Non-destructive detection of the fruit firmness of Korla fragrant pear based on electrical properties

In order to achieve the non-destructive detection of the firmness of Korla fragrant pear during the ripening period, the characteristic variables integrating the parallel equivalent inductance (Lp), quality factor (Q), parallel equivalent capacitance (Cp), dissipation factor (D), parallel equivalent resistance (Rp) and impedance (Z) were formulated through principal component analysis (PCA). Further, based on the characteristic variables, the models were established for predicting the firmness of Korla fragrant pear by using the generalized regression neural network (GRNN) and back-propagation neural network (BPNN). The results showed that firmness has significant correlations with the six electrical parameters. The first two principal components (PCs) were selected as the characteristic variables of the electrical parameters. GRNN exhibited the best performance in predicting firmness (R2=0.9628, RMSE=0.383). The results could provide important references for non-destructive detection of the quality of Korla fragrant pear.