Ricker wavelet LS-SVM and its parameters setting for seismic prospecting signals denoising

Xiaoying Deng; Yue Li; Baojun Yang

doi:10.1117/12.773771

10 November 2007 Ricker wavelet LS-SVM and its parameters setting for seismic prospecting signals denoising

Xiaoying Deng, Yue Li, Baojun Yang

Proceedings Volume 6795, Second International Conference on Space Information Technology; 67951H (2007) https://doi.org/10.1117/12.773771
Event: Second International Conference on Spatial Information Technology, 2007, Wuhan, China

Abstract

LS-SVM (Least Squares-Support Vector Machines) are applied to seismic prospecting signals denoising so as to suppress the stochastic noise in this paper. Firstly, we propose and prove a new admissible support vector kernel-Ricker wavelet kernel, which is superior to the popular RBF (radial basis function) kernel in terms of the waveform retrieved and SNR (Signal to Noise Ratio) gained when applied to the noise reduction of seismic prospecting signals. LS-SVM embed two tuning parameters which may diminish the overall performance of LS-SVM if not well chosen, so we investigate the selection of LS-SVM parameters including kernel parameter and regularization parameter, respectively. We can conclude that Ricker wavelet kernel parameter should be set as the predominant frequency of seismic signal and regularization parameter γ can be accepted in a wide range. Our denoising experimental results show that the performance of Ricker wavelet LS-SVM using the aforementioned parameters setting outperforms Wiener filtering, median filtering and LS-SVM based on RBF kernel in terms of the definition of seismic prospecting event retrieved and SNR gained.

Citation Download Citation

Xiaoying Deng, Yue Li, and Baojun Yang "Ricker wavelet LS-SVM and its parameters setting for seismic prospecting signals denoising", Proc. SPIE 6795, Second International Conference on Space Information Technology, 67951H (10 November 2007); https://doi.org/10.1117/12.773771

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available