Non-destructive measurement of fracturability and chewiness of apple by FT-NIRS

In order to assess quickly and non-destructively the fracturability and chewiness of apple fruit by FT-NIR spectra in the wavelength range of 4000 cm−1–12000 cm−1, multivariate models were built using multiple linear regression (MLR), partial least squares regression (PLSR), and principal component regression (PCR). Fracturability and chewiness reference data were instrumentally measured using a Texture Profile Analysis (TPA) test. The effects of various pre-processing methods of the spectroscopic data on the performance of the multivariate models were analyzed. Standard normal variate transformation (SNV), multiplicative scatter correction (MSC), Min-Max normalization(MMN) and first derivative (FD) were tested. The performance of the fracturability prediction models was better for the PLSR model (R2 = 0.91, RMSEP = 101.90) than for the PCR and MLR models. With regard to chewiness, the performance of the PCR model (R2 = 0.88, RMSEP = 13.46) was similar to the one of the PLSR model but better than the one of the MLR model. The results demonstrated that NIR spectra together with stoichiometry could determine precisely fracturability and chewiness of apple, and the predictive ability of the models developed by other methods may be improved in the future.