| High quality seismic data plays an important role in the identification of thin layers by time-frequency analysis methods,so it is necessary to suppress random noise in seismic data.In this paper,a noise suppression method based on ICEEMDAN shear-wave threshold for seismic data is constructed.This method solves the modal aliasing problem of empirical mode decomposition and makes the characteristics of the modal components decomposed by ICEEMDAN more obvious.Then,the noise suppression of the modal components is carried out by shear-wave threshold and the processed modal components are reconstructed.Compared with the traditional wavelet threshold noise suppression method,the effectiveness of suppressing random noise in seismic data is verified.Thin layer prediction has always been a key and difficult point in seismic exploration.The existing thin layer detection methods are mainly based on the tuning theory and spectral decomposition theory,as well as the seismic attributes of thin layer single method,multi-method combination and multi-discipline and method integration to study thin layer.Time-frequency analysis is one of the effective methods for seismic signal processing,which maps the seismic signal from time domain to frequency domain,provides more valuable information for thin layer detection,and thus improves the ability to describe thin layer space.The conventional time-frequency analysis method is affected by fixed time window,window function and other factors,so the recognition effect of thin layer is not good.Therefore,a Continuous Wavelet Transform based on Tunable Factor Gabor Wavelet(TFGW)is used in this paper.The time-frequency analysis method of CWT(TFGW-CWT)is used to process the seismic data.In TFGW-CWT method,Gabor wavelet transform with adjustable factor is adopted,and then the continuous wavelet coefficients of each adjustable factor are combined by the minimum absolute value projection method to reduce the cross interference of adjacent frequencies and improve the local time-frequency resolution.Further,Non-negative Matrix Factorization(NMF)is used to reduce the dimension of time-frequency data and mine the main frequency features in the spectrum,which has a good correspondence with the thin layer.In this paper,the common time-frequency analysis method is used to analyze the seismic signal,and the high resolution time-frequency analysis method and its application in thin layer identification are mainly studied. |
