Non-destructive egg freshness determination: an electronic nose basedapproach

Abstract

An electronic nose (EN) based system, which employs an array of four inexpensivecommercial tin-oxide odour sensors, has been used to analyse the state offreshness of eggs. Measurements were taken from the headspace of four sets ofeggs over a period of 20–40 days, two ‘types of egg data’ being gathered using ourEN; one type of ‘data’ related to eggs without a hole in the shells and the othertype of ‘data’ related to eggs wherein we made tiny holes in the shells. Principalcomponent analysis, fuzzy C means, self-organizing maps and 3D scatter plotswere used to define regions of clustering in multisensor space according tothe state of freshness of the eggs. These were correlated with the ‘useby date’ of the eggs. Then four supervised classifiers, namely multilayerperceptron, learning vector quantization, probabilistic neural networkand radial basis function network, were used to classify the samples intothe three observed states of freshness. A comparative evaluation of theclassifiers was conducted for this application. The best results suggest thatwe are able to predict egg freshness into one of three states with up to95% accuracy. This shows good potential for commercial exploitation.