Classification for GM and Non-GM Maize Kernels Based on NIR Spectra and Deep Learning

The classification of GM and non-GM maize kernels is fundamental for further analysis the gene action in maize. Therefore, a complete novel detection scheme based on near-infrared spectra was designed to distinguish kernels. Hyperspectral images (935–1720 nm) 777 from 3 kinds were captured, average extracted modeling analysis. classical methods feature engineering first studied, backpropagation neural network–genetic algorithm model showed best performance with prediction accuracy 0.861. Then, deep learning developed. To dig out interactive information between different bands match application scenarios, original transformed into two-dimensional matrices before establishing models. A modified convolution network (i.e., VGG net) dilated finally constructed classify kernels, reached 0.961. This research provides referential way detect Future will improve monitoring unauthorized organisms by introducing visualization technology learning.