A conversation between Molly Nelson and her dad about how sheep could be better scanned and identified was the seed of an idea that has managed to turn cutting-edge facial recognition technology into a farmer-friendly tool to aid stock management.
“We had heard about how a Dunedin company Iris Data Science was developing facial recognition for sheep, Molly said.
“With some research, we figured it could be possible to get a lot more information built into an app from that technology.
“We were also thinking about animal welfare around putting tags in ears and how we could reduce plastic use.”
The girls were fortunate to have Libby Deadman, one of their development team members, also taking a computer course as a subject, as she had a good understanding of app building. With some help from her teacher, the girls also sought some advice from Hamilton-based software development company Rezare.
“They gave us an idea of the costs and process to get from a prototype to an operating app,” she said.
The resulting app has a focus on simplicity, with the girls acknowledging farmers’ technical capabilities but understanding that when using it in the field, they need something accessible and simply laid out for ease of operation.
“And we knew they wanted the information accessible all on one platform, and not to have any double ups,” she said.
At a practical level a farmer wanting information on one particular sheep would photograph the particular animal and via wi-fi or mobile data, the image would be uploaded to a cloud server, accessing the Iris image and confirming the animal’s ID number.
Significant amounts of additional data could then be accessed on that sheep through the app.
“The main information farmers would seek out about their sheep were weight, condition score, animal health treatments, including drenching, shearing and any antibiotic use,” she said.
However, because of the almost unlimited ability to add more information to the app, the girls are also excited about other information that could be inputted in the future.
“We are thinking it could include behavioural aspects like mothering ability, or even mob position – not a lot of research has been done on this yet but it is thought sheep occupy different positions in a flock, which may affect their performance as they are moved around,” Libby explained.
Ultimately, the girls foresee flocks getting their facial images photographed and digitised the way contractors come and conduct other tasks like dipping or shearing.
They are also confident the app interface may also prove useful for other stock types, and understand facial recognition for cattle is already in use in the United States.
An Indian firm is already claiming 95% accuracy with its Mooofarm algorithm for identifying cattle in the Punjab district.
Being able to easily identify sheep was a stumbling block to getting them included in the NAIT scheme when it was launched over 14 years ago.
Today, the facial recognition technology may provide a cheap means of getting the high head numbers registered and capable of being identified.
“Farmers often have the information the app captures, but so often it is held in several different places, so this brings it all together,” Lucy Fullerton-Smith said.
The business model for the app has been cost on a per sheep-per year basis, initially proposed at $1,40 a year, coming in just under the $1,50 a sheep tag price.
This would include the link to the Iris digitised technology that sits behind the app, and they have already had some companies express an interest in the app.
The girls’ award comes as they form a consolidated minority among the largely male contingent studying agri-business at St Pauls.
The year 13 students are also keen to pursue careers in agri-business, with Molly considering a diploma in agriculture and Lucy looking at a business degree as a pathway to an agri-business career.
Dean Williamson, co-owner of GlobalHQ publishers and a key B.linc competition sponsor said the girls’ effort stood out in what was a high-class field of high school entries. He was inspired by the commitment and innovation their approach had delivered.
“We are in a period where the rule book has gone out the window about how to approach business, and this competition inspires some innovative and nimble ways to continue growing, harvesting and processing primary products,” he said.
B.linc Innovation has a proven model of collaboration that ‘joins up’ people and existing knowledge to start new conversations led by business and industry to generate innovative solutions.
Julia Henson of B.linc Innovation says that all winners have shown incredibly innovative solutions alongside actionable pathways to improve outcomes for primary industries in NZ. “They have ensured that their solutions have taken a truly collaborative ecosystem into account and have shown how their solution is applicable in the real world, considering challenges specific to New Zealand primary industries. “