Strawberry-picking robots because we really don't have enough people for that

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.

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.

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.

http://www.bbc.com/news/business-43816207