Researchers at the University of Cambridge have developed an iceberg lettuce harvesting robot dubbed the ‘Vegebot.’ Using machine learning algorithms, the robot can learn how to recognize and harvest the lettuce despite any challenges it may face. Now being used as a prototype, Vegebot showcases how machinery, like robotics with algorithms, can progress in agriculture, even for crops that may be difficult to harvest for robots, like iceberg lettuce. Originally trained in a lab environment, the robot has been put to the test in many different field conditions with approval from G’s Growers, a local fruit and vegetable co-operative based in the United Kingdom.


Vegebot could help reduce labor times, making it easier on human workers to pick lettuce and other crops in fields. (Image Credit: Cambridge University)


Other crops like potatoes and wheat have been harvested at a scale level with the use of mechanical equipment for a few decades now, but other crops like iceberg lettuce have resisted automation. It becomes challenging for robots to harvest due to how iceberg lettuce grows flat to the ground and gets damaged easily.

“Every field is different, every lettuce is different,” said co-author Simon Birrell from Cambridge’s Department of Engineering. “But if we can make a robotic harvester work with iceberg lettuce, we could also make it work with many other crops.”

“At the moment, harvesting is the only part of the lettuce life cycle that is done manually, and it’s very physically demanding,” said co-author Julia Cai, who worked on the computer vision components of the Vegebot while she was an undergraduate student in the lab of Dr. Fumiya Iida.

At first glance, the robot identifies the type of crop it will be harvesting within its vision range, then determines whether or not the crop is healthy and mature enough to be picked, and finally, cuts off the lettuce from the plant gently enough so it doesn’t become damaged and can be sent to supermarkets.

Vegebot contains two main components as part of its harvesting functionality, a computer vision system and a cutting system. The camera acts as an identifying image system by capturing a photo of the lettuce field, identifying each lettuce in the image and decides on whether or not the lettuce is ready for harvesting. In some cases, lettuce may not be picked because it has not yet fully grown, or it may have become infested with a disease that could be carried over to the next lettuce.

Researchers also developed and trained a machine learning algorithm on images of lettuces to help identify whether or not a lettuce head was healthy. Once the classifications were made in a lab environment, it was then trained in a lettuce field, with differing weather conditions, on thousands of lettuces. Vegebot also has a second camera attached close to the blade, ensuring a clean cut with better handing.

The robots gripping arm was also adjusted to improve the pressurization when handing the lettuce. Adjustments allowed the robot to handle the lettuce with care so that it doesn’t drop to the ground or become crushed by the gripping power. It can also be adjusted to suit other crops.

“We wanted to develop approaches that weren’t necessarily specific to iceberg lettuce so that they can be used for other types of above-ground crops,” said Iida, research team leader.

Robotic use in agriculture could help solve issues revolving around labor shortages and could also help with reducing food waste. At present, every field is only harvested once, and any fruits or vegetables that aren’t mature for picking at harvest time are generally thrown away.

With a robotic harvester in place, it could be trained to pick only the ripe fruits or vegetables and since the robot works tirelessly on the field, it could perform multiple passes on the same field, making a return at a later date to pick vegetables that weren’t ripe in previous passes.

Have a story tip? Message me at: cabe(at)element14(dot)com