The system of an electronic nose for non-destructively monitoring the fruit changing process is presented.Based on gas sensor array and neural-based pattern recognition techniques

the olfactory system designed is able to classify the smell of three different states（good

bruise

blet） of ripe apples with very good accuracy.The experiment results indicate clearly that the rapid and exact identification measure of artificial neural network（ANN） combining principal component analysis（PCA） and the accuracy rate of differentiating apples is above 83.33 %.