Sparse decomposition of ground penetration radar (GPR) signals facilitatesthe use of compressed sensing techniques for faster data acquisition andenhanced feature extraction for target classification. In this paper, weinvestigate the application of an online dictionary learning (ODL) technique inthe context of GPR to bring down the learning time as well as improveidentification of abandoned anti-personnel landmines. Our experimental resultsusing real data from an L-band GPR for PMN/PMA2, ERA and T72 mines show thatODL reduces learning time by 94\% and increases clutter detection by 10\% overthe classical K-SVD algorithm. Moreover, the proposed methodology could behelpful in cognitive operation of the GPR where the system adapts the rangesampling based on the learned dictionary.
展开▼