刘洋,,
刘财,
郑植升
吉林大学地球探测科学与技术学院, 长春 130026
基金项目: 国家重点研发计划课题（2018YFC0603701），国家自然科学基金项目（41774127，41974134）和吉林大学高层次科技创新团队建设项目（2017TD-14）资助
详细信息
作者简介: 刘璐, 女, 博士研究生, 主要从事地震数据处理工作.E-mail: liul20@mails.jlu.edu.cn
通讯作者: 刘洋, 男, 教授, 博士生导师, 主要从事非平稳地球物理数据处理和地质-地球物理综合研究等工作.E-mail: yangliu1979@jlu.edu.cn
中图分类号: P631收稿日期:2021-01-18
修回日期:2021-07-20
上线日期:2021-12-10
Iterative seismic random noise suppression method based on compressive sensing
LIU Lu,LIU Yang,,
LIU Cai,
ZHENG ZhiSheng
College of Geo-exploration Science and Technology, Jilin University, Changchun 130026, China
More Information
Corresponding author: LIU Yang,E-mail:yangliu1979@jlu.edu.cn
MSC: P631--> Received Date: 18 January 2021
Revised Date: 20 July 2021
Available Online: 10 December 2021
摘要
摘要:复杂地表和复杂介质条件下，随机噪声往往严重影响着复杂地震信号的信噪比，同时深层地球物理目标探查中弱地震信号总是被随机噪声所掩盖，如何有效地压制随机噪声干扰、恢复有效地震信号仍然是高精度地震勘探中的关键问题.压缩感知理论突破了奈奎斯特采样定理的限制，利用有效地震信号的可压缩性和稀疏性，提供了从不可压缩随机噪声中进行有效信号分离的数据原理.本文系统分析压缩感知框架下地震随机噪声压制的稀疏优化反问题，提出了基于迭代软阈值算法的"采集-重建-修复"方案对该问题进行求解.在实现高度稀疏表征的基础上进行地震数据的压缩感知随机观测，通过迭代反演对有效地震信号进行重构，有效提高复杂地震数据的信噪比，同时，当求解稀疏优化问题时，如果出现正则化项引起重构信号衰减现象，可以匹配除偏对衰减的有效信号进行修复.通过与工业标准f-x预测滤波方法进行比较，理论模型和实际数据处理的结果表明，压缩感知迭代噪声压制方法对复杂地震数据中的随机噪声有较好的压制效果，可以有效恢复出被较强非平稳随机噪声干扰的时空变同相轴信息.
关键词: 随机噪声压制/
压缩感知/
基追踪降噪/
迭代软阈值/
除偏
Abstract:Random noise can affect the signal-to-noise ratio of complex seismic signals, especially under complex surface and subsurface conditions. Meanwhile weak signals are usually covered by random noise in geophysical exploration at depth. How to effectively suppress such random noise interference and recover the effective seismic signal remains a key problem in high-precision seismic exploration. Compressive sensing breaks though the limitation of the Nyquist sampling theory, providing an approach of effective signal separation from uncompressible random noise based on the compressibility and sparsity of effective signals. In this paper, we analyze the inverse problem of sparse optimization that corresponds to seismic random noise suppression within the compressive sensing framework. An "acquisition-reconstruction-repair" workflow based on an iterative soft threshold is proposed to resolve the inverse problem. With the highly sparse representation of seismic signals, one can obtain random observation to seismic data based on compressive sensing and rebuild only useful signals by using iterative inversion, which improves the signal-to-noise ratio of complex seismic signals. Furthermore, we use a debiasing method to reduce the attenuation of signals due to the presence of the regularization term. This debiasing step improves the proposed method by recovering the attenuated amplitude of effective signals. Compared with the industrial standard FXDECON method, the proposed method shows a better denoising result in model and field data examples and can reasonably recover nonstationary events from noisy data.
Key words:Random noise suppression/
Compressive sensing/
Basis pursuit denoising/
Iterative soft threshold/
Debiasing
PDF全文下载地址:
http://www.geophy.cn/data/article/export-pdf?id=f0a58dce-8ff1-4016-b4aa-1d58b0e02b76
