Coherent Noise Attenuation in the GPR Data via the Karhunen-Loéve Transform

This paper deals with application of the Karhunen-Loéve (KL) transform for system ringing attenuation in GPR data. The system ringing usually originates from strong shallow reflections or bad ground-antenna coupling, and appears as horizontal banding through a dataset. System ringing can be a serious obstacle in case of GPR surveys aimed at geological/hydrogeological mapping, where this type of coherent noise can coincide with geological horizons. The Karhunen-Loéve transform is investigated here as one of the filtration methods capable of attenuating the coherent noise. The KL transform has been used in the field of seismic data processing, mainly for dipping coherent noise attenuation and multiples suppression on CDP gathers, but is seldom used in GPR data processing. This paper presents a brief outline of the theoretical background, highlight the main features of the KL transform, and gives two examples of its application, followed by the comparison of the results with those obtained using other techniques. As it is shown, the KL transform can be an effective mean of suppressing the coherent noise, contained in GPR data.