Showing posts from February, 2018

Remote sensing of vineyard performance

New Winetech funded project 2018 - This project’s objective is to allow researchers to develop a remote sensing - machine learning framework for rapid, cost-effective, real-time monitoring of vineyard performance. The framework could form the basis for further development and extend the utility of the framework that could then be employed in a myriad of applications. These include yield estimation, fruit quality assessment, stress (pests and disease, drought, etc.) detection, and monitoring nutrient status. This study aims to generate models for vineyard performance monitoring using terrestrial LiDAR and terrestrial hyperspectral imaging. Models will be created using advanced machine learning techniques and will be tested across cultivars to assess their operational potential. Additionally, the utility of combining terrestrial LiDAR and terrestrial hyperspectral imaging will be evaluated as a means to improve the assessment of vineyard performance. It is envisaged that a com

Development of a DNA-based method to identify mites

New Winetech funded project 2018 - The aim of this study is to provide the wine industry with a quick, accurate and cost-effective diagnostic test for the presence/absence of grapevine mites in plant material. The test will help with identification of the specific species, strains and populations. In turn, correlating specific genetic groups with plant symptoms and ecology will help to improve management of the pest. Some important grapevine varieties (e.g. Sauvignon Blanc and Cabernet Sauvignon) seem to be particularly susceptible to mite attack. In vineyards, bud infertility, leaf and bunch malformation and shortened internodes may indicate grapevine budmite infestation. However, these symptoms may also have other causes, and microscopy is currently the only confirmation method. Due to taxonomic uncertainty, this process is difficult and often inconclusive. The objective of this project is to design and validate a DNA-based test that will help to identify grapevine mite

On-line web based application for phenolic analysis

New Winetech funded project 2018 -  The main objective of this project is to establish a web based application where efficient management of phenolic data, generated through spectral information, can be achieved. It is common practice today to generate a large amount of data from chemical and physical analyses throughout the winemaking process. The value of this generated data is often not fully exploited and it is therefore not optimally used in winemaking decisions. Handling large and complex datasets is in most cases not possible due to time availability, absence of dedicated personnel, or limited knowledge on data processing. This project will create a user friendly web application that will be used as a tool to obtain phenolic data in real time. This will be achieved through prediction calibrations using UV-Visible spectral data developed at the DVO in previous Winetech funded projects. This web based application will comprise of a phenolic estimation tool in combinat

Analysis and Visualisation of Merlot Tasting Notes

New Winetech funded project 2018 -  Merlot is a widely planted grape varietal in South Africa. In the 2017 Platter Guide however, there are only two five star Merlots and only a few 4.5 star Merlots. South African producers, tasters and consumers are unsure and divided about the taste characteristics of high quality Merlots, and don’t know what to expect from a Merlot. The objective of this project is to identify from tasting notes and wine reviews, the taste characteristics of high, average and low quality single-varietal Merlots. The researchers will use data mining techniques to determine if they can obtain a better description or definition of Merlot characteristics. They will use data from Platter's Wine Guide primary data source, which offers an almost complete coverage of South African wines. They will be analysing data from various editions, due to the small number of highly-ranked Merlots. The researchers will complement this with reviews extracted from accessible ov