Estimation of vineyard vegetative growth: analysis of 3D point cloud from unmanned aerial vehicle imagery

One of the crucial elements for precision viticulture and site-specific management is to assess the spatial variability of vegetative growth for an accurate characterization of vigor and further estimation of yield forecast. Two of the main parameters related to vegetative growth are External Leaf Area (SA) and weight of pruning wood, and both have been traditionally estimated by using methods rely on manual sampling. These methods are time-consuming making it difficult to handle the intrinsic spatial variability of vineyards. The application of remote sensing based on photogrammetric techniques and OBIA (object-based-image-analysis) to images acquired with an Unmanned Aerial Vehicle (UAV) has shown to be an efficient way to derive accurate three-dimensional (3D) canopy information in woody crops such as vineyard, olive or almond. In this context, a set of dense 3D point clouds of every vine was generated using photogrammetric techniques on images acquired by an RGB sensor onboard an UAV in two vineyards with 'Pedro Ximenez' variety drip-irrigated, trellis-trained and managed under organic system. Point clouds were then analyzed by using an OBIA automatic algorithm to accurately assess SA and to study the relationship between weight of pruning wood and vine volume. Results from a non-destructive field sampling and estimated by UAV-imagery were compared. Significant correlations between observed and estimated data were recorded indicating the utility of the procedure developed for an accurate characterization of every vine vegetative growth. This opens the door to progress in digitizing applications in vineyards.