一种基于负熵的Bussgang地震盲反褶积方法
|
摘要
独立分量分析作为一种新兴的信号处理工具,许多学者对其在地震信号处理领域中的应用进行了初步探讨。本文在Santamaria等提出的自适应高斯混合盲反褶积的基础上,结合独立分量分析中Bussgang类反褶积算法,通过引入约束项,对基于负熵准则的目标函数进行了改进,构造出新的非线性算子;同时采用修正的期望最大化(EM)算法优化反射系数高斯混合模型参数,成功地实现了地震信号的盲反褶积。数值模拟和实际资料处理的结果表明,该方法能较好地适应非最小相位系统,得到原始反射系数的最优估计,而且算法收敛快、精度高,是提高地震资料分辨率的有效手段。
Analysis of independent components,as a new signal processing tool, its application in seismic signal processing region has been preliminarily discussed by many scholars. Based on adaptive Gauss hybrid blind deconvolution presented by Mr. San-tamaria et al and in combination with Bussagang deconvolution algorithm in analysis of independent components, the paper improved negative entropy rule-based objective function by introducing constraint item and constructed new non-linear operator; meanwhile, using the updated expected maximization (EM) algorithm to optimize the parameters of Gauss hybrid model of reflection coefficient , successfully implementing the blind deconvolution of seismic signals. The results of numeric simulation and practical data processing showed that the approach could be better suitable for non-minimum phase system,obtain the optimum evaluation of raw reflection coefficient and be characteristic of fast convergence and high precision, which is effective tool for improving the resolution of seismic data.
Analysis of independent components,as a new signal processing tool, its application in seismic signal processing region has been preliminarily discussed by many scholars. Based on adaptive Gauss hybrid blind deconvolution presented by Mr. San-tamaria et al and in combination with Bussagang deconvolution algorithm in analysis of independent components, the paper improved negative entropy rule-based objective function by introducing constraint item and constructed new non-linear operator; meanwhile, using the updated expected maximization (EM) algorithm to optimize the parameters of Gauss hybrid model of reflection coefficient , successfully implementing the blind deconvolution of seismic signals. The results of numeric simulation and practical data processing showed that the approach could be better suitable for non-minimum phase system,obtain the optimum evaluation of raw reflection coefficient and be characteristic of fast convergence and high precision, which is effective tool for improving the resolution of seismic data.
引文
[1]刘喜武,刘洪,李幼铭.独立分量分析及其在地震信息处理中应用初探.地球物理学进展,2003,18(1):90～96
[2]Ignacio Santamaria et al.Deconvolution of seismic data using adaptive gaussian mixtures.IEEE Trans on Geoscience and Remote Sensing,1999,37(3):855～859
[3]Canadas G.A mathematical framework for blind deconvolutioninversion problems.SEGInt'l Exposo-tion and72nd Annual Meeting,2002
[4]Anthony Larue et al.Blind frequency deconvoluti-on:A new approach using mutual information rate.SEGInt'l Exposition and74th Annual Meeting,2004,10
[5]Bazi w E,Ulrych T J.Principle phase decomposition:a newconcept in blind seismic deconvolution.IEEE Trans on Geoscience and Remote Sensing,2006,44(8):2271～2281
[6]刘喜武,刘洪.地震盲反褶积综述.地球物理学进展,2003,18(2):203～209
[7]杨福生,洪波著.独立分量分析的原理及应用.北京:清华大学出版社,2006
[8]Walden A T.Non-Gaussianreflectivity,entropy,and de-convolution.Geophysics,1985,50(12):2862～2888
[9]李春峰.从多尺度反射系数的角度评估反褶积模型的适用性.勘探地球物理进展,2005,28(4):250～257
[10]Tenorio L.Modeling non-Gaussian reflectivities:gen-eralizing Wiener-Levinson deconvolution.Geophysics,2001,66(6):1913～1920
[11]Dempster P et al.Maxi mumlikelihood fromincomplete data via the EM algorithm.Royal Stat Soc,1977,39(1):1～38
[2]Ignacio Santamaria et al.Deconvolution of seismic data using adaptive gaussian mixtures.IEEE Trans on Geoscience and Remote Sensing,1999,37(3):855～859
[3]Canadas G.A mathematical framework for blind deconvolutioninversion problems.SEGInt'l Exposo-tion and72nd Annual Meeting,2002
[4]Anthony Larue et al.Blind frequency deconvoluti-on:A new approach using mutual information rate.SEGInt'l Exposition and74th Annual Meeting,2004,10
[5]Bazi w E,Ulrych T J.Principle phase decomposition:a newconcept in blind seismic deconvolution.IEEE Trans on Geoscience and Remote Sensing,2006,44(8):2271～2281
[6]刘喜武,刘洪.地震盲反褶积综述.地球物理学进展,2003,18(2):203～209
[7]杨福生,洪波著.独立分量分析的原理及应用.北京:清华大学出版社,2006
[8]Walden A T.Non-Gaussianreflectivity,entropy,and de-convolution.Geophysics,1985,50(12):2862～2888
[9]李春峰.从多尺度反射系数的角度评估反褶积模型的适用性.勘探地球物理进展,2005,28(4):250～257
[10]Tenorio L.Modeling non-Gaussian reflectivities:gen-eralizing Wiener-Levinson deconvolution.Geophysics,2001,66(6):1913～1920
[11]Dempster P et al.Maxi mumlikelihood fromincomplete data via the EM algorithm.Royal Stat Soc,1977,39(1):1～38
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心