Background matrix subtraction (BMS): A novel background removal algorithm for GPR data

• We present the BMS as a novel algorithm to attenuate background noise in GPR data.
• The BMS algorithm is tested using both synthetic and real GPR data.
• The tests show the superiority of BMS technique compared with traditional technique.
• Future studies will focus on full automation of the BMS noise attenuation method.

Background noise is a common type of coherent noise that severely compromises the integrity of the high-resolution images provided by ground penetrating radar survey. Several existing techniques employ different approaches to attenuate background noise. In this study, we present the background matrix subtraction (BMS) as an alternative technique to remove horizontal background noise and we compare its efficiency to that of the conventional background removal technique. Instead of calculating an average trace that is subtracted from the GPR data in the conventional background removal methods, the BMS technique is based on calculating a complete background matrix of the same size of the GPR section. The background matrix is created through a series of windowing, sample exclusion, weighting, and iteration. This series of processes guarantees that the background matrix is least affected by target response and is composed purely of horizontal background noise. The computed background matrix is then subtracted from the GPR data to remove horizontal events. Results of experiments conducted on both synthetic and real GPR data show that the BMS technique yields better results than the commonly used background removal technique.