Using the power of AI to predict crops and reduce waste

Researchers at the National Robotarium in Edinburgh have developed a new artificial intelligence system that can accurately count the number of flowers on fruit trees, helping farmers increase crop productivity.

The research team explained that the system's information, which it derives from smartphone images, can be used to predict crop yields 6 months before harvest, enabling farmers to optimize the allocation of resources such as water and labor, plan harvest and distribution, and reduce waste.

Experiments conducted in peach orchards in Spain showed that the system was 90% accurate in counting the number of flowers, compared to counts conducted manually, where the error rate ranged between 30-50%.

“Farmers around the world often rely on manual methods for estimating yield, which can have a large margin of error,” said Fernando Oat Chien, associate professor of autonomous systems development at the National Robotarium. “By harnessing the power of artificial intelligence and off-the-shelf technology like smartphones, Our approach integrates seamlessly with traditional agricultural practices, making it easier for farmers to adopt and benefit from innovative solutions.”

He added that during the trial in Spain, farmers expressed their “appreciation for the simplicity and accuracy of AI in flower counting, and its ability to help them make more informed decisions about crop management, such as targeted pruning and herbicide use.”

“By focusing efforts on areas that may produce the most fruit, farmers can optimize resources and maximize the quantity and quality of the crop,” Qin continued.

The researchers aim to compare the AI's predictions with the actual peach crop in September 2024.

If these methods turn out to be effective, they believe the approach could be adapted to crops such as apples, pears and cherries.