Shearlet域深度残差CNN用于沙漠地震信号去噪
|
摘要
由于沙漠地震信号中含有较强的随机噪声,从而给沙漠地震数据的处理和解释带来了很大的困难。针对上述问题,提出了一种基于Shearlet变换的深度残差卷积神经网络(ST-CNN:Deep Residual Convolutional Neural Network for Shearlet Transform)模型,实现沙漠地震信号的随机噪声压制。在训练阶段,将沙漠地震信号经Shearlet分解后的系数作为输入,将随机噪声经Shearlet分解后的系数作为标签,通过卷积神经网络(CNN:Convolutional Neural Network)学习输入和标签之间的映射关系;在测试阶段,利用此映射关系即可从沙漠地震信号系数中预测出噪声系数,并间接地获得有效信号系数,最后通过Shearlet反变换获得有效信号。通过与传统的Shearlet硬阈值去噪算法对比,发现该算法可把沙漠地震信号的信噪比从-4. 48 d B提高到14. 15 d B,具有更好的去噪效果。
Desert seismic signals contain strong random noise,which brings great trouble to the processing and interpretation of desert seismic signals. In order to solve this technical problem,Deep Residual Convolutional Neural Network for Shearlet Transform model is proposed for the implementation of the desert seismic signal random noise suppression. In training phase,the Shearlet coefficients of desert seismic data are taken as inputs,and the Shearlet coefficients of random noise are taken as labels. Through network training,the mapping relationship between them could be learned by a deep CNN( Convolutional Neural Network). In test phase,the coefficients of random noise can be predicted from the coefficients of desert seismic data by the mapping relationship,and thereafter the effective signals coefficients is obtained indirectly. Finally the effective signals can be reconstructed by inverse Shearlet transform. By comparing with the traditional Shearlet hard threshold denoising algorithm,the proposed algorithm has improved the SNR of the desert seismic signals from-4. 48 d B to 14. 15 dB and has better denoising performance.
Desert seismic signals contain strong random noise,which brings great trouble to the processing and interpretation of desert seismic signals. In order to solve this technical problem,Deep Residual Convolutional Neural Network for Shearlet Transform model is proposed for the implementation of the desert seismic signal random noise suppression. In training phase,the Shearlet coefficients of desert seismic data are taken as inputs,and the Shearlet coefficients of random noise are taken as labels. Through network training,the mapping relationship between them could be learned by a deep CNN( Convolutional Neural Network). In test phase,the coefficients of random noise can be predicted from the coefficients of desert seismic data by the mapping relationship,and thereafter the effective signals coefficients is obtained indirectly. Finally the effective signals can be reconstructed by inverse Shearlet transform. By comparing with the traditional Shearlet hard threshold denoising algorithm,the proposed algorithm has improved the SNR of the desert seismic signals from-4. 48 d B to 14. 15 dB and has better denoising performance.
引文
[1]ZHANG C,LI Y,LIN H B,et al. Seismic Random Noise Attenuation and Signal-Preserving by Multiple Directional TimeFrequency Peak Filtering[J]. Comptes Rendus Geoscience,2015,347(1):2-12.
[2]THOMAS R,BRAM K,FERTIG J,et al. Acquisition and Processing of High-Resolution Reflection Seismic Data from A Survey within the Complex Terrain of The Bavarian Folded[J]. Geophysical Prospecting,2002,50(4):411-424.
[3]KONG J B,ZHUANG D C,GAO Y,et al. A Study of Static Correction and Denoising Method for Low Signal-to-Noise Ratio Seismic Data from The Desert Area of Tarim Basin and Its Application[J]. Geophysical&Geochemical Exploration,2005,29(3):257-260.
[4]MA Y Y,LI G F,WANG Y J,et al. Random Noise Attenuation by f-x Spatial Projection-Based Complex Empirical Mode Decomposition Predictive Filtering[J]. Applied Geophysics,2015,12(1):47-54.
[5]GAN S,CHEN Y,ZU S,et al. Structure-Oriented Singular Value Decomposition for Random Noise Attenuation of Seismic Data[J]. Journal of Geophysics&Engineering,2015,12(2):262-272.
[6]EISNER L,ABBOTT D,BARKER W,et al. Noise Suppression for Detection and Location of Microseismic Events Using A Matched Filter[J]. Seg Technical Program Expanded Abstracts,2008,27(1):1431-1435.
[7]JONES I F,LEVY S. Signal-to-Noise Ratio Enhancement in Multichannel Seismic Data Via The Karhunen-Loeve Transform[J]. Geophysical Prospecting,1987,35(1):12-32.
[8]XU Y,CAO S,YUAN H,et al. Prestack Seismic Random Noise Attenuation with a Hyperbolic Radon-ASVD[J]. Oil Geophysical Prospecting,2017,52(3):451-457.
[9]SADREAZAMI H,AHMAD M O,SWAMY M. A Study on Image Denoising in Contourlet Domain Using the Alpha-Stable Family of Distributions[J]. Signal Processing,2016,128:459-473.
[10]王建花,王守东,刘燕峰.基于Contourlet系数相关性的地震噪声压制方法[J].中国海上油气,2016,28(1):35-40.WANG Jianhua,WANG Shudong,LIU Yanfeng. Seismic Noise Suppression Method Based on Correlation of Contourlet Coefficients[J]. China Offshore Oil&Gas,2016,28(1):35-40.
[11]GUO K,LABATE D. Optimally Sparse Multidimensional Representation Using Shearlets[J]. SIAM Journal on Mathematical Analysis,2007,39(1):298-318.
[12]EASLEY G R,LABATE D,COLONNA F. Shearlet-Based Total Variation Diffusion for Denoising[J]. IEEE Transactions on Image Processing,2009,18(2):260-268.
[13]朱俊鹏,赵洪利,杨海涛.基于卷积神经网络的视差图生成技术[J].计算机应用,2018,38(1):255-259.ZHU Junpeng,ZHAO Hongli,YANG Haitao. Disparity Map Generation Technology Based on Convolutional Neural Network[J]. Journal of Computer Applications,2018,38(1):255-259.
[14]董峻妃,郑伯川,杨泽静.基于卷积神经网络的车牌字符识别[J].计算机应用,2017,37(7):2014-2018.DONG Junfei,ZHENG Bochuan,YANG Zejing. Character Recognition of License Plate Based on Convolution Neural Network[J]. Journal of Computer Applications,2017,37(7):2014-2018.
[15]KRIZHEVSKY A,SUTSKEVER I,HINTON G. Imagenet Classification with Deep Convolutional Neural Networks[C]∥Conference on Neural Information Processing Systems. New York:Association for Computing Machinery, 2017,60(6):84-90.
[16]SZEGEDY C,LIU W,JIA Y,et al. Going Deeper with Convolutions[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE Computer Society,2015:1-9.
[17]HE K,ZHANG X,REN S,et al. Deep Residual Learning for Image Recognition[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE Computer Society,2016:770-778.
[18]KUTYNIOK G,LIM W Q. Compactly Supported Shearlets Are Optimally Sparse[J]. Journal of Approximation Theory,2011,163(11):1564-1589.
[2]THOMAS R,BRAM K,FERTIG J,et al. Acquisition and Processing of High-Resolution Reflection Seismic Data from A Survey within the Complex Terrain of The Bavarian Folded[J]. Geophysical Prospecting,2002,50(4):411-424.
[3]KONG J B,ZHUANG D C,GAO Y,et al. A Study of Static Correction and Denoising Method for Low Signal-to-Noise Ratio Seismic Data from The Desert Area of Tarim Basin and Its Application[J]. Geophysical&Geochemical Exploration,2005,29(3):257-260.
[4]MA Y Y,LI G F,WANG Y J,et al. Random Noise Attenuation by f-x Spatial Projection-Based Complex Empirical Mode Decomposition Predictive Filtering[J]. Applied Geophysics,2015,12(1):47-54.
[5]GAN S,CHEN Y,ZU S,et al. Structure-Oriented Singular Value Decomposition for Random Noise Attenuation of Seismic Data[J]. Journal of Geophysics&Engineering,2015,12(2):262-272.
[6]EISNER L,ABBOTT D,BARKER W,et al. Noise Suppression for Detection and Location of Microseismic Events Using A Matched Filter[J]. Seg Technical Program Expanded Abstracts,2008,27(1):1431-1435.
[7]JONES I F,LEVY S. Signal-to-Noise Ratio Enhancement in Multichannel Seismic Data Via The Karhunen-Loeve Transform[J]. Geophysical Prospecting,1987,35(1):12-32.
[8]XU Y,CAO S,YUAN H,et al. Prestack Seismic Random Noise Attenuation with a Hyperbolic Radon-ASVD[J]. Oil Geophysical Prospecting,2017,52(3):451-457.
[9]SADREAZAMI H,AHMAD M O,SWAMY M. A Study on Image Denoising in Contourlet Domain Using the Alpha-Stable Family of Distributions[J]. Signal Processing,2016,128:459-473.
[10]王建花,王守东,刘燕峰.基于Contourlet系数相关性的地震噪声压制方法[J].中国海上油气,2016,28(1):35-40.WANG Jianhua,WANG Shudong,LIU Yanfeng. Seismic Noise Suppression Method Based on Correlation of Contourlet Coefficients[J]. China Offshore Oil&Gas,2016,28(1):35-40.
[11]GUO K,LABATE D. Optimally Sparse Multidimensional Representation Using Shearlets[J]. SIAM Journal on Mathematical Analysis,2007,39(1):298-318.
[12]EASLEY G R,LABATE D,COLONNA F. Shearlet-Based Total Variation Diffusion for Denoising[J]. IEEE Transactions on Image Processing,2009,18(2):260-268.
[13]朱俊鹏,赵洪利,杨海涛.基于卷积神经网络的视差图生成技术[J].计算机应用,2018,38(1):255-259.ZHU Junpeng,ZHAO Hongli,YANG Haitao. Disparity Map Generation Technology Based on Convolutional Neural Network[J]. Journal of Computer Applications,2018,38(1):255-259.
[14]董峻妃,郑伯川,杨泽静.基于卷积神经网络的车牌字符识别[J].计算机应用,2017,37(7):2014-2018.DONG Junfei,ZHENG Bochuan,YANG Zejing. Character Recognition of License Plate Based on Convolution Neural Network[J]. Journal of Computer Applications,2017,37(7):2014-2018.
[15]KRIZHEVSKY A,SUTSKEVER I,HINTON G. Imagenet Classification with Deep Convolutional Neural Networks[C]∥Conference on Neural Information Processing Systems. New York:Association for Computing Machinery, 2017,60(6):84-90.
[16]SZEGEDY C,LIU W,JIA Y,et al. Going Deeper with Convolutions[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Boston:IEEE Computer Society,2015:1-9.
[17]HE K,ZHANG X,REN S,et al. Deep Residual Learning for Image Recognition[C]∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas:IEEE Computer Society,2016:770-778.
[18]KUTYNIOK G,LIM W Q. Compactly Supported Shearlets Are Optimally Sparse[J]. Journal of Approximation Theory,2011,163(11):1564-1589.