基于结构化匹配追踪的地震资料去噪技术研究  

Seismic data de-noising technology based on structured matching pursuit

作　　者：余为维 冯磊[1] 

机构地区：[1]河南理工大学资源环境学院,焦作454000

出　　处：《地球物理学进展》2015年第1期304-311,共8页Progress in Geophysics

基　　金：国家自然科学基金项目(41202084);高等学校博士学科点专项科研基金新教师类课题(20114116120003)联合资助

摘　　要：地震数据往往受噪声干扰,影响有效反射波的识别,因此提高地震资料信噪比在数据处理中尤为重要,我国在此方面研究已经取得进展.结构化匹配追踪算法是一种采用原子字典分组式及树状结构搜索的全局结构化算法,具有运算效率高、速度快的特点.本文在分析结构化匹配追踪算法的基础上提出基于该算法的地震资料去噪技术,根据反射波信号与噪声信号的频段差异,对分解得到的时频原子设置阈值频段,去除含噪成分,从而达到去噪目的.与常规去噪技术相比,该技术精准、快速地分解地震反射波信号与噪声信号,最大限度地减小对地震反射波信号的损害.Seismic data is often affected by noise,affecting the recognition of effective reflection wave.Thus it＇s particularly important to improve signal-noise ratio of seismic data in data processing.Structured matching pursuit algorithm is a global structured algorithm by adopting packet-mode atom dictionary and tree-structured search,which is characterized by high efficiency and speed.By analyzing structured matching pursuit algorithm,this paper presents a de-noising technology of seismic data based on this algorithm.According to the difference of frequency range between reflection wave signal and noise signal,we set the threshold frequency range on timefrequency atoms after decomposition,remove the noisy component,and thus achieve the purpose of de-noising.Compared with conventional de-noising technology,this technology is accurate and rapid in decomposition of seismic reflection wave signal and noise signal,minimizing the damage of seismic reflection wave signal.

关 键 词：信噪比 结构化 匹配追踪 去噪 地震资料 

分 类 号：P631[天文地球—地质矿产勘探]