Differentiation of Tomato Seedlings with Different Mechanical Damage Severities Using E-Nose and GC-MS

Abstract

Abstract. The volatile organic compounds (VOCs) emitted by plants that suffer mechanical damage of different severities are different, which can be used as a principle to determine if a plant has suffered mechanical damage and predict the damage severity. This study used both electronic nose (E-nose) and gas chromatography-mass spectrometry (GC-MS) techniques to identify mechanically damaged tomato seedlings and predict the damage severity. E-nose performed well in qualitative classification and quantitative prediction. Both principal component analysis (PCA) and linear discriminant analysis (LDA) indicated the feasibility of discriminating tomato seedlings with different mechanical damage severities. A total of 26 VOCs were identified using GC-MS. The concentration of each constituent was calculated, and the trends of their concentrations with the increase in mechanical damage severity were determined. Multiple regression analysis was used to explore the relationship between gas sensors and VOCs and indicated a good correlation. A back-propagation neural network (BPNN) was developed to predict mechanical damage severity based on the E-nose data, with 100% correct discrimination for the training set and 91.7% for the testing set. This study demonstrated that E-nose could predict the mechanical damage severity of tomato seedlings.