Research Associate

Julie will be speaking in Session 9. Fresh Science B. Click here to see more.

Abstract

Grapevine trunk diseases (GTD) are slow progressing fungal infections that cause vines to become less productive and produce fruit of poor quality due to uneven ripening.

GTD is known to heavily affect major varieties grown in Australia and New Zealand such as Shiraz and Sauvignon Blanc. If GTD has progressed significantly, cost of remediation is substantial as entire vines may need to be replaced. Therefore, early detection of GTD is desired to reduce remediation costs. Current identification of GTD requires diagnosis by experts to determine the extent of infection.

Given a number of visual identifiers present for GTD, computer vision and image-based analysis have the potential to provide insights on the GTD extent in a block or vineyard. The above reasons give rise to a SAWIDS funded project as a collaboration between SARDI/PIRSA and UNSW to determine three things in order of level of challenge: Firstly, is detection of GTD from imagery feasible? Secondly, to what extent can we detect GTD? Thirdly, can we detect GTD using on-farm edge computing?

This talk provides an update on this project and how we leveraged manual assessments on more than 13000 vines completed by industry experts.

Biography

Dr Julie Tang has a PhD in Mechatronics engineering focused on viticulture and agritech. Her background is in computer vision and machine learning, for both vineyards and apple orchards as she has spent the last three years as a researcher in a project for variable timing of chemical thinning for apple flowers. She is passionate about technology to support businesses in improving processes for food sustainability. She continues her involvement within the AgriTech space and is currently employed as a data scientist at Bitwise Agronomy and a casual academic at the University of New South Wales.