Abstract
Ringing is a common type of coherent noise that degrades the quality of ground-penetrating radar data. Conventional ringing attenuation or a background removal algorithm works by generating an average ensemble that is then subtracted from all traces in a GPR section. This study presents four alternative algorithms, usually used in stacking-seismic reflection data, for generating such an ensemble. These algorithms include median stack, diversity stack, alpha-trimmed stack and smart stack. The traditional algorithm and the four alternative algorithms are tested using two real GPR data sets. The outcomes of the five background removal methods are compared to each other and to the original data using statistical analyses and visual inspection. The results show that all the four alternative techniques are more efficient in background removal than the conventional technique, with the smart ensemble yielding the best results followed by the alpha-trimmed ensemble.