Exploring Independent Component Analysis for GPR Signal Processing
Anxing Zhao, Yansheng Jiang, Wenbing Wang · PIERS Online · 2005
Independent component analysis (ICA) is a kind of blind signal processing (BSP) or blind source/signal separation (BSS) technique that develop quickly in recent years. It has found many applications in a variety of fields and been successfully used to acoustical signal processing, biomedical signal processing, feature extraction, edge detection and face recognition, etc. In this paper, based on the analysis to the GPR signal characteristics, we have primarily explored ICA for GPR signal processing using FastICA. A primary study on FDTD simulation and experimental data being processed by FastICA has indicated ICA technique in GPR signal processing will be promising. In our approach, the separation of target signal and background signal has been implemented automatically. We also primarily settled the ambiguities of the sign of the separated target signal components. Our primary results have showed it can effectively remove the background in the raw GPR data and keep the fineness of target signal simultaneously. Additionally, our method may be used on-line. We reckon ICA can become a powerful technique in GPR signal processing.