摘要
独立成分分析方法(ICA)是近几年才发展起来的一种新的基于信号高阶统计特性的分析方法。对于相互统计独立的信号源经线性组合而产生的混合信号,采用ICA能从其中分离出各独立的信号分量。在地震勘探中,地震道信号和随机噪音通常具备非高斯信号的特点,统计性相互独立。在对地震数据和随机噪音信号特点进行分析的基础上,采用基于最小互信息的快速ICA方法实现地震数据去噪。模型分析和实例计算表明,该方法具有一定的可行性。
Independent component analysis(ICA)is a recently devel- oped method based on higher order statistical analysis.ICA can separate signals that are statistically independent but linearly mixed.Seismic signals and random noises are generally non- gaussain and are statistically independent.Based on an analysis of the characteristics of seismic signals and random noises,the author adopted fast ICA that are based on minimum mutual in- formation to eliminate random noises from seismic data.The feasibility of the method is verified with model and real data.
出处
《勘探地球物理进展》
2007年第1期30-32,68,共4页
Progress in Exploration Geophysics
基金
四川省重点学科建设项目（SZD0414）资助.
关键词
独立成分分析
快速算法
地震数据
随机噪音
independent component analysis
fast algorithm
seismic data
random noise
