Using RGB images and LiDAR data to characterise fruit-to-leaf ratios in grapevine collections

Abstract

One of the main effects of global warming is an increase in the sugar concentration of grapes at harvest time, resulting in wines with a high alcohol content and an unbalanced structure. The fruit to leaf ratio is a key factor in determining the final sugar concentration, and training systems and management techniques can help to control this parameter. There is also variability in the ability of grapevine genotypes to accumulate sugars, which depends on the ripening date, the fruit/leaf ratio, but also on unclear mechanisms at berry level. Indeed, some varieties have been described as ‘low sugar accumulators’, but the genetic determinism(s) of such a trait remains to be described.

When describing genetic variability for sugar accumulation, it is easy to take into account veraison dates and yield levels, but it is difficult to characterise leaf volume and activity across a large number of genotypes.

To gain access to such a trait, we have developed a high-throughput digital phenotyping system. Image analysis based on RGB images and cloud points from LiDAR data is used to obtain estimates of canopy characteristics.

This system has been used to characterise several varieties in the germplasm collection and in breeding populations at INRAE, Colmar, as well as the kinetics of sugar accumulation.

We will present how these data can be used to estimate the effects of fruit/leaf ratio on the rate of sugar accumulation and to identify genetic effects that are independent of development stage and fruit/leaf ratio.