Time-frequency Analysis And Its Applications In Seismic Data Extraction Of Frequency And Denoising

Time-frequency analysis is a powerful tool for non-stationary signal analysis over time,has become a hot topic of modern signal analysis and processing research. Seismic signal isnon-stationary signals, this analysis provides information about the joint distribution of thetime domain and frequency domain, we use this method to describe the spectrum of seismicsignal time-varying relationships to reveal the seismic signal how frequently change over time.In this paper, Hilbert-Huang transform research focus, systematic study of the varioustime-frequency analysis methods and properties of concepts discussed several methodscommonly used when analyzing the frequency, such as short-time Fourier transform,Wigner-Ville distribution, wavelet transform, S transform, Hilbert-Huang Transform, it doesresearch in the time-frequency analysis on the seismic data extraction of frequency anddenoising.Through analysis and discussion, a systematic exposition of the reasons for the end effectof EMD algorithm problems, the proposed two methods to treat analysis signal symmetricextension and extension to the period of the signal in the first extension after its two endpointsinterception, these two methods to improve the end effect of a certain effect. After herein bysignals given practice verify the practicality of the two methods, both of which algorithm issimple, it is easy programming, but also the distortion generated at the end of the EMDalgorithm has certain improvement, a good end to eliminate the effect of the original signal ina given frequency information have reservations. Based on analysis of short-time Fouriertransform and Winger-Ville distribution, the proposed combination of the two methods whencalculating the signal frequency distribution, and Winger-Ville distribution, elimination ofcross-interference term has a certain effect. This method is to remove the cross-terms, whilealso retaining the original signal time-frequency aggregation good, and in the case of lowsignal and noise ratio can remove most of the noise, which can be well applied to seismic datadenoising, the increase in the signal to noise ratio of the original data. In view of the seismicsignal extraction of attribute parameters and denoising analysis, using the Matlab language towrite the corresponding program, will be used in the theoretical model data and reached theexpected goal.