Australian computer scientists plan to develop technology that can accurately estimate the yield of cereal plants.
The team from Adelaide University, together with plant physiologists and an industry partner, will construct computerised 3D models using images of plants as they grow to match the plants’ changing ‘shape’ with its biological properties and predict yield.
‘We are using image analysis to understand the shape of plants so that we can automatically and rapidly measure plant structural properties and how they change over time,’ said research leader Prof Anton van den Hengel, director of the Australian Centre for Visual Technologies (ACVT).
‘We want to be able to predict yield based on a collection of measurable plant attributes early in the plant’s lifespan, rather than having to wait for the plant to mature and then measuring the yield.’
Van den Hengel said this image-based approach would enable the detailed, accurate and rapid estimation of large numbers of plants’ potential yields under various growing conditions, for example high salinity or drought.
‘By expediting the development of plant varieties capable of delivering increased yield under harsh environmental conditions, this project will help to improve Australia’s agricultural efficiency and competitiveness,’ he said. ‘It will help Australian agriculture prepare for the impact of climate change and the need to produce more food for a growing population.’
The image-based analysis will be incorporated into the Plant Accelerator at the university’s Waite Campus. Opened last year, the Plant Accelerator houses more than 1km of conveyor systems that deliver plants automatically to the imaging and other stations.
The project involves industry partner LemnaTec, which provided some of the equipment used in the Plant Accelerator and which will help to commercialise the technology.