A 2D line from a georadar (GPR) dataset was collected at the Houston Coastal Center in 2012, aiming to image a series of culverts underneath a road bridge. We processed the dataset using a Gabor deconvolution workflow instead of traditional processing methods. We find that Gabor deconvolution is able to correct for attenuation and greatly improve resolution and signal bandwidth at late arrival times. Diffractor imaging velocities were estimated by hyperbola fitting and the dataset was imaged using zero-offset Gazdag migration to collapse the hyperbolic events of the section. We then used shot record Gazdag PSDM to improve the image. Since we are dealing with circular reflector geometry that is not a point diffractor, we apply a correction factor to the diffractor velocity to obtain a more accurate result. Certain features are visible in the deconvolved section, including the steeply dipping flanks of the edge of the bridge. In the future we hope to extend our velocity model to take into account lateral velocity variations as well to refine the velocities deeper in the subsurface.
View full article as PDF (3.79 Mb)