论文详情

曲波变换及全变差最小化技术联合去噪

石油物探

2012年 51卷第No. 4-期

阅读：69

Title

Joint noise attenuation technique based on curvelet transform and total variation minimization

单位

中国石油化工股份有限公司江苏油田分公司物探技术研究院，江苏南京210046

Organization

Yu Hua, Geophysical Research Institute, Jiangsu Oilfield Branch Company, SINOPEC, Nanjing 210046, China

摘要

为了更好地发挥曲波变换在地震资料去噪处理方面的优势，同时克服其固有缺陷，在地震资

料去噪处理中应用了基于曲波变换和全变差最小化技术的联合衰减随机噪声技术。该技术在应用

曲波变换时，对选取的阈值使用非线性阈值法，同时对小系数采用全变差最小化技术。理论模型

与实际地震资料的测试结果表明，该技术不仅可以很好地衰减随机噪声，有效地克服单独使用曲

波变换带来的伪曲线现象，同时较好地保护了有效信号。

Abstract

In order to take advantage of curvelet transform in seismic data denosing and overcome its inherent

shortcomings, we applied the joint random noise attenuation technique based on curvelet transform and

total variation minimization on seismic data. This technique uses the nonlinear-thresholding approach to

realize the curvelet transform and adopts total variation minimization for small coefficients. Theoretical

model and real data processing results show that this technique can greatly attenuate random noise,

effectively overcome aliased curves when merely using curvelet transform, and preserve the significant

signal in seismic data as well.

关键词：

曲波变换; 全变差; 随机噪声; 非线性阈值法;

Keywords:

curvelet transform; total variation; random noise; nonlinear-thresholding approach;

基金项目

DOI

10.3969/j.issn.1000-1441.2012.04.005