Classification of different kinds of pesticide residues on lettuce based on fluorescence spectra and WT–BCC–SVM algorithm

In order to improve the reliability of the spectrum feature extracted by wavelet transform, a method combining wavelet transform (WT) with bacterial colony chemotaxis algorithm and support vector machine (BCC–SVM) algorithm (WT–BCC–SVM) was proposed in this paper. Besides, we aimed to identify different kinds of pesticide residues on lettuce leaves in a novel and rapid non-destructive way by using fluorescence spectra technology. The fluorescence spectral data of 150 lettuce leaf samples of five different kinds of pesticide residues on the surface of lettuce were obtained using Cary Eclipse fluorescence spectrometer. Standard normalized variable detrending (SNV detrending), Savitzky–Golay coupled with Standard normalized variable detrending (SG–SNV detrending) were used to preprocess the raw spectra, respectively. Bacterial colony chemotaxis combined with support vector machine (BCC–SVM) and support vector machine (SVM) classification models were established based on full spectra (FS) and wavelet transform characteristics (WTC), respectively. Moreover, WTC were selected by WT. The results showed that the accuracy of training set, calibration set and the prediction set of the best optimal classification model (SG–SNV detrending-WT–BCC–SVM) were 100%, 98% and 93.33%, respectively. In addition, the results indicated that it was feasible to use WT–BCC–SVM to establish diagnostic model of different kinds of pesticide residues on lettuce leaves.