Using an unmanned platform and VIS-NIR cameras to determine biophysical and geometrical parameters of olive, grapevine and citrus canopies

In the last decade, unmanned aerial vehicles (UAVs/UAS/RPAS) have been proposed for agricultural applications due to their great flexibility in flight scheduling and the increasing availability of dedicated miniaturized sensors. The capability of these systems in determining biophysical and geometrical parameters was tested in olive (Olea europaea L., ‘Frantoio’), vineyard (Vitis vinifera L., ‘Sangiovese’) and citrus orchards (Citrus clementina Hort. ex Tan. × Citrus sinensis L. Osb., ‘Mandared’) located at Venturina (LI, Suvereto (LI) and Lentini (SR), Italy, respectively. Images were acquired from an altitude of 50 m above ground level using an octocopter equipped with a consumer camera (RGB) and a multispectral camera (NIR-RG). The normalized difference vegetation index (NDVI) was calculated by means of the map algebra technique. Tree volume and canopy height were derived from the digital surface model (DSM) obtained from the images by UAV and structure from motion technique. Results showed that NDVI was linearly related to LAI (leaf area index) in all the tested species (R2=0.63-0.88). The best performance for tree canopy volume estimation was obtained in olive, with an R2 value of 0.90 and a slope of 0.85. The slight underestimation observed in the UAV-estimated volume of olive trees might be related with its irregularly shaped canopy, which is hardly assessed from the ground. The UAV-RGB camera technique was able to correctly estimate the canopy height of the three species (R2=0.79-0.88; RMSE=0.08-0.22 m). Our results confirm that the combined use of VIS-NIR cameras and UAV platforms is a rapid and reliable technique to determine canopy structure and LAI in orchards and vineyards.