基于不同背景噪声的联合检测技术  

Joint Detection Technique Based on Different Background Noise

作　　者：周夕良[1] 

机构地区：[1]湖州职业技术学院机电工程分院,浙江湖州313000

出　　处：《计算机与现代化》2011年第3期153-155,158,共4页Computer and Modernization

摘　　要：提出一种简单、实用的联合检测技术,将经典的基于能量和过零率的双门限法与基于小波变换的子带能量法巧妙地结合起来进行语音信号的端点检测,充分利用双门限法算法简单、实时性好、准确率高,小波子带能量对高斯白噪声区分性能强的特性。仿真结果表明新算法能提高检测系统在噪声环境下端点检测的准确率。The paper puts forward a simple and practical joint detection techniques,combines double-thresholds method based on energy and zero-crossing rate with subband energy method based on wavelet transform skillfully for voice signal endpoints detection,makes the best of double-thresholds method＇s simple algorithm,good real-time,high accuracy rate and wavelet-zone energy＇s good discrimination function from white Gaussian noise.Simulation result shows that the new algorithm can improve the endpoints detection accuracy rate of the detection system in the noise environment.

关 键 词：双门限法 端点检测 子带能量 联合检测 信噪比 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]