Minimally destructive assessment of mangosteen translucency based on electrical impedance measurements

• We study electrical impedance spectroscopy of the pericarp of mangosteen.
• Electrical impedance parameters are used to classify translucent mangosteen.
• The model based on relative parameters provided the best classification accuracy.

Electrical impedance spectroscopy in a frequency range of 1 kHz–200 kHz was studied to develop a classifying model for translucent mangosteen. The optimal configuration of the measurement was investigated. Transverse alignment of two measuring needles with the stem–calyx axis and with the measured position on the part of the pericarp pertinent to the largest flesh segment proved to be the optimal configuration. The optimal electrical parameters were selected at frequencies of 1, 4, 7, 8, 14, 47, 73, and 81 kHz as the classifying variables based on the student t-test analysis for a significant difference between the normal and translucent mangosteen and the largest difference of the average values of the electrical parameters. The differences in the electrical parameters and their reciprocals were the optimal classifying variables. The model constructed from the samples from two seasons was robust in terms of seasonality, providing a classification accuracy of 82.7%. The difference in the initial moisture content of the pericarp was justifiably compensated by the differences in the electrical parameters. The EIS technique was suitable for measurement of mangosteen samples at the maturity color stage in which the sample contained no yellow latex in the pericarp.