Bandwidth Enhancement by Constrained Convex Optimisation of Spectral Scalars

Bandwidth enhancement is a solution that is used to increase the vertical resolution of seismic images. The objective is to improve the interpretation of thin geological structures. The enhancement is achieved by increasing the temporal frequency content of the seismic data using approaches such as de-convolution and spectral scaling. The main problem with these approaches is that they also tend to boost the level of noise. In this abstract, we propose a window-based spectral enhancement solution that has a minimal side effect of noise boosting. The solution first determines, from the data, the useful signal bandwidth that needs to be extended. Then, spectral scalars are computed using a constrained convex optimisation process to obtain a flat output spectrum. The method is computationally cheap, works with few parameters and does not produce any spectral boosting when the data has a poor SNR.