相位编码多次波逆时偏移成像

Phase-encoding-based reverse time migration of multiples

随着勘探地质体复杂度日益增大,反射波成像面临诸多挑战,作为有潜力提升成像精度的重要补充或替代手段的多次波成像愈发受到重视.然而,串扰噪声的产生阻碍了多次波成像的实际应用.近年来,作为压制串扰噪声的有效技术之一,可控阶多次波成像取得了较大研究进展,但因其需重复计算各阶多次波成像结果,计算成本昂贵.为此,本文引入随机相位编码技术,并与多次波分阶思想结合,提出相位编码多次波逆时偏移方法:首先,对各阶多次波进行随机相位编码(含随机时间延迟与极性反转);其次,叠加编码后的各阶多次波,产生多次波超道集;最后,以编码后0至(N-1)阶多次波组成的超道集为虚拟震源进行正向延拓,同时反传编码后1至N阶多次波所组成的超道集,并进行互相关,得到各阶多次波联合成像结果.本文所提方法能够同步使用各阶多次波,实现各阶多次波联合成像,避免各阶多次波单独进行成像,从而成倍提升计算效率.本文用两套模拟数据与一套实际数据算例对所提方法进行了测试,成像结果验证了该方法的可行性、有效性与应用前景.

Reverse time migration of multiples(RTMM)can supply migration results with wider illumination and broader wavenumber spectra than conventional reverse time migration(RTM)of primaries.However,crosstalk artifacts caused by interferences between unrelated multiples severely degrade the imaging quality of multiples.To address the crosstalk problems,RTM using controlled-order multiples(RTM-CM)was proposed.In RTM-CM,different-order multiples are first isolated using the multiples decomposition strategy,followed by migration procedures of consecutive-order multiples.By exploiting RTM-CM,most energetic crosstalks can be significantly eliminated.However,N individual migrations need to be conducted for staking in RTM-CM,thereby causing an N-fold increase in computational costs,in which N is the maximum order of considered multiples.To reduce computation costs and inherit the benefits of controlled-order multiple imaging,we developed a phase-encoding-based migration approach for multiples.This method first phase encodes different-order multiples by randomly modifying time shifts and polarity reversals.And then the encoded multiples with orders of 0 to N-1 are stacked to form the forward-propagating supergathers,whereas the backward-propagating supergathers comprise encoded multiples with orders of 1 to N.Finally,by using supergathers,the proposed approach can simultaneously accomplish migrations of all isolated-order multiples,which significantly improves computational efficiency.Numerical examples on two synthetic and one field datasets corroborate that the proposed approach can considerably enhance imaging quality by suppressing crosstalks and reducing computational costs.