Application of the ONVMS model to discriminate challenging UXO targets

Discrimination of challenging targets, such as small and deep targets in highly cluttered environments, is still an enormous problem for UXO industry. One way to improve target classification is to enhance the sensor hardware, while another way is to fully utilize the data provided by current EMI sensors, by deploying advanced signal processing approaches. To address the latter, our group developed a new ortho-normalized volume magnetic source (ONVMS) technique for representing subsurface targets responses, and for discriminating between UXO and non-UXO targets. The technique has been applied to live-site UXO data sets and demonstrated to be robust and tolerant to noise. In this paper the ONVMS is applied to the data from a small 20 mm projectile, acquired using an advanced, commercially available EMI system “MetalMapper” (MM). The target intrinsic parameters are first extracted from the data using both the ONVMS and simple dipole models, and then analyzed from library matching classification perspective.