A two-stage framework for detection of pesticide residues in soil based on gas sensors

Abstract

The analysis of pesticide residues in soil is essential for the ecological environment, food safety and human health. Focused on the real time detection of pesticide residues in soil by using E-nose technique, this paper proposes a two-stage framework to identify the categories and concentrations of pesticide residues. This method is based on multi-task learning and transfer learning, which improves the generalization ability of the model. Experimental results show that the proposed method outperforms the comparing existed methods, and verify the effectiveness of two-stage stratege and transfer learning as well. It can be potentially applied to the in situ identification of soil pesticide residues, which is of great significance for soil improvement and agricultural production.