Use of very high-resolution airborne images to analyse 3d canopy architecture of a vineyard

Differencing between green cover and grape canopy is a challenge for vigour status evaluation in viticulture. This paper presents the acquisition methodology of very high-resolution images (4 cm), using a Sensefly Swinglet CAM unmanned aerial vehicle (UAV) and their processing to construct a 3D digital surface model (DSM) for the creation of precise digital terrain models (DTM). The DTM was obtained using python processing libraries. The DTM was then subtracted to the DSM in order to obtain a differential digital model (DDM) of a vineyard. In the DDM, the vine pixels were then obtained by selecting all pixels with an elevation higher than 50 [cm] above the ground level. The results show that it was possible to separate pixels from the green cover and the vine rows. The DDM showed values between −0.1 and + 1.5 [m]. A manually delineation of polygons based on the RGB image belonging to the green cover and to the vine rows gave a highly significant differences with an average value of 1.23 [m] and 0.08 [m] for the vine and the ground respectively. The vine rows elevation is in good accordance with the topping height of the vines 1.35 [m] measured on the field. This mask could be used to analyse images of the same plot taken at different times. The extraction of only vine pixels will facilitate subsequent analyses, for example, a supervised classification of these pixels