A Method for Detecting Irrigation Needs Using a Uav

Since leaf temperature observations in fields have become easier due to the development and popularization of unmanned aerial vehicles (UAVs), new regional irrigation management systems could be developed based on a crop water stress index (CWSI) map created using UAV thermal images. Developing such a system requires a method of calculating CWSI from thermal images. In addition, since CWSI is an irrigation indicator, the UAV must fly when the CWSI reflects suction in the root zone. Moreover, the UAV must fly several times to mitigate the effect of fluctuations in wind speed and solar radiation. Using observations from a maize field in Kununurra, we developed a method to estimate mean temperatures of leaves receiving direct sunlight, which enabled us to estimate leaf temperatures with a mean absolute error (MAE) of 0.5°C compared to observed temperatures. The error of CWSI led by this temperature MAE was <0.07. In the method, pixels in a thermal image of a field were considered leaves receiving direct sunlight if the pixel temperature were between the highest and lowest theoretical temperatures of leaves receiving direct sunlight. The highest theoretical temperature of leaves receiving direct sunlight was calculated based on the energy balance, while the lowest theoretical temperature was calculated based on the ratio of leaves receiving direct sunlight. For the dry season in Kununurra, the optimal flight time of the UAV was 15:00, with a flight frequency of six at that time estimating CWSI within the MAE of 0.02 to the one-hour mean CWSI.