Quantitative evaluation of impact damage to apples using NIR hyperspectral imaging

Impact damage to apples is one of the most crucial quality factors and needs be detected in postharvest sorting processes. In this study, impact ‘Red Fuji’ apple fruit was investigated quantitatively by hyperspectral imaging technology. A total 240 samples were prepared with six groups for different degrees. The technique based on near-infrared (NIR) spectrometry range 900–1700 nm used measure mechanical parameters, such as average pressure, contact load, damaged area, absorbed energy, firmness. Four types spectral pre-treatment, including standard normal variate, multiplicative scatter correction, first-order derivative, second-order adopted improve model’s predictive performance. quantitative relationships between spectra parameters successfully modeled partial least squares (PLS) regression. For apples, raw data without pre-treatment performed better than those after pre-treatments. model, characteristic wavelengths selected Savitzky–Golay derivative (SG 2nd Der) competitive adaptive reweighted sampling (CARS) method. results indicate that CARS-PLS regression model produced SG Der-PLS model. good prediction performances presented coefficient determination (RP2) root mean square errors (RMSEP) values. RP2 RMSEP firmness are 0.66 0.02 MPa, 0.86 53.80 N, 0.83 116.37 mm2, 0.81 0.24 J, 0.64 0.19 respectively. This study demonstrates potential NIR a highly accurate way predict apples.