Non-destructive quality assessment method for Korla fragrant pears based on electrical properties and adaptive neural-fuzzy inference system

Abstract

This study aimed to achieve the rapid and non-destructive quality assessment of Korla fragrant pears, thereby maximising the value of damaged pears and improving the market supply–demand balance. The electrical characteristics of damaged Korla fragrant pears were measured using a fruit electrical property detector and the trends in the quality of the fragrant pears within their shelf life were analysed. The quality of damaged Korla fragrant pears was predicted by using an adaptive neural-fuzzy inference system (ANFIS). Electrical properties were used as inputs to the system and quality was the output of the system. A total of eight functional models with different degrees of membership were constructed and the optimal one was selected. The results demonstrated that the equivalent parallel capacitance (Cp) of the damaged Korla fragrant pears increased gradually throughout their shelf life, while equivalent parallel resistance (Rp) and complex impedance (Z) decreased. Additionally, the green–red difference (a*) increased gradually, while hardness and soluble solid content (SSC) decreased continuously. Based on the combination of electrical properties and the ANFIS, gbellmf was the optimal model to determine the SSC of the damaged Korla fragrant pears (R2 = 0.9123) and gaussmf was the optimal model to determine the hardness and a* of the damaged Korla fragrant pears (R2 = 0.9112 and 0.8853, respectively). It was concluded that the combined use of electrical properties and the ANFIS method enabled the successful quality assessment of fragrant pears. This approach was conducive to facilitating the optimisation and updating of non-destructive quality assessment technologies for fragrant pears, promoting quality and efficiency improvement in the fragrant pear industry.