Evaluation of Rapeseed Quality by the Electronic Nose

Abstract

The quality of rapeseed, which is supposed to be free of any detectable foreign aromas, was evaluated by an electronic nose equipped with an array of 8 quartz microbalance sensors and 4 metal oxide semiconductor sensors. Signals generated by the sensors were analysed by Principal Component Analysis (PCA) and Discriminant Functions Analysis (DFA). The aim of the analysis was to distinguish between good quality rapeseed and rapeseed that contained small percentages of mouldy and/or burnt rapeseed. All sound rapeseed samples (7 % moisture content) could be distinguished from samples with high moisture content (14.4 % weight by weight (w/w)) that were incubated for different numbers of days, and which consequently became mouldy. Another set of sound rapeseed samples was mixed with different percentages (1 % to 10 % w/w) of mouldy rapeseed which was obtained by incubating rapeseed with high moisture content for six days. All in all 95.6 % of the samples were correctly classified by DFA of the sensor signals. Also 96.1 % of samples which contained different amounts of burnt rapeseed were correctly classified.