A team of scientists at the University of Cambridge has developed a machine learning robot handy in picking vegetables and identifying harvest. Named as the ‘Vegebot’, the machine was previously trained to identify iceberg lettuce in a lab. According to reports, the robot has successfully passed the test in various fields, conducted in collaboration with G’s Growers, a co-operative of local fruits and vegetables.
The robot’s competency is not yet as fast and as efficient as that of a human, but the experiment has paved way for the advent of robotics in agriculture, especially in harvesting a challenging crop like iceberg lettuce. Iceberg lettuce being one of the commonly grown crops in the UK, it is vulnerable to damage and grows flat to the ground, this has been a major challenge to robotic harvesters.
“Every field is different, every lettuce is different, but if we can make a robotic harvester work with iceberg lettuce, we could also make it work with many other crops,” commented co-author Simon Birrell from Cambridge’s Department of Engineering.
According to reports, the robot first identifies the target crop within its premise of vision. It then determines whether the particular lettuce is healthy and ready for harvest. Finally the Vegebot cuts the lettuce, without crushing it and thus makes it ‘supermarket ready’. This is mainly because of a second camera placed near the cutting blade. This process only takes a couple of seconds for a human, however it is still challenging for a robot, according to experts.
Two significant components of the robot are the vision and the cutting system. An overhead camera on the Vegebot captures the image of the lettuce field and identifies the lettuce from the captured image. It then individually classifies whether the lettuce is worth harvesting. The machine may reject the crop for either not being mature or due to diseases that could spread among other lettuces.
Based on the example images of the lettuce, the team has developed a machine learning algorithm. Only after the Vegebot could identify harvest lettuces in the lab, the test was carried out in the field, along with a variety of weather conditions and with real lettuces.
Moreover, the scientists have worked on adjusting the pressure on the gripping arm of the robot, in order to help the robot hold the lettuce firmly, without dropping or crushing it.
“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 Dr Fumiya Iida, one of the leaders of the research team.
The research has thus laid groundwork for future labour shortage problems in agriculture and could help in curbing food wastage.