基于EEMD的地声信号单通道盲源分离算法

The single channel seismic-acoustic signal blind-source separation method based on EEMD

针对只有一个观测通道时,基于矩阵运算的盲源分离算法将会失效的问题,提出一种适用于单观测通道的地声信号盲源分离方法.首先采用总体经验模态分解方法将观测信号分解为固有模态矩阵,使单通道的欠定问题转化为多通道的正定问题,再利用已有的盲源分离算法进行分离.仿真实验说明该方法可以抑制宽频及瞬态干扰,有效地提取源信号,而且对频带有交叠的信号也有一定的分离效果.实验数据处理显示该方法可以在地声环境噪声干扰下,有效地提高目标信号的信噪比,增加检测性能,证明了该方法在地声信号处理中的有效性.

With only a single observing channel,the blind source separation methods based on matrix calculating do not work.To deal with this problem,a single channel blind-source separation method for seismic-acoustic signal processing was presented.First,the observing signal was decomposed into intrinsic mode matrix by ensemble empirical mode decomposition,transforming the single channel underdetermined problem into the multi-channel positive definite problem.Then the input matrix for blind source separation was obtained.Finally,the source was estimated by the blind source separation method.The simulation results show that this method can separate the source signal from broad frequency and impulsive noise,and works well when the frequencies of different source signals overlap.Experimental results show that the signal-to-noise ratio can also be improved for seismic-acoustic background noise.It proves that this method is effective for processing seismic-acoustic signal.