一种利用子波变换多尺度分辨特性的信号消噪技术

A Noise Removal Technique with Wavelet Transform Multir esolution Properties

本文基于传统信号处理方法对非平稳信号的局限性，利用子波变换下奇异信号和随机噪声在多尺度空间中不同的模极大值传播特性设计了一种实用的子波消噪方法，对不同信号的仿真结果表明该方法既能提取一定的信噪比改善增益，同时又能保持对突变信息的良好分辨率，适合于一类非平稳信号进行时变处理．

Based on the limitation of classical processing method in case of nonstational signal a practical noise removal algorithm is presented by using of the defference of WT modulus masima evolution behaviors between singular signal and random noise in multiscale space The simulation results show this method is effective in processing a class of nonstational signals with a useful SNR gain while having a good resolution in irregulraity information