Spatial Mapping of Soluble Solids Content Variability in Kiwifruit Using Hyperspectral Imaging and Chemometrics

The Soluble Solids Content (SSC) of kiwifruit is a vital quality attribute influencing its edibility and shelf life. This study investigates the feasibility of mapping SSC across five internal regions (Regions I to V) of kiwifruit, from center to outer pericarp. Hyperspectral reflectance images (400-1000 nm) of 500 kiwifruits were collected, and an algorithm was developed for region-based spectral point selection. A thirdorder tensor framework organized spectral data by region. Chemometric modeling used Partial Least Squares Regression (PLSR) with preprocessing methods: Standard Normal Variate (SNV), Multiplicative Scatter Correction (MSC), and second derivative (D2). Region III, representing the fleshy part, yielded the best performance, with MSC+D2+PLSR combination achieving RC2=0.485, RMSEC=1.216, and RP2=0.349, RMSEP=1.349. This highlights the potential of hyperspectral imaging and regional analysis for SSC prediction in kiwifruit.