Assessment of plum quality changes during postharvest storage using multispectral NIR technique

Abstract

Plums are perishable fruit, and agriculture demands rapid nondestructive methods to be able to make decision on the storability and optimal postharvest technology. This study aimed to apply a multispectral technique utilizing selected NIR wavelengths to assess the quality of two plum cultivars popular on the Hungarian market. Samples of ‘Stanley’ and ‘Elena’ were stored at 1 °C, 5 °C, 10 °C and 15 °C for up to 24 days. The five bands at 906, 1060, 1336, 1454, and 1680 nm were selected based on the standard deviation values of the normalized spectra. These wavelengths are associated with vibrational overtones and combinations of C-H and O-H bonds. ANOVA confirmed the sensitive response of selected wavelengths to storage conditions, while the correlation analysis found significant relationship among absorbance readings, weight loss (WL) and soluble solids content (SSC). The SVM models achieved better performance in predicting WL and SSC in terms of R2, RMSEP, and RPD compared to PLSR. The validation revealed that SSC can be predicted using single SVM model for both cultivars with R2 = 0.984 and RMSEP = 0.196 %, while WL reached R2 = 0.853 and RMSEP = 1.188 %. Results suggest that prediction of WL requires individual model for each cultivar while prediction of SSC might work with one model for these cultivars.

Keywords

• Plum
• Postharvest storage
• NIR technique
• Wavelength selection
• Multivariate analysis