Fourier-based seismic data reconstruction algorithm has several significant advantages such as the improvement of efficiency by using FFT or Non-uniform Fast Fourier Transform (NFFT) and the regularization feature in some of the algorithms which avoids the geometry errors caused by binning the input seismic traces into a regular spatial grid. However, in some cases, only when spatial aliasing problem is overcome properly can the algorithm play its powerful role. In order to address the aliasing problem, an additional constraint item which characterizes the spatial continuity of the inversion result is put into the conventional optimization formula. The weight matrix in the constraint item can be approximately degenerated as a diagonal matrix which makes nearly an equal computational complexity to that of conventional optimization formula but almost no difference for the function of anti-aliasing. A 1D synthetic data is used to illustrate the basic principle of the method. A 2D synthetic data example demonstrates that the algorithm performs well on spatial aliasing data. The algorithm is also tested on a 5D real data in which the relative regular but aliasing dimension of receiver lines is successfully interpolated.
Presentation Date: Wednesday, September 27, 2017
Start Time: 4:45 PM
Location: 360A
Presentation Type: ORAL