摘要
多孔算法克服了Mallat算法二抽样的问题,具有平移不变性,而且每个尺度下的数据长度都和原始数据长度一样,便于对每个尺度下的细节和概貌进行频谱分析。本文采用多孔算法进行小波变换,对每个尺度下的细节进行频谱分析,继而确定小波变换中的分解层数,并引入一种新的阈值函数Garrote函数做阈值,经过理论合成信号和实际地震数据测试表明,基于多孔算法的小波去噪是有效的。
A＇trous algorithm solves the problem of the two sampling algorithm with Mallat and it has the feature that the translation is invariant and that each scale length of the data has the same as the original data＇s so as to make spectrum analysis on them. In this paper, a＇trous wavelet transform algorithm is used, the details of each scale spectrum is analyzed, and then the wavelet transform decomposition level is determined, and besides, a new threshold function as function Garrote threshold is introduced. Finally, the practical seismic data show that wavelet denoising based on a＇trous algorithm is effective.
出处
《工程地球物理学报》
2008年第4期444-448,共5页
Chinese Journal of Engineering Geophysics
基金
武汉市科技局项目(编号:20061009134-07)资助
关键词
小波变换
小波去噪
多孔算法
Garrote函数
wavelet transform
algorithm wavelet denoising
A＇trous algorithm
Garrote function
