A two-dimensional correlation spectroscopy analysis-based approach for asymptomatic rot detection in stored potatoes using hyperspectral imaging

Abstract

Fusarium dry rot (FDR), which is caused by several Fusarium species, is a major disease affecting potatoes during storage. The study aimed to identify the gleyic stage and monitor rot progression in stored potatoes using a hyperspectral imaging (HSI) system. We evaluated the susceptibility parameters and quality attributes during the infection process and monitored starch, soluble protein, malondialdehyde, and aerobic bacterial contents in all samples. To further characterize the infection process, we collected spectral data on different storage days and then mapped these data using two-dimensional correlation spectroscopy. The results revealed 20 peaks related to these component contents. Then, the quantitative analysis models of these indicators were established based on the 2D correlation synchronization spectrum. The optimal correlation coefficients of the validation set were 0.9273, 0.9634, 0.9470, and 0.9487 for these indicators. Visual analysis was implemented to these indicators, and the content distribution can be effectively observed on hyperspectral images.