Hocking believed there had to be a more efficient way to ensure harvesters collected more ripe blueberries and fewer green ones.
“The fruit moves along a conveyor belt after it’s harvested,” he said.
“Our AI model scans each individual berry and determines its ripeness: whether it’s unripe, partially ripe, or overripe.
“That information then feeds into an algorithm that determines whether the harvester is shaking too hard and needs to slow down, or if it’s shaking too lightly and not collecting enough berries.”
Hocking said the goal was to have around 80% ripe, blue fruit on the conveyor belt – if more than 20% of the berries are green, the machine is likely damaging the plant and harvesting fruit too early.
While his AI model was very accurate, Hocking said that wasn’t the hard part.
“The real challenge is developing the algorithm that can adjust the speed of the beaters in real time.”
Hocking said that in the current market there were either cheaper, older harvesters with no automation or fully automated harvesters costing around $700,000 or more.
“A lot of farms can’t afford to buy a new harvester,” he said.
“There are a lot of older models in New Zealand, and many farms are still using them.”
Hocking said his product could be attached to older machinery, avoiding the hefty price tag of a brand-new harvester.
The goal is to have the product ready for next season.
The hardware is expected to cost around $5000, with software subscriptions costing approximately $20,000 a year.
University of Waikato third-year mechanical engineering student Thomas Hocking.
“We will provide analytics, IoT services, connectivity, and a range of additional features alongside the software.”
Hocking said orchard managers currently had little real-time feedback on harvester performance and often had to drive 15 minutes across the orchard on a quad bike to check for themselves.
By installing Wi-Fi-enabled technology on harvesters, key performance data can be sent directly back to the office, allowing managers to monitor operations in real-time and either instruct drivers to make adjustments or make the changes remotely themselves.
Blueberries move along a conveyor belt after harvesting.
The technology could save orchards thousands of dollars while also helping reduce mistakes caused by worker fatigue as people spend up to 14 hours a day harvesting.
It also has broader agricultural applications, with potential use for crops such as raspberries, olives, wine grapes and coffee beans.
“A lot of these crops are harvested in the same way by shaking them, so the produce falls on to two centralised conveyor belts,” Hocking said.
“We can go down that route, or we can take a more generic IT capability approach where we simply provide farms with ways to collect their data and present it in a user-friendly way.”
Hocking and his product are entered in the Prototype Award and will be at the Innovation Hub at Site IN21 at Fieldays.