Size matters when it comes to apple thinning

Matt Stevens comes from an orchard background and his robotics company is developing technology he hopes will be able to provide better thinning processes for his fellow apple growers. ~ Luke Edwards photo

By Luke Edwards

For anyone worried about the robot takeover, fear not. Matt Stevens has a perfectly simple solution should the machines try to rise up against humanity.

Bamboo.

Stevens, president and head grower at Finite Robotics, presented at last month’s Ontario Fruit and Vegetable Convention, showing off robotic orchard thinning technology his team has developed. While playing a video of the machine in action, the robot came up against a bamboo stake in the orchard. The stake baffled the robot, leading Stevens to joke to the audience that should the robots ever rise up, just get your hands on some bamboo.

But while the robot takeover talk is all presumably with tongue firmly placed in cheek, it gets to a significant point Stevens makes about the robotic thinning technology.

“There are things robots are incredibly good at, and there are things humans are incredibly good at, and they’re not the same thing,” he said.

To further the point, Stevens had the audience perform a bend and twist motion with their arms. Unsurprisingly, every human managed the task with ease. Had he been in a room with robots, however, he said they would have overwhelmingly failed.

Gerbe Botden is the vice president of the Blue Mountain Fruit Company in Thornbury, where Finite Robotics did some testing. Better thinning processes, completed earlier, will mean a better, more consistent end product, he said. Earlier thinning can also improve the chances of return blooms, improving yields for future years.

While humans might be able to contort their arms in ways robots can’t, human eyes fail in detecting millimetre differences in fruit size.

The solution, therefore, is fairly straightforward.

“Get them (the robots) to do 50, 60, 70 per cent of the work, and then get them out,” said Stevens.

“We want this to be a tool, not something that makes it more complex for the grower,” added Botden.

Finite Robotics took their prototype robot through what Stevens called a skills bootcamp. The skills included movement, vision, decision making, reach, removal and disposal.

The first step, simply getting the robot moving efficiently through the orchard was a challenge.

“It did this drunken sailor walk where it would stop, and then it would forget where it was pointing,” Stevens said.

Initially, the company purchased the movement technology from an outside vendor. After the failures in Botden’s orchards, they scrapped those plans and developed their own solution.

Thanks to plenty of hard work earlier, however, the vision portion of the bootcamp was quite successful. The robot was able to see, and measure the fruits. There were some issues with lighting, but Stevens said they’ve been able to overcome those challenges.

With solid vision systems, decision making becomes interesting.

“This is where there’s a difference. When you’re using a robot, they can do sizing very accurately. You can start to do some logic that you can’t do as a human, because I can’t see the differentiation between minor changes in size,” Stevens said.

For removal, they developed an apparatus that utilizes vacuum pressure to suck the fruitlet in and a device that rotates and cuts it off then pushes it back into the bin for disposal.

Stevens said the plan is to start offering the technology in a service model for the first few years, and eventually give growers an opportunity to purchase devices outright.

For more information, visit finiterobotics.com.