GPR Data Enhancement via Multiple Reflections’ Filtering

Based on a classification methodology for multiple reflections’ (MR) identification recently published by the authors, this work proposes a filtering technique for GPR
data to be integrated within the layer stripping (LS)approach. In LS, the reconstruction of the dielectric stack is obtained by recursively inverting the signals backscattered by each layer, from the top to the bottom. According to the state-of-the-art, to simplify the solution
of the problem, MR are commonly neglected. The scope of this article is therefore to first discuss the implementation of the MR filtering procedure within a LS algorithm and evaluate from the quantitative point of view which improvements this will yield to the overall performances.