A faster method to reconstruct seismic data using anti-leakage Fourier transform

Abstract

Despite the simplicity of the anti-leakage Fourier transform (ALFT), and its ability to regularize pure non-uniform data sets and handle higher dimension seismic data, ALFT still remains less appealing because of its iterative algorithm and so its costly computation. During this study a set of parameters and techniques were studied to reduce needed computation and speed up ALFT. Having a good weighting function and cautiously selecting more Fourier coefficients at each iteration could improve the resolution of each iteration and subsequently enhance the converge speed of ALFT. Here a precomputed matrix by exponential term of Fourier transform is used to compute Fourier components at each iteration of the ALFT. This proposed matrix dramatically reduces the needed complex multiplication and speeds up the ALFT. By merging this process with a proper weighting function, the fast-ALFT will be quicker than the original ALFT, while it will have almost the same output.