Rapid Detection Of Storage Time Of Chicken、Egg And Goat Milk By Electronic Nose

The animal food such as chicken、egg and goat milk are the basic material for survival、work and creating civilization which provide superior nutrient than vegetable. In the processof storing the freshness and quality of them will have a great change as time goes on. Thetraditional national standards such as TVB-N, pH and total plate count have significantrelationship with storage time, however, it will takes longer time for detection and morecomplicated preprocess is needed. At the same time samples were destroyed so it is hard tomeet the market demand in a short time. Therefore it is very critical to develop quick andaccurate technology for rapid detection of the chicken、egg and milk.In this experiment, chicken、egg and goat milk stored at ambient temperature and frozentemperature were the object detected by electronic nose at the same time. PCA, LDA,significance test, neural network were applied to analyze the result. PCA, LDA and Loadingswere used to the chicken and the egg. Based on PCA, LDA and Loadings, neural network andsignificance test of acidity were applied to the goat milk, specific results were as follows:(1)The results showed that the classification of chickens during day1-day6storing atambient and frozen temperature by PCA method were better than LDA method which couldclassify different samples clearly. It was indicated by Loadings analysis that sensors W1W(sulfur compounds) and W5S (Nitrogen oxides) played a central role in evaluating differentstorage time of chicken.(2) The results of classification of ambient and frozen eggs during day1-day6wereobviously which could be seen in the PCA and LDA pictures. It was indicated by Loadingsanalysis that sensors W5S (Nitrogen oxides) played a central role in evaluating differentstorage time of egg.(3) The results showed that electronic nose was able to classify fresh goat milk duringday1-day6storing at ambient temperature and refrigerated temperature by PCA and LDA, andLDA method could demonstrate variation trend of volatile composition. Biplot of sensorresponses and samples scores had a great agreement with acidity, which could divide allsamples into3groups, fresh、slight corruption and corruption. The results showed that thedifferent samples could be discriminated according to the storage condition by PCA method.It also illustrated the mutual relationships between samples and sensors finding that W1C,W3C and W5C sensors showed a decreasing response over time; however, W5S, W2W and W1W sensors were characterized by an increasing trend. Frozen temperature could weakenthe extent of sensors responses.Better prediction values were obtained by PSO-BP neural network which was97.78%than BP which was90.00%at ambient temperature,95.56%and87.78%at refrigeratedtemperature respectively. PSO was adopted to optimize the parameters of BP. The betteralgorithm would give better combinational parameters of BP.All of these suggested that electronic nose has good potential to quantitative andqualitative detect chicken、egg and goat milk rapidly and non-destructively.