Meet Fruitometry - the Kiwi AI imaging startup helping NZ grow better kiwifruit 2

In the early 2000s, software and electronics engineer Christopher Miller was developing world-leading imaging technology to help Emirates Team New Zealand win the America’s Cup. 

By 2018, he had sold his St Mary’s Bay home and moved to Katikati, Bay of Plenty, to spend his days pushing a computer on a wheelbarrow around a kiwifruit orchard.

There’s little discernible connection between high-tech competitive sailing and the humble kiwifruit. But to Miller, who worked with Team New Zealand for almost 10 years, the transition was very logical.

The head of Team New Zealand handed me a photo of a sail and said ‘can you write a computer programme to measure the shape?’, and I said, ‘I think so.’”

“My day job for Team New Zealand was doing stuff that had never been done before with electronics or software. When I first started working with them, the head of Team New Zealand handed me a photo of a sail and said ‘can you write a computer programme to measure the shape?’, and I said, ‘I think so.’”

Christopher Miller working in his garage (Photo: Supplied)

By the 2007 America’s Cup in Valencia, Miller and his team had developed the world’s first sail image processing system that was reliable and accurate enough to be used during the actual competition, which New Zealand narrowly lost to Alinghi.

While he wanted to build the next generation for the 2017 America’s Cup, he says by that time the focus had shifted away from sales and onto foils. It was then that Miller decided to venture down a different path.

“A mechanical engineer by the name of Loren Poole, whom I worked with at Team New Zealand, had a kiwifruit farm in Katikati. But he left the America’s Cup because he said he was making more money in kiwifruit.”

He asked me, ‘can you make a system that measures kiwifruit?’ And I said, ‘Yeah, I think I can.’”

“He asked me, ‘can you make a system that measures kiwifruit?’ And I said, ‘Yeah, I think I can.’”

In the garage of his St Mary’s Bay home, Miller got to work designing his first prototype, which would eventuate as the aforementioned computer that was carried around orchards on a wheelbarrow. However, the project quickly evolved after he sold his home for $2m, moved to Katikati, and invested the capital in what would become Fruitometry. Since that time, it has grown into a commercial agritech business with six staff, and has scanned 1000 hectares of kiwifruit orchards across the Bay of Plenty.

Meet Fruitometry - the Kiwi AI imaging startup helping NZ grow better kiwifruit 3
A visualisation of canopy coverage on a kiwifruit orchard (Photo: Supplied)

So what does Fruitometry actually do?

Essentially, Miller says, it produces density maps of orchards to help orchardists to grow the healthiest, best quality and therefore most valuable fruit.

“The most important thing for a kiwifruit grower to do is prune and thin so that they get an optimal crop,” Miller says. “Kiwifruit is prolific, and if every cane became a flower and every flower turned into fruit, you’d just end up with a huge crop of small crap fruit that you can’t even give away.”

Because growers want to optimize both the fruit output and quality – sugar content, taste, size and color – they need to focus their pruning on areas that will yield the best results, and therefore the best return.

“So what we do is scan entire blocks and generate density maps that pinpoint exactly where orchard managers should direct their labour,” Miller says.

With the wheelbarrow prototype no longer in service, Miller’s team now uses a fleet of quad bikes that have been retrofitted with cameras, three kilowatt generators, two kilowatts of LED lights and custom-built computers. They move about the orchard, taking about 50,000 pictures per hectare. A satellite inertial navigation system lets them know exactly where they are, while an AI convolution neural network automatically detects the fruit within the images. Optical algorithms allow them to extract spatial information when they’re moving about the orchard. The resulting data gets translated into a “moving density of features”, which helps growers improve their orchards’ performance.

Miller says the accuracy of the technology is still improving as it’s being introduced to kiwifruit growers, most of whom are still relying on manual methods of assessing pruning locations.

“They’re still learning how to use it, because they’re interested in making sure that the numbers we give them through digital estimation match up with their Victorian-era manual counts that they acquire by taping off 25 square meters and having some labourers walk around with clicker counters. In reality, there’s a lot of human error in those numbers.”

“We’ve had a lot of good feedback… but the operating expenses are non-trivial. So it’s about getting them to budget for that. And it’s building up a good story and customer viewpoint.”

With a $2.6m investment by kiwifruit grower Seeka in July 2021, Miller hopes that clients will see gains to their bottom lines in the next year or two. However, he says the long term goal is to take the subjectivity out of the measurement and allow growers to maximize their returns in an industry already worth $3b.

“I did consider other crops. But the reality is, if you’re going to be doing digital scanning in New Zealand horticulture, kiwifruit is the highest value crop to be scanning. That’s where the money is.”

But one of the biggest benefits, Miller says, is that unlike in America’s Cup racing, which places strict rules on technology, he is now free to create, build and develop with impunity.

“In the America’s Cup, you can’t make the yachts fly, you can’t run them by robots, you can’t use AI on them. But in the kiwifruit industry, there’s way more opportunity for open-ended technology to start a company and build equity and build a customer base.”

But what about his new lifestyle? Does he miss the high life and prestige of America’s Cup sailing?

“It’s a lot more gratifying to be living and working in Katikati, helping the nation grow food more efficiently than helping rich white sailors go around in circles on the water.”

“But still, it’s scary as fuck at times because I bet my home on it!”