With strawberry picking season well under way - but migrant labour in short supply in several countries - we look at the various robots being developed around the world to help producers harvest this most popular fruit.
Next time you buy strawberries take a look a good look in the punnet. Do the berries still have the stem attached or has it been plucked off leaving only the green hat of leaves called the calyx?
You may not think that matters, but it's a key consideration for growers as they contemplate the merits of a range of robotic prototypes that promise to pick strawberries as fast and as carefully as humans.
Whether the berry is plucked or whether the stalk is snipped through and kept attached is one critical difference between the concepts that Spanish, Belgian, British and US engineers are testing, ready to roll out in fields as soon as next year.
Harvesting soft fruit mechanically represents a huge challenge - each berry needs to be located, even if it's behind a leaf, assessed for ripeness and then harvested and boxed with enormous care to avoid bruising.
But recent developments in visual sensor technology, machine learning and autonomous propulsion have brought the goal within reach.
"If you can put a man on the moon you can get a machine to pick a strawberry," says Tom Coen, founder of Octinion, a Belgium-based start-up conducting a final phase of field trials this summer in partnership with growers in the UK and continental Europe.
"Today we can say we have a [robotic] arm that is competitive with a human in terms of price and speed," he says.
Octinion's arm is mounted on a self-driving trolley. It reaches up from below and, using 3D vision, grips a ripe berry between two cushioned plastic paws. The gripper then turns the fruit by 90 degrees to snap it off its stalk, mimicking the technique a human picker would use.
The prototype is picking one strawberry every four seconds, says Mr Coen, and depending on the cultivar, will collect between 70% and 100% of the ripe fruit - results that he says make it competitive with human pickers.
The berry is left with only the calyx, which is the way European consumers are accustomed to buying their berries.
"We don't believe in cutting," he says. Stalks risk bruising other berries in the punnet, he argues.
But Cambridge-based start-up Dogtooth is taking a different approach.
Founders Duncan Robertson and Ed Herbert have just returned from Australia where they've been testing a picker that delivers berries with a centimetre or so of stem still attached, the way UK retailers prefer, because it extends shelf life.
Dogtooth is cautious about giving away too much about how its robot works, but like Octinion it is based around robotic arms mounted on a mobile platform.
It uses computer vision to identify ripe fruit and machine learning to evolve efficient picking strategies. After picking, the robot grades berries to determine their size and quality, and places them directly into punnets.
Dogtooth also prides itself in working around the needs and current practices of UK growers.
So while Octinion's machine will only work on fruit grown on raised platforms, usually in polytunnels, Dogtooth's will pick traditional British varieties in the field.
"Adopting robotic practice will be a big ask, so I don't want to ask growers to pull out existing infrastructure to support our robot," says Mr Robertson.
"We're trying to reinvent an important part of how soft fruit is grown, not reinvent the whole thing."
Robots can operate at all times of the day or night - harvesting during the chillier night hours can dramatically lengthen shelf life and avoid bruising.
But developers emphasise the motivation is not to replace migrant labour with cheaper, more efficient robots. In fact, it's not proving easy to replicate the standards that human pickers deliver.
Strawberry farmers say they are increasingly struggling to find people to do the work. They need the robots.
In the UK, the fall in the value of sterling following the EU referendum vote has made it increasingly difficult to recruit overseas workers. UK citizens seem reluctant to do such seasonal, physically laborious work.
In the US, growers say they have had to let fruit rot in the fields. Tighter immigration rules, a rise in the minimum wage, and a dwindling birth rate in Mexico, have all meant there just aren't the workers available to harvest them.
So producers will have to scale up robotic picking if they're to survive, many argue.
Agrobot, built by Spanish entrepreneur Juan Bravo, should be commercially available in California next year. And in Florida, the Harvest Croo team, led by former Intel engineer Bob Pitzer, is also close to launch.
Both of these are much bigger than the robotic arms being developed by Dogtooth and Octinion.
Both boast tractor-like vehicles spanning several rows of plants with arms that reach down to locate and pick fruit.
Each Agrobot arm grabs a stalk and snips, carrying the fruit off by its stem to minimise damage.
The Harvest Croo (Computerized Robotic Optimized Obtainer) uses paddles to gather up the plant's leaves to expose the fruit. Rotating grippers then grasp and snap the berries off the stalk. Americans, like most Europeans, are accustomed to stemless fruit.
Mr Pitzer says two thirds of the country's strawberry production is backing the move to mechanisation.
"Growers advertise and pay a lot of money - a good picker can make $30 (£22) an hour [in Florida]; in California it's $50 an hour," he says.
But they still can't recruit enough workers.
"People like to say if you paid them more they'd do the job, but it's just not true. We know in future that labour won't be available."