The rise of the apple picking robots

Crosscut archive image.

Washington State University Prof. Manoj Karkee. Credit: Washington State University

Three out of five apples in the United States come from Washington. That’s 10-12 billion apples if you’re doing the math – enough to wrap around the earth 29 times. The $2.25 billion earned in 2012 was nearly double the revenue of wheat, the state’s second most lucrative crop. The 100 million crates totaled 4 billion pounds of Red Delicious, Golden Delicious, Gala, Fuji, Granny Smith, Braeburn, Honeycrisp, Pink Lady and Cameo. The amazing thing is that every apple in Washington, in fact every apple worldwide, is hand-picked. Researchers at Washington State University are working to change that.

The PR people of large orchards like Monkey Ridge near the Snake River or Pride Packing Co. in Yakima paint the phrase ‘Hand-Picked’ in a positive light. Like vine-ripened, hand-picked is a buzz phrase for retailers. But the reason they’re hand-picked is, unlike cherries, wheat, corn or grapes, there are no machines capable of harvesting an apple. That’s right – we carry in our pockets computers more powerful than NASA had on their first trip to the moon, but no one has successfully created a machine that can harvest an apple.

The industry may see a prototype within the year.

As the U.S. has pulled out of the Great Recession and access to more lucrative jobs has increased, fewer people travel for work. In combination with the decrease in immigration from Mexico, farm labor has been lacking for years. “Labor has been a challenge,” says David Douglas of Douglas Fruit. “There aren’t enough local workers. And it’s no secret that the majority of the work has been done by Hispanics.”

J. Edward Taylor, a professor of Agricultural and Resource Economics at UC Davis, estimates that 10-12 thousand fewer Mexican immigrants come to the U.S. to work every year. “The Mexican economy,” says Taylor, “is moving towards a service based economy. The children of farm workers are not going into farming.” Mexican farm labor has decreased so much, in fact, that Mexico is importing labor from Guatemala.

The 2014 harvest was a bumper crop, orchards producing record numbers of apples. And yet, with this shortage of workers, many orchards struggled to keep up. “We got our fruit picked,” says Douglas, “but not when we wanted to.” Apples that stay on the trees for too long pass their peak and are no longer fit for fresh consumption. This fruit is sold for juice or applesauce, for a much smaller profit.

What’s more, apple picking is hard work. Replacing labor takes more than posting ads. Orchard expectations are that workers clear a tree of 95-98 percent of its apples and without damaging them. To make it worth the minimum wage (assuming the workers are paid fairly), workers need to pick about one apple every 2 seconds.

For now, says Douglas, “the amount of fruit we grow is based on demand.” The worry, however, is that if labor does not grow with demand – new records for production are set every year – then they will grow what they can pick, not what they can sell.

When you consider these issues, it makes you wonder: When almost every other major crop has a machine harvester, why hasn’t the apple industry caught up?

Manoj Karkee is an assistant professor with Washington State University’s Center for Precision & Automated Agricultures in Prosser, Washington. He and his department recently received a half a million dollars from the USDA to develop a machine to pick apples. As he explains, picking an apple is very complicated. “The human being is still the most effective picker.” Most crops grow in a way that present relatively simple solutions: corn and wheat can be whipped up in a combine, cherries can be shaken from a tree, even delicate grapes hang in a way that allow for easy machine plucking. But apples are both delicate and hard to get to. The trees have long, spindly branches and for years they were staggered and without rows.

To say there are no machines automating the apple harvest is actually technically not true. There is a sort of vacuum machine on the market. But this machine is more of an aid to pickers – they can drop their apples into tubes rather than baskets – and does not serve to automate the industry as some hope. There are also machines that violently shake the tree. Of course, as the apples fall, they become badly bruised and are therefore only suitable for the less lucrative cider and juice industries.

As Karkee sees it, there are two main obstacles to overcome: not damaging fruit and working as fast or faster than humans. “If the machine cannot pick 95-98 percent of apples or cannot pick an apple in two seconds then it does not make sense.” So in an eerily logical decision, he and his lab decided they would model the machine off of the human worker. If you consider what makes a human picker effective, it comes down to their eyes and hands: they see an apple, they pick it. Says Karkee “The basic components [of the robot] are a camera system to identify and locate apples and an understanding of how human workers pick apples: how much force, how many fingers and how they rotate the apple.” We’re way beyond a combine harvester now.

Crosscut archive image.
The robotic vision system used by Karkee's machine. Credit: Vision Robotics Corporation

For the machine to recognize apples without the aid of a person, it needs a computer that can identify correct colors – shades of red, green and yellow specific to apples. Screen shot examples make it look like a fighter jet locking in on targets. Once the machine has located all the apples, the arm follows and grabs them from the tree. Karkee imagines an octopus-like machine with many arms moving down the orchard, all picking at once.

“We’re not claiming that [the robot] is novel and new," Karkee says, "But it hasn’t made it to the orchards because of limitations.” Outside of their effort, a few companies have projects in Europe. In Florida, similar technology is being developed for citrus. But so far, machines can only pick about 80 percent of the fruit at a rate of an apple or orange every 4-10 seconds. Not fast enough to dethrone migrant labor.

While Karkee and his team hope to have a prototype soon, he doesn’t see the technology hitting the market for at least 5 years, maybe 10. But are orchards ready for it? Says Douglas, “Will we use technology when it comes? Yes. If we can find ways to do the work more efficiently it would help minimize the unknown. We won’t need to match numbers to fit labor.”

That, of course, raises the question of the worker. “That comes up every time,” says Karkee. “People who make a living will have concerns.” Karkee argues that the machines will only fill the gaps in labor. Eventually, he says, the low paying, low-skill jobs will be replaced with high skill, high paying jobs, namely operating and maintaining these complicated machines.

Despite what Karkee says, it would be easy to jump to the conclusion that once the apple harvester hits the market, the labor force would be put out of work. When the tomato harvester hit the market, a huge legal battle erupted between employers and the unions.

Taylor sees things differently. “We live in a much different time," he says. "I don’t envision a doomsday scenario in which there is no longer work for people.” More likely, he believes, is that some orchards would adopt the technology and some would not. “Those put out of work would move to the orchards still using people.” So, while there would eventually be a smaller need for labor, the process would be anything but sudden. And Taylor believes that at the same time, we can expect to see the number of Mexicans working U.S. farms decrease by more every year.

Considering the 5-10 year estimate on the apple harvester, that gives farm families in Yakima a few years before the labor demand decreases.

“1,500 years ago,” argues Karkee, “50 percent of the population worked in agriculture. Now, only 2 percent do. That enables us to make advancements in other fields.” Human advancements or no, this machine is about, as Douglas said, minimizing the unknown. If successful, it will undoubtedly move the industry beyond migrant labor, which will in turn benefit orchard owners and apple consumers. Karkee and Taylor paint an image of a win, win, win scenario, in which a skilled workforce produces a greater number of cheaper apples. And in the end, says Karkee, this especially benefits workers because "nobody likes to pick apples."

Unfortunately, no orchard pickers could be located for comment on this story.


Please support independent local news for all.

We rely on donations from readers like you to sustain Crosscut's in-depth reporting on issues critical to the PNW.


About the Authors & Contributors

David Kroman

David Kroman

David Kroman is formerly a reporter at Crosscut, where he covered city politics.