摘要
传统的小波阈值降噪算法会使信号的不连续点附近产生伪吉布斯(pseudo Gibbs)现象。为了抑制这种现象,本文提出了信号幅值跳变小波降噪方法。该方法首先对原始信号进行幅值跳变处理,消除其不连续性;然后利用小波阈值降噪方法进行降噪;最后对得到的信号进行幅值逆跳变处理,获得原始信号的降噪信号。分别采用信号幅值跳变小波降噪方法、平移不变小波降噪方法和传统的小波阈值降噪算法对具有不连续点的信号进行降噪处理,结果发现,与另外两种方法相比,信号幅值跳变小波降噪方法能够有效抑制pseudo Gibbs现象,且降噪后的信号具有较高的信噪比。
Signal de-noising using conventional orthogonal wavelet transform often exhibits pseudo Gibbs phenomenon in the vicinity of signal discontinuities.To suppress such defect,signal amplitude changing wavelet de-noising method was proposed in this paper.In this method,the original corrupted signal was pre-processed before the wavelet de-noising stage and post-processed after that stage.In this way,for the discontinuities were removed from the signal,the cause for the appearance of visual artifacts was eliminated and the de-noised signal without pseudo Gibbs phenomenon was obtained.In order to testify the effectiveness of the method,its de-noising capability was compared with that of translation-invariant de-noising method and conventional wavelet de-noising method by simulations.The results indicate that the de-noising pedormance by the method proposed in this paper is improved in terms of pseudo Gibbs phenomenon suppress and output signal to noise ratio.Therefore,signal amplitude changing wavelet de-noising method provides an effective solution for signal de-noising.
出处
《信号处理》
CSCD
北大核心
2014年第4期443-449,共7页
Journal of Signal Processing
基金
国家自然科学基金资助项目(51179135
51379160)
湖北省自然科学基金重点项目(2011CDA046)
关键词
平移不变
小波降噪
伪吉布斯
信号处理
幅值跳变
translation invariant
wavelet de-noising
pseudo Gibbs
signal processing
amplitude changing
