Fusion of hyperspectral imaging and electronic nose for identification of green vegetable in egg pancakes

Egg pancake (EP) is commonly consumed breakfast in Chinese cuisine, and the identification of its type holds significance for applications such as intelligent food production and self-service purchasing. To enhance the accuracy of distinguishing green vegetables in EPs, fusion of hyperspectral and electronic nose information was employed. Spectral and texture information were extracted from hyperspectral images, and electronic nose responsive data were collected. Subsequently features were extracted by applying Competitive Adaptive Reweighted Sampling (CARS), Pearson's correlation analysis, and Histogram Statistics (HS) tailored for corresponding data types. These data types were then input into four classification models: Linear Discriminant Analysis (LDA), Convolutional Neural Network (CNN), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN). Comparative analysis revealed that the most promising results were obtained utilizing LDA with fused datasets with 97.50% accuracy, 92.98% recall and 95.12% F1-score. Hence, a novel method was proposed to accurately predict different green vegetables in EPs.