For up to £250 Bonus for sports, use our exclusive bet365 Bonus www.abonuscode.co.uk Claim your bonus and start betting at bet365 now.

Fruit-loving robots

The way fruit is harvested in the orchard hasn’t changed much since the Garden of Eden: it’s still a manual process, by and large.

Many fruit and vegetables, like strawberries, apples, peppers and cucumbers are harvested by hand because there are no good mechanical harvesting methods. The problem is not just that robot hands might damage the delicate crops: machines must also learn to observe when they are ripe. New techniques, such as artificial neural networks to process camera images and direct robot arms, are now being applied in the agricultural sector to solve those problems.

 

The European project CROPS (Clever Robots for Crops), that ran from 2010 to 2014, has delivered several prototypes of agricultural robots, among which harvesters for cucumber, pepper and tomato. Those were not very efficient yet: the pepper robot for example, needed an average of 94 seconds to recognize and pick a ripe pepper. A human employee does that in 6 seconds.

Since then, researchers at Wageningen University, The Netherlands, have been working very hard on an improved version. At the end of March, the SWEEPER (Sweet Pepper Harvesting Robot) was tested in a Belgian greenhouse, and more tests will follow this summer.

In the U.S., California based startup Abundant Robotics Inc. is developing autonomous, robotic apple pickers. They employ computer vision and a fruit identification algorithm that “learns” in each orchard so that the harvester can adapt to the specific cultivar and canopy structure. For power, the robots are plugged into the small tractors already used pervasively in fruit farming. Abundant Robotics has already conducted field trials with orchards in Washington state, and in Australia.

“You direct this robot to go someplace, see and pick an apple without damaging it, and go again. It’s a very non-trivial engineering challenge. To detect apples very precisely you have to see down at the millimeter level in real time. That requires software, and on the hardware side, chips that allow you to do real time image processing on the fly,” according to Abundant Robotics’ CEO and co-founder Dan Steere.

Read: AI and Machine Learning: can we build an artificial brain?

In the UK, another startup named Dogtooth Technologies is developing intelligent robot strawberry farmers, each equipped with several cameras. The company is working with researchers at the National Institute of Agricultural Botany in Cambridge to improve the robots’ computer vision and machine learning algorithms and enable them to make better decisions about whether to pick a fruit or not. In countries like Belgium and Japan, similar technologies are being developed, all with their own angle.

“Picking is almost an excuse to get cameras in the field,” said Ed Herbert, founder and chief operating officer of Dogtooth Technologies. “You have got these robots moving up and down the fields with several cameras on, and they’ll be imaging the crop in minute detail. So you can look for signs of disease or pest, and that means that you can intervene earlier, increasing your yield.” 

Orchards are getting ready for mechanical harvesting. Now the machines are nearly ready, too, and they’re on their way. No matter the type of machine, however, harvesting 100 percent of crop by automated harvester probably isn’t in the cards. More realistically, machines will be able to pick 80 to 90 percent of fruit on a trellis row, with the rest left for human picking crews.

Worry over labor shortages will be a thing of the past, but at the same time, many workers worldwide are still depending on jobs in agriculture for their subsistence. How they are going to reap the fruits of all these new developments, remains an open question.

Image: Harvest Automation OmniVeyor HV-100 Robot