地震信号分析中具有波形形态信息约束的超完备字典构造方法及其应用

项目名称： 地震信号分析中具有波形形态信息约束的超完备字典构造方法及其应用

项目编号： No.41274125

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 天文学、地球科学

项目作者： 陈文超

作者单位： 西安交通大学

项目金额： 80万元

中文摘要： 利用形态分量分析（Morphological Component Analysis, MCA）理论，根据河道等地层横向非均质体与稳定沉积地层对应地震波形形态特征的差异性，研究适合于地震信号处理的MCA的超完备字典构造方法；根据地震信号的特点，选择或构造基于MCA稀疏分解的评价方法；选择或构造适合海量高维地震信号处理和分析的快速MCA求解算法；研究利用MCA对3D地震资料中河道等地层横向非均质体的精细刻画方法。该项研究具有重要的实际意义，也将发展稀疏表示理论与方法。

中文关键词： 稀疏；波形字典；形态成分分析；超完备字典；块坐标松弛算法

英文摘要： This project utilizes the Morphological Component Analysis (MCA) theory to study the construction method of overcomplete dictionaries for MCA that are suitable for seismic signal processing. The morphological diversity of the waveforms between lateral heterogeneity of seismic strata such as channels and stable sedimentary strata is especially studied. According to the features of seismic signals, this project selects or constructs the evaluation rules based on MCA sparse decomposition and also the fast numerical solution method for processing and analyzing massive high dimensional seismic signal. Furthermore, by utilizing MCA, this project studies precise characterization method for lateral heterogeneity of seismic strata such as channels in 3D seismic data. This study has important practical significance and may develop the theory and method of sparse signal representation.

英文关键词： sparse；waveform dictionary；Morphological Component Analysis；overcomplete dictionary；Block-Coordinate Relaxation