Monitoring maize canopy chlorophyll density under lodging stress based on UAV hyperspectral imagery

Abstract

Lodging causes severe decreases in crop yield, reduces grain quality, and increases the difficulty of mechanical harvesting. Obtaining the spatial distribution information of maize lodging grades in a timely and accurate manner is essential for yield loss assessment, post-stress management, and insurance claims settlements. The purpose of this study is to explore the ability of unmanned aerial vehicle (UAV) imaging technology to monitor maize lodging stress. With the support of maize lodging control experiments, the canopy chlorophyll density (CCD) of maize populations under stress from different lodging grades was used as the characterization index. The responses between hyperspectral characteristic parameters and CCD with different lodging grade stresses were analyzed. The monitoring model of the maize CCD under lodging stress was constructed using the sensitive characteristic parameters of original canopy spectra (OCS), first-order differential (FOD), wavelet coefficient (WC), and vegetation index (VI). The results showed that the reflectance of the stalk was significantly higher than that of the leaf in hyperspectral imagery, which was the main reason for the change in the original canopy spectra under lodging stress. The original canopy spectral reflectance increased with the severity of lodging stress. The accuracy of the CCD model was VI > WC > FOD > OCS (R2 = 0.63, 0.61, 0.59, 0.57, respectively), in which the accuracy of VI was the highest (R2 = 0.63, RMSE = 0.36 g/m3). This is because CCD considers not only the change in canopy spatial structure after maize lodging, but also the change in physiological activity of maize plants under lodging stress. The maize lodging grades were evaluated according to the CCD model based on the UAV hyperspectral imagery.