水声信号动态阈值正交匹配追踪降噪方法

发布时间：2018-04-01 14:13

本文选题：降噪　切入点：稀疏表示　出处：《声学技术》2017年04期

【摘要】：为了对实时采集的水声信号进行数据压缩的同时实现信号降噪,提出了一种动态阈值正交匹配追踪方法(Dynamic Threshold Orthogonal Matching Pursuit,DTOMP)。该方法将稀疏分解原理应用于水声信号的预处理,通过在正交匹配追踪算法中引入阈值约束,并根据噪声分布特性将其分为两部分,用以控制预设置的参数。通过对加噪正弦信号、实测鲸鱼叫声和舰船辐射噪声信号的降噪实验,表明该方法能够在对原始水声信号进行压缩的同时提高信噪比,且在较宽的信噪比变化范围内比小波方法具有更好的降噪性能。
[Abstract]:In order to compress the data of underwater acoustic signals collected in real time and reduce the noise of signals, a dynamic threshold orthogonal matching tracking method is proposed. The sparse decomposition principle is applied to the pretreatment of underwater acoustic signals. By introducing the threshold constraint into the orthogonal matching tracking algorithm and dividing it into two parts according to the characteristics of the noise distribution, it is used to control the preset parameters. The experimental results of whale calls and ship radiated noise signals show that the proposed method can compress the original underwater acoustic signals and improve the signal-to-noise ratio (SNR) at the same time, and it has better noise reduction performance than the wavelet method in a wide range of signal-to-noise ratio (SNR) variations.
【作者单位】：西北工业大学航海学院;
【分类号】：TB56

【相似文献】