AutoWeed technology on display at Goondiwindi field day

Brendan Calvert, Peter Ridd and Alex Olsen with the AutoWeed technology at Friday's trial at "Willaroo".
Brendan Calvert, Peter Ridd and Alex Olsen with the AutoWeed technology at Friday's trial at "Willaroo".

Local farmers were able to witness a "game-changer" in the war on weeds at a field day near Goondiwindi on Friday.

The AutoWeed robotic spot-spraying technology is a world first in many regards, described as using similar features to Facebook's facial recognition, but to identify weeds.

The project by James Cook University began by investigating new robotic tools for efficient weed control via herbicide application. They then developing a prototype robotic spot-sprayer for pass-over foliar spraying of in-fallow weed regrowth.

In partnership with North West Local Land Services, the team have been taking aim at Harrisia Cactus in the region. This saw the collection of over 2000 images of Harrisia Cactus from the "Mumbulu" and "Willaroo" using their WeedLogger dataset collection instruments.

After a brief talk about the project at the field day last week there was a demonstration spraying of Harrisia Cactus in a heavily infested paddock at "Willaroo". The machine successfully sprayed all targets with the local landholders looking on. The audience included local landholders from the Goondiwindi and Boggabilla region, North West LLS representatives, and some representatives from other tech startup companies.

Systems engineer, Alex Olsen, said the feedback from locals was great. "It was the first time they'd seen a robotic spot spraying system accurately spray Harrisia Cactus automatically. They were excited to see when the 'green-on-green' technology would be available as a product for them to use on their crops," he said.

AutoWeed aims to deliver a robust green-on-green detection and spray system that is applicable to any crop or grazing weed species; and that can be retro-fitted to available agricultural machinery at a reasonable price point for landholders. "We are seeking investment to accelerate the delivery of this platform and are hopeful of delivering a product to market in 2020," Mr Olsen said.

He said many trials across different locations had been conducted targeting eight nationally significant weed species: Chinese apple, Lantana, Parkinsonia, Parthenium, Prickly acacia, Rubber vine, Siam weed and Snake weed.

In partnership with the QLD Department of Agriculture and Fisheries AutoWeed demonstrated theirr robotic detection and spray system on Navua Sedge, a grass-like pasture weed, in Malanda, QLD. "Our successful spraying trial of Navua Sedge, a difficult weed for even the most experienced landholders to identify, gives us the confidence that our detection system is capable of targeting any weed in any environment," he said.

You can watch the AutoWeed trials on YouTube.