Monitoring of Farmland Soil Moisture Based on Unmanned Aerial Vehicle Multispectral Data

In order to improve the efficiency of precision farmland management, a method to quickly monitor the soil moisture content of farmland is designed, based on remote sensing data transmitted and obtained in real time with unmanned aerial vehicle (UAV). Firstly, multispectral images of farmland were obtained through UAV and a representative area of farmland was selected to detect the soil moisture content of random samples. Then, the perpendicular drought index (PDI) method, combined with the soil moisture data of samples, were used to construct an inversion model of soil moisture, with which the data of the soil moisture of a wide range of farmland were finally obtained. UAV images and soil moisture data in 6 different periods were used for the inversion accuracy analysis and the method validation. The results showed that this method is of high precision in soil moisture monitoring, with the coefficient of determination (R²) being more than 0.8 in 6 periods and the root mean square error (RMSE) and the systematic error (SE) being both less than 0.1 in 5 periods. This method of soil moisture monitoring based on UAV data had its validity and feasibility proved and could provide reference for rapid monitoring of large-scale farmland soil moisture.