Enhanced feature extraction for landmine detection using handheld ground penetrating radar (GPR) based on full wave inversion (FWI)

This paper reports the exploration of the potential of enhanced target classification through feature extraction for anti-personnel (AP) mine detection using handheld ground penetrating radar (GPR). Principal component analysis (PCA) using singular value decomposition (SVD) of the Jacobian matrix is used to determine the ability of a bistatic handheld GPR/metal detector system to detect the presence of air space or vacuum in an AP mine preceded by initial detection by the metal detector and successful full wave inversion (FWI). The results are promising and show that under the right conditions of accurate sub-surface parameter estimation through FWI and clutter mitigation, GPR can detect air space in a mine, treating it as a kind of `container', and enable improved target classification for mine detection.