地球物理学进展 ›› 2021, Vol. 36 ›› Issue (2): 730-735.doi: 10.6038/pg2021EE0116

• 应用地球物理学Ⅱ(海洋、工程、环境、仪器等) • 上一篇    下一篇

多道奇异谱分析压制强振幅窄带随机噪声研究

姜占东1(), 范彩伟1, 黎孝璋1, 卢忠严1, 吴云鹏1, 严浩雁2   

  1. 1.中海石油(中国)有限公司湛江分公司,湛江 524057
    2.中海油能源发展股份有限公司工程技术特普公司,湛江 524057
  • 收稿日期:2020-05-05 修回日期:2020-12-06 出版日期:2021-04-20 发布日期:2021-05-09
  • 作者简介:姜占东,男,1985年生,硕士研究生,工程师,主要从事海洋石油勘探相关地球物理工作.E-mail: jiangzhd@cnooc.com.cn
  • 基金资助:
    国家专项基金(2016ZX05024-005)

Study on suppressing strong amplitude narrow band random noise based on multichannel singular spectrum analysis

JIANG ZhanDong1(), FAN CaiWei1, LI XiaoZhang1, LU ZhongYan1, WU YunPeng1, YAN HaoYan2   

  1. 1. Zhanjiang Branch Company of China National Offshore Oil Corporation, Zhanjiang 524057, China
    2. Data Processing Company of CNOOC Energy Development Co., Ltd., Zhanjiang 524057, China
  • Received:2020-05-05 Revised:2020-12-06 Online:2021-04-20 Published:2021-05-09

摘要:

地震数据残留噪声往往不满足白噪假设,而是与地震数据频带一致的窄带非白噪随机噪声,传统的去噪方法难以进行彻底的压制,其残留的噪声会造成地质解释认识的多解性.文中就多道奇异谱分析降秩去噪、阻尼多道奇异谱分析降秩去噪、三维曲波变换去噪方法对此类强振幅噪声压制效果进行了研究,阻尼多道奇异谱分析去噪方法在降秩过程中通过对噪声进行二次分解引入阻尼算子,对另外两种方法存在的强振幅窄带随机噪声残留问题进行了较好地解决,显著提高了此类强振幅噪声压制效果.

关键词: 多道奇异谱分析, 随机噪声, 降秩, 阻尼算子, 窄带

Abstract:

The residual noise of seismic data often does not satisfy the white noise hypothesis, but is a narrow-band random non white noise consistent with the frequency band of seismic data, which is difficult to suppress completely by raditional denoising methods, and the residual noise will lead to the multi solution of geological interpretation. In this paper, the suppression effect on this kind of strong amplitude noise by multi-channel singular spectrum analysis rank reduced denoising, damped multichannel singular spectrum analysis rank reduced denoising, 3D curve transform denoising are studied. In the process of rank reduction, damped multichannel singular spectrum analysis adhibits damping operators by second decomposition of these noise, effectively solving residue problems of strong amplitude narrow band random noise produced by the other two metheds, which improves denoising effect on such strong amplitude noise.

Key words: Multichannel singular spectrum analysis, Random noise, Rank reduction, Damping operators, Narrow band

中图分类号: