PGNAA Spectral Classification of Metal with Density Estimations

For environmental, sustainable economic and political reasons, recycling processes are becoming increasingly important, aiming at a much higher use of secondary raw materials. Currently, for the copper aluminium industries, no method non-destructive online analysis heterogeneous materials available. The Prompt Gamma Neutron Activation Analysis (PGNAA) has potential to overcome this challenge. A difficulty when using PGNAA classification arises from small amount noisy data, due short-term measurements. In case, classical evaluation methods detailed peak by fail. Therefore, we propose view spectral data as probability distributions. Then, can classify material maximum log-likelihood with respect kernel density estimation discrete sampling optimize hyperparameters. measurements pure alloys achieve near perfect under 0.25 second.