一种基于ADALINE网络的自适应滤波语音降噪方法  

Method for adaptive noise canceller in voise signal based on ADALINE

作　　者：韩玥[1] 张亦军 贾晓岚[3] 

机构地区：[1]上海师范大学机械与电子工程学院 [2]兰州军区68026部队 [3]郑州航空工业管理学院计算机科学与应用系

出　　处：《郑州轻工业学院学报（自然科学版）》2007年第2期85-87,共3页Journal of Zhengzhou University of Light Industry:Natural Science

摘　　要：针对噪声环境下通话困难,而传统的低通滤波器对语音伤害比较大,且无法满足环境变化要求的问题,设计了基于ADALINE(线性自适应神经网络)的自适应滤波语音降噪方法.语音信号分析证明,经过用Widrow-Hoff学习方法训练的自适应滤波器可以较好地适应噪声环境,得到满意的语音信息.It is very difficult to communicate in the circumstances that filled with noise. And the traditional low pass has great effect on the antonym, which is not fit for the changing environment. ADALINE is used to reduce the noise in sound signal. According to the analysis of the sound signal, this filter trained by Widrow-Hoff can be fit for the noise environment properly and get satisfied sound signal.

关 键 词：线性自适应神经网络 语音信号 自适应滤波 

分 类 号：TP29[自动化与计算机技术—检测技术与自动化装置]