Classification of fruit species based on effective sensor array selection

In this paper, the PEN3 electronic nose system was used to collect six kinds of fruit odor data. The smell print of each fruit is composed of 10 sensor data collected by the electronic nose. The combination of different sensor data was selected to classify the kinds of fruits. The principal component analysis (PCA) and linear discriminant analysis (LDA) were used to classify the data. Then, a combination of the sensor data with large instantaneous amplitude response is selected, and the six kinds of fruits were classified under the same classification method. The experimental results show that, in the steady state, the classification result using the sensor data with array optimization is better than that of direct sensor array data. In the case of sensor array optimization, the classification result using the extracted features is better than the result without feature extraction.