The Influence Of Low-Frequency Components Of Seismic Data On Velocity Inversion And A Long Wavelength Velocity Inversion Method

The velocity,which can be estimated from well log data or the seismic records,is a crucial attribute parameter in seismic exploration.Nevertheless,if low-frequency information of the used seismic data is missing,the conventional inversion can only provide medium and short wavelength components of velocity,namely a relative velocity.Therefore,this paper focuses on the usage and recovery of the low frequencies.First of all,based on the analyses of velocity reconstruction,we found that missing high-frequency information of the used seismic data can only affects the resolution,while missing low-frequencies may cause the inaccurate of absolute value and resulting in the wrong judgement of the lithologic interpretation and structural profile.Secondly,the seismic reflectivity inversion is based on the hypothesis that the strong reflection coefficient is distributed sparsely,and there is a resolution limit problem.In the traditional concept,resolution limitation is based on analysis of double pulse model and do not focus on low frequencies.Thus,we design a series examinations of different scales of multiple pulses reflection coefficient model to illustrate the precision and stability of the low frequencies recovery.Based on those results,when the reflectivity model is extremely non-sparse,although full-band reflectivity inversion is unstable,low frequencies can also be stably and accurately recovered.Finally,based on the above the analyses of the limitation of low-frequency recovery from a band-limited signal,we present a long wavelength velocity inversion method based on SBL reflectivity inversion.In order to improve the quality of seismic data,we propose a method to reduce random noise basing on the sparse representation theory and a method to compensate for the effect of formation absorption by using swarm intelligence optimization search method.