'Vegebot' can harvest lettuce thanks to machine learning

Thanks to a machine learning crash course in the lab, a robot can now recognize and harvest lettuce.

A team of engineers at the University of Cambridge have developed “Vegebot,” a robot that was trained to spot and harvest iceberg lettuce in their lab. They’ve since taken Vegebot on a test run in a variety of local fields, with the help of fruit and vegetable coop G’s Growers.

While Vegebot is not as fast or efficient as human workers, it was successfully able to recognize the lettuce and retrieve it.

It’s big news for farmers who grow crops like lettuce that have not been successfully automated in the past.

“Every field is different, every lettuce is different,” co-author Dr. Simon Birrell said in a press release. “But if we can make a robotic harvester work with iceberg lettuce, we could also make it work with many other crops.”

Unlike crops like grapes, wheat and potatoes, iceberg and other types of lettuce have not been good candidates for automated harvesting in the past. Iceberg lettuce, for example, grows close to the ground and is easily damaged, making it difficult to use an automated process without damaging the crop.

“At the moment, harvesting is the only part of the lettuce life cycle that is done manually, and it’s very physically demanding,” co-author Julia Cai said.

The Vegebot has a computerized vision system that lets it analyze plants, identify them as lettuce, and determine whether they’re ready to be harvested. It also has a complex cutting system that allows it to harvest each head of lettuce without crushing it, and so it is “supermarket ready.” 

Machine learning was used to teach Vegebot how to recognize different types of lettuce in the lab. Then, the engineers took it out in the field and taught it to recognize lettuce in a number of different settings.

“For a human, the entire process takes a couple of seconds, but it’s a really challenging problem for a robot,” co-author Dr. Josie Hughes said.

The experimental prototype shows machine learning can be used to develop robotic harvesters for other crops, like asparagus, peppers and sweet cherries. That could help farmers deal with rising costs and labor shortages – though it may also put migrant workers out of a job.

It could also help reduce food waste. Machine learning can be used to teach Vegebot and its successors to only recognize and harvest ripe fruit, and it could make multiple passes over several weeks in the same field, unlike human labor teams.

“We’re also collecting lots of data about lettuce, which could be used to improve efficiency, such as which fields have the highest yields,” Hughes said. “We’ve still got to speed our Vegebot up to the point where it could compete with a human, but we think robots have lots of potential in agri-tech.”

The study has been published in the Journal of Field Robotics.

—

By Kyla Cathey, Earth.com staff writer