Greater computing power, more connected data and AI as a partner are all part of the evolution of digitally networked agriculture
The grains industry is on the cusp of making major productivity gains with the nexus of autonomous technologies now coming together, according to Ben Kelly, marketing manager for John Deere Australia. Speaking at the GRDC’s recent Grains Research Update in Perth, Mr Kelly said while many loved the “really cool tools” of autonomous technologies, their primary aim was to make growers more profitable and sustainable.
Machines communicating with each other and gathering data were already putting more power into growers’ hands for improved decision-making. This would continue to accelerate with the increase in data, along with improved computing capacity to utilise this data. He expected the on-board computing on machinery to double over the next six years.
This would enable increasingly fine-scale real-time camera and sensor data collection to inform automated and precision activities such as automated on-the-go variable-rate applications, and potentially depth of seed at sowing.
Improved resolution
Over the past 30 years, precision agriculture has advanced from an overarching farm management level to the paddock level, to zones within paddocks and, most recently,
to the plant level.
Over the past decade, there have been significant improvements in camera speed and resolution, which have advanced from capturing one frame per second to 15 frames per second.
“That makes a massive difference when we’re looking for weeds, or these tiny variations we’re trying to solve. And the processing time to look at that image is much quicker,” Mr Kelly said.
“So, sensors are becoming higher resolution, they’re generating more data, we’re able to be connected with the machines, and get more and more of that data off multiple machines.”
AI in agriculture
He said artificial intelligence had also taken a leap forward, allowing a systems-based approach, connecting all the data together to deliver insights that growers could act on.
“The concept is that we’re able to connect people, equipment and data across the whole production system.
“Today, we’re optimising the harvester ground speed. Maybe we can start to look for weeds and map them at harvest.
“We can share that weed map with the sprayer to ensure we hit those weeds. But maybe if we already sprayed those weeds, and now we’re seeing them at harvest, we are being alerted to a resistance problem.”
Mr Kelly said all this data could be run through algorithms that could deliver intelligence that would continually make machines ‘smarter’.
“There are also ways in which this data will become linked. All these technologies are coming together and AI can aid in the analysis of all these data layers.”
Mr Kelly said the aggregation of anonymous datasets from many growers enabled the creation of massive data pools for training AI technologies, which could then also be applied by the individual grower to their own data.
Training an AI partner
He explained that AI systems would allow growers to query their data. For example, what if growers could ask questions about the relationship between protein, yield, nitrogen and soils on their farm, or where they could optimise nitrogen applications.
“And it will look at all that data and start to come back with suggestions. It may not be right all of the time today, but I guarantee it’s going to start getting better pretty quickly.
I think that is going to be a game changer for not only agriculture but every industry
out there.”
With the increase in automated equipment, Mr Kelly said connected networks would also help to manage the movement of vehicles across farm operations: when fuel was needed, or spray tanks needed refilling, where and when the next mother bin or silo bags were needed at harvest.
And increasingly, robotic equipment would not only provide this information to growers, but also would act on it autonomously.
Five steps for future farming
In summary, Mr Kelly said there were five steps for growers to follow to prepare for automation.
- Start digitising the farm. It’s critical. Autonomous machines will need a digital twin of the farm.
- Leverage connectivity. Whether that’s cellular or satellite connectivity, start leveraging it, but continuous connectivity will come.
- Maximise connectivity in technology. We have some of the best adoption of technology in Australia. We have a huge opportunity to utilise this effectively.
- Establish boundaries. Autonomous equipment is going to use a boundary to know where it is, what it’s doing, what’s there.
- Start documenting field activities. These layers of data and having big sacks of data later in the field are going to become hugely valuable.
