Multispectral imagery acquired from a UAV to assess the spatial variability of a Tempranillo vineyard

Abstract

Unmanned aerial vehicles (UAV) are a novel and flexible tool in precision viticulture, as they can be used to acquire useful information to evaluate the spatial variation of vegetative growth, yield components and grape quality. In this work, the capability of multispectral imagery acquired by a UAV and the derived spectral information to assess the spatial variability of a Tempranillo (Vitis vinifera L.) vineyard has been explored. The study was conducted in a vertical shoot positioned (VSP) Tempranillo vineyard of 3.5 ha located in Navarra, Spain. With the aim of establishing relationships between field variables and the remotely sensed information, a grid of 74 experimental blocks at 20 m intervals was built. Several variables related to the vegetative growth and yield were measured prior to harvest in each block. Multispectral imagery was acquired at 17 cm spatial resolution with a Mini-MCA 6 (Tetracam inc, USA) camera which contained information on 6 different spectral bands located in the visible and near infrared spectral regions. After geometric and radiometric corrections, the images were combined to generate a mosaic of the whole vineyard and pixel values across bands were extracted for the corresponding experimental blocks in the field. Spectral bands and indices derived from the aerial images were correlated with vine vigor and yield variables and their statistical significance was tested. Overall, the study has shown that, while spectral information and derived indices were effectively correlated to vine vigor and yield parameters, the correlation was lower than expected. We hypothesize that a specific post-processing chain will overcome the inherent systematic and random noise and distortions of the original images thus improving correlation with field data.