Study on Rapid Detection of Pesticide Residues in Shanghaiqing Based on Analyzing Near-Infrared Microscopic Images.

Aiming at guiding agricultural producers to harvest crops at an appropriate time and ensuring the pesticide residue does not exceed the maximum limit, the present work proposed a method of detecting pesticide residue rapidly by analyzing near-infrared microscopic images of the leaves of Shanghaiqing (Brassica rapa), a type of Chinese cabbage with computer vision technology. After image pre-processing and feature extraction, the pattern recognition methods of K nearest neighbors (KNN), naïve Bayes, support vector machine (SVM), and back propagation artificial neural network (BP-ANN) were applied to assess whether Shanghaiqing is sprayed with pesticides. The SVM method with linear or RBF kernel provides the highest recognition accuracy of 96.96% for the samples sprayed with trichlorfon at a concentration of 1 g/L. The SVM method with RBF kernel has the highest recognition accuracy of 79.16~84.37% for the samples sprayed with cypermethrin at a concentration of 0.1 g/L. The investigation on the SVM classification models built on the samples sprayed with cypermethrin at different concentrations shows that the accuracy of the models increases with the pesticide concentrations. In addition, the relationship between the concentration of the cypermethrin sprayed and the image features was established by multiple regression to estimate the initial pesticide concentration on the Shanghaiqing leaves. A pesticide degradation equation was established on the basis of the first-order kinetic equation. The time for pesticides concentration to decrease to an acceptable level can be calculated on the basis of the degradation equation and the initial pesticide concentration. The present work provides a feasible way to rapidly detect pesticide residue on Shanghaiqing by means of NIR microscopic image technique. The methodology laid out in this research can be used as a reference for the pesticide detection of other types of vegetables.