Data-Driven Dormant Apple Tree Pruning and Tree Vigour Models to Improve Farm Outcomes

Project Lead

Vivid Machines, Toronto, ON

Project Partners

Tall Grass Ventures, Calgary, AB

Gibson Orchards, Newcastle, ON

Blue Mountain Fruit, Thornbury, ON

Project Description

Vivid Machines’ system combines modern deep learning and computer vision approaches with state-of-the-art edge processing technology, providing accurate, plant-level, large-scale assessment of specialty crops in real time, in the field. As requested by its agricultural partners, Vivid Machines is collaborating to build advanced predictive models that will further increase the volume of quality fruit grown on existing acreages, while reducing food waste and the impacts of commercial agriculture on soil and water quality.

Project Innovation

• Develop proprietary datasets: Collect and annotate an initial data set of multispectral representative image frames and ground-truth counts.
• Develop proprietary machine learning algorithms: Create trained, tested, and benchmarked models that can achieve a high level of accuracy in vigour estimation, in bud-type classification, and in bud counting.
• Develop a customer dashboard displaying the vigour and bud types and counts on an individual tree level from processed data along, as well as guidance for the farmer based on predictions for vigour and bud types and counts.

Project Investment

CAAIN Contribution
$820,164

Total Project Value
$2,443,006

Project Contact

Jenny Lemieux
Co-Founder and CEO
Vivid Machines
jenny@vivid-machines.com