Information technology developments outlined at agronomy conference

New wave of innovation in data collection to improve connectivity and lower costs

Innovation
Austrailan Academy of Technology and Engineering president, Professor Hugh Bradlow, said the 5G mobile network should improve connection reliability in the future. PHOTO Nicole Baxter

Austrailan Academy of Technology and Engineering president, Professor Hugh Bradlow, said the 5G mobile network should improve connection reliability in the future. PHOTO Nicole Baxter

Aa

Future tools to gather ag knowledge at Australia's most recent agronomy conference.

Aa

The latest Australian Society of Agronomy's conference highlighted some of the emerging information technologies that may potentially speed plant breeding efforts and reduce future farm costs.

Professor Hugh Bradlow, the president of the Australian Academy of Technology and Engineering, told the Wagga Wagga, New South Wales, conference the next wave of information technology will enable the creation of a digital facsimile of the real world on a computer that allows researchers to non-destructively test new concepts.

"That would release an explosion of innovation because instead of having to physically try things on animals and plants, researchers will be able to experiment on their digital twins," Professor Bradlow said.

He said the 'Internet of Things' was already using low-powered sensors that can communicate across networks to record all types of field biophysical data such as rainfall and temperature.

"Going forward, the fifth-generation mobile network, known as 5G, will be the nerve centre for everything because it is built to enable sensor networks as well as human networks," he said.

"Poor mobile phone coverage won't be an issue in five years because the 5G network will be hooked into a constellation of low-orbiting satellites that will improve connection reliability."

Going forward, the fifth-generation mobile network, known as 5G, will be the nerve centre for everything because it is built to enable sensor networks as well as human networks. - Australian Academy of Technology and Engineering president Professor Hugh Bradlow

Professor Bradlow said mobile phones were already packed full of small, low-cost and energy-efficient sensors to detect sound, light and location, using machine learning to process and store data 'in the cloud' that is gathered by robots, drones and satellites.

"We have an abundance of data and computing power to enable machine learning, which uses algorithms that can be trained," he said.

"While machines can automate tasks with so-called 'vision' technologies, they don't actually 'learn' but use algorithms that recognise and use static and repeatable patterns.

"Machines using algorithms need to be trained - and if they're updated they require retraining - and provide no explanation as to how they may have arrived at a particular result."

Professor Bradlow said researchers at Iowa State University were searching for corn genotypes that were more drought-tolerant using wearable sensors or 'graphene oxide tattoos' applied underneath leaves to measure electrical resistance.

"Plant moisture uptake is determined by electrical resistance, so the researchers were able to tell which plants were more tolerant to drought than others, which speeds the breeding process," he said.

"Another example of a future tool is the use of technology to detect aphids to find aphid-resistant plants more quickly."

Professor Bradlow said drones capable of sensing light, temperature, biomass and moisture would be important future tools to assist in making large-scale measurements quickly.

One application where these can be used is in precision spraying. However, at present, drone sprayers were too expensive because they rely on a pilot and line of sight.

System automation will in future allow drones to fly precision routes, Professor Bradlow said, and manufacturers are now developing these with machine vision and object recognition capabilities.

He said the other way of collecting broadacre measurements is using satellites that provide 100 per cent visual coverage of the earth, using scanning cameras and spot images.

"In fact, spot images now have such high resolution that they've had to de-tune them to avoid recognising the faces of individuals," he said.

One application that everybody is aiming for in the future, Professor Bradlow said, is new ways to manufacture nitrogen fertiliser.

"The current process consumes two per cent of the entire world's energy," he said.

"If catalysts can be found to lower the temperatures or pressures needed to make fertiliser, the cost of production could be reduced with potential flow-on price reductions on-farm."

GRDC Research Code: COD1905-002AWX

More information: Professor Hugh Bradlow, 03 9864 0900, @hughbradlow (Twitter)

Aa