Enhancing Transferability of Near-Infrared Spectral Models for Soluble Solids Content Prediction across Different Fruits

Near-infrared (NIR) spectroscopy is widely used for non-destructive detection of fruit quality, but the transferability NIR models between different fruits still a challenge. This study investigates from strawberry to grape and apple using two case studies. A total 94 strawberry, 80 grape, 125 samples were measured their soluble solids content (SSC) spectra. Partial least squares (PLS) regression was establish model predicting SSC, with an acceptable root mean square error prediction (RMSEP) correlation coefficient (R) 0.53 °Brix 0.91, respectively. Directly applying spectra significantly degrades performance, increasing RMSEP up 3.47 16.40, Spectral preprocessing can improve predictions all three fruits, bias cannot be eliminated. Global modeling produces generalized model, degrades. Calibration transfer SS-PFCE PLS correction, which are calibration methods without standard samples, found effective way model. Therefore, may feasible improving multiple fruits.