Detection of Winter Wheat Panicle Relative Water Content Using Hyperspectral Vegetation Indexes

Timely assessment of panicle relative water content in wheat is critical for harvest diagnosis and precision management of wheat to generate higher yield and better quality. The objective of this study was to determine the most suitable index using wheat panicle hyperspectral data for the assessment of wheat panicle relative water content. Ground‐based hyperspectral datasets were obtained during the heading‐maturity growth stage under eastern and western Guanzhong wheat eco‐sites in Shaanxi Province using the FieldSpec3 Pro spectrometer. This study systematically analysed the performance of three newly developed types of two‐band hyperspectral spectral indexes and previously published spectral indexes to estimate wheat panicle relative water content. Results showed that the newly developed two‐band index NDVI (R 1890 , R 2134 ) in eastern Guanzhong wheat eco‐sites and NDVI (R 1225 , R 1302 ) in western Guanzhong wheat eco‐sites performed best in estimating wheat panicle relative water content at the panicle scale in wheat, generating the coefficients of determination, root mean square error and residual prediction deviation values between the measured and predicted values of 0.98 and 0.98, 2.95% and 2.69%, 6.98 and 6.45, respectively. This study indicates that the relative water content of winter wheat panicles can be precisely estimated using panicle hyperspectral vegetation indexes, which aids in monitoring the wheat maturation phase.