摘要
本报告综述了近年来发展起来的真振幅单程波方程偏移理论,对各种基于单程波方程的偏移振幅作出了系统的分析介绍.为了得到正确的偏移振幅,叠后偏移之前必须进行球面扩散校正.但是,作为叠后相位移偏移算法的推广,Dubrulle氏的共炮检距偏移除了零炮检距和平层两种特例外,它不能给出正确的偏移振幅.通过分析,我们发现传统的单程波单炮偏移不是保振幅算法.利用真振幅单程波方程分解和校正地表的初值条件,我们可以将常规偏移方法改造为真振幅算法并且证明在高频渐近的意义下,新方法的偏移振幅等价于Kirchhoff反演的结果.进一步研究发现可以利用单平方根算子和双平方根算子输出真振幅共反射角剖面.我们的分析指出,这时正确的成像条件为乘积pUpD*.与真振幅共炮偏移所需的商型成像条件pU/pD相比,共反射角偏移的计算稳定性大为改善并且在实际地震资料处理中有重要的应用.
In this technical report, we review the recent development of true amplitude one-way wave equation migration and summarize the migrated amplitude performance from different one-way wave equations. To produce correct amplitude from postack phase-shift migration, we have to apply accurate geometrical spreading compensation in preprocessing. Although Dubrulle's common-offset phase-shift migration gives correct imaging position, it is not a true amplitude migration except for zero offset or flat reflectors. We have shown that the conventional common-shot one-way wave equation migration does not produce correct migration amplitude. Base on the true amplitude one-way wave equations and the surface boundary condition corrections, we developed true amplitude common-shot migration and have proved that it produces an inversion output that agrees asymptotically to Kirchhoff inversion. In the next step, we proposed a way to obtain true amplitude common-angle image gathers from both single-square-root and double-square-root wave equation migrations. Our analysis indicates that the PUPD imaging condition is a good candidate to produce true amplitude angle domain common imaging gathers. Compared to the PU/PD imaging condition which is required for true amplitude common-shot migration, the new approach suggests a more stable way of doing inversion and is more attractive to the real seismic data processing.
出处
《地球物理学报》
SCIE
EI
CAS
CSCD
北大核心
2006年第5期1410-1430,共21页
Chinese Journal of Geophysics
关键词
真振幅偏移
单程波方程
共炮偏移
共反射角偏移
单平方根算子
双平方根算子
True amplitude migration, One-way wave equaiton, Common-shot migration, Common-angle migration, Single-square-root operator, Double-square-root operator