Shortwave infrared hyperspectral imaging for detection of pH value in Fuji apple

pH value is regarded as one of the most important attributes that affect sensory characteristics and edible quality of apple. The objective of the research was to explore the feasibility of applying shortwave infrared hyperspectral imaging system to detect pH value of apple. A shortwave infrared hyperspectral imaging system was developed over the wavelength region of 1 000-2 500 nm and used to acquire hyperspectral images of apple samples. After reflectance calibration, mean reflectance spectral was calculated by averaging the intensity of all pixels within the roundness region of interest (ROI). Synergy interval partial least squares (siPLS) algorithms as an effective multivariable method was conducted on the calibration of regression model to estimate the pH value in Fuji apple. The performance of the final model was back-evaluated according to root mean square error of calibration (RMSEC) and correlation coefficient (Rc) in calibration set, and tested in prediction set. The optimal prediction siPLS model was obtained with correlation coefficient (Rp) of 0.8474 and mean square error of prediction (RMSEP) of 0.0398. The results indicated that shortwave infrared hyperspectral imaging combined with siPLS chemometrics could be an accurate and fast method for nondestructive prediction of pH value in Fuji apple.