基于BP神经网络的车辆声音信号识别  

Recognition of Vehicle Acoustic Signals Based on BP Neural Network

作　　者：白琳[1] 黄梓瑜[1] 叶程[1] 姜莹莹[1] 

机构地区：[1]西南石油大学电气信息学院,四川成都610500

出　　处：《自动化技术与应用》2014年第2期64-66,86,共4页Techniques of Automation and Applications

摘　　要：声音信号包含着很多的信息量,且获取比较方便。近年来,声音识别技术被广泛应用于各个领域,有着重要的意义。本文主要介绍了小波理论和BP神经网络识别算法。通过对采集到的不同类型车辆的声音信号进行分析和预处理,采用小波变换提取低维的特征参数,构建并成功训练一个三层BP神经网络,从而有效地识别出了车型。The acoustic signals contain a great amount of information and easy to be obtained. In recent years, acoustic recognition technology has been widely used in various fields and it has important significance. This paper mainly introduces wavelet theory and BP neural network recognition algorithm. It analyzes and preprocesses by collecting acoustic signals of different types of vehicles, and extracts low dimensional characteristic parameters by using wavelet transform, then it builds and trains a three layers BP neural network successfully, this method can recognize the types of vehicles effectively.

关 键 词：声音信号 小波变换 BP神经网络 车型识别 

分 类 号：TN912.34[电子电信—通信与信息系统]