Soluble Solids Content and pH Prediction and Maturity Discrimination of Lychee Fruits Using Visible and Near Infrared Hyperspectral Imaging

Abstract

Hyperspectral imaging (HSI) technique has shown promise as a rapid and nondestructive tool to evaluate various internal quality attributes of fruits and vegetables. The objective of this study was to investigate the nondestructive prediction of soluble solids content (SSC) and pH of lychees and maturity discrimination. Two hyperspectral imaging systems of visible/short-wave near infrared range (600–1000 nm, Spectral Set I) and long-wave near infrared range (1000–2500 nm, Spectral Set II) were employed. Results showed that Spectral Set II (SSC: r p = 0.877, RMSEP = 0.911 °Brix; pH: r p = 0.745, RMSEP = 0.291) performed better than Spectral Set I (SSC: r p = 0.790, RMSEP = 1.279 °Brix; pH: r p = 0.701, RMSEP = 0.308) for the internal quality prediction of litchi and maturity discrimination. The partial least square discriminant analysis (PSL-DA) model had a discrimination rate of 90.63 % for Spectral Set I and 96.88 % for Spectral Set II. β-Coefficients of partial least squares regression (PLSR) models were used to choose optimal wavelengths for quality predictions. The performance of optimized PLSR in both spectral sets were comparable to the models developed using the whole spectral range.