低信噪比地震数据高分辨率Seislet加权叠加方法研究  

Seislet weighted stacking method with high resolution for low SNR seismic data

作　　者：彭堂 刘洋[1] 刘财[1] 陈常乐[1] 陈佳伟 PENG Tang;LIU Yang;LIU Cai;CHEN ChangLe;CHEN JiaWei(College of Geo-exploration Science and Technology,Jilin University,Changchun 130026,China)

机构地区：[1]吉林大学地球探测科学与技术学院,长春130026

出　　处：《地球物理学报》2025年第3期1087-1101,共15页Chinese Journal of Geophysics

基　　金：国家科技重大专项课题(2024ZD1002704);国家自然科学基金项目(42474144);吉林省自然科学基金(20200201045JC)联合资助。

摘　　要：地震数据叠加充分利用多次覆盖观测系统的优势,能够有效提高地震数据的信噪比(SNR),然而,经典动校正叠加中的动校拉伸现象会引入低频假象,降低叠加剖面的分辨率,而常规切除处理往往导致浅层覆盖次数减少.Seislet变换是一种特殊针对地震数据特点的压缩技术,可直接沿同相轴方向进行压缩叠加,通过避免动校正过程提高叠加分辨率,但其难以解决低信噪比数据叠加问题.本文提出一种新的VD-Seislet加权叠加(VDSWS)方法,系统构建了Seislet加权叠加的理论框架并设计相应的权系数组合方法,在Seislet叠加方法中引入速度相关(VD)倾角以及高阶Seislet权系数,对低信噪比数据同相轴时空轨迹进行准确的预测,保证加权叠加的高分辨率特征;选取处理道与标准道的局部相似性作为权系数,重新分配数据权重来实现高信噪比叠加功能;对不同深度下反射同相轴的覆盖次数进行动态统计,计算归一化有效权系数来实现Seislet加权叠加结果的振幅保真性.通过理论模型和实测数据的处理,并与传统等权动校正叠加及加权叠加结果对比,验证了VD-Seislet加权叠加技术能够合理地恢复低信噪比数据浅中深部反射层位信息,获得兼顾高分辨率、高信噪比和高保真度的叠加剖面,最终实现低信噪比地震数据的高分辨率叠加.Seismic data stacking makes full use of the advantages of seismic data acquisition with multifold coverage,which can effectively improve the Signal-to-Noise Ratio(SNR)of seismic data.However,the classical Normal Moveout(NMO)stretching will lead to low-frequency artifacts,which can reduce the resolution of stacking profiles,meanwhile,conventional mute processing often leads to a reduction of folds in shallow layers.Seislet transform is a compression technology tailored specifically for seismic data,which can directly compress and stack data along the direction of events.By avoiding the classical NMO process,Seislet transform can improve the resolution of stacking,however,it is difficult to solve the problem of low SNR data stacking.This paper proposes a new VD-Seislet Weighted Stacking(VDSWS)method,which systematically constructs the theoretical frame of Seislet weighted stacking and designs the corresponding combination method of weight coefficients.In the Seislet stacking method,Velocity-Dependent(VD)slopes and high-order Seislet weight coefficients are introduced to accurately predict the time-space trajectories of seismic events in the low SNR data,which ensures the high-resolution features of weighted stacking.Furthermore,the local similarity between the processing traces and the standard traces is selected as another weight coefficients,which will reassign the data weight to achieve high SNR stacking function.We also dynamically calculate the fold number of the reflection events at different depths,and calculate the normalized effective weight coefficients to achieve amplitude fidelity of the Seislet weighted stacking results.Compared with the traditional equal weight NMO stacking and weighted stacking,the VD-Seislet weighted stacking technique can reasonably recover the information of shallow,middle,and deep reflection events in low SNR data,and it obtains the stacking profile with high resolution,high SNR and high fidelity in the synthetic and field tests.Ultimately,the proposed method can achieve high-res

关 键 词：低信噪比 动校拉伸 VD-Seislet加权叠加 高分辨率 

分 类 号：P631[天文地球—地质矿产勘探]