Comparison Study on Migration Methods of GPR Scanning Images for Concrete Inspection

Large infrastructures like dams, power stations, cooling towers and bridges use concrete as the main building material. Concrete structures that offer critical services require routine inspection to prevent the occurrences of catastrophic failure. Ground Penetrating Radar (GPR) is a nondestructive technique particularly well adapted to the inspection of concrete structures and widely adopted for sub-surface imaging to assess the structural condition and to locate inner defects. However, in the B-scan GPR image, the subsurface scatterers (reinforcing bars, inner defects etc.) shows up as a hyperbola distortion. A standard migration is therefore needed in order to migrate the unfocused B-scan image and to identify the scatterer’s correct placement and dimension. Two image migration techniques: FK and Kirchhoff methods were evaluated for GPR Bscan image processing in this study. The developed GPR system operating at central frequency of 1600 MHz has been applied for the inspection of a variety of concrete samples with known internal structural features, such as reinforcing bars, plastic ducts and fabricated air-filled voids. 2D radar data were acquired along single parallel profile lines of all samples in order to locate the position and depth of the embedded elements and defects. All the objects in the concrete samples can be focused by three methods. By comparison, the Kirchhoff method focused the image slightly better than F-K algorithm with higher resolution in terms of sizes and locations of the objects.