Application of digital images and multivariate calibration for the quantification of anthocyanin and antioxidant activity in grape juice

Background and Aims The current methods for measuring the anthocyanin concentration (AC) and antioxidant activity (AA) of grape juice have limitations. The objective of this work was therefore to evaluate the application of digital imaging as a simple, fast and low-cost method for determining AC and AA in juice. Methods and Results The AC and AA of 75 grape juice samples were measured using reference methods and then correlated against digital images of 200 dpi resolution obtained with a scanner, using partial least squares regression. The samples were separated with the Kennard-Stone algorithm into calibration sets (50 samples) and external validation sets (25 samples). The models used for the quantification of AC and AA presented a root mean square error of calibration of 8.8 mg/L and 9.8 mu mol and a root mean square error of prediction of 7.5 mg/L and 8.0 mu mol, respectively. The correlation coefficients for the adjustment were 0.8479 and 0.9258 for the models in the determination of AC and AA, respectively. Conclusions Digital imaging and multivariate calibration can be used to measure efficiently selected parameters of grape juice composition. Significance of the Study Analytical methods based on digital imaging offer promising alternatives to conventional methods, since they are fast, simple, non-destructive and employ little or no chemical reagents, thereby reducing analysis costs and facilitating in situ analysis.