含噪信号的特征基表示与信号去噪重构

Characteristic Base Representation of Noisy Signal and Reconstruction of Denoised Signal

提出了一种利用信号幅度谱的去噪重构算法，并作了仿真实验研究．该算法利用含噪信号在幅频特性上的特征，通过特征基来进行信号的去噪处理．仿真研究结果表明，采用该种算法作合噪信号的重构，可以明显地去除信号的宽带噪声．该算法可以应用于噪声严重污染信号的离线去噪处理．

A denoising reconstruction algorithm using the signal magnitude spectrum is proposed and the simulation studies have been performed. The magnitude-frequency characteristics of the noisy signal is utilized in the algorithm and the denoising is realized by the use of characteristic bases. The simulation results have shown that the wide-band noise can be largely cancelled using the algorithm proposed to perform the noisy signal reconstruction. The algorithm can used for the off-line denoising of the strong noise corrupted signals.