Fred Baret

　　INRA EMMAH, UMR 1114 Domaine Saint-Paul, Site Agroparc, 84914 Avignon Cedex 9, France

　　The parallel development of civil UAV and miniaturized sensors relevant for vegetation sensing opens the road for high resolution images (0.1 mm to 10 cm) of field crop experiments. The processing of these images generates estimates of variables such as plant density, height or bio-volume of each micro plot, Green Area Index, canopy chlorophyll content or canopy temperature. To facilitate the use of this new technique, INRA (French national institute for agronomical research) and its research partners (UMT CAPTE, PHENOME project) develop a UAV system as well as an image processing pipeline. The UAV is a hexacopter with a maximum payload of 800g and about 20 minutes autonomy that can carry 1) a high resolution RGB camera or 2) a specially developed 6 band multispectral camera (Airphen, Hi-Phen) with an optional thermal camera (Tau2, FLIR). The data processing pipeline uses photogrammetry for precise georeferencing of raw images. For each micro plot, the process extracts the relevant part of the raw images for each micro plot and then computes vegetation indices or invert radiative transfer models such as Pro-Sail to estimate key variable of the canopy. The accuracy with which the canopy characteristics can be retrieved is finally illustrated by comparison with independent measurements.