FastICA算法在低信噪比爆破振动信号信噪分离中的应用研究  

Separating noise from a blasting vibration signal based on fast ICA

作　　者：路亮[1,2] 龙源[1] 钟明寿[1,2] 谢全民[1] 李兴华[1] 

机构地区：[1]解放军理工大学工程兵工程学院 [2]中国人民解放军72351部队

出　　处：《振动与冲击》2012年第17期33-37,共5页Journal of Vibration and Shock

基　　金：国家自然科学基金(51178460);国家自然科学基金(10772198);解放军理工大学预研基金项目(20110305)

摘　　要：传统信噪分离方法对低信噪比爆破振动信号的细节提取能力变弱或近乎失效,因此,提出应用快速独立分量分析(FastICA)方法对信号中的独立分量进行非高斯性度量以完成含噪信号的信噪分离。结合数值仿真算例及实测信号分离试验,验证了FastICA用于爆破振动信号信噪分离的有效性,结果表明,FastICA可以准确地从信噪较低的含噪信号中,分离出贴合源信号的逼近信号,从而为将FastICA引入到爆破振动信号分析领域,实现爆破振动特征的快速提取提供了借鉴。Traditional separating methods have little ability of detail extraction for blasting vibration signal with low SNR. Here, fast ICA was used for non-Gaussian measuring of independent components in a blasting vibration signal to separate noise from it. Efficiency of fast ICA was validated with numerical simulations and separation tests of the real measured signal. Experimental results showed that, a signal close to the original signal can be separated from a blasting vibration signal with low SNR, using fast ICA. The study results provided a reference for rapid detail extraction of a blasting vibration signal.

关 键 词：低信噪比 爆破振动 独立分量分析 信噪分离 分离性能 

分 类 号：TD235.1[矿业工程—矿井建设] O384[理学—流体力学]