Quantitative Assessment of Hayward Kiwi Soluble Solids Content Prediction Using Hyperspectral Imaging

Soluble Solids Content (SSC) of kiwi fruit is an important internal quality attribute that influences its consumption quality and post-harvest shelf life. In this study, a non-invasive measurement and quantitative prediction of SSC indices in Hayward kiwi cultivars was conducted using hyperspectral imaging techniques. Hyperspectral images of 495 kiwi fruits were collected in reflectance mode between the wavelengths of 400 and 1000 nm. This study employed Partial Least Square Regression (PLSR) for calibration and prediction, combined with various preprocessing methods. Amongst, combination of SGS+MSC+PLS obtained better calibration and prediction outcomes, with ${\text{R}}_C^2 = 0.42$, RMSEP = 1.3023 and ${\text{R}}_P^2 = 0.31$, RMSEP = 1.3211 respectively. This study also looked at the feasibility of employing a Moving Window PLSR (MWPLSR) approach to find the best spectral interval and reduce the impact of multicollinearity in the spectral data. Then, MWPLSR approach delivered the best calibration and prediction values in a spectral region spanning 628-947 nm.