In partnership with Botanical Resources Australia and Horticulture Innovation Australia.
Unless you’re an avid gardener, the humble weed may not seem like much of a problem, but to the farming families of Australia, weeds cost the Australian grains industry a staggering $3.27 billion annually in control measures and lost production, according to the Grains Research and Development Corporation.
UniSQ’s Centre for Agricultural Engineering develop solutions for a sustainable and profitable rural sector with research across a broad range of agricultural challenges faced by the growers on the ground. Seeing the need to conquer the humble weed, a field-ready prototype was developed in 2014 which distinguished between the composition characteristics of crops and those of broadleaf weeds in the sugarcane and pyrethrum industries. The project was undertaken in collaboration with Botanical Resources Australia and a provisional patent was lodged for innovative machine vision techniques developed in the project.
Australia’s cotton growers produce enough cotton annually to clothe 500 million people.
‘The new machine-vision-based system is able to distinguish one plant from another,’ UniSQ Researcher Dr Rees says. ‘It uses a range of primary identification techniques including colour, shape, texture, lines, plant height and other 3-D characteristics to differentiate weeds from crop. This means it can accurately identify a plant as a weed or a crop.’
One 250kg bale of cotton lint can produce 215 pairs of jeans, 1200 t-shirts or 4300 pairs of socks.
Australia is the third largest exporter of cotton to the world and the possibilities offered through increased crop yield by intelligent weed detection are valued in the billions.
In an industry that contributes $2 billion annually to Australia’s economy and supports the livelihoods of predominantly family owned and operated growers, the technology offers a low-cost and high-accuracy alternative to existing commercial technologies for spot spraying weeds in a fallow situation.
