适于泥石流除噪的EMD联合小波阈值除噪方法

次声传感器采集到的泥石流次声信号中包含有大量的无关干扰信号,严重影响信号的分析与评估。针对含噪泥石流信号中无法准确确定噪声频段的特点,以及传统经验模态分解(Empirical Mode Decomposition, EMD)联合小波阈值去噪方法无法智能分辨噪声所在频段的缺点,提出了信号经EMD分解后,基于相关性选择噪声频段的方法。首先利用EMD分解获取信号的固有模态函数(Intrinsic Mode Function, IMF)分量,然后计算各个IMF分量与原始信号的相关性,根据相关性大小确定IMF噪声频段,然后采用小波阈值去噪方法对噪声频段进行处理,最后对处理后的信号进行重构得到去噪泥石流信号。通过模拟实验分析,证明该方法具有智能选择噪声频段的能力,是一种更适于泥石流信号的去噪方法。

EMD decomposition and wavelet threshold denoising method for removing noise from debris flow signals

Infrasound signals collected by infrasound sensor contain a large number of irrelevant interference signals, which seriously affect the analysis and evaluation of the signals. In view of the characteristics that the noise frequency band can not be accurately identified in the noisy debris flow signal and the shortcoming that the traditional method of empirical mode decomposition (EMD) combined with wavelet threshold denoising can not intelligently distinguish the frequency band where the noise is located, a correlation based method of selecting the noise frequency band is proposed after decomposing the EMD signal. Firstly, the EMD decomposition is used to obtain the intrinsic mode function (IMF) components of the signal, and then the correlation between each IMF component and the original signal is calculated. The frequency band of IMF noise components is selected according to the level of correlation, and then processed by the wavelet threshold de-noising method. Finally, the processed IMF components are reconstructed to get the denoised infrasound signal of debris flow. The simulation results show that this method has the ability to select the noise frequency band intelligently, and is a more suitable denoising method for debris flow signals.