Main content area

Drought phenotyping in Vitis vinifera using RGB and NIR imaging

Briglia, Nunzio, Montanaro, Giuseppe, Petrozza, Angelo, Summerer, Stephan, Cellini, Francesco, Nuzzo, Vitale
Scientia horticulturae 2019 v.256 pp. 108555
Vitis vinifera, canopy, carbon dioxide, color, drought, image analysis, irrigation, leaf area, leaf water potential, petioles, phenotype, photosystem II, stomatal conductance, vines, water stress
This study examined whether morphophysiological traits (i.e., leaf area, plant water consumption, leaf water potential) of drought stressed grapevines (Vitis vinifera L.) might be determined through the use of non-destructive RGB and NIR image-based analysis techniques for possible implementation of affordable phenotyping. The study was carried out at a centre which is part of the European Plant Phenotyping Network (EPPN) also aiming at contribute to the standardisation of phenotyping protocols. Four groups of 20 potted vines each were subjected to various irrigation treatments restoring 100% (control), 75% (IRR75%), 50% (IRR50%) and 25% (IRR25%) of their daily water consumption within a 22-day period of drought imposition. Leaf gas exchanges, leaf water potential (Ψ), photosystem II efficiency (Fv/Fm), RGB and NIR data were simultaneously collected during drought imposition. Values of Ψ in IRR25% vines reached −1.2 MPa pre-dawn, in turn stomatal conductance and net photosynthetic rate reached values as low as approx. 0.02 mol H2O m−2 s–1 and 1.0 μmol CO2 m−2 s–1, respectively. Through a cross-validation analysis, this study modelled (R2 = 0.78) the estimation of plant canopy area based on the number of pixel of RGB images of vines under various drought levels. Estimated leaf area was employed to calculate water consumption per unit leaf area, which resulted correlated (R2 = 0.86) with Ψ. Results revealed a correlation between Ψ and Dark Green colour class (R2 = 0.71) and suggest a new working hypothesis concerning the phenotyping of leaf (or petiole) angle. NIR and Dark Green colour fraction decreased with increasing levels of drought while the Yellow one increased. The outcomes presented may strengthen the role of RGB and NIR based images to identify the occurrence of water-stress in Vitis spp. and contribute to both the standardisation of phenotyping protocols pursued by the global phenotyping community and the possible development of new tools for precision irrigation in a HTP domain.