Abstract
The visible light vegetation indices (VIs) derived from the red, green, and blue spectral bands of UAV (unmanned aerial vehicle) imagery play a vital role in precision agriculture applications. Nevertheless, the effects of solar elevation angle variations across different flight times remain poorly understood. The DJI Phantom 4 RTK high-precision positioning aerial survey UAV was used to conduct a timed flight over cotton plots with both weak growth without nitrogen application and strong growth with nitrogen application. The visible light VIs for 13 UAVs at 12 different flight times were extracted, and a one-dimensional linear regression model established. By comparing the difference significance and slope values of the models, to evaluate the influence degree of solar elevation angle and cotton growth on the visible light VIs of 13 kinds of UAV, so as to provide a reference for the reasonable planning of UAV flight time under the background of precision agriculture. The results show that: (1) No matter in the test plots with relatively weak or prosperous cotton growth, Solar elevation angle was always significantly positively correlated with the excess red vegetation index (ExR) and red-green ratio index (RGRI). There was a significant linear negative correlation with the excess green minus excess red vegetation index (ExGR), excess green vegetation index (ExG), red-green-blue vegetation index (RGBVI), modified green-red vegetation index (MGRVI), green leaf index (GLI), normalized green-red difference index (NGRDI), and visible atmospherically resistant index (VARI), but there was no significant linear regression relationship with the Kawashima index (IKAW). (2) In the test plot without nitrogen application with relatively weak cotton growth, the linear regression model of visible light vegetation index (green-blue ratio index GBRI, excess blue vegetation ExB, normalized green-blue difference index NGBDI) and Solar elevation angle of cotton field is more likely to reach a significant level. (3) The effect of the solar elevation angle on ExG, ExGR, GLI, ExB, RGBVI, NGBDI and GBRI can be reduced in the test plot with nitrogen application with relatively strong cotton growth, and the effect on ExR, NGRDI and MGRVI can be increased. (4) Solar elevation angle had the greatest influence on ExGR, RGBVI, and MGRVI, and IKAW was the least influential. Therefore, it is suggested in this study that when the visible light VIs of ExGR, RGBVI, MGRVI are applied to the growth monitoring and evaluation of cotton fields, the flight time (or solar elevation angle) of UAVs should be as consistent as possible.