By Matthew Denholm, rural and regional correspondent
This coming harvest, pulses such as lentils and chickpeas will be commercially graded in Australia with the aid of artificial intelligence, in a world-first.
The breakthrough – pioneered by a husband and wife start-up company – promises major efficiency gains for the nation’s $22bn pulse and grain industry, as well as tech export opportunities.
It replaces a laborious process – grading pulses and grains by hand using tweezers and a magnifying glass – with a machine-learned system that does the job in a fraction of the time, more accurately.
“We’re solving a real global challenge, making the whole supply chain faster and more efficient and delivering trust for everybody,” said co-founder Anna Falkiner.
Across the globe every year, samples are taken from billions of tonnes of grain, pulses and legumes and graded by hand. This determines the quality – and therefore the value and usage – of each shipment.
Taken into account are defects, such as damage, insect impact or mould, poor colour, wrinkling or contamination with weed seed.
In Australia, this process typically occurs at a silo facility, with the grain truck driver waiting outside while a trained grader takes 15 to 45 minutes to assess the sample.
The outcome can add or subtract hundreds of dollars per tonne but is subject to human subjectivity and error under time pressure, requiring samples to be kept for years in case of dispute.
Using Cropify’s technology, recently endorsed by the Grain Trade Australia trading standards committee, the sample is placed on a tray inside a light box, about the size of a desktop photocopier. The machine, dubbed Opal, takes high resolution images of the sample, sends them to the cloud for assessment by machine-learned models and the results returned in five minutes.
These are stored electronically, ending the need to retain physical samples.
Removes human subjectivity
The technology does not eliminate the need for a human, but reduces the training required – from two weeks to less than two days – while removing human subjectivity.
“It speeds up the supply chain at a critical time of the year when growers and bulk handlers are looking to move as much grain as fast as possible,” said Falkiner’s husband and company co-founder Andrew Hannon.
Customers already on board include GrainCorp, Louis Dreyfus, Wimpak and Shannon Bros, with new deals being hatched.
The company started with lentils and has expanded to chickpeas for this season’s harvest across NSW and southern Queensland, while planning for other crops.
“We have a whole suite of pulses on our development pipeline, but also we will be doing sorghum and wheat,” Falkiner said.
“We want our product to be a good fit for our customers so they can use it across the spectrum of commodities that they in-take.”
Adelaide-based Cropify believes its technology stands apart due its high accuracy, based on a bank of images used to “teach” the model.
There was interest in the technology from Canada, a major red lentil producer, after a demonstration visit, and the couple was seeking capital investment to expand.
“Any small agricultural product is ultimately in scope for us – we see an opportunity to grow the business across all Australian agricultural commodities that fit into that scope, and then globally,” Hannon said.
This article first appeared in The Australian as Harvesting AI for more efficient pulse, grain crop supply chains.
Related Topics #
UNLOCK INSIGHTS
Discover the untold stories of emerging ASX stocks.
Daily news and expert analysis, it's free to subscribe.
By proceeding, you confirm you understand that we handle personal information in accordance with our Privacy Policy.