未来的发展方向.Most of the current seismic data processing techniques make use of primaries for imaging or inversion, the existence of multiples has a severe impact on seismic imaging, inversion and interpretation results. Therefore, multiples are usually treated as noise which need to be removed in a preprocessing step. For improving seismic data S/N ratio and fidelity by making a rational choice between distinctive multiple elimination methods, we systematically review the current status and progress of multiple removal techniques, and briefly introduce three categories of demultiple methods, including filtering methods, wave field prediction-and-subtraction methods as well as the recently developed EPSI （Estimation of primaries by sparse inversion）, furthermore, we analyze the applicability together with pros and cons of the methods in terms of prior assumptions and practical prerequisites. The prediction-and-subtraction methods based on wave equation can adapt to complex seismic data because they overcome the limitations of filtering methods introduced by assuming that primaries and multiples have different statistical characteristics, but least square subtraction potentially distorts primaries during the adaptive subtraction. In contrast, EPSI estimates primaries and multiples in a full waveform inversion process where adaptive subtraction is avoided and primaries are better preserved. This paper highlights the breakthrough of EPSI over feedback methods on multiple elimination through two field datasets, meanwhile, it points out the limitations and prospects of EPSI.

关 键 词：多次波压制 研究进展 滤波法 预测相减法 稀疏反 演法 

分 类 号：P631[天文地球—地质矿产勘探]