Research On Quality Factor Estimation And Deconvolution Based On Nonstationary Signal

The anelasticity of underground medium leads to the amplitude decay and the phase distortion of seismic wave field,which reduces the resolution of the seismic data.The high resolution seismic data processing mainly includes deconvolution and inverse Q filtering two common methods.However,conventional deconvolution methods are usually based on stationary trace model,neglecting the filtering when traveling in the subsurface.In addition inverse Q filtering methods have some limitations in practical applications because Q estimation is challenging in field data.The article mainly studies an estimation method of Q value based on nonstationary trace model and deconvolution method based on nonstationary trace.We put forward a novel nonstationary deconvolution algorithm is proposed based on hyperbolic smoothing with variable-step sampling.During the calculating of nonstationary deconvolution factor,grids partitioning is adaptive to energy distribution of the seismic data,thus the proposed method improve the reliability and stability of estimating nonstationary deconvolution factor and the accuracy of the estimating Q value in Gabor time-frequency domain.In the actual reflection coefficient as blue spectrum as the characteristics,we put forward colored Gabor deconvolution,which could better compensate the high frequency information in the deep.The model and the actual data results show that an estimation method of Q value based on Gabor time-frequency domain is reliable,so the imaging quality of the seismic data have been improved significantly after inverse Q filtering.And tests on synthetic and real data demonstrate that,nonstatioanry deconvolution in Gabor time-frequency domain is more effective in compensating for the large offsets data and deep data,with events being more continuous laterally.Colored Gabor deconvolution has further improved.