The Research Of Time-varying Deconvolution Based On S-Transform Domain

With the rapid development of oil exploration technology,requires for seismic data processing techniques are increasingly improved,such as requiring seismic processing more precise,finely reflects the information of stratum structure.This requires that we try to improve the resolution of seismic profiles,eliminating the effects of seismic waves.There are some common methods to improve the resolution of seismic data,such as deconvolution,spectral whitening and wavelet transform,etc.These methods are based on stable convolution model which Robison proposed.But,in the reality,seismic wavelet is constantly changing by its energy or amplitude because of the non-uniform and non-elastic underground.That means the wavelet is unstable,which make the received seismic record is unstable too.Therefore,we are processing the unstable seismic record by the methods which based on stable seismic record and always fail to improve the resolution of these unstable seismic records.Conventional deconvolution method has many limitations,such as seismic wavelet is minimum phase and stable,the reflection coefficient sequence is white noise,and so on.S-transform and its extensions(generalized S transform)to provide fine time-frequency spectrum,has been widely used in seismic data processing,such as: random noise attenuation of seismic records,repression of surface wave,attenuation parameter extraction,detect the initial seismic waves,extract sedimentary cycles and so on.In order to adapt to the time-varying characteristics of the seismic wavelet,first,we extend the conventional convolution model to time-varying convolution model,time-frequency analysis the seismic records by S transform.Then estimate source wavelet and attenuation factor,calculate the time-varying wavelet.Finally,time-varying deconvolution in S domain.Make the improve resolution processing in the S domain can compensated energy of seismic wave attenuated by propagation,can get higher resolution and energy balanced data.Model and real data to verify the correctness of the proposed method,reflects a certain practicality.