利用三种滤波器对语音信号进行增强  

To Enhance Speech Signal Based on Three Kinds of Filters

作　　者：张翼洲 郜参观[1] 俞月娇 

机构地区：[1]伊犁师范学院电子信息工程学院,伊宁835000

出　　处：《信息化研究》2017年第1期22-28,共7页INFORMATIZATION RESEARCH

摘　　要：本文基于RLS(最小均方滤波器)、LMS(递归最小二乘滤波器)、Kalman(卡尔曼滤波)的原理,通过Matlab仿真,将真实语音信号及噪声信号转换成软件可识别的存储数据,将两种信号合成后分别用RLS算法、LMS算法、Kalman算法进行仿真去噪,得到三种滤波器去噪后的仿真结果,比较分析滤波时间、滤波效果,区别出三种滤波器的优缺点,判断其外部噪音是否去除,语音信号是否增强。Based on the RLS （least mean square filter）, LMS （recursive least square filter）, Kalman filte-ring principle, through Matlab simulation, the real speech signal and noise signal is transformed into a data storage software and the two kinds of signals are combined respectively with RLS algorithm, LMS algorithm, Kalman algorithm simulation noise. Thus we get three kinds of filter after denoising simulation results, comparative analysis of the filtering time, filtering effect, and distinguish the advantages and disadvantages of the three kinds of filters, that the external noise is removed, and whether speech signal is enhanced.

关 键 词：最小均方滤波器 递归最小二乘滤波器 卡尔曼滤波器 MATLAB仿真 

分 类 号：TN713.7[电子电信—电路与系统]