Using a SVD-based algorithm for T2spectrum calculation in TD-NMR application to detect hidden defects in hazelnuts

Hazelnuts are widely used into the confectionary industry for their flavour and taste. Nondestructive techniques are required to detect the quality of the nuts in shell with the aim of avoiding putting a poor product into the market chain. In this work, Time Domain-NMR (TD-NMR) has been employed to detect hidden defects in hazelnuts in shell. In particular, the Carr-Purcell-Meiboom-Gill (CPMG) technique has been used to evaluate the transverse relaxation times, T2, for healthy and unhealthy hazelnuts, highlighting the differences between them. An index that takes in account the different time constants of the transverse relaxation curve has been defined to establish a classification method for the hazelnuts. The results of this work provide a basis for the development of a measurement system for inline inspection of hazelnuts in shell.