Corn kernel moisture content is the most frequently used indicator for kernel maturity prediction and determination of the appropriate mechanized harvesting time. Corn varieties with lower moisture contents and thus faster dehydration of the corn kernels after physiological maturity are suitable for high plant density and mechanized harvesting. However, nondestructive measurement of the dynamic moisture content of the husk-covered corn kernel layer before physiological maturity in the field remains challenging at present. Therefore, in this study, a nondestructive measurement method based on the vector reflection coefficient measurement technique is proposed and developed for in situ monitoring of the husk-covered corn kernel layer moisture in the field. Based on the one-port calibration procedure used for the handheld moisture detection device, quadratic regression models were established for the corn kernel layer moisture using impedance parameters from 51 husk-covered corn samples. Results showed that this method enabled nondestructive estimation of the husk-covered corn kernel layer moisture ($R^2=0.62$) with a 3.53% root mean square error ($\mathit{RMSE}$) using a serial capacitor $C_s$ and the husk moisture ($R^2=0.85$) with a $RMSE$ of 7.12%. Additionally, intact corn ears were measured in the field from the dent stage to the maturity stage using this device to investigate the dynamic moisture contents of the kernels and husks. Results showed that the measured moisture content of kernels before physiological maturity followed exponential decay functions with increasing numbers of days, and the mean difference between the measured kernel moisture contents of the intact corn ears and the oven-drying moisture contents of these kernels is 3.72%. In conclusion, the handheld moisture detection device allows us to measure the moisture of husk-covered corn kernel layers nondestructively before maturity in the field. This study thus provides a new way to obtain corn kernel dehydration curves in situ before maturity and estimate the corn maturity time.

玉米籽粒含水量是预测谷物成熟度和确定合适的机械收割时间的最常用指标。水分含量较低，因此生理成熟后玉米籽粒脱水更快的玉米品种适用于高植物密度和机械化收获。但是，目前在田间进行生理成熟之前，对果皮覆盖的玉米籽粒层的动态水分含量进行无损测量仍然是一项挑战。因此，在本研究中，提出并开发了一种基于矢量反射系数测量技术的无损测量方法，用于现场监测被稻壳覆盖的玉米仁层水分。根据用于手持式水分检测设备的单端口校准程序，使用来自51个被玉米皮覆盖的玉米样品的阻抗参数，建立了玉米籽粒层水分的二次回归模型。结果表明，该方法能够对果皮覆盖的玉米籽粒层水分进行无损估计（${\mathrm{[R}}^2=0.62$），均方根误差为3.53％（$\mathit{RMSE}$）使用串联电容器 $C_s$ 和果壳水分（${\mathrm{[R}}^2=0.85$）与 $\mathrm{[R}\mathrm{中号}\mathrm{小号}Ë$占7.12％。此外，使用该设备在从凹痕阶段到成熟阶段的田间测量完整的玉米穗，以研究籽粒和果壳的动态水分含量。结果表明，在生理成熟之前测得的籽粒水分含量随指数衰减函数随着天数的增加而增加，并且完整玉米穗的测得籽粒水分含量与这些籽粒的烘干干燥水分含量之间的平均差为3.72％ 。总之，手持式水分检测设备使我们能够在田间成熟之前非破坏性地测量被稻壳覆盖的玉米仁层的水分。因此，这项研究提供了一种新方法，可以在成熟之前就地获得玉米籽粒脱水曲线，并估算玉米成熟时间。