Ground penetration radar target classification via complex natural resonances

Ground penetration radar (GPR) requires not only target detection but also identification. Many such targets have simple geometry and their late time impulse responses are dominated by only a few complex natural resonances (CNR). An example is the detection of unexploded ordnance. These CNR´s represent an attractive way for target identification. The low SNR, spurious poles and multiple targets problems associated with Prony´s method when extracting the CNR´s are discussed. New techniques leading to the improvement are proposed. Calculated and measured examples of using CNR´s for target identification are also presented.