TECHNOLOGY is a space in livestock research which tends to pique producer interest whenever the topic is brought up at industry events.
Such was the case at last month's Farm Smart Showcase at Muresk Institute, Northam, when Mark Ferguson, from New Zealand-based neXtgen Agri, gave a presentation on unlocking potential in the sheep industry through technology.
Dr Ferguson is one of a team of researchers and organisations collaborating on developing technology for use in the livestock space and he took the opportunity last month to lay out some of the projects he and his WA-based colleagues at Muresk, Murdoch, Australian Wool Innovation and Department of Primary Industries and Regional Development are working on.
"I think it's the single most exciting time to be involved in agriculture, particularly livestock," Dr Ferguson said.
"We've got little bits of data coming off farms at the moment, but that figure is about to increase exponentially and there is going to be massive amounts of information available soon.
"We don't need to be scared about this - I think the augmented era is something we should lean into.
"We're only limited by our imagination in terms of what technology can do but we have to be careful about getting carried away with how cool the tech can be because it has to solve real problems.
"We are 100 per cent focused on the big ticket items that we think will make farmers money."
The research Dr Ferguson discussed with attendees at last month's event was largely focused on applying machine learning to sheep enterprises by feeding data to computers which then develop algorithms which can make accurate predictions.
Researchers built a crate with four cameras installed which took images of a sheep from side, front, above and behind and then ran 4000 head through the crate which also had scales installed to collect 400 images per animal.
"We had a few questions we wanted to ask with those images at our fingertips," Dr Ferguson said.
"Could we do facial recognition, could we weigh those sheep via camera and could we record other things, like face cover or wrinkle score via camera?
"The answer to could the computer identify an individual sheep from those images was yes, with nearly 100pc accuracy.
"That opens up a whole heap of options such as being able to identify a sheep from a distance which is exciting because as yet we've only been able to identify a sheep at close range, so as this research develops, the machine will get smarter and smarter, which will take us into some pretty exciting spaces."
The machine could also predict liveweight via camera within 10pc of the actual liveweight.
"That's a pretty encouraging piece of data because of the opportunities that could provide on-farm if, for example, we could predict the liveweight of sheep while they're at a water trough," Dr Ferguson said.
He is also interested in subjective wool traits and live intramuscular fat scanning in sheep.
"We're starting to build a database of those things going forward as well," he said.
Dr Ferguson's presentation also touched on smart tag and sensor technology.
"The tags being used have bluetooth capability and an accelerometer which is basically a fit bit so we're getting lots and lots of data about the movement of sheep," he said.
"There's heaps of potential there because so much data is being released every second and we can see patterns in the data thanks to machine learning.
"Through that accelerometer data and machine learning, we can start to predict things like grazing behaviour and feed intake which is ongoing work at this stage."
Dr Ferguson said the focus was getting all the base measurements worked out through the sensor technology, that is when sheep are grazing, sitting, ruminating and lambing.
"Once we've got that base data, then we can start thinking about how we would then apply that to a sheep production system where we might have alerts for lambing or when they're hungry or wormy," he said.
"There are a whole lot of opportunities there."
If you were wondering what sort of impact this work could have on the industry economically, Dr Ferguson quoted some modelling work which valued the benefit to industry at $300 million.
"One of the big ticket items is managing stocking rate and getting it up as every consultant will tell you stocking rate is an important driver of profit," he said.
"Actually managing stocking rate in an effective way is tough though, mainly because we haven't got infinite labour but sensors give you another set of eyes and bring a lot of other opportunities with them.
"So I think we'll see a future where we'll have all this data and it won't come to us in spreadsheets, it'll come to us in answers."
GRAZING SHEEP IN THE AUGMENTED ERA
GRAZING Bytes is one of the projects Mark Ferguson referred to in his presentation, which is work being lead by Andrew Thomson and his team at Murdoch University.
The project, which kicked off in April of this year, is aimed at providing the necessary data and analysis to allow Digibale smart tags to become an integral part of data-based decisions around grazing and sheep management on-farm.
The idea being that by collecting data from the smart tags, researchers can train machine learning algorithms which will be able to accurately predict feed on offer and grazing behaviour of sheep.
Functional application of those algorithms will help producers optimise grazing strategies in real time, therefore achieving optimum pasture and animal performance.
Researchers have flagged many other potential applications and benefits to this sophisticated use of farm data, from economics to animal welfare, but increasing the appeal of sheep and wool production through the easier management which comes with the ability to remote monitor stock via smart tag technology has also been raised.
"Any system that enables improvements in labour use efficiency or technologies that remove the need for human intervention will have a major impact on sheep farm profitability and increase the appeal of sheep and wool production to producers that may have displaced sheep, especially Merinos, during the past 10-20 years," Dr Thompson said.
* If you're interested in learning more about predicting grazing behaviour and the Grazing Bytes project, team members working on the project will be at Southern Dirt TECHSPO at Katanning next month.
