Identification of Pesticide Residues of Lettuce Leaves Based on LVQ Neural Network

Abstract

In order to ensure the healthy safety of edible lettuce, an identification way of pesticide residues of lettuce leaves is studied. Soilless cultivation of lettuce sample was adopted. At tillering stage, various levels of fenvalerate were sprayed to different groups of lettuce, lettuce leaves accompanied by a collection of pesticide residues and lettuce leaves not sprayed pesticide residues were collected, Hyperspectral Information of each blade leaf were collected and lettuce leaves sample spectral library was set up. Randomly 90 samples were selected as training samples, and the classification models based on BP neural network and LVQ neural network were set up. The remaining samples were taken to forecast classification test. Results showed that, the prediction identification correct rate of LVQ neural network model is 98.3607%. The algorithm is suitable for pesticide residues diagnosis of lettuce leaves.