摘要
本文研究了形态分量分析(MCA)算法含噪地震数据的重构问题.针对传统形态分量分析算法在含噪地震数据重构精度不足问题,提出组合方法以提高重构精度.给定随机缺失坏道地震数据,加入不同比率的高斯噪声,研究数据的重构效果.采用离散余弦、离散曲波和离散小波作为字典集合,对数据形态分量进行稀疏表达,并迭代优化各分量,以优化分量进行重构.分析不同随机缺失比率、不同信噪比模型仿真模拟结果.实际地震数据处理结果表明,组合MCA算法对随机缺失含噪地震数据具有较好的重构效果,并具有明显地随机噪声压制效果.
In this paper,the reconstruction of seismic data with noise by Morphological Component Analysis(MCA)algorithm is studied.In order to improve the reconstruction accuracy of seismic data with noise,a combination method is proposed in this paper to solve the problem of insufficient reconstruction accuracy of traditional morphological component analysis algorithm.For a given random seismic data with bad track missing,Gaussian noise of different ratio is added to study the reconstruction effect of data.In this paper,discrete cosine,discrete curved wave and discrete wavelet are used as dictionary sets to sparsely express the morphological components of data,and each component is iteratively optimized to reconstruct the optimized components.In the combined morphological component algorithm,the key feature components are collected intensively and the redundant components are selected sparingly to achieve the feature optimization and integrity.The optimized component reconstruction can reduce the noise interference and improve the accuracy of seismic data reconstruction.In this paper,the combined algorithm is used to effectively separate the noise component from the in-phase axis component,which can not only reduce the noise interference in the effective component reconstruction,but also improve the stability of the morphological component algorithm to achieve high resolution reconstruction.In this paper,the simulation results of different random miss ratios and different SNR models are analyzed in detail.The actual seismic data processing results show that the MCA algorithm has a better reconstruction effect on the random seismic data with missing noise,with continuous in-phase axis,energy balance,high resolution,and obvious random noise suppression effect.
作者
李思翰
刘洪林
LI SiHan;LIU HongLin(Northeast Petroleum University School of Earth Sciences,Daqing 163318,China)
出处
《地球物理学进展》
CSCD
北大核心
2021年第4期1547-1553,共7页
Progress in Geophysics
关键词
离散曲波变换
离散小波变换
观测矩阵
绝对中值偏差
数据重建
Discrete cruvelet transform
Discrete wavelet transform
Observation matrix
Absolute median deviation
Data reconstruction
