Processing strategies for high-resolution GPR concrete inspections

摘要

A high-resolution multi-sensor and multi-polarization Ground Penetrating Radar (GPR) dataset was acquired on a concrete retaining wall. This dataset was characterised as a low pass filter with the help of a moving window spectral analysis. In order to examine the benefits and limits of innovative processing strategies, the dataset was processed with three different methods: classical 2-D processing, full 3-D processing followed by data fusion and inverse scattering followed by data fusion. A comparison of the results for two layers of rebar present in the wall shows that the innovative approaches improve the results for near surface structures when compared to classical 2-D processing. For deeper structures, the benefits of the innovative approaches are limited because of the low pass properties of the concrete.