Olive Tree Canopy Assessment by UAV Multispectral Images Before and After Pruning

Abstract

Precision olive growing applications have been increasing in recent years to manage spatial variability and achieve high quality and quantity. Pruning is one of the most important agronomic operations affecting plant vigour conditions. This study aims to provide a new approach for estimating the influence of pruning biomass on the vigour conditions, using multispectral images acquired by Unmanned Aerial Vehicle (UAV). The study was carried out in 2022 winter season, in an olive orchard located in western Sicily (Segesta, Italy). Multispectral images were acquired using a UAV Phantom 4 (DJI, China); in automatic flight configurations. Two flights were performed before (Day of Year 339) and after the pruning (DOY 364) with an altitude at 70 m a.g.l. generating a ground surface distance (GSD) of 3.6 cm/pixel. Side overlap and forward overlap were at 70%, flight paths were generated to minimize the time of flight, gimbal pitch was set at 90°. Starting from the Digital Elevation Model (DEM) and orthomosaic the Object-Based Image Analysis (OBIA) was applied. OBIA was used to perform canopy pixel segmentation from the soil and extract geometric and spectral canopy conditions. Pruning weight was measured in selected plants at the same time of the second flight. Correlation coefficients between geometric (canopy volume and area), vegetation indices (VIs) and pruning weight were calculated. VIs showed high correlation with canopy volume and area providing good prediction. The OBIA methodology demonstrated high accuracy in olive canopy reconstructions and detection in changes of vigour conditions. The trial provided interesting results to support olive growers to improve smart pruning management.