Classification detection of saccharin jujube based on hyperspectral imaging technology

In order to identify saccharin jujube more quickly and effectively, this article used hyperspectral imaging technology to explore a new detection method. About 240 winter jujubes under four gradients of saccharin stress were prepared, and the original spectra of them were collected. First, a hybrid method was used to preprocess original spectra, which improved the signal-to-noise ratio of spectra. Then, three feature extraction methods were adopted to extract feature wavelengths from the spectra after pretreatment, respectively. Subsequently, support vector machines (SVM) was employed to form the classification models based on feature spectra and raw spectra. Because the random parameters (C, g) of SVM affected the ability of model, three intelligent algorithms were used to tune the parameters of SVM, respectively. Finally, the classification accuracy of the final model (VISSA-GWO-SVM) reached best result with accuracy of 91.67%. Therefore, hyperspectral imaging technology combined with VISSA-GWO-SVM model is a feasible method to identify saccharin jujube. Practical applications Traditional methods for identification of saccharin jujube mainly depended on the experience of consumers and destructive and time-consuming testing. In order to overcome the shortcomings of traditional detection methods, hyperspectral imaging technology was used to identify saccharin jujube. Then, the optimal model (VISSA-GWO-SVM) was established in this paper, and the recognition accuracy reached a best value. Thus, hyperspectral imaging technology combined with the VISSA-GWO-SVM model could be a rapid, precise, and nondestructive method for food safety and market supervision departments to ensure food safety.