AI Tool Helps Fruit Farmers Predict Crops

A new AI system developed by the National Robotarium is revolutionizing how fruit farmers predict their harvests.

This innovative tool uses images from standard smartphones to count flowers on fruit trees, enabling farmers to forecast crop sizes months in advance.

Tested in Catalonia, Spain, the AI achieved a 90% accuracy rate, surpassing traditional manual counting methods that can have error rates of 30-50%.

This advancement allows for more precise yield forecasts up to six months before harvest, helping growers optimize water usage, allocate resources efficiently, and better plan for harvesting and distribution. The agriculture sector, which consumes 65% of the world’s fresh water and sees nearly half of its produce wasted, stands to benefit significantly from these improvements.

The AI system, developed with scientific partners in Chile and Spain, recognizes patterns and features in flower images, even when petals overlap. Researchers will validate the AI’s predictions against the actual peach harvest in September 2024, with plans to adapt the technology for other fruits like apples, pears, and cherries.

The National Robotarium, funded by the UK and Scottish governments, aims to make Edinburgh the data capital of Europe and contribute to sustainable and productive agriculture worldwide.