فهرست نشریات علمی دسترسی آزاد ایران

To address the inherent ill-posedness of the geophysical inverse problems, it is necessary to include a suitable regularization function in the corresponding optimization framework. Typically, the choice of the regularization function depends on prior assumptions about the geometric characteristics of the unknown model parameters, e.g., smoothness or blockiness. First-order total variation regularization (TV) allows the reconstruction of well-defined edges and models exhibiting block-like characteristics. However, it is associated with the generation of undesirable staircase artifacts. This study applies a novel approach for removing staircase artifacts using a combined second-order non-convex total variation with overlapping group sparse regularizer. This regularizer aims to smooth out the staircase effect while still keeping the edges of the model. Moreover, the study applies the proposed method for the nonlinear seismic cross-hole tomography problems, where the goal is to reconstruct both smooth and blocky features of the model and avoid staircase artifacts of the TV regularization. The numerical examples indicate the efficiency of the proposed regularization method.