Total-variation improved split Bregman method for ground penetrating radar image restoration

• Our method is achieved by combining isotropic and anisotropic total-variation model.
• Using isotropic diffusion in flat region and anisotropic for object boundaries.
• Our method can reserve edges and features for GPR data interpretation effectively.

In this paper, an improved data processing procedure is proposed for the purpose of ground penetrating radar (GPR) image restoration. The image processing method is achieved by combining the isotropic and anisotropic total-variation (TV) model of split Bregman, suppressing the staircasing effect and circumventing the edges blurred. Numerical experiments indicate that the proposed approach can well recover edges and most of the details of a (GPR) image. Hence, the proposed method is efficient in (GPR) image restoration.