Journal
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
Volume 69, Issue 11, Pages 9157-9169Publisher
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIM.2020.2994603
Keywords
Moisture; Ear; Electrodes; Moisture measurement; Kernel; Sensors; Probes; Handheld device; impedance model; intact corn ear; moisture measurement; radio frequency reflection coefficient detection
Funding
- National Natural Science Foundation of China [31771671]
- National Key Research and Development Program of China [2016YFD0300606, 2016YFD0300304]
Ask authors/readers for more resources
Kernel moisture content is an important factor in mechanized corn harvesting. However, the rapid and nondestructive in situ detection of the kernel moisture content within an intact corn ear in the field remains challenging. In this work, we designed a novel handheld device with a double-ring sensing probe for nondestructive and convenient kernel moisture detection in intact corn ears based on an active reflection bridge operating in the radio frequency range. By proposing an equivalent first-order resistor-capacitor impedance model for a load composed of a corn ear and the double-ring sensing probe, we found that the reflection coefficient's phase angle theta was monotonic with the moisture content and also found that the measurement signal frequency affected the corn grain moisture measurement sensitivity. Before using the handheld device in the field, we investigated its optimal signal frequency range and the effects of temperature variations on the measurement results. In the frequency range of 140-165 MHz, the measured voltages from the device showed a good linear relationship and high sensitivity with respect to both the dielectric constants and the corn ear temperature (9 degrees C-22 degrees C). Using this device at a signal frequency of 160 MHz, we conducted corn ear moisture measurements in a total of 73 samples and then established a linear model. The results showed that, for corn ear moisture contents in the range of 15.7%-31.5%, the R values of the training set and the prediction set were 0.7027 and 0.8317, respectively, with a prediction error of 3.54% at the 95% confidence interval.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available