SPGL1算法在地震资料随机噪声压制中的应用  

Application of SPGL1 in Random Noise Suppression of Seismic Data

作　　者：张博伦[1] 王林飞[1] 蔡正辉 黄云笛 ZHANG Bo-Lun WANG Lin-Fei CAI Zheng-Hui HUANG Yun-Di(The Key Lab of Submarine Geosciences and Prospecting Techniques, Ministry of Education, Ocean University of China College of Law and Politics, Ocean University of China, Qingdao 266100 ,China)

机构地区：[1]中国海洋大学海底科学与探测技术教育部重点实验室,山东青岛266100 [2]中国海洋大学法政学院,山东青岛266100

出　　处：《中国海洋大学学报（自然科学版）》2016年第12期81-86,共6页Periodical of Ocean University of China

基　　金：国家自然科学基金青年科学基金项目(41204087);教育部高等学校博士学科点专项科研基金新教师类课题(20120132120030)资助~~

摘　　要：为了提高地震资料的信噪比,本文基于稀疏表示理论,将SPGL1算法应用到地震资料随机噪声的压制中,在构建过完备原子库时,以Morlet子波为基础,应用Hilbert变换确定待分解数据的瞬时属性,并将此作为过完备原子库时频参数选取范围,提高计算效率。通过合成数据和实际资料的处理分析,证明了该方法具有良好的去噪效果。In seismic exploration,random noise seriously distorts and interferes with seismic signal.In this study,the SPGL1 algorithm of seismic data denoising is adopted based on the sparse representation,which can efficiently suppress random noise and certainly enhance the SNR.During the creation of overcomplete dictionary of atoms,instantaneous attributes of undecomposed data are ascertained by HilbertTransform with the help of Morlet wavelet,which are regarded as the range of time-frequency parameters of the over-complete dictionary of atoms.Both theoretical analysis and actual results show that,comparison with wavelet threshold denoising method,the proposed method is of strong denoising ability and fast processing rate.

关 键 词：SPGL1 稀疏分解 去噪 Morlet子波 过完备原子库 HILBERT变换 

分 类 号：P631.4[天文地球—地质矿产勘探]