相空间重构单通道ICA算法及其在变形监测数据去噪中的应用  

Single channel ICA algorithm based on phase space reconstruction and its application in deformation monitoring data de-noising

作　　者：黄大伟[1,2] 戴吾蛟[1,2] 刘斌[1,2] 

机构地区：[1]中南大学测绘与遥感科学系,湖南长沙410083 [2]湖南省精密工程测量与形变灾害监测重点实验室,湖南长沙410083

出　　处：《测绘工程》2014年第6期38-42,共5页Engineering of Surveying and Mapping

基　　金：国家自然科学基金资助项目(41074004);中南大学研究生自主探索创新基金项目(2012zzts073)

摘　　要：常用的去噪方法需对数据序列进行多尺度分解实现去噪,而未考虑信号本身的关联和特征。相空间重构单通道ICA算法依据相空间重构将一维信号扩展到高维相空间,从信号本身获取其隐含的动力学特征,从而利用独立分量分析分离有用信号和噪声。实验证明相空间重构的单通道ICA算法能够实现信号去噪,其去噪能力与多尺度分解去噪方法相当。此外,算法无须多尺度分解,自适应强,且适用于变形监测信号去噪。For general multi-scale decomposition de-noising methods, the internal correlation andcharacteristic of signal is not taken into consideration. The single channel ICA algorithm based on phasespace reconstruction expands a single channel to multi-dimensional phase space to get the implied dynamiccharacteristic, and separate the useful signal and noise by independent component analysis. Theexperiments demonstrate that a single channel ICA algorithm based on the phase space reconstruction isable to achieve as good effect of signal de-noising as multi-scale decomposition methods. In addition, thealgorithm is much more adaptive without multi-scale decomposition, and suitable for deformationmonitoring data de-noising.

关 键 词：相空间重构 独立分量分析 相空间重构的单通道ICA算法 变形监测信号去噪 

分 类 号：P228.4[天文地球—大地测量学与测量工程]