Detection and classification of pesticide residues in dandelion (Taraxacum officinale L.) by electronic nose combined with chemometric approaches

Abstract

In this study, for the first time establish a suitable pesticide residue detection system for dandelion (Taraxacum officinale L.) based on electronic nose to determine and study the concentration of pesticide residue in dandelion. Dandelions were sprayed with different concentrations of pesticides (avermectin, trichlorfon, deltamethrin, and acetamiprid), respectively. Data collection was performed by application of an electronic nose equipped with 12 metal oxide semiconductor (MOS) sensors. Data analysis was conducted using different methods including BP neural network and random forest (RF) as well as the support vector machine (SVM). The results showed the superior effectiveness of SVM in discrimination and classification of non-exceeding maximum residue limits (MRLs) and exceeding MRLs standards. Moreover, the model trained by SVM has the best performance for the classification of pesticide categories in dandelion, and the classification accuracy was 91.7%. The results of this study can provide reference for further development and construction of efficient detection technology of pesticide residues based on electronic nose for agricultural products.