基于奇异值差分序列的信号降噪处理方法研究

Research on Signal Denoising Method Based on Singular Value Difference Sequence

奇异值分解是一种十分有效的信号降噪处理方法,但奇异值个数的选择对信号降噪的影响巨大。文章由奇异值序列构造奇异值差分序列,寻找差分序列突变位置,结合不同信号差分序列的特征提出了确定奇异值个数的方法。将提出的选择奇异值个数的方法应用于三种模拟信号降噪中,降噪结果表明提出的方法可以达到有效滤除噪声的目的。

Singular value decomposition is a very effective signal denoising method,but the selection of the number of singular values has a great impact on signal denoising.In this paper,the singular value difference sequence is constructed from the singular value sequence to find the mutation position of the difference sequence.Combined with the characteristics of different signal difference sequences,a method to determine the number of singular values is proposed.The proposed method of selecting the number of singular values is applied to the noise reduction of three analog signals.The noise reduction results show that the proposed method can effectively filter the noise.