基于改进变分模态分解的Ф-OTDR信号去噪方法

De-Noising Method of Ф-OTDR Signal Based on Improved Variational Mode Decomposition

摘要:

针对相位敏感光时域反射仪(Ф-OTDR)信号信噪比过低的问题,提出了一种基于改进变分模态分解(VMD)结合独立成分分析(ICA)的去噪方法。首先,采用模拟退火方法(SA)对VMD进行优化；然后,采用SA-VMD将预处理后的Ф-OTDR信号分解成一系列本征模态分量(IMF),并根据相关准则选取IMF分量进行虚拟噪声重构；最后,将原始信号与虚拟噪声作为ICA的输入,去除信号中的噪声,提高信号信噪比。采用自行设计的相干Ф-OTDR系统进行实验验证,结果表明,该方法能够有效去除噪声,与EMD-ICA和SA-VMD方法相比,信噪比提高了4dB,这对系统的实际应用具有重要意义。

Abstract:

Aiming at the problem of low SNR of phase-sensitive optical time domain reflectometer (Ф-OTDR), a de-noising method based on improved variational mode decomposition (VMD) and independent component analysis (ICA) is proposed. First, the simulated annealing method (SA) was used to optimize the VMD. Then, SA-VMD was used to decompose the pre-processed Ф-OTDR signal into a series of intrinsic mode function (IMF) components, and IMF components were selected for virtual noise reconstruction according to relevant criteria. Finally, the original signal and virtual noise were used as the input of ICA to remove noise and improve the SNR of the signal. Experiments were carried out to verify the proposed method on self-designed coherent Ф-OTDR system, and the results show that the method can effectively remove noise. Compared with the SA-VMD and EMD-ICA methods, the SNR is improved by 4dB, which is of great significance to the practical application of the system.