Eyes on the ground

May 04, 2015

Automation and advanced technologies hold the key to a more productive, biodiverse and safe Australian landscape, Jude Dineley reports.

Dog ‘Facebook’ to manage Aussie pest problem

Facial recognition technology is  being used by the Invasive Animals CRC  to identify, track and control  wild dog populations, which cause  significant damage to Australian farms.
Facial recognition technology is being used by the Invasive Animals CRC to identify, track and control wild dog populations, which cause significant damage to Australian farms.

It’s estimated that wild dogs cost Australian farmers more than $65 million each year – a small part of the estimated $1 billion annual price of animal pests to agriculture. Pest monitoring is an important part of ensuring control strategies are effective, and automated technologies that promise more efficient and detailed monitoring are under investigation.

Southern Downs Regional Council in Queensland is working with Australian agricultural tech company Ninox Robotics to spot wild dogs and other pests in their region. The project involves using unmanned aerial vehicles (drones) equipped with thermal imaging cameras, which can map dozens of square kilometres of countryside in a few hours.

The Invasive Animals CRC (IA CRC), NSW Department of Primary Industries and CRC partners have developed camera trap technology with facial recognition software – similar to that used by Facebook to tag your friends – to identify individual dogs and help combat the wild dog problem. Initial tests in northern NSW were able to pinpoint individual dogs with 87% accuracy. The researchers are seeking further funding to turn the technology into user-friendly software for widespread use.

Future versions could monitor other pests including feral cats, and threatened species, says IA CRC researcher Paul Meek. “Technology is providing us with new opportunities to carry out research and management,” says Meek. “And it’s already changing the way we do things.”


Drones streamline cattle musters

iStock_000035347982_LargeMustering cattle on large Australian stations is a time consuming, expensive and sometimes dangerous operation. Before mustering can begin, graziers need to locate livestock using helicopters, horses, quadbikes and motorbikes, sometimes setting up remote camps.

By mapping the cattle’s location, drone technology under development by the CSIRO could potentially halve mustering costs, says project leader and farming systems specialist Dr Dave Henry. Using an off-the-shelf drone and thermal camera, the researchers accurately located cattle on the Lansdown Research Station near Townsville in 2013, and they are seeking funding for large-scale trials – the next step towards a marketable product.

“Technology is providing us with new opportunities to carry out research and management.”

Using sensors, drones could also monitor feed in paddocks, optimising animal production and minimising environmental impact. “Ultimately, graziers and land managers could manage cattle and their environment, and their whole farm business, in a more precise, timely and informed manner,” says Henry.


Satellites drive precision tractors

Precision agriculture uses sensing technologies, from satellites to drones, to help automate tasks like sowing and harvesting. The benefits of satellite positioning in agriculture are substantial, with an analysis by Allen Consulting predicting it will pump up to $28 billion into the Australian economy by 2030.

Improved satellite positioning in agriculture will yield greater navigational accuracy for unmanned farming vehicles such as drones and automated tractors.
Improved satellite positioning in agriculture will yield greater navigational accuracy for unmanned farming vehicles such as drones and automated tractors.

A collaboration including the CRC for Spatial Information (CRCSI) and the Japan Aerospace Exploration Agency has developed positioning technology for a driverless tractor using GPS and the Japanese Quasi-Zenith Satellite System (QZSS). In summer trials in the Riverina, NSW, the tractor navigated rows of crops to an accuracy of 5 cm.

Existing technologies rely on mobile phone coverage and a costly, dense network of ground-based antennas called reference stations. These improve the accuracy of the machinery’s satellite-derived position from several metres to a few centimetres.

But mobile coverage and expensive antennas “are barriers to adoption in remote Australia,” says Dr Phil Collier, CRCSI research director. The researchers’ alternative requires fewer reference stations, instead transmitting position corrections to the tractor via a satellite communication channel unique to QZSS. This approach promises multiple benefits for farmers in remote areas.
Traversing the same ground each time, the tractors use less fuel and reduce erosion. The day may even come where fleets of robotic tractors work overnight, says Collier.


Managing bushfire threat

Automation can also play a major role in predicting and managing the threat of bushfires. Typically, emergency services and researchers rely upon observations by satellites, from aircraft and on the ground.

Drones could provide valuable extra data, says Dr Thomas Duff, a Bushfire & Natural Hazards CRC researcher at the University of Melbourne who specialises in simulations that predict fire behaviour. In contrast to helicopters, unmanned vehicles eliminate risks to pilots, and are cheaper and more manoeuvrable, enabling more detailed observations.

With Country Fire Authority Victoria, researchers at the CSIRO
are using drones to make observations of controlled fires for use in bushfire simulations. The RISER (Resilient Information Systems for Emergency Response) collaboration based at the University of Melbourne is monitoring grasslands to better understand how they dry out each year. Duff says this research is critical to more accurate predictions of fire behaviour.

invasiveanimals.com

crcsi.com.au

bnhcrc.com.au

Related stories

Leave a Reply

Your email address will not be published. Required fields are marked *