计算机科学 ›› 2017, Vol. 44 ›› Issue (5): 33-36.doi: 10.11896/j.issn.1002-137X.2017.05.006

• 网络与通信 • 上一篇    下一篇

基于单位超球面上Mean Shift聚类的地震子波盲估计

萧蕴诗,赵彦青,程成   

  1. 同济大学电子与信息工程学院 上海201804,同济大学电子与信息工程学院 上海201804,苏州科技大学电子与信息工程学院 苏州215009
  • 出版日期:2018-11-13 发布日期:2018-11-13

Blind Estimation of Seismic Wavelet on Unit Hypersphere with Mean Shift

XIAO Yun-shi, ZHAO Yan-qing and CHENG Cheng   

  • Online:2018-11-13 Published:2018-11-13

摘要: 石油勘探领域中,地震信号可以看作地震子波与地震反射系数的褶积。由于缺乏先验知识,地震反褶积本质上是一个盲过程。针对带状独立分量分析方法估计子波的多解性,以及地震子波的单位模长约束。对子波空间进行了单位超球面建模,进而研究了这种特定几何空间的黎曼度量及梯度,并由此构造了单位超球面上的Mean Shift聚类算法,最后依据聚类结果求取子波平均。模型实验与实际资料应用结果表明,与带状独立分量分析方法估计的地震子波相比,通过该方法估计的地震子波保真度更高,与设计子波相似度更高,反褶积处理后能够有效提高地震资料的分辨率。

关键词: 盲反褶积,地震子波,单位超球面,黎曼梯度,Mean Shift聚类

Abstract: Seismic data can be described by convolution of seismic wavelet and reflectivity sequence in seismic exploration.Seismic deconvolution is essentially a blind process due to lack of priori knowledge.Wavelet estimation by BICA has multiple solutions and the solutions conform to unit norm constraint.In this paper,a wavelet space model was established on unit hypersphere and its riemannian metric and gradient were studied,then a Mean Shift clustering algorithm on unit hypersphere was presented and an average wavelet can be estimated by the clustering result.Results of synthetic data and real data show that,compared with the BICA method,estimated wavelet has higher fidelity and similarity to the designed wavelet,and real data have higher resolution after deconvolution.

Key words: Blind deconvolution,Seismic wavelet,Unit hypersphere,Riemannian gradient,Mean Shift clustering

[1] LIU X W,LIU H.Survey on seismic blind deconvolution[J].Progress in Geophysics,2003,8(2):203-209.(in Chinese) 刘喜武,刘洪.地震盲反褶积综述[J].地球物理学进展,2003,18(2):203-209.
[2] BELLINI S.Bussgang techniques for blind deconvolution andequalization[D].Englewood Cliffs,Prentice Hall,1994:8-52.
[3] LIU J,HE Z Y.A survey of blind source separation and blind deconvolution[J].Acta Electronica Sinica,2002,30(4):570-576.(in Chinese) 刘琚,何振亚.盲源分离和盲反卷积[J].电子学报,2002,30(4):570-576.
[4] NSIRI B,CHONAVEL T,BOUCHER J M,et al.Blind submarine seismic deconvolution for long source wavelets[J].IEEE Journal of Oceanic Engineering,2007,2(3):729-743.
[5] SANTAMARIA I,PANTALEON C J,IBANEZ J,et al.Deconvolution of Seismic Data Using Adaptive Gaussian Mixtures[J].IEEE Transactions on Geoscience and Remote Sensing,1999,7(2):855-859.
[6] LUO H,LI Y D.The application of blind channel identification techniques to prestack seismic deconvolution[J].Proceedings of the IEEE,1998,6(10):2082-2089.
[7] LI Y,ZHANG G.Blind seismic deconvolution using variational Bayesian method[J].Journal of Applied Geophysics,2014,0:82-89.
[8] SUN X K,SUN Z D,XIE H W.Nonstationary sparsity-con-strained seismic deconvolution[J].Applied Geophysics,2014,11(4):459-467.
[9] KAZEMI N,BONGAJUM E,SACCHI M D.Surface-Consistent Sparse Multichannel Blind Deconvolution of Seismic Signals[J].IEEE Transactions on Geoscience and Remote Sensing,2016,4(6):3200-3207.
[10] MENG X H,WU H Z,LIU G F.Study of blind deconvolution and application of method[J].Oil Geophysical Prospecting,2005,40(6):642-645.(in Chinese) 孟小红,吴何珍,刘国峰.盲源反褶积方法与应用研究[J].石油地球物理勘探,2005,40(6):642-645.
[11] CAI L F,TIAN X M.Seismic blind deconvolution method based on kurtosis maximization[J].Geophysical Prospecting for Petroleum,2012,1(1):30-36.(in Chinese) 蔡连芳,田学民.基于峭度最大化的地震盲反褶积方法[J].石油物探,2012,1(1):30-36.
[12] ZHANG F C,LIU J,YIN X Y.Modified Cauchy constrained seismic blind deconvolution[J].Oil Geophysical Prospecting,2008,3(4):391-396.(in Chinese) 张繁昌,刘杰,印兴耀.修正柯西约束地震盲反褶积方法[J].石油地球物理勘探,2008,43(4):391-396.
[13] KAPLAN S T,ULRYCH T J.Blind deconvolution and ICAwith a banded mixing matrix[C]∥4th International Symposium on ICA and Blind Signal Separation.2003:223-228.
[14] LIU X W,GAO W,ZHANG N,et al.ICA with banded mixing matrix based seismic blind deconvolution[J].Progress in Geophysics,2007,22(4):1153-1163.(in Chinese) 刘喜武,高伟,张宁,等.基于带状混合矩阵ICA实现地震盲反褶积[J].地球物理学进展,2007,22(4):1153-1163.
[15] CHENG C,XIAO Y S,YUE J G.Clustering algorithm for two classes on unit Hypersphere[J].Control and Decision,2011,6(1):80-84.(in Chinese) 程成,萧蕴诗,岳继光.单位超球面上的二元聚类算法[J].控制与决策,2011,6(1):80-84.
[16] NEILL B O.Elementary differential geometry[M].New York:Academic,1966.
[17] COMANICIU D,MEER P.Mean shift:a robust approach toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619.

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!