次声信号的时频信号能量统计检测方法  

An infrasonic signal detection method based on the time-frequency domain signal energy statistic

作　　者：唐伟[1] 刘俊民[1] 王晓明[1] 石建芳[1] 王海军[1] 

机构地区：[1]禁核试北京国家数据中心,北京100085

出　　处：《声学学报》2014年第1期85-92,共8页Acta Acustica

基　　金：国防预研基金资助

摘　　要：次声信号检测是次声事件监测的重要环节,常用的短时平均与长时平均之比STA/LTA(Short Term Average to Long Term Average)和著名的逐次多通道互相关方法PMCC(the Progressive Multiple Cross-Correlation)误检率都比较高.为此,提出一种时频域信号能量统计的次声信号检测方法,其基本原理是根据目标信号与噪声的频域分布差异,采用窄带分频技术将待检波形变换到时频域,按时频格点计算局部平均能量与全局平均能量之比,当比值大于设定阈值时,认为该格点属于目标信号的一部分。以国际监测系统次声台站监测数据为对象进行了实验研究,并与PMCC检测结果进行了比较,结果表明所提出的信号检测方法对微弱次声信号检测能力强、误检信号少.Infrasound signal detection is a major procedure in events monitoring, The commonly used STA/LTA （Short Term Average to Long Term Average） and PMCC （the Progressive Multiple Cross-Correlation） methods have too many false detections. So, a algorithm based on TFSES （the Time-Frequency domain Signal Energy Statistic） is proposed according to the distribution difference between the target infrasound signal and the noise in the frequency domain. Firstly, the narrow band frequency division method is used to transform the waveform into the Time-Frequency domain. Then, calculate the ratio of local average energy to that of the global. The signal is detected at the grid while the ratio is greater than the threshold. The method is tested by using the actual monitoring data from the International Monitoring System infrasound stations, and the experimental results are compared with PMCC. It is obvious that the TFSES method in this paper has a good performance in weak infrasound signal detection, can remarkably reduce the amount of false detections.

关 键 词：声信号检测 信号能量 统计检测 时频域 PROGRESSIVE MULTIPLE 平均能量 目标信号 

分 类 号：TB52[理学—物理] TN911.23[理学—声学]