计算机工程与应用 ›› 2018, Vol. 54 ›› Issue (5): 57-60.DOI: 10.3778/j.issn.1002-8331.1609-0447

• 理论与研发 • 上一篇    下一篇

同步挤压小波变换对随机噪声抑制的研究

张志禹1,李向月1,李向阳2   

  1. 1.西安理工大学 自动化与信息工程学院,西安 710048
    2.中国空间技术研究院 西安分院,西安 710100
  • 出版日期:2018-03-01 发布日期:2018-03-13

Research on random noise suppression by synchrosqueezing wavelet transform

ZHANG Zhiyu1, LI Xiangyue1, LI Xiangyang2   

  1. 1.School of Automation and Information Engineering, Xi’an University of Technology, Xi’an 710048, China
    2.Xi’an Branch, China Academy of Space Technology, Xi’an 710100, China
  • Online:2018-03-01 Published:2018-03-13

摘要: 针对低信噪比信号在噪声去除中存在的问题,提出了利用同步挤压小波变换对随机噪声的处理研究。该变换是一种新的小波变换方法,可将时域信号转化成高分辨率的时频谱,再结合时频谱重排的思想,通过挤压任一中心频率附近区间值,从而得到同步挤压小波变换量值。研究发现,同步挤压小波变换可将随机噪声压缩为点状噪声或者颗粒噪声,并聚集分布,基于噪声这一分布特点,从而选用中值滤波,可达到很好地抑制噪声的目的;而传统小波变换后,随机噪声在大小尺度上均有分布,比较分散,用中值滤波不能达到好的滤波效果。对比结果表明:该方法具有较好的抑制随机噪声的能力。

关键词: 信噪比, 同步挤压小波变换, 随机噪声, 中值滤波, 小波变换

Abstract: Aiming at lowering the signal-to-noise ratio of a noisy signal, a synchrosqueezing wavelet transform-based denoising method is proposed. The transformation is a new method of wavelet transform, which can convert the time-domain signal into high-resolution time-frequency spectrum. Then, combining the idea of spectrum rearrangement to obtain the value of the synchrosqueezing transform by squeezing the interval value of any center frequency. The study finds that the synchrosqueezing wavelet transform can be used to compress the random noise into the point like noise or the particle noise, and the aggregation distribution. Based on the distribution of noise characteristics, the median filter can be used to achieve the purpose of suppressing noise, but through the traditional wavelet transform, the random noise is distributed in the size of the scale and relatively decentralized, with median filter can not achieve a good filtering effect. The comparison results show that the method has the ability of restraining noise.

Key words: signal to noise ratio, synchrosqueezing wavelet transform, random noise, median filter, wavelet transform