Seismic data denoising based on learning-type overcomplete dictionaries*
Tang Gang1,2, Ma Jian-Wei3, and Yang Hui-Zhu1
1. Institute of Seismic Exploration, School of Aerospace, Tsinghua University, Beijing 100084, China.
2. Geophysical Department, RIPED, PetroChina, Beijing 100083, China.
3. Institute of Applied Mathematics, Harbin Institute of Technology, Harbin 150001, China.
Abstract The transform base function method is one of the most commonly used techniques for seismic denoising, which achieves the purpose of removing noise by utilizing the sparseness and separateness of seismic data in the transform base function domain. However, the effect is not satisfactory because it needs to pre-select a set of fixed transform-base functions and process the corresponding transform. In order to find a new approach, we introduce learning-type overcomplete dictionaries, i.e., optimally sparse data representation is achieved through learning and training driven by seismic modeling data, instead of using a single set of fi xed transform bases. In this paper, we combine dictionary learning with total variation (TV) minimization to suppress pseudo-Gibbs artifacts and describe the effects of non-uniform dictionary sub-block scale on removing noises. Taking the discrete cosine transform and random noise as an example, we made comparisons between a single transform base, non-learning-type, overcomplete dictionary and a learning-type overcomplete dictionary and also compare the results with uniform and nonuniform size dictionary atoms. The results show that, when seismic data is represented sparsely using the learning-type overcomplete dictionary, noise is also removed and visibility and signal to noise ratio is markedly increased. We also compare the results with uniform and nonuniform size dictionary atoms, which demonstrate that a nonuniform dictionary atom is more suitable for seismic denoising.
This work was in part of fi nancially supported by The National 973 program (No. 2007 CB209505), the Basic Research Project of PetroChina's 12th Five Year Plan (No. 2011A-3601) and RIPED Youth Innovation Foundation (No. 2010-A-26-01).
Cite this article:
TANG Gang,MA Jian-Wei,YANG Hui-Zhu. Seismic data denoising based on learning-type overcomplete dictionaries*[J]. APPLIED GEOPHYSICS, 2012, 9(1): 27-32.
[1]
Broadhead, M., 2008, The impact of random noise on seismic wavelet estimation: The Leading Edge, 27(2),
[2]
6 - 230.
[3]
Candes, E., and Donoho, D., 2002, New tight frames of curvelets and optimal representations of objects
[4]
with smooth singularities: Technical Report, Stanford University.
[5]
Deng, C. Z., 2008, Research on image sparse representation theory and its applications: PhD Thesis, Huazhong
[6]
University of Science and Technology.
[7]
Elad, M., and Aharon, M., 2006, Image denoising via sparse and redundant representations over learned dictionaries:
[8]
IEEE Trans. Image Process, 15(12), 3736 - 3745.
[9]
Herrmann, F., and Hennenfent, G., 2008, Non-parametric seismic data recovery with curvelet frames: Geophys. J.
[10]
Int., 173, 233 - 248.
[11]
Meyer, F. G., 1999, Fast compression of seismic data with local trigonometric bases, in Aldroubi, A., Laine, A., and
[12]
Unser, M., Eds., Wavelet VII: Proc. SPIE 3813, 648 - 658.
[13]
Protter, M., and Elad, M., 2009, Image sequence denoising via sparse and redundant representations: IEEE Trans.
[14]
Image Process, 18(1), 27 - 35.
[15]
Shan, H., Ma, J. W., and Yang, H. Z., 2009, Comparisons of wavelets, contourlets and curvelets in seismic denoising:
[16]
Journal of Applied Geophysics, 69, 103 - 115.
[17]
Tang, G., and Ma, J. W., 2009, Application of total variation based curvelet shrinkage for three-dimensional seismic
[18]
data denoising: IEEE Geosci. Remote Sensing Lett. 8(1), 103 - 107.
[19]
van den Berg, E., and Friedlander, M., 2008, Probing thePareto frontier for basis pursuit solutions: SIAM J.
[20]
Scientific Computing, 31(2), 890 - 912.
[21]
Wang, Y., and Wu, R., 2000, Seismic data compression by an adaptive local cosine/sine transform and its effects on
2012, Vol. 9 
