Research On Application Of Time-frequency Analysis In High-precision Seismic Data Processing

Time-frequency analysis technology is a key technology for seismic data processing and analysis.As a typical non-stationary signal,seismic signals are processed by time-frequency analysis,which can quickly and efficiently acquire relevant information of underground reservoirs,and which provides references and basis for the analysis and interpretation of subsequent seismic data.Based on the time-frequency analysis method,this paper studies the application of time-frequency analysis in high-precision seismic data processing.Automatic first-break picking is foundamental for seismic data processing,and it is related to the success of subsequent seismic data processing such as static correction and the establishment of near-surface velocity model.With the seismic exploration developing toward omnidirection,high-density,multi-dimension,multi-component,high-resolution and high-fidelity,the seismic exploration data is increasingly more and more.Therefore,it is very necessary to study a method suitable for accurate and efficient first-break picking of seismic data.This paper proposes an automatic first-break picking CEEMD-based method and an automatic first-break picking constraint-based method.Compared with the conventional method,this new processing method is simple and efficient,which can provide accurate first-break information,particularly,suitable for massive data processing.Formation absorption effect causes seismic wave energy attenuation and resolution decrease.Inverse Q filter can effectively restore attenuation energy and improve resolution.Regarding the issue that the gain function of conventional inverse Q filtering cannot adapt to signal-to-noise ratio(SNR),this paper proposes a time-space-adaptive inverse Q filtering method,in gain function which sets a gain parameter with time-space variable characteristics related to the local SNR of seismic data.The gain parameter adaptively compensates for the high-frequency components according to SNR.Time frequency analysis proves that time-space-adaptive inverse Q filtering method balances resolution and SNR,which well adapts to actual data and has good application effect.Spectral decomposition method can reveal the hidden valuable information in seismic data.The second-order synchronous compression transformation based on the local estimation of instantaneous frequency second-order decomposes the seismic signal,which can obtain the time-frequency distribution with good aggregation and high definition,and which can separate and reconstruct signals.This method can clearly describe the seismic properties and display the liquid nature and geological characteristics of reservoirs.Thus,this method has a broad application prospect.Multiple will cause the distortion of frequency,amplitude,and phase of the effective reflected wave in seismic data,reduce the SNR of seismic data,result in errors in post-processing such as seismic data migration and velocity analysis,affect the prediction of target reservoirs in the middle and deep layers,and then affect the accuracy and reliability of seismic data interpretation.Aiming at the problem of multiple,especially internal multiple,this paper proposes an internal multiple suppression method combining adaptive Shearlet transform and high-resolution Radon transform.This method can not only effectively remove internal multiple,but also improve the SNR,fidelity and resolution of seismic data,which has good application effects and prospects.