Search

Estimation of true resistivity and thickness from vertical electrical sounding (VES) data is considered a highly nonlinear geophysical inverse problem. Different local and global optimization methods have been used for the inversion of VES data. The success of local optimization largely depends on the level of noise present in the data and how close the initial selected model parameters are to the true model parameters. Although global optimization is not very sensitive to the selection of initial model parameters, it carries significant computational cost and time. Moreover, a common problem in any global optimization method is the frequent occurrence of premature convergence when the exact position of the global minimum cannot be known a priori in a complex geophysical error surface. To reduce these limitations, we propose a novel method based on the Walsh transform (WT) technique with integration of hybrid optimization for the inversion of VES data. Within the framework, we test Walsh-based selection of initial model parameters individually for local, global, and hybrid optimization. We show how the WT technique combined with the hybrid optimization method can offer an excellent alternative scheme for automatic inversion of VES data. Tests on inversion of synthetic and field VES data demonstrate that the Walsh transform can be very effective in selecting the initial model parameters naturally from the data for rapid and robust inversion of VES data. [ABSTRACT FROM AUTHOR]