Evaluation of Crop Water Stress Using Drone Images and Numerical Weather Prediction Model Data

Water conditions in soil are measured with soil moisture sensors such as tensiometer and time-domain reflectometry.  However, installed soil moisture sensors may not fully represent the entire cultivation area due to factors such as topography, meteorological conditions, and irrigation systems.The purpose in this study is to identify spatial variations of crop growth and moisture conditions using drone images and weather data. The drone, equipped with multi-spectral, hyper-spectral, and infrared cameras, captured images, and precipitation information up to 3 days later was automatically collected from numerical weather prediction model. Thermal images of crops and soil responded immediately depending on the presence or absence of irrigation. In irrigated crops, leaf temperature decreased due to transpiration. The hyper-spectral images, including short-wave infrared wavelengths, proved sensitive to soil water conditions. However, reflectance-based water indices showed no immediate differences for crops unless soil moisture fell below the wilting point. There was a difference in crop growth depending on the level of irrigation, which was clearly revealed in the vegetation index. Crop growth was poor in areas where irrigation was low. When soil moisture sensor values decrease and no rainfall is expected in the near future, drone images can be utilized to identify specific areas experiencing crop moisture stress. This suggests the potential for drones to support irrigation decision-making.Acknowledgments: This research was funded by the Rural Development Administration, grant number RS-2022-RD009999.