Investment
Investment
GRDC Code: UOQ2002-008RTX
Machine learning applied to high-throughput feature extraction from imagery to map spatial variability
This project uses Machine Learning to develop high-throughput phenotyping (HTP) of crop canopy features. Plant images from the many project partners train machine learning models for this. The model training is done using the Weiner supercomputer at UQ. These models will be part of edge-computing units, which can take and process images offline. These can be mounted on farm machinery and UAVs, or at the side of the paddock. This investment also works with the GRDC 'CropPhen' investment.
- Project start date:
- 17/02/2020
- Project end date:
- 30/06/2022
- Crop type:
-
- Wheat, (Cereal)
- Barley, (Cereal)
- Sorghum, (Cereal)
- Canola/Rapeseed, (Oilseed)
- Organisation
- The University of Queensland
- Region:
- North, South, West
- Project status
-
Completed
GRDC News
GroundCover Supplement
Enticing data scientists to take on grain...
1725285600000
A global competition has proved to be a novel means of engaging data scientists to...
Deep-learning imaging to picture crop health
1649599200000
Using a technique that has become known for internet ‘deepfakes’, researchers are working to improve...