At Inde Technology, we pride ourselves on pushing the boundaries of what's possible with technology. As we expand our horizons and as specialists in AI, our recent partnership with PGG Wrightson and its Livestock team on the groundbreaking SkyCount project highlights how we're working with customers to create meaningful AI.
Re-syndicated article from Microsoft News Centre.
Livestock counting has traditionally been a bugbear for farmers, requiring a lot of time, cost and effort to confirm exactly how many animals there are for audit purposes. Leading agricultural services provider PGG Wrightson partnered with New Zealand AI specialists Inde to take the stress out of the process for farmers and animals alike. Combining Microsoft Azure, Power Apps and some very clever drones, they built a solution that has brought this age-old task into the 21st century.
Counting sheep may be the definition of relaxing, but in real life, it’s an extremely time-consuming, costly and potentially disruptive time for business. For audit and insurance purposes, farming enterprises need to know how many animals they actually have, performing stock counts every year. However, when your large property covers varied or challenging terrain, doing a headcount is logistically demanding.
This manual process requires rounding up and bringing in all the animals from the paddocks, which can take up to a week for large properties with multiple mobs of cattle or sheep. For dairy cows, stock counts can range from 200 to 2,000. However, getting the actual numbers right is essential, as any discrepancies can lead to significant financial losses.
“In today’s environment, farmers are counting every penny, so every animal that’s missing is a big hole in your books,” explains Roger Lamb, National Operations Manager – Livestock at PGG Wrightson, which commissions stock agents to audit the process. “Some animals might be worth up to $20,000, so having any of them missing means your accounting is out by a lot.”
Farmers are also keen to promote the wellbeing and productivity of their animals, and lower the risk of potential injuries and stress, by reducing the need to herd them. That applies to farm workers as well – the key is to ensure that regular farming operations can go on undisrupted.
“Traditional farm audits can take five or six staff off the road for a week,” said Roger Lamb, Project Owner at PGG Wrightson. “I thought: there has to be a better way.”
Roger and his colleagues looked into drones and AI-powered stock counts as a solution, thinking that instead of bringing the animals to them, they could meet them wherever they were. However, no one was doing exactly what they were looking for. Instead, the team decided to build the solution themselves. With the help of fellow Kiwi business and AI partner, Inde, SkyCount™ was born.
Training flying cameras to count sheep (and cows)
Inde Chief Technology Officer, Rik Roberts, threw himself into the mission with gusto, first working out which drones and cameras were right for the task of filming animals to capture the images needed to train the AI platform.
“It was about finding a balance. Larger drones can carry bigger cameras, but these posed issues with flight planning. If you’re flying them near controlled airspace, you need to get permission from the Civil Aviation Authority. However, smaller drones flying too low can frighten the animals,” said Rik.
Testing also brought up some interesting challenges. Infrared cameras couldn’t “see” sheep through their thick fleece. Still cameras required lots of shots to be stitched together, making the final product blurry, resulting in lost or duplicated animals. Also, while beef cattle were unbothered by the sound of the drones, dairy cattle didn’t like it at all, stressing the animals.
After months of testing, they had determined that mid-sized, quieter drones could operate closer to animals without causing distress, landing on DJI Mavic Enterprise Series drones that shot regular video through HD cameras rather than infra-red.
The next hurdle the team had to overcome was training the AI models.
“Our team spent countless hours collecting high-quality aerial images we could feed our model, including the characteristics of different breeds, to photographing them in various topographies and at different times of the day, to ensure shadows, trees and rocks were not mistaken for animals and vice versa,” said Rik.
“There’s a lot of variability in landscape types between regions and animal breeds. Getting enough data was the real challenge, including grass types, terrain types and weather conditions.”
Tens of thousands of images were captured and processed in Microsoft Azure to train the AI model.
They also needed to establish automated mission planning. As the drones would be operating in remote rural areas with limited connectivity, it was paramount to programme them to execute flight plans consistently and efficiently.
“If the property does not already have a digital map, we’ll do an initial mapping mission. We can define all the drones’ flight altitudes and their directions, and angle the cameras so that the second time, the third time, fourth time we go out for a stock count, they’re flying the same missions,” Rik explains.
Finally, the SkyCount™ drones were ready to launch.
”We knew we were onto something”
“We trialled the solution with a high-country North Canterbury farmer. He said there were two paddocks he didn’t think he had any stock in – but when flying our drone, we identified a number of livestock were in the paddock. So, we knew we were onto something,” says Roger.
Instead of taking a week to muster and count animals, requiring a team of people, the whole process has been reduced drastically. It can be completed by a single drone operator, with farmers able to review and approve the results in a few minutes. The result is much more accurate than any human count – around 97% first time.
The finalised count is submitted to SkyCount™, which uses Microsoft Power Apps to combine the farm’s information automatically with the count to generate the audit report.
Roger is delighted. “Now, when we come out and do audits, we don’t have to move a single animal or interrupt your operation. We can come and do our thing, and the people on the farm don’t necessarily know we’ve even been there. This is a significant time reduction, allowing staff on the farm do other work.”
Rik is proud of how user-friendly the drones are, no matter how remote the setting.
“This allows users to quickly become experts and ensures consistent, repeatable results, minimising the need for recounts,” he says.
The sky is the limit
Inde and PGG Wrightson already have big plans to expand the capabilities of the AI model to even more livestock types and use cases. Rik believes there’s an opportunity to monitor pests such as wallabies and deer, or monitor pasture conditions.
Vanessa Sorenson, Chief Partner Officer for Microsoft ANZ, says this is one of the best uses of Microsoft Azure and AI technology she’s seen yet.
“New Zealand is famous as a farming nation, but the technology we’re developing in this country is next-level – and this solution has the potential to make a huge difference to the agriculture sector around the world. Inde’s really gone above and beyond to demonstrate just what AI can do, and I’m really excited to see where SkyCount™ goes next. Pardon the pun, but the sky’s the limit and I can see huge interest from farming communities and conservation initiatives everywhere.”
The success of SkyCount™ has also inspired PGG Wrightson to investigate further technological innovations.
“On the back of how successful this project has been, we’re now looking into ways of integrating IoT devices and farm management software to create a comprehensive, automated system for monitoring and managing operations that will drive even further efficiencies for our customers,” says Roger.
“As for SkyCount™, I’m hopeful that within another few years, we’ll have another 20 benefits we don’t even know about yet.”
COMMENTS