A phone-based imaging tool to measure fruit volume to optimise harvest time
A. Deloire, B. Pailthorpe, J. Gao, N. Bordes, K. Suklje, L. Schmidtke
Background
Prof. Alain Deloire has showed the characterisation of grape berry volume and skin colour in white cultivars could help the harvest decision process, and achieve desirable wine styles (Deloire 2013 - pdf).
Vineyards in the Barossa valley (Photo: N. Bordes)
The Dyostem grape maturity monitoring system helps create a more consistent wine profile year after year. It measures sugar loading of grapes over time, and optimizes picking date based on the desired wine profile. The technique however is expensive, slow and destructive. It is not portable and cannot be deployed in the vineyard.
Dyostem system (photo: vivelys)
More information on Dyostem is available on the vivelys website in French or from the vivelys technical data sheet (pdf in English)
The Challenge
The proposed project addresses the question of accurate, rapid, non-invasive measures of ripeness at the berry and bunch levels. The delivery of information in real time will expedite vineyard management decisions, such as irrigation to limit berry heterogeneity, dehydration, crop loss and eventually fruit shrivelling. This will enable better decision on harvest times and fruit grading, benefiting both the growers and wineries. Improved irrigation management based upon in field fruit measures of colour and size distributions, the latest as indicators of plant water budgets could be obtained saving water and electricity.
Grape for wine production
Objectives
This project aims to develop a portable, phone-based imaging tool that allows growers and winemakers to measure the volume and the colour of grape berries (red and white varieties) in the vineyard, in a non-destructive manner. The growers/winemakers will be able to assess and prevent berry dehydration (to limit crop loss and fruit heterogeneity), and assess the maturation degree of the white grape. Anticipation of harvest in conjunction with weather forecasts, disease pressure, winery and vineyard logistics, and the potential predetermined style of wine based upon berry maturation phase are advantages to the real time in field objective assessment of berry development.
We will develop a mobile phone based software application that will measure the:
- colour of berries (hue angle), calibrated against a standard colour scale for red and white varieties;
- number of berries;
- sizes/volumes of berries; and
- distribution of fruit colour and volume.
We will develop measurement protocols to ensure that the data collection in the field is reproducible and meaningful. Data management/ collection/ storage, traceability across vineyards and vintages will be helpful to gain an understanding of the potential impact of the weather (vintage effect; heat waves), the cultural practices and the potential disease issues on grape colour and volume evolution. This will lead, over years, to the construction of a powerful database on Australian vineyard ripening and balance (bunch homogeneity versus heterogeneity) by using fruit distribution across bunches, vines, vineyards and sites.
Report Update
Contacts
- A. Deloire
- National Grape and Wine Industry Centre, Charles Sturt University
- B. Pailthorpe
- CM3: Research Unit for Machine Learning, Charles Sturt University
- J. Gao
- School of Computing and Mathematics, Charles Sturt University