Autofocusing Techniques for GPR Data from RC Bridge Decks

Abstract

According to the Federal Highway Administration, more than a quarter of the bridge decks in the United States are obsolete or deficient and most of them are made of reinforced concrete (RC). Precise decisions for bridge maintenance are in high demand due to the limited budget. Among various nondestructive evaluation techniques, ground penetrating radar (GPR) has gained its popularity for reasons such as high speed and fine resolution. Migration is an important intermediate process for GPR data analysis and its accuracy depends on the velocity or dielectric permittivity estimation of the subsurface. Velocity analysis for commonly adopted bistatic GPR systems is based on an iterative trial-and-error process, which involves a human decision process to obtain the properly migrated data. Besides, for heavily polluted data, it can be hardly possible to perform the visual inspection. Autofocusing techniques are rarely discussed in the field of GPR. In this paper, potential autofocusing metrics are nominated and evaluated by both simulation and experimental data. The effects of noise, aperture width, and crossing rebar signals on the metric performance are investigated and the results demonstrate that the higher-order metrics are the most robust and sensitive autofocusing metrics for the migration of GPR data from RC bridge decks.