改进的FastICA算法在地震数据去噪中的研究  

Research on the Improved FastICA Algorithm in Seismic Data Denoising

作　　者：秦飞龙 颜文勇[1] 刘剑[3] 成和平[1] 张永峰 王科[1] QIN Feilong;YAN Wenyong;LIU Jian;CHENG Heping;ZHANG Yongfeng;WANG Ke(School of Big Data and Artificial Intelligence,Chengdu Technological University,Chengdu 611730,China;College of Mathematics and Science,University of Electronic Science and Technology of China,Chengdu 611731,China;School of Automobile and Communications,Chengdu Technological University,Chengdu 611730,China)

机构地区：[1]成都工业学院大数据与人工智能学院,成都611730 [2]电子科技大学数学科学学院,成都611731 [3]成都工业学院汽车与交通学院,成都611730

出　　处：《重庆理工大学学报（自然科学）》2020年第6期182-188,共7页Journal of Chongqing University of Technology：Natural Science

基　　金：中国地质调查局项目(12120114002001);四川省科技厅计划项目(2019YJ0375);成都工业学院博士基金项目(2018RC022);数学地质四川省重点实验室开放基金资助项目(scsxdz2018yb03);成都工业学院实验室开放基金项目(2018ZH10)。

摘　　要：为了有效去除地震数据的干扰噪声,采用FastICA算法进行降噪处理。地震数据的有效信号和噪声是非线性混合的,不满足FastICA算法使用的线性混合条件,提出了一种改进的FastICA算法进行地震数据降噪。该方法首先利用小波变换将地震数据分解到不同尺度空间上,然后设计了一种新的阈值降噪函数,并将FastICA和小波变换进行有效结合,对不同尺度空间上的噪声进行消除,通过小波重构恢复有效信号,利用仿真实验确定了新算法的分解层数为3。将改进的方法进行实际地震数据降噪处理,结果证明改进的方法能有效去除地震数据的各类干扰噪声,通过信噪比得出改进的FastICA较原始FastICA降噪效果更理想。In order to remove the interference noise of seismic data effectively,the paper used FastICA algorithm to seismic data denoising.However,the mixing of effective signal and noise of seismic data is non-linear characteristics,which does not satisfy the condition that the data is linear mixture of FastICA algorithm.Therefore,this paper proposes an improved FastICA algorithm to noise reduction of seismic data to solve defects of FastICA algorithm in seismic data denoising.Firstly,the original seismic data can be decomposed into different scales by using wavelet transform,Secondly,this paper proposed a new threshold method and combined the FastICA algorithm and wavelet transform to remove noise of seismic data.Then,the effective signal is reconstructed by wavelet transform.Through simulation experiments,it can be obtained the best decomposition levels is 3 of improved FastICA.Finally,the improved FastICA algorithm is applied to denoising of the actual seismic data.The results show that the improved FastICA algorithm can effectively remove all kinds of interference noise of seismic data.The improved method can eliminate noise of seismic data effectively,which is evaluated by signal-to-noise ratio(SNR).

关 键 词：改进的FastICA 小波变换 分解层数 地震数据 去噪 

分 类 号：P631.443[天文地球—地质矿产勘探]