Research on Recognition System of Agriculture Products Gas Sensor Array and its Application

AbstractThis paper details the work towards an electronic nose system (e-nose) for detecting the quality of agricultural products and its application in analysis of different fruit jams. In this study, we focused on optimizing the design of gas sensor array and gas channels, as well as power circuit and collection circuit of solid electrolyte sensor; PC monitoring program was designed by Delphi. The sensor array composed of twelve sensors with the type of metal oxides and solid electrolyte. Principle Component Analysis (PCA) was employed to extracted useful information from original electronic signals; then characteristic variables of signals from each sensor were extracted, they were maximum, minimum and means values of signal. Diverse PCs were used as input vectors for identification models. Linear Discriminant Analysis (LDA) achieved best distinguished results; the testing rate of predication set was 95%. This work provided an innovative method for detecting and monitoring the quality of jam