稀疏表示噪声抑制算法提升BOTDR系统信噪比的研究

Research on Improving Signal to Noise Ratio of BOTDR System with Sparse Representation Denoising Algorithm

针对布里渊光时域反射仪(BOTDR)系统信噪比低的问题,利用三维布里渊增益谱(3D-BGS)内部存在的相似性及冗余性,提出基于稀疏表示的BOTDR信噪比提升算法。通过分析稀疏表示基本原理和3D-BGS的空间域相似性,可知随机噪声能够在稀疏重构过程中作为残差被丢弃,阐明了降噪算法的可行性。利用离散余弦变换(DCT)构建初始字典,同时采用K-奇异值分解(K-SVD)和正交匹配追踪(OMP)算法进行字典及系数矩阵更新,进而实现三维布里渊增益谱(3D-BGS)的重构和降噪。仿真结果分析表明,当添加10 dBm高斯白噪声时,信噪比提升了5.26 dB。研究证明,稀疏表示算法为应用图像处理方法提高BOTDR系统信噪比供了新的思路及理论依据。

Aiming at the problem of low signal-to-noise ratio in the Brillouin optical time domain reflectometer(BOTDR)system,a sparse representation algorithm is proposed to improve the signal-to-noise ratio of BOTDR by using the similarity and redundancy existing in the three-dimensional Brillouin gain spectrum(3 D-BGS).By analyzing the basic principle of sparse representation and the spatial domain similarity of 3 D-BGS,random noise can be discarded as residual in the process of sparse reconstruction,which illustrates the feasibility of the denoising algorithm.In order to realize the reconstruction and noise reduction of 3 d-BGS,the discrete cosine transform algorithm is used to construct the initial dictionary,and the k-singular value decomposition(K-SVD)and orthogonal matching pursuit(OMP)algorithm are used to update the dictionary and coefficient matrix.The simulation results show that the SNR increases by 5.26 dB when 10 dBm white Gaussian noise is added.This research has proved that the sparse representation algorithm provides new ideas and theoretical basis for applying image processing methods in improving the SNR of BOTDR systems.