平稳音频信号中非稳态噪声的时频分析检测方法  

Time-Frequency Detection and Analyze Method of Abnormal Noise in Steady Speech Signal

作　　者：王茹斐 徐帆云 Wang Rufei;Xu Fanyun(School of Electronic Engineering, University of Electronic Science and Technology, Chengdu 611731, China)

机构地区：[1]电子科技大学电子工程学院,四川成都611731

出　　处：《电声技术》2018年第2期76-80,共5页Audio Engineering

摘　　要：音频信号异常监控广泛运用于军用和民用生产生活中,它是公共安全、生产安全预警领域的重要检测手段。实现对枪声、爆炸声、玻璃破碎声等突发异常声音的时频特性检测,有利于准确高效地确定其发生的时间信息和频率信息,便于对紧急情况采取应对措施。传统的傅里叶变换等频率分析方法能够检测音频信号和长时间存在噪声的频谱变化,并构建滤波器将噪声滤除,但是对于异常出现的高斯白噪声或异常频率信号无法获取其所在的时间段,去除难度大,效率低,不利于实时音频信号处理。本文提出了一种针对音频信号的基于维格纳分布WVD(Wigner-Ville Distribution)的时频域分析方法,能够实时快速精确地检测异常噪声所在的时域范围和频率所在时域范围。实际测试实验仿真分析验证了本方法的可行性和有效性。Monitoring of abnormal audio signal is widely used in military and civilian life,it is an important means of detection of public safety,production safety alert. Realizing the detection of the time-frequency characteristics of sudden abnormal sound such as gunshot sound,explosion sound and glass breaking sound is helpful to accurately and efficiently determine the time information and frequency information that occurs. And it is convenient to take measures on emergencies. The traditional Fourier transform and other frequency analysis methods can detect the spectrum change of audio signal and noise always exist,and then construct filter to filter out the noise. However,for white Gaussian noise or noise with abnormal frequency appearance,we can not get the time of the period,and it is difficult and inefficient to remove. It is also inconvenient to real-time audio signal processing. In this paper,we propose a time-frequency domain analysis method based on Wigner-Ville Distribution（ WVD） for audio signals,which can quickly and accurately detect the time domain range and the time domain range where the abnormal noise lies. The actual test experiment simulation analysis verified the feasibility and effectiveness of this method.

关 键 词：平稳频率 非稳态噪声 Wigner-Ville变换 时频分析 检测 

分 类 号：TN912.3[电子电信—通信与信息系统]