Designed by a team at Cambridge University, the ' Vegebot ' was originally developed to recognize and harvest iceberg lettuce in a laboratory setting. The robot was also successfully tested in field conditions in collaboration with G's Growers, a local fruit and vegetable cooperative.
Not as fast or efficient as a human, this prototype robot demonstrates how it can be developed even for products such as iceberg lettuce, which are particularly difficult to harvest mechanically.
Crops such as potatoes and wheat have been mechanically harvested at scale for decades, while many other crops have resisted automation to this day. Iceberg is one of the products that resist automation in lettuce. The most widely grown lettuce type in the UK, iceberg lettuce is a crop that grows flat on the ground and is easily damaged at harvest. For this reason, iceberg lettuce is a challenging product for robotic harvesters.
Simon Birrell from Cambridge Engineering Department says, “Every field and every lettuce is different; "But if we can successfully harvest iceberg lettuce with the robotic harvester, we can successfully adapt it to many other crops."
Dr. Julia Cai , who worked on Vegebot 's computer vision components while a graduate student in Fumiya Iida's lab; “The lettuce harvest is done by hand and it is very physically demanding,” he said.
First, the vegebot identifies the target crop in sight, then determines whether the lettuce is ready to be harvested, and finally cuts the lettuce without crushing or damaging it, making it ready for the supermarket. Josie Hughes , one of the team members, points out that "this process, which takes a few seconds for a human, is a really challenging problem for a robot."
Related Products
Vegebot has two main components, a computer vision system and a cutting system. The robot takes an image of the lettuce field with the overhead camera on it and identifies all the lettuce in the image, classifies them according to whether they can be harvested or not, and harvests them.
The researchers developed a machine learning algorithm on the sample lettuce images and identified it to the robot. Vegebot, which can recognize healthy lettuce in the laboratory, performed its identification on thousands of lettuce in the field under various weather conditions.
A second camera, located next to the cutting blade on the Vegebot, helps cut the plant neatly. The researchers were also able to adjust the pressure on the robot's grip arm so that it could grip the lettuce firmly enough not to drop or crush it. The grip strength can also be adjusted for other products.
In the future, robotic harvesters could help tackle labor shortages in agriculture and reduce food waste. Today, fields are harvested once at harvest time. During the harvest, immature products such as ripe products are harvested and then immature products are discarded.
A robotic harvester, programmed to collect only ripe produce, can harvest at any time of the day, and by making multiple passes on different dates from locations on the same field, it can harvest the immature produce from its previous pass.
Hughes; “With Vegebot, they also collected a lot of data that could be used to increase the productivity of lettuce,” he says. Speaking of the need to increase the harvest speed of Vegebot to the point where it can compete with a human, Hughes also says that robots have a lot of potential in agricultural technology.