Machine vision is helping to make “automation resistant” produce easier to harvest. Crops like lettuce are particularly difficult to harvest mechanically because they are fragile and low to the ground. Now, an agricultural robot featuring machine vision, named Vegebot, has been tested in a variety of field conditions, and it’s demonstrating how robotic farming might be expanded.
The Vegebot and Its Benefits
The Vegebot was developed by a team at the University of Cambridge. It has two main components: a computer vision system and a cutting system. The Vegebot identifies the target crop within its field of vision with the use of its overhead machine vision camera. It then images each and every lettuce and determines whether a lettuce is healthy and ready to be harvested. Once it finds a healthy lettuce that is ready to be harvested, it cuts the lettuce from the rest of the plant without damaging it so that it is “supermarket ready.”
Researchers had to develop and train the agricultural robot’s machine learning algorithm with example images of lettuces. The robot had to identify healthy lettuces in the lab first, then it was training in the field in a variety of weather conditions with thousands of lettuces. The Vegebot’s second camera near the cutting blade helps to make sure of a smooth cut. The agricultural robot’s gripping arm had to be adjusted so it held the lettuce firmly, but not firmly enough to crush it. The force of the grip can be adjusted for other crops as well.
The Future of Robotic Harvesters
The Vegebot has a ways to go before it can keep up with the speed of an experienced human harvester, but the potential is there. Growers hope that robotic farming can help overcome problems with labor shortages. Experts say that immigration concerns and a lack of interest in agriculture jobs will result in an ever-growing labor shortage for farms in the United States.
Robotic harvesters will also likely reduce food waste. Fields are usually harvested once and unripe fruits and vegetables are often discarded, but a robotic harvester would be capable of picking only the ripe vegetables. Since it could also harvest 24/7, it could perform multiple passes on a field. It could return at a later date to harvest any vegetables that were unripe when it made the first pass.
Discover the possibilities of machine vision technology in agricultural robots by reading How Life Science Vision Systems Increase the Accuracy of Crop Yield Estimation.