Seismic high-resolution processing method based on spectral simulation and total variation regularization constraints.

There is little low-and-high frequency information on seismic data in seismic exploration, resulting in narrower bandwidth and lower seismic resolution. It considerably restricts the prediction accuracy of thin reservoirs and thin interbeds. This study proposes a novel method to constrain improving seismic resolution in the time and frequency domain. The expected wavelet spectrum is used in the frequency domain to broaden the seismic spectrum range and increase the octave. In the time domain, the Frobenius vector regularization of the Hessian matrix is used to constrain the horizontal continuity of the seismic data. It effectively protects the signal-to-noise ratio of seismic data while the longitudinal seismic resolution is improved. This method is applied to actual post-stack seismic data and pre-stack gathers dividedly. Without abolishing the phase characteristics of the original seismic data, the time resolution is significantly improved, and the structural features are clearer. Compared with the traditional spectral simulation and deconvolution methods, the frequency distribution is more reasonable, and seismic data has higher resolution. [ABSTRACT FROM AUTHOR]