Detection method for the degree of damage to Korla fragrant pears based on electrical properties

To solve challenges in the automated and rapid, non-destructive detection of damage to Korla fragrant pears, this study explored the laws influencing the impact of load on damage to fragrant pears during ripening. A comparative analysis of the external features and microscopic structural changes before and after fragrant pear damage was performed. The electrical parameters of fragrant pears were collected using the fruit electrical parameter detection system, including parallel equivalent capacitance (Cp), parallel equivalent resistance (Rp), and complex impedance (Z). The correlations between the electrical parameters of fragrant pears and the degree of damage were analyzed. A detection model for the degree of damage to Korla fragrant pears was constructed using partial least squares regression (PLSR), support vector regression (SVR), and particle swarm optimization-least squares support vector regression (PSO-LSSVR), and the optimal model was determined and screened. The results showed that, in the same ripening, the damaged area of fragrant pears increased as the falling height increased. Given equal impact loads, the damage area of fragrant pears increased as the picking time increased. Cp, Rp, and Z were strongly correlated with the damaged area of fragrant pears. When the test frequency was 1 kHz, the PSO-LSSVR model showed the optimal detection performance (R² = 0.9172, RMSE = 117.56) for the damaged area of fragrant pears. These research results provide a theoretical reference for the quality assessment and storage regulation of fragrant pears.