Abstract:Pitaya is a fruit that exhibits health benefits. The FieldSpec 3 spectroradiometer was used to collect spectral data of pitaya in the wavelength range of 350–2500 nm. Multiple pretreatments, successive projections algorithm (SPA), and partial least squares regression (PLSR) were adopted to establish the active acidity prediction model for pitaya. The experimental results showed that the optimal partial least squares (PLS) model was established after the original spectrum was processed by using the Savitzky-Golay convolution smoothing method (SGS). The correlation coefficients of cross-validation (RCV) and root mean square error of cross-validation (RMSECV) were found to be 0.8862 and 0.1535, respectively. In combination with SPA algorithm, the preferentially selected 25 variables were used to establish the PLS model with a correlation coefficient of prediction (RP) of 0.8702 and a root mean square error of prediction (RMSEP) of 0.1682. The predictive accuracy of the model was higher than that of the model constructed by using 2151 variables from the original spectrum. The effect of the fruit peel on the model accuracy was analyzed. After the optimal normalization pretreatment of spectral data, the RP of the whole fruit PLS model was 0.8151 and that of the RP of the fruit flesh PLS model was 0.8583, which showed that the fruit peel affected the model, but the effect could be reduced by spectral optimization. The results obtained indicate that it is feasible to use diffuse reflectance based on near infrared spectroscopy combined with SPA for the non-destructive measurement of the active acidity of pitaya.
