This project is supported by funding from the Australian Government Department of Agriculture, Water and the Environment as part of its Rural R&D for Profit program.
Vineyards can often be busy places, but two blocks in McLaren Vale have seen abnormal levels of activity in recent months.
The local Accolade Wines viticulturists have hosted waves of CSIRO scientists who are determined and equipped to measure all manner of things in all manner of ways.
It’s part of two complementary projects, supported by Wine Australia and the Australian Government’s Rural R&D for Profit program, designed to explore how to effectively and efficiently use modern sensors to provide data to inform vineyard management strategies.
The leaders are Dr Everard Edwards and Dr Mark Thomas from CSIRO Agriculture and Food. Edwards is focusing on vine canopy, including size, composition and management, while Thomas is interested in issues around the fruit, including composition, disease and yield estimation.
But there are more than a dozen CSIRO people involved to some degree, including specialists from its Manufacturing and Energy business units and from the Quantitative Imaging Group and the Robotics and Autonomous Systems Group within Data 61, which is Australia’s largest data innovation group.
The early work – primarily in McLaren Vale but also in the Rymill and Katnook Estate vineyards in Coonawarra – has focused on gathering as much data as possible using an array of options, with a view to selecting and fine-tuning those that warrant more detailed testing next season.
Some trials have been specific to one or other project but others, including work looking at the potential of hyperspectral imaging, have been important to both.
Edwards has had a particular focus on comparing different techniques for assessing canopy size and structure. That has seen him alternating between two different types of LiDAR (Light Detection and Ranging) systems – one newly developed by CSIRO – a drone fitted with cameras and even a basic Go Pro.
‘They all worked, they all produced results, and now we are in the process of comparing them and working out the trade-off between cost, difficulty, resolution and quality of data’, he said.
One of Dr Edwards’ aims is to be able to provide growers with data on their canopy size and density in order to better manage it, particularly around assessments of vigour across the vineyard.
‘Down in Coonawarra’, he adds, ‘they might also be interested in using that data to inform on any leaf plucking and thinning they do. Down the track the potential is that you could use this information to automatically turn a leaf plucking machine on or off or adjust it to get what you want. But that is a way off yet.’
Dr Thomas has similarly been doing intensive early work on yield analysis, trialling a range of camera and monitoring options, and crossing both projects taking imagery of different plant parts and categorising them.
‘If you take a hyperspectral image it can identify each pixel on that image as to whether it’s fruit, flower, leaf, cane or tendril. That has the potential for being used in yield assessment’, Dr Edwards said.
The first phase of work has also focused on how to move more of the trials from the lab to the field in the next phase. A question of light is one such example.
‘In the lab we have an artificial light source and plenty of light across all wavelength ranges that are used, whereas in the field we either need an artificial light source as well, which requires more power and is a bit more difficult, or to rely on sunlight, which has lots of gaps where the atmosphere absorbs those bands.
‘One thing we are doing is testing some of the data analysis with those bands taken out of the spectra, so we can better assess before going into field whether sunlight will be OK.’
And has it been getting a bit crowded out in the field? ‘It’s not too bad. We don’t have any permanent structures and we fit around them in terms of things like spraying when they need space.’