一种曲波域蒙特卡罗阈值去噪算法  

An Algorithm of Monte Carlo Threshold Denoising Based on Curvelet Domain

作　　者：陈思雅[1] 谢凯[1] 阮宁君[1] 张龙[1] 龚康奕 沈政春 李纪成[1] 

机构地区：[1]长江大学电子信息学院,湖北荆州434023

出　　处：《长江大学学报（自科版）（上旬）》2016年第5期34-38,3,共5页JOURNAL OF YANGTZE UNIVERSITY (NATURAL SCIENCE EDITION) SCI ＆ ENG

基　　金：中国石油创新基金项目(2010D-5006-0304);湖北省大学生创新创业训练项目(2014012)

摘　　要：为促进油气勘探发展,有效识别地震有用信号,提出了一种曲波域蒙特卡罗阈值去噪算法。该算法利用曲波变换的多尺度与多方向性的特点,结合蒙特卡罗阈值滤波来去除随机噪声,再利用循环平移法抑制吉布斯现象,消除人为干扰,增强去噪效果。用该算法对合成地震数据与实际地震剖面分别进行去噪试验,并与小波变换和曲波循环平移算法的试验结果进行对比,证明了该算法能有效地保持有效波信息,增加了弱反射信号能量,很好识别有效信号,优于小波变换去噪算法与原始曲波变换去噪算法。In order to promote the development of oil and gas exploration and effectively identify the useful seismic signals,an algorithm of Monte Carlo threshold denoising based on curvelet transform is put forward. Combined with the Monte Carlo threshold filtering,the characteristics of multi-scales and multi-directions of curvelet transform of the algorithm are applied. For the elimination of human disturbance,circular shift is used to suppress Gibbs phenomenon,thus the denoising effect is enhanced. The algorithm is applied for the denoising experiment of the simulation data and the actual seismic profile. Both of the experimental results are compared with the original curvelet algorithm and wavelet transform,The comparison proves that the algorithm can effectively maintain the effective wave information,increase the energy of weak reflection signals,it can be used for effective identification of signals,it is better than that of the wavelet transform and the original curvelet algorithm.

关 键 词：循环平移 曲波变换 随机噪声 蒙特卡罗阈值 

分 类 号：P631.4[天文地球—地质矿产勘探]