Norfolk research finds farm robots may misjudge strawberry ripeness
Researchers find improved imaging, meaning better fruit selection by farm robots
During one of the busiest times of the year, strawberry farmers face increased pressures to supply perfectly ripe fruits for events like Wimbledon.
For many farmers, they rely on robotic technology to ensure only the ripest fruits are harvested.
However, a new study from the University of East Anglia (UEA) reveals these farming robots may be misled by the colours captured by their cameras.
The study suggests that these cameras and varying lighting conditions can trick robots into picking strawberries that look ripe but aren’t.
Instead, the researchers found that using smarter image calibrations can ensure robots pick tastier berries for shoppers.
This research marks an important step towards more intelligent robotic farming systems.
One of the researchers, James Bennett said:
"Cameras can sometimes get this wrong, they can trick and distort the perception of colour, which means counting or estimating the health of the plant can be wrong,
So we've done a calibration approach and tested this in the field, and we got much more consistent images, which means when we're counting the strawberries and predicting how many there are, this is much more accurate for the farmer."
James believes this will be beneficial to the farmer in more ways than one:
"The question on every farmer's mind is, will they have enough to meet the demand?
being able to count the number of strawberries and tell them how many they've got and forecast how many they'll have in a week is really helpful for them,
It means that fruit doesn't go to waste in the field and it means we don't have empty shelves at the supermarket"
He also said it will mean a better price for the customer.
The research was funded by AgriFoRwArdS, the world's first EPSRC Centre for Doctoral Training in Agri-Food Robotics, and will be presented at the London Imaging Meeting 2026.