作　　者： (何玉凤）;

机构地区： 山东科技大学计算机科学与工程学院,山东青岛266590

出　　处： 《软件导刊》 2017年第9期28-31,共4页

摘　　要： 针对矿山微震与爆破信号难以识别的问题,提出一种基于局部均值分解(LMD)和奇异值分解(SVD)的微震信号特征提取方法。首先对矿山微震信号和爆破信号进行LMD分解,将多分量的调频信号分解成一系列频率由高到低的乘积函数(PF)分量;其次,借助相关系数和方差贡献率筛选出包含信号主要信息的PF分量;最后利用SVD计算所选的PF分量构成矩阵奇异值,以此作为区分矿山微震与爆破信号的特征向量。实验结果表明,LMD和SVD相结合的特征提取方法能准确、有效地提取矿山微震和爆破信号特征,为信号识别研究提供了一种新方法。 A method of microseismic signal feature extraction based on local mean decomposition （I.MD） and singular value de composition （SVD） is proposed for the identification of microseismic and blasting signals in mines. Firstly,the mine microseis mie signal and blasting signal are decomposed by LMD, and the FM multi-component signal is decomposed into a series of fre quency from high to low product function （PF） components~ secondly, the PF component contains the main information signal are selected by correlations and variance contribution ratios, finally the singular values of selected PF component matrix are eal culated by SVD, which are used as a characteristic vector to distinguish the microseismic and blasting signals of the mine. The experimental results show that the feature extraction method combined with LMD and SVD can accurately and effectively mine the characteristics of microseismic and explosion signals, and provides a new method for signal recognition research.

关 键 词： 局部均值分解 奇异值分解 微震信号 方差贡献率 特征向量