Artificial intelligence has reached the banana sector. The Tekniker technology center, a member of the Basque Research and Technology Alliance (BRTA) that has extensive knowledge in the use of artificial intelligence techniques, has developed within the framework of the eFOOD 4.0 project a novel tool for the characterization of bananas and optimization of its maturation process.
"We have developed a deep learning model that allows detecting characteristics or defects in images through image processing based on Artificial Intelligence," stated Aitor Gutierrez, a researcher at Tekniker. They created a 'virtual analyst' that identifies the characteristics of bananas based on a set of more than 2,000 images obtained from the facilities of Eurobanan Logistica Norte and Grupo Uvesco.
Tekniker has been responsible for pre-processing the entire data set, 'training' the Artificial Intelligence model, and launching the solution at the Uvesco Group facilities.
The tool makes it possible to optimize the ripening process of bananas to offer a product with the best quality at any time of the year, applying optimized ripening conditions with the consequent efficiency gains and reduction of food waste.
"Our 'virtual analyst' assists producers in making decisions about product management: monitoring the status of bananas in real-time and allowing them to obtain estimates on the evolution of the banana," Aitor stated.