Performance evaluation of the neural networks for moisture detection using GPR

Abstract

Ground penetrating radar (GPR) is a highly researched area; however, despite this, there is a lack of knowledge about the well-known problem of moisture distorting the results of GPR surveys. This research analyses the results of a GPR survey on a Case Study Bridge structure in order to analyse this effect, specifically when checking for the positioning of rebar. The expected distortions of the GPR results due to the presence of moisture were indeed present, as further evidenced by subsequent destructive testing and velocity analysis. Furthermore, neural networks were also utilised to detect moisture ingress from the GPR raw data.

Keywords::

• GPR and data processing
• bridge structures
• non-destructive
• moisture ingress
• neural networks
• back-propagation learning algorithm