Clutter reduction and detection of landmine objects in ground penetrating radar data using singular value decomposition (SVD)

Abstract

This paper demonstrates that the Singular Value Decomposition (SVD) B an efficient method for the reduction of clutter in Stepped-Frequency Ground Penetrating Radar (SFGPR) data. In this contribution, two algorithms have been applied for clutter reduction. In the first algorithm, SVD has been applied after mean subtraction and the target has been estimated, The suggestion of applying a threshold and excogitate a new formula to calculate the threshold has been presented on the first algorithm. Signal-to-Noise ratio (SNR) has been analyzed. In the second algorithm, a background estimate has been done directly using SVD. The technique has been experimentally validated using PMN landmine at different depths.