Author(s)

Liu Yanfeng

Corresponding Author

Liu Yanfeng

ABSTRACT

Near-infrared spectroscopy (NIRS) is used to estimate Mancozeb residues in vegetables. In order to ensure high recognition rate and remove the disturbing factor in chlorophyll, headspace gas method was established to for determining Mancozeb residues. CS2 gas was generated with acid stannous chloride for 20mins in oven at 95℃. The infra-red absorption peaks of raw spectra at 1520cm-1 match well with each other. So, the wavenumber from 1504cm-1 to 1530cm-1 were selected for subsequent analysis after baseline correction method. Least Squares Support Vector Machines (LS_SVM) was used to establish a model to estimate Mancozeb residues. The results show that in this experiment the accuracy of Mancozeb residues can be believable when the Mancozeb residues content was more than 1.2mg/kg.

KEYWORDS

Near-infrared spectroscopy, Numerical Analysis, LS_SVM