Early prediction of the vine yield by non-destructive method

The vineyard production has always been stamped by its irregular yearly harvest quantity. The yield control is decisive for the harvest organisation and the wine quality.
Different methods try already to make a prediction of the yield by estimating the grape volume at harvest on geographical areas more or less widespread; or counting the inflorescences and the bunch-per-vine or the berry-per-bunch; or the weight of ripe berries or ripe bunches using berry growth parameters…
Control yield is actually done by operators at a different stages of the plant growth but it is very lengthy and boring. Furthermore, the estimation is based on the sampling method of berries in the parcel.

Fundamental requirements and constraints From the existing methods, we notice that the major problems are :
1/ they need historic data acquisition on a long period of time, which is not always available,
2/ they need a too heavy implementation, manually counting grapes, bunches or berries,
3/ the quality of the results provided is limited and not really reliable: no more precise than 10%. This can be explained by the fact that the prediction is realised probably too early (July) because of the lack of human availability later
4/ the robustness of the sensors and equipment
5/ the price of the solutions

• Be an action living tester
• Host a demonstration project
• Interact with potential solutions providers

• Digital Economy

VINOVALIE is a winemaker in the south-west of France. This is a cooperative which aim is to collect, transform and commercialize wine from 400 wine-growers in Cahors, Fronton and Gaillac areas.