Rapid Non-destructive Testing of Wheat Moisture Content Based on Dielectric Characteristics and Machine Learning
Hualu Song · 2024
In order to achieve rapid and non-destructive testing of wheat moisture content, this paper measured the capacitance C and conductance G of wheat grains at different moisture contents and temperatures in the hot air drying process. This was done using an impedance analyzer and a dielectric parameter measurement sensor with an excitation voltage in the frequency range of 1 MHz to 10 MHz. The capacitance C, conductance G, and grain temperature T of 10 measurement frequency points from 1 MHz to 10 MHz were selected as input variables of the model. The multiple linear regression model (MLR) and partial least squares regression model (PLS) were employed to establish the prediction model of grain moisture content. The GWO-SVM was employed to construct a prediction model for grain moisture content, with a set $R_{\mathrm{p}}{ }^{2}$ of 0.998, an RMSEP of $0.416 \%$, and an RPD of 20.4. This study offers new research avenues and fundamental data for the development of rapid non-destructive testing of wheat moisture content based on dielectric characteristics and support vector machines.