DETECTION OF PLANT WATER STRESS USING UAV THERMAL IMAGES FOR PRECISION FARMING APPLICATION

Adequate and real-time monitoring of water stress is critical to enhance productivity, crop quality, as well as water use efficiency. This study contributes to the new approach of precise and rapid estimation of real-time water stress using thermal images taken with an Unmanned aerial vehicle. Different physiological parameters stomatal conductance (gs), leaf area, and ground reality parameter (Tc) were calculated between 11.30 and 13.30 (Chinese standard time) on sampling day. The volumetric water content (theta, m(3) M-3) of the soil at different depths of (20, 40, and 60 cm) was measured. Data processing steps were implemented in MATLAB for thermal images to calculate the canopy temperature T-1. Empirical (CWSle) and statistical (CWSIs) methods of CWS1 were applied for model calibration. Results showed that different spectral indices (TCARI, NDVI, OSAVI TCARI/OSAVI) had a high correlation with stomatal conductance (gs) (R-2 = 0.590) and transpiration rate (tr) (R-2 = 0.602) as compared to CWSle and CWSls. Volumetric water content (theta) and CWSIsi have a high correlation coefficient (0.872). However, the transpiration rate shows a week correlation with spectral indices (TCARI, NDVI, OSAVI, and TCARI/OSAVI) as compared with CWSI. The plotted high-resolution map shows the distribution of water stress in different irrigation treatments and potentially applied in precision irrigation management.