基于模态分析的机械声信号识别  

Mechanical Signal Recognizng Based on Semi-supervised Manifold Learning

作　　者：贺文红[1] HE Wenhong(Naval Equipment Department,Beijing 100071)

机构地区：[1]海军装备部

出　　处：《舰船电子工程》2019年第11期207-212,共6页Ship Electronic Engineering

摘　　要：论文在一维信号处理的框架下,结合语音识别和机械声音信号处理技术,提取了机械声信号本征模态中的梅尔倒谱特征,并使用半监督模糊粗糙拉普拉斯特征映射对该二级高维特征进行流行降维以及类别区分以得到更深度的特征,最后利用该特征,构建有效分类器,从而实现对蕴涵了机械转速信息的机械声音信号的高精度识别和分类。In this paper,the meier cepstrum features in Intrinsic Mode Functions(IMFs)of mechanical signal are extracted,under the framework of speech recognition technology and the mechanical signal processing.Besides,the semi-supervised rough fuzzy Laplace Eigenmapmothed is presented to reduce the dimensionality of the features and to construct deep feature.In this way,the classifier is built to recognize the underwater signal more accurately.

关 键 词：半监督学习 流行学习 梅尔倒谱 模糊C均值 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]