Detection of early decayed oranges by using hyperspectral transmittance imaging and visual coding techniques coupled with an improved deep learning model

The effective detection of decayed citrus fruit is challenging because infected fruit show few visual symptoms. A new method was proposed for early decayed orange detection using spectral visualization, data augmentation, and deep learning. First, the Borderline-Synthetic minority oversampling technique (SMOTE) algorithm was used to expand the decayed sample data, which solves the problem of a large gap in the number of samples from different categories (sound and decayed). Next, the spectral data was encoded into Gramian angle summation field (GASF) images, and the GASF image features were extracted using AlexNet-SVM for classification when the spectra of different samples were highly similar. Experimental results showed a classification accuracy greater than 95 %, full-wavelength features optimized, and the time for generating GASF images reduced by 59.43 % compared to the pre-optimization time. Thus, the proposed methodology provided a novel solution for defect detection in fruit and vegetables based on spectral analysis and indicated that combining one-dimensional spectral data with image processing using visual coding techniques is feasible.